aboutsummaryrefslogtreecommitdiffstats
path: root/libarchive
diff options
context:
space:
mode:
authorMartin Matuska <mm@FreeBSD.org>2011-12-20 22:47:56 +0000
committerMartin Matuska <mm@FreeBSD.org>2011-12-20 22:47:56 +0000
commit35fa5e2f58abd428bff0de1238364880447ef095 (patch)
tree07565a115ca59ea7230cd38e88901b0290750cde /libarchive
downloadsrc-35fa5e2f58abd428bff0de1238364880447ef095.tar.gz
src-35fa5e2f58abd428bff0de1238364880447ef095.zip
Vendor import of libarchive (release/2.8, r3824)
Notes
Notes: svn path=/vendor/libarchive/dist/; revision=228753
Diffstat (limited to 'libarchive')
-rw-r--r--libarchive/CMakeLists.txt139
-rw-r--r--libarchive/archive.h748
-rw-r--r--libarchive/archive_check_magic.c134
-rw-r--r--libarchive/archive_crc32.h66
-rw-r--r--libarchive/archive_endian.h162
-rw-r--r--libarchive/archive_entry.3433
-rw-r--r--libarchive/archive_entry.c2202
-rw-r--r--libarchive/archive_entry.h524
-rw-r--r--libarchive/archive_entry_copy_bhfi.c74
-rw-r--r--libarchive/archive_entry_copy_stat.c77
-rw-r--r--libarchive/archive_entry_link_resolver.c405
-rw-r--r--libarchive/archive_entry_private.h184
-rw-r--r--libarchive/archive_entry_stat.c118
-rw-r--r--libarchive/archive_entry_strmode.c87
-rw-r--r--libarchive/archive_entry_xattr.c158
-rw-r--r--libarchive/archive_hash.h309
-rw-r--r--libarchive/archive_platform.h165
-rw-r--r--libarchive/archive_private.h124
-rw-r--r--libarchive/archive_read.3714
-rw-r--r--libarchive/archive_read.c1249
-rw-r--r--libarchive/archive_read_data_into_fd.c93
-rw-r--r--libarchive/archive_read_disk.3308
-rw-r--r--libarchive/archive_read_disk.c198
-rw-r--r--libarchive/archive_read_disk_entry_from_file.c570
-rw-r--r--libarchive/archive_read_disk_private.h62
-rw-r--r--libarchive/archive_read_disk_set_standard_lookup.c303
-rw-r--r--libarchive/archive_read_extract.c183
-rw-r--r--libarchive/archive_read_open_fd.c190
-rw-r--r--libarchive/archive_read_open_file.c165
-rw-r--r--libarchive/archive_read_open_filename.c272
-rw-r--r--libarchive/archive_read_open_memory.c156
-rw-r--r--libarchive/archive_read_private.h199
-rw-r--r--libarchive/archive_read_support_compression_all.c60
-rw-r--r--libarchive/archive_read_support_compression_bzip2.c353
-rw-r--r--libarchive/archive_read_support_compression_compress.c444
-rw-r--r--libarchive/archive_read_support_compression_gzip.c465
-rw-r--r--libarchive/archive_read_support_compression_none.c40
-rw-r--r--libarchive/archive_read_support_compression_program.c459
-rw-r--r--libarchive/archive_read_support_compression_rpm.c287
-rw-r--r--libarchive/archive_read_support_compression_uu.c637
-rw-r--r--libarchive/archive_read_support_compression_xz.c708
-rw-r--r--libarchive/archive_read_support_format_all.c51
-rw-r--r--libarchive/archive_read_support_format_ar.c584
-rw-r--r--libarchive/archive_read_support_format_cpio.c777
-rw-r--r--libarchive/archive_read_support_format_empty.c93
-rw-r--r--libarchive/archive_read_support_format_iso9660.c2961
-rw-r--r--libarchive/archive_read_support_format_mtree.c1309
-rw-r--r--libarchive/archive_read_support_format_raw.c185
-rw-r--r--libarchive/archive_read_support_format_tar.c2418
-rw-r--r--libarchive/archive_read_support_format_xar.c3151
-rw-r--r--libarchive/archive_read_support_format_zip.c903
-rw-r--r--libarchive/archive_string.c453
-rw-r--r--libarchive/archive_string.h151
-rw-r--r--libarchive/archive_string_sprintf.c164
-rw-r--r--libarchive/archive_util.3160
-rw-r--r--libarchive/archive_util.c392
-rw-r--r--libarchive/archive_virtual.c94
-rw-r--r--libarchive/archive_windows.c1236
-rw-r--r--libarchive/archive_windows.h347
-rw-r--r--libarchive/archive_write.3629
-rw-r--r--libarchive/archive_write.c466
-rw-r--r--libarchive/archive_write_disk.3375
-rw-r--r--libarchive/archive_write_disk.c2712
-rw-r--r--libarchive/archive_write_disk_private.h38
-rw-r--r--libarchive/archive_write_disk_set_standard_lookup.c262
-rw-r--r--libarchive/archive_write_open_fd.c144
-rw-r--r--libarchive/archive_write_open_file.c109
-rw-r--r--libarchive/archive_write_open_filename.c166
-rw-r--r--libarchive/archive_write_open_memory.c126
-rw-r--r--libarchive/archive_write_private.h122
-rw-r--r--libarchive/archive_write_set_compression_bzip2.c408
-rw-r--r--libarchive/archive_write_set_compression_compress.c492
-rw-r--r--libarchive/archive_write_set_compression_gzip.c477
-rw-r--r--libarchive/archive_write_set_compression_none.c257
-rw-r--r--libarchive/archive_write_set_compression_program.c347
-rw-r--r--libarchive/archive_write_set_compression_xz.c439
-rw-r--r--libarchive/archive_write_set_format.c72
-rw-r--r--libarchive/archive_write_set_format_ar.c550
-rw-r--r--libarchive/archive_write_set_format_by_name.c76
-rw-r--r--libarchive/archive_write_set_format_cpio.c344
-rw-r--r--libarchive/archive_write_set_format_cpio_newc.c295
-rw-r--r--libarchive/archive_write_set_format_mtree.c1050
-rw-r--r--libarchive/archive_write_set_format_pax.c1386
-rw-r--r--libarchive/archive_write_set_format_shar.c625
-rw-r--r--libarchive/archive_write_set_format_ustar.c587
-rw-r--r--libarchive/archive_write_set_format_zip.c670
-rw-r--r--libarchive/config_freebsd.h154
-rw-r--r--libarchive/cpio.5325
-rw-r--r--libarchive/filter_fork.c161
-rw-r--r--libarchive/filter_fork.h41
-rw-r--r--libarchive/filter_fork_windows.c113
-rw-r--r--libarchive/libarchive-formats.5355
-rw-r--r--libarchive/libarchive.3331
-rw-r--r--libarchive/libarchive_internals.3365
-rw-r--r--libarchive/mtree.5269
-rw-r--r--libarchive/tar.5831
-rw-r--r--libarchive/test/.cvsignore10
-rw-r--r--libarchive/test/CMakeLists.txt159
-rw-r--r--libarchive/test/README63
-rw-r--r--libarchive/test/main.c2130
-rw-r--r--libarchive/test/read_open_memory.c167
-rw-r--r--libarchive/test/test.h302
-rw-r--r--libarchive/test/test_acl_basic.c229
-rw-r--r--libarchive/test/test_acl_freebsd.c255
-rw-r--r--libarchive/test/test_acl_pax.c517
-rw-r--r--libarchive/test/test_archive_api_feature.c76
-rw-r--r--libarchive/test/test_bad_fd.c41
-rw-r--r--libarchive/test/test_compat_bzip2.c85
-rw-r--r--libarchive/test/test_compat_bzip2_1.tbz.uu24
-rw-r--r--libarchive/test/test_compat_bzip2_2.tbz.uu11
-rw-r--r--libarchive/test/test_compat_cpio.c106
-rw-r--r--libarchive/test/test_compat_cpio_1.cpio.uu19
-rw-r--r--libarchive/test/test_compat_gtar.c119
-rw-r--r--libarchive/test/test_compat_gtar_1.tar.uu232
-rw-r--r--libarchive/test/test_compat_gzip.c97
-rw-r--r--libarchive/test/test_compat_gzip_1.tgz.uu24
-rw-r--r--libarchive/test/test_compat_gzip_2.tgz.uu11
-rw-r--r--libarchive/test/test_compat_lzma.c155
-rw-r--r--libarchive/test/test_compat_lzma_1.tlz.uu10
-rw-r--r--libarchive/test/test_compat_lzma_2.tlz.uu9
-rw-r--r--libarchive/test/test_compat_lzma_3.tlz.uu9
-rw-r--r--libarchive/test/test_compat_solaris_tar_acl.c128
-rw-r--r--libarchive/test/test_compat_solaris_tar_acl.tar.uu61
-rw-r--r--libarchive/test/test_compat_tar_hardlink.c108
-rw-r--r--libarchive/test/test_compat_tar_hardlink_1.tar.uu39
-rw-r--r--libarchive/test/test_compat_xz.c84
-rw-r--r--libarchive/test/test_compat_xz_1.txz.uu13
-rw-r--r--libarchive/test/test_compat_zip.c80
-rw-r--r--libarchive/test/test_compat_zip_1.zip.uu15
-rw-r--r--libarchive/test/test_empty_write.c120
-rw-r--r--libarchive/test/test_entry.c903
-rw-r--r--libarchive/test/test_entry_strmode.c71
-rw-r--r--libarchive/test/test_extattr_freebsd.c172
-rw-r--r--libarchive/test/test_fuzz.c166
-rw-r--r--libarchive/test/test_fuzz_1.iso.Z.uu495
-rw-r--r--libarchive/test/test_link_resolver.c205
-rw-r--r--libarchive/test/test_open_failure.c198
-rw-r--r--libarchive/test/test_open_fd.c128
-rw-r--r--libarchive/test/test_open_file.c108
-rw-r--r--libarchive/test/test_open_filename.c109
-rw-r--r--libarchive/test/test_pax_filename_encoding.c333
-rw-r--r--libarchive/test/test_pax_filename_encoding.tar.uu118
-rw-r--r--libarchive/test/test_read_compress_program.c84
-rw-r--r--libarchive/test/test_read_data_large.c125
-rw-r--r--libarchive/test/test_read_disk.c172
-rw-r--r--libarchive/test/test_read_disk_entry_from_file.c80
-rw-r--r--libarchive/test/test_read_extract.c168
-rw-r--r--libarchive/test/test_read_file_nonexistent.c37
-rw-r--r--libarchive/test/test_read_format_ar.ar.uu12
-rw-r--r--libarchive/test/test_read_format_ar.c87
-rw-r--r--libarchive/test/test_read_format_cpio_bin.c64
-rw-r--r--libarchive/test/test_read_format_cpio_bin_Z.c61
-rw-r--r--libarchive/test/test_read_format_cpio_bin_be.c55
-rw-r--r--libarchive/test/test_read_format_cpio_bin_be.cpio.uu8
-rw-r--r--libarchive/test/test_read_format_cpio_bin_bz2.c58
-rw-r--r--libarchive/test/test_read_format_cpio_bin_gz.c61
-rw-r--r--libarchive/test/test_read_format_cpio_bin_lzma.c60
-rw-r--r--libarchive/test/test_read_format_cpio_bin_xz.c70
-rw-r--r--libarchive/test/test_read_format_cpio_odc.c68
-rw-r--r--libarchive/test/test_read_format_cpio_svr4_bzip2_rpm.c127
-rw-r--r--libarchive/test/test_read_format_cpio_svr4_bzip2_rpm.rpm.uu47
-rw-r--r--libarchive/test/test_read_format_cpio_svr4_gzip.c61
-rw-r--r--libarchive/test/test_read_format_cpio_svr4_gzip_rpm.c127
-rw-r--r--libarchive/test/test_read_format_cpio_svr4_gzip_rpm.rpm.uu46
-rw-r--r--libarchive/test/test_read_format_cpio_svr4c_Z.c62
-rw-r--r--libarchive/test/test_read_format_empty.c47
-rw-r--r--libarchive/test/test_read_format_gtar_gz.c60
-rw-r--r--libarchive/test/test_read_format_gtar_lzma.c78
-rw-r--r--libarchive/test/test_read_format_gtar_sparse.c318
-rw-r--r--libarchive/test/test_read_format_gtar_sparse_1_13.tar.uu1370
-rw-r--r--libarchive/test/test_read_format_gtar_sparse_1_17.tar.uu1370
-rw-r--r--libarchive/test/test_read_format_gtar_sparse_1_17_posix00.tar.uu1597
-rw-r--r--libarchive/test/test_read_format_gtar_sparse_1_17_posix01.tar.uu1370
-rw-r--r--libarchive/test/test_read_format_gtar_sparse_1_17_posix10.tar.uu1370
-rw-r--r--libarchive/test/test_read_format_gtar_sparse_1_17_posix10_modified.tar.uu1370
-rw-r--r--libarchive/test/test_read_format_iso.iso.Z.uu26
-rw-r--r--libarchive/test/test_read_format_iso_2.iso.Z.uu35
-rw-r--r--libarchive/test/test_read_format_iso_Z.c99
-rw-r--r--libarchive/test/test_read_format_iso_joliet.iso.Z.uu66
-rw-r--r--libarchive/test/test_read_format_iso_joliet_long.iso.Z.uu71
-rw-r--r--libarchive/test/test_read_format_iso_joliet_rockridge.iso.Z.uu68
-rw-r--r--libarchive/test/test_read_format_iso_multi_extent.c94
-rw-r--r--libarchive/test/test_read_format_iso_multi_extent.iso.Z.uu67
-rw-r--r--libarchive/test/test_read_format_iso_rockridge.iso.Z.uu206
-rw-r--r--libarchive/test/test_read_format_iso_rockridge_ce.iso.Z.uu63
-rw-r--r--libarchive/test/test_read_format_iso_rockridge_new.iso.Z.uu208
-rw-r--r--libarchive/test/test_read_format_iso_rockridge_rr_moved.iso.Z.uu304
-rw-r--r--libarchive/test/test_read_format_iso_zisofs.iso.Z.uu63
-rw-r--r--libarchive/test/test_read_format_isojoliet_bz2.c135
-rw-r--r--libarchive/test/test_read_format_isojoliet_long.c141
-rw-r--r--libarchive/test/test_read_format_isojoliet_rr.c160
-rw-r--r--libarchive/test/test_read_format_isorr_bz2.c203
-rw-r--r--libarchive/test/test_read_format_isorr_ce.c223
-rw-r--r--libarchive/test/test_read_format_isorr_new_bz2.c204
-rw-r--r--libarchive/test/test_read_format_isorr_rr_moved.c270
-rw-r--r--libarchive/test/test_read_format_isozisofs_bz2.c187
-rw-r--r--libarchive/test/test_read_format_mtree.c186
-rw-r--r--libarchive/test/test_read_format_mtree.mtree.uu13
-rw-r--r--libarchive/test/test_read_format_pax_bz2.c66
-rw-r--r--libarchive/test/test_read_format_raw.c89
-rw-r--r--libarchive/test/test_read_format_raw.data.Z.uu5
-rw-r--r--libarchive/test/test_read_format_raw.data.uu5
-rw-r--r--libarchive/test/test_read_format_tar.c480
-rw-r--r--libarchive/test/test_read_format_tar_empty_filename.c66
-rw-r--r--libarchive/test/test_read_format_tar_empty_filename.tar.uu39
-rw-r--r--libarchive/test/test_read_format_tbz.c59
-rw-r--r--libarchive/test/test_read_format_tgz.c60
-rw-r--r--libarchive/test/test_read_format_tlz.c60
-rw-r--r--libarchive/test/test_read_format_txz.c63
-rw-r--r--libarchive/test/test_read_format_tz.c61
-rw-r--r--libarchive/test/test_read_format_xar.c697
-rw-r--r--libarchive/test/test_read_format_zip.c92
-rw-r--r--libarchive/test/test_read_format_zip.zip.uu14
-rw-r--r--libarchive/test/test_read_large.c93
-rw-r--r--libarchive/test/test_read_pax_truncated.c288
-rw-r--r--libarchive/test/test_read_position.c94
-rw-r--r--libarchive/test/test_read_truncated.c149
-rw-r--r--libarchive/test/test_read_uu.c134
-rw-r--r--libarchive/test/test_tar_filenames.c186
-rw-r--r--libarchive/test/test_tar_large.c312
-rw-r--r--libarchive/test/test_ustar_filenames.c191
-rw-r--r--libarchive/test/test_write_compress.c102
-rw-r--r--libarchive/test/test_write_compress_bzip2.c228
-rw-r--r--libarchive/test/test_write_compress_gzip.c252
-rw-r--r--libarchive/test/test_write_compress_lzma.c250
-rw-r--r--libarchive/test/test_write_compress_program.c118
-rw-r--r--libarchive/test/test_write_compress_xz.c257
-rw-r--r--libarchive/test/test_write_disk.c332
-rw-r--r--libarchive/test/test_write_disk_failures.c72
-rw-r--r--libarchive/test/test_write_disk_hardlink.c219
-rw-r--r--libarchive/test/test_write_disk_perms.c457
-rw-r--r--libarchive/test/test_write_disk_secure.c215
-rw-r--r--libarchive/test/test_write_disk_sparse.c280
-rw-r--r--libarchive/test/test_write_disk_symlink.c117
-rw-r--r--libarchive/test/test_write_disk_times.c167
-rw-r--r--libarchive/test/test_write_format_ar.c209
-rw-r--r--libarchive/test/test_write_format_cpio.c194
-rw-r--r--libarchive/test/test_write_format_cpio_empty.c75
-rw-r--r--libarchive/test/test_write_format_cpio_newc.c214
-rw-r--r--libarchive/test/test_write_format_cpio_odc.c241
-rw-r--r--libarchive/test/test_write_format_mtree.c155
-rw-r--r--libarchive/test/test_write_format_pax.c146
-rw-r--r--libarchive/test/test_write_format_shar_empty.c58
-rw-r--r--libarchive/test/test_write_format_tar.c114
-rw-r--r--libarchive/test/test_write_format_tar_empty.c92
-rw-r--r--libarchive/test/test_write_format_tar_ustar.c347
-rw-r--r--libarchive/test/test_write_format_zip.c180
-rw-r--r--libarchive/test/test_write_format_zip_empty.c56
-rw-r--r--libarchive/test/test_write_format_zip_no_compression.c304
-rw-r--r--libarchive/test/test_write_open_memory.c76
250 files changed, 76948 insertions, 0 deletions
diff --git a/libarchive/CMakeLists.txt b/libarchive/CMakeLists.txt
new file mode 100644
index 000000000000..459a5ddbb903
--- /dev/null
+++ b/libarchive/CMakeLists.txt
@@ -0,0 +1,139 @@
+
+############################################
+#
+# How to build libarchive
+#
+############################################
+
+# Public headers
+SET(include_HEADERS
+ archive.h
+ archive_entry.h
+)
+
+# Sources and private headers
+SET(libarchive_SOURCES
+ archive_check_magic.c
+ archive_endian.h
+ archive_entry.c
+ archive_entry.h
+ archive_entry_copy_stat.c
+ archive_entry_link_resolver.c
+ archive_entry_private.h
+ archive_entry_stat.c
+ archive_entry_strmode.c
+ archive_entry_xattr.c
+ archive_hash.h
+ archive_platform.h
+ archive_private.h
+ archive_read.c
+ archive_read_data_into_fd.c
+ archive_read_disk.c
+ archive_read_disk_entry_from_file.c
+ archive_read_disk_private.h
+ archive_read_disk_set_standard_lookup.c
+ archive_read_extract.c
+ archive_read_open_fd.c
+ archive_read_open_file.c
+ archive_read_open_filename.c
+ archive_read_open_memory.c
+ archive_read_private.h
+ archive_read_support_compression_all.c
+ archive_read_support_compression_bzip2.c
+ archive_read_support_compression_compress.c
+ archive_read_support_compression_gzip.c
+ archive_read_support_compression_none.c
+ archive_read_support_compression_program.c
+ archive_read_support_compression_rpm.c
+ archive_read_support_compression_uu.c
+ archive_read_support_compression_xz.c
+ archive_read_support_format_all.c
+ archive_read_support_format_ar.c
+ archive_read_support_format_cpio.c
+ archive_read_support_format_empty.c
+ archive_read_support_format_iso9660.c
+ archive_read_support_format_mtree.c
+ archive_read_support_format_raw.c
+ archive_read_support_format_tar.c
+ archive_read_support_format_xar.c
+ archive_read_support_format_zip.c
+ archive_string.c
+ archive_string.h
+ archive_string_sprintf.c
+ archive_util.c
+ archive_virtual.c
+ archive_write.c
+ archive_write_disk.c
+ archive_write_disk_private.h
+ archive_write_disk_set_standard_lookup.c
+ archive_write_private.h
+ archive_write_open_fd.c
+ archive_write_open_file.c
+ archive_write_open_filename.c
+ archive_write_open_memory.c
+ archive_write_set_compression_bzip2.c
+ archive_write_set_compression_compress.c
+ archive_write_set_compression_gzip.c
+ archive_write_set_compression_none.c
+ archive_write_set_compression_program.c
+ archive_write_set_compression_xz.c
+ archive_write_set_format.c
+ archive_write_set_format_ar.c
+ archive_write_set_format_by_name.c
+ archive_write_set_format_cpio.c
+ archive_write_set_format_cpio_newc.c
+ archive_write_set_format_mtree.c
+ archive_write_set_format_pax.c
+ archive_write_set_format_shar.c
+ archive_write_set_format_ustar.c
+ archive_write_set_format_zip.c
+ filter_fork.c
+ filter_fork.h
+)
+
+# Man pages
+SET(libarchive_MANS
+ archive_entry.3
+ archive_read.3
+ archive_read_disk.3
+ archive_util.3
+ archive_write.3
+ archive_write_disk.3
+ cpio.5
+ libarchive.3
+ libarchive_internals.3
+ libarchive-formats.5
+ mtree.5
+ tar.5
+)
+
+IF(WIN32 AND NOT CYGWIN)
+ LIST(APPEND libarchive_SOURCES archive_entry_copy_bhfi.c)
+ LIST(APPEND libarchive_SOURCES archive_windows.c)
+ LIST(APPEND libarchive_SOURCES archive_windows.h)
+ LIST(APPEND libarchive_SOURCES filter_fork_windows.c)
+ENDIF(WIN32 AND NOT CYGWIN)
+
+# Libarchive is a shared library
+ADD_LIBRARY(archive SHARED ${libarchive_SOURCES} ${include_HEADERS})
+TARGET_LINK_LIBRARIES(archive ${ADDITIONAL_LIBS})
+SET_TARGET_PROPERTIES(archive PROPERTIES SOVERSION ${SOVERSION})
+
+# archive_static is a static library
+ADD_LIBRARY(archive_static STATIC ${libarchive_SOURCES} ${include_HEADERS})
+SET_TARGET_PROPERTIES(archive_static PROPERTIES COMPILE_DEFINITIONS
+ LIBARCHIVE_STATIC)
+# On Posix systems, libarchive.so and libarchive.a can co-exist.
+IF(NOT WIN32 OR CYGWIN)
+ SET_TARGET_PROPERTIES(archive_static PROPERTIES OUTPUT_NAME archive)
+ENDIF(NOT WIN32 OR CYGWIN)
+
+# How to install the libraries
+INSTALL(TARGETS archive archive_static
+ RUNTIME DESTINATION bin
+ LIBRARY DESTINATION lib
+ ARCHIVE DESTINATION lib)
+INSTALL_MAN(${libarchive_MANS})
+INSTALL(FILES ${include_HEADERS} DESTINATION include)
+
+add_subdirectory(test)
diff --git a/libarchive/archive.h b/libarchive/archive.h
new file mode 100644
index 000000000000..140df548b738
--- /dev/null
+++ b/libarchive/archive.h
@@ -0,0 +1,748 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD: src/lib/libarchive/archive.h.in,v 1.50 2008/05/26 17:00:22 kientzle Exp $
+ */
+
+#ifndef ARCHIVE_H_INCLUDED
+#define ARCHIVE_H_INCLUDED
+
+/*
+ * Note: archive.h is for use outside of libarchive; the configuration
+ * headers (config.h, archive_platform.h, etc.) are purely internal.
+ * Do NOT use HAVE_XXX configuration macros to control the behavior of
+ * this header! If you must conditionalize, use predefined compiler and/or
+ * platform macros.
+ */
+#if defined(__BORLANDC__) && __BORLANDC__ >= 0x560
+# define __LA_STDINT_H <stdint.h>
+#elif !defined(__WATCOMC__) && !defined(_MSC_VER) && !defined(__INTERIX) && !defined(__BORLANDC__)
+# define __LA_STDINT_H <inttypes.h>
+#endif
+
+#include <sys/stat.h>
+#include <sys/types.h> /* Linux requires this for off_t */
+#ifdef __LA_STDINT_H
+# include __LA_STDINT_H /* int64_t, etc. */
+#endif
+#include <stdio.h> /* For FILE * */
+
+/* Get appropriate definitions of standard POSIX-style types. */
+/* These should match the types used in 'struct stat' */
+#if defined(_WIN32) && !defined(__CYGWIN__)
+#define __LA_INT64_T __int64
+# if defined(_SSIZE_T_DEFINED) || defined(_SSIZE_T_)
+# define __LA_SSIZE_T ssize_t
+# elif defined(_WIN64)
+# define __LA_SSIZE_T __int64
+# else
+# define __LA_SSIZE_T long
+# endif
+# if defined(__BORLANDC__)
+# define __LA_UID_T uid_t
+# define __LA_GID_T gid_t
+# else
+# define __LA_UID_T short
+# define __LA_GID_T short
+# endif
+#else
+#include <unistd.h> /* ssize_t, uid_t, and gid_t */
+#define __LA_INT64_T int64_t
+#define __LA_SSIZE_T ssize_t
+#define __LA_UID_T uid_t
+#define __LA_GID_T gid_t
+#endif
+
+/*
+ * On Windows, define LIBARCHIVE_STATIC if you're building or using a
+ * .lib. The default here assumes you're building a DLL. Only
+ * libarchive source should ever define __LIBARCHIVE_BUILD.
+ */
+#if ((defined __WIN32__) || (defined _WIN32) || defined(__CYGWIN__)) && (!defined LIBARCHIVE_STATIC)
+# ifdef __LIBARCHIVE_BUILD
+# ifdef __GNUC__
+# define __LA_DECL __attribute__((dllexport)) extern
+# else
+# define __LA_DECL __declspec(dllexport)
+# endif
+# else
+# ifdef __GNUC__
+# define __LA_DECL __attribute__((dllimport)) extern
+# else
+# define __LA_DECL __declspec(dllimport)
+# endif
+# endif
+#else
+/* Static libraries or non-Windows needs no special declaration. */
+# define __LA_DECL
+#endif
+
+#if defined(__GNUC__) && __GNUC__ >= 3
+#define __LA_PRINTF(fmtarg, firstvararg) \
+ __attribute__((__format__ (__printf__, fmtarg, firstvararg)))
+#else
+#define __LA_PRINTF(fmtarg, firstvararg) /* nothing */
+#endif
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+ * The version number is provided as both a macro and a function.
+ * The macro identifies the installed header; the function identifies
+ * the library version (which may not be the same if you're using a
+ * dynamically-linked version of the library). Of course, if the
+ * header and library are very different, you should expect some
+ * strangeness. Don't do that.
+ */
+
+/*
+ * The version number is expressed as a single integer that makes it
+ * easy to compare versions at build time: for version a.b.c, the
+ * version number is printf("%d%03d%03d",a,b,c). For example, if you
+ * know your application requires version 2.12.108 or later, you can
+ * assert that ARCHIVE_VERSION >= 2012108.
+ *
+ * This single-number format was introduced with libarchive 1.9.0 in
+ * the libarchive 1.x family and libarchive 2.2.4 in the libarchive
+ * 2.x family. The following may be useful if you really want to do
+ * feature detection for earlier libarchive versions (which defined
+ * ARCHIVE_API_VERSION and ARCHIVE_API_FEATURE instead):
+ *
+ * #ifndef ARCHIVE_VERSION_NUMBER
+ * #define ARCHIVE_VERSION_NUMBER \
+ * (ARCHIVE_API_VERSION * 1000000 + ARCHIVE_API_FEATURE * 1000)
+ * #endif
+ */
+#define ARCHIVE_VERSION_NUMBER 2008005
+__LA_DECL int archive_version_number(void);
+
+/*
+ * Textual name/version of the library, useful for version displays.
+ */
+#define ARCHIVE_VERSION_STRING "libarchive 2.8.5"
+__LA_DECL const char * archive_version_string(void);
+
+#if ARCHIVE_VERSION_NUMBER < 3000000
+/*
+ * Deprecated; these are older names that will be removed in favor of
+ * the simpler definitions above.
+ */
+#define ARCHIVE_VERSION_STAMP ARCHIVE_VERSION_NUMBER
+__LA_DECL int archive_version_stamp(void);
+#define ARCHIVE_LIBRARY_VERSION ARCHIVE_VERSION_STRING
+__LA_DECL const char * archive_version(void);
+#define ARCHIVE_API_VERSION (ARCHIVE_VERSION_NUMBER / 1000000)
+__LA_DECL int archive_api_version(void);
+#define ARCHIVE_API_FEATURE ((ARCHIVE_VERSION_NUMBER / 1000) % 1000)
+__LA_DECL int archive_api_feature(void);
+#endif
+
+#if ARCHIVE_VERSION_NUMBER < 3000000
+/* This should never have been here in the first place. */
+/* Legacy of old tar assumptions, will be removed in libarchive 3.0. */
+#define ARCHIVE_BYTES_PER_RECORD 512
+#define ARCHIVE_DEFAULT_BYTES_PER_BLOCK 10240
+#endif
+
+/* Declare our basic types. */
+struct archive;
+struct archive_entry;
+
+/*
+ * Error codes: Use archive_errno() and archive_error_string()
+ * to retrieve details. Unless specified otherwise, all functions
+ * that return 'int' use these codes.
+ */
+#define ARCHIVE_EOF 1 /* Found end of archive. */
+#define ARCHIVE_OK 0 /* Operation was successful. */
+#define ARCHIVE_RETRY (-10) /* Retry might succeed. */
+#define ARCHIVE_WARN (-20) /* Partial success. */
+/* For example, if write_header "fails", then you can't push data. */
+#define ARCHIVE_FAILED (-25) /* Current operation cannot complete. */
+/* But if write_header is "fatal," then this archive is dead and useless. */
+#define ARCHIVE_FATAL (-30) /* No more operations are possible. */
+
+/*
+ * As far as possible, archive_errno returns standard platform errno codes.
+ * Of course, the details vary by platform, so the actual definitions
+ * here are stored in "archive_platform.h". The symbols are listed here
+ * for reference; as a rule, clients should not need to know the exact
+ * platform-dependent error code.
+ */
+/* Unrecognized or invalid file format. */
+/* #define ARCHIVE_ERRNO_FILE_FORMAT */
+/* Illegal usage of the library. */
+/* #define ARCHIVE_ERRNO_PROGRAMMER_ERROR */
+/* Unknown or unclassified error. */
+/* #define ARCHIVE_ERRNO_MISC */
+
+/*
+ * Callbacks are invoked to automatically read/skip/write/open/close the
+ * archive. You can provide your own for complex tasks (like breaking
+ * archives across multiple tapes) or use standard ones built into the
+ * library.
+ */
+
+/* Returns pointer and size of next block of data from archive. */
+typedef __LA_SSIZE_T archive_read_callback(struct archive *,
+ void *_client_data, const void **_buffer);
+
+/* Skips at most request bytes from archive and returns the skipped amount */
+#if ARCHIVE_VERSION_NUMBER < 2000000
+/* Libarchive 1.0 used ssize_t for the return, which is only 32 bits
+ * on most 32-bit platforms; not large enough. */
+typedef __LA_SSIZE_T archive_skip_callback(struct archive *,
+ void *_client_data, size_t request);
+#elif ARCHIVE_VERSION_NUMBER < 3000000
+/* Libarchive 2.0 used off_t here, but that is a bad idea on Linux and a
+ * few other platforms where off_t varies with build settings. */
+typedef off_t archive_skip_callback(struct archive *,
+ void *_client_data, off_t request);
+#else
+/* Libarchive 3.0 uses int64_t here, which is actually guaranteed to be
+ * 64 bits on every platform. */
+typedef __LA_INT64_T archive_skip_callback(struct archive *,
+ void *_client_data, __LA_INT64_T request);
+#endif
+
+/* Returns size actually written, zero on EOF, -1 on error. */
+typedef __LA_SSIZE_T archive_write_callback(struct archive *,
+ void *_client_data,
+ const void *_buffer, size_t _length);
+
+#if ARCHIVE_VERSION_NUMBER < 3000000
+/* Open callback is actually never needed; remove it in libarchive 3.0. */
+typedef int archive_open_callback(struct archive *, void *_client_data);
+#endif
+
+typedef int archive_close_callback(struct archive *, void *_client_data);
+
+/*
+ * Codes for archive_compression.
+ */
+#define ARCHIVE_COMPRESSION_NONE 0
+#define ARCHIVE_COMPRESSION_GZIP 1
+#define ARCHIVE_COMPRESSION_BZIP2 2
+#define ARCHIVE_COMPRESSION_COMPRESS 3
+#define ARCHIVE_COMPRESSION_PROGRAM 4
+#define ARCHIVE_COMPRESSION_LZMA 5
+#define ARCHIVE_COMPRESSION_XZ 6
+#define ARCHIVE_COMPRESSION_UU 7
+#define ARCHIVE_COMPRESSION_RPM 8
+
+/*
+ * Codes returned by archive_format.
+ *
+ * Top 16 bits identifies the format family (e.g., "tar"); lower
+ * 16 bits indicate the variant. This is updated by read_next_header.
+ * Note that the lower 16 bits will often vary from entry to entry.
+ * In some cases, this variation occurs as libarchive learns more about
+ * the archive (for example, later entries might utilize extensions that
+ * weren't necessary earlier in the archive; in this case, libarchive
+ * will change the format code to indicate the extended format that
+ * was used). In other cases, it's because different tools have
+ * modified the archive and so different parts of the archive
+ * actually have slightly different formts. (Both tar and cpio store
+ * format codes in each entry, so it is quite possible for each
+ * entry to be in a different format.)
+ */
+#define ARCHIVE_FORMAT_BASE_MASK 0xff0000
+#define ARCHIVE_FORMAT_CPIO 0x10000
+#define ARCHIVE_FORMAT_CPIO_POSIX (ARCHIVE_FORMAT_CPIO | 1)
+#define ARCHIVE_FORMAT_CPIO_BIN_LE (ARCHIVE_FORMAT_CPIO | 2)
+#define ARCHIVE_FORMAT_CPIO_BIN_BE (ARCHIVE_FORMAT_CPIO | 3)
+#define ARCHIVE_FORMAT_CPIO_SVR4_NOCRC (ARCHIVE_FORMAT_CPIO | 4)
+#define ARCHIVE_FORMAT_CPIO_SVR4_CRC (ARCHIVE_FORMAT_CPIO | 5)
+#define ARCHIVE_FORMAT_SHAR 0x20000
+#define ARCHIVE_FORMAT_SHAR_BASE (ARCHIVE_FORMAT_SHAR | 1)
+#define ARCHIVE_FORMAT_SHAR_DUMP (ARCHIVE_FORMAT_SHAR | 2)
+#define ARCHIVE_FORMAT_TAR 0x30000
+#define ARCHIVE_FORMAT_TAR_USTAR (ARCHIVE_FORMAT_TAR | 1)
+#define ARCHIVE_FORMAT_TAR_PAX_INTERCHANGE (ARCHIVE_FORMAT_TAR | 2)
+#define ARCHIVE_FORMAT_TAR_PAX_RESTRICTED (ARCHIVE_FORMAT_TAR | 3)
+#define ARCHIVE_FORMAT_TAR_GNUTAR (ARCHIVE_FORMAT_TAR | 4)
+#define ARCHIVE_FORMAT_ISO9660 0x40000
+#define ARCHIVE_FORMAT_ISO9660_ROCKRIDGE (ARCHIVE_FORMAT_ISO9660 | 1)
+#define ARCHIVE_FORMAT_ZIP 0x50000
+#define ARCHIVE_FORMAT_EMPTY 0x60000
+#define ARCHIVE_FORMAT_AR 0x70000
+#define ARCHIVE_FORMAT_AR_GNU (ARCHIVE_FORMAT_AR | 1)
+#define ARCHIVE_FORMAT_AR_BSD (ARCHIVE_FORMAT_AR | 2)
+#define ARCHIVE_FORMAT_MTREE 0x80000
+#define ARCHIVE_FORMAT_RAW 0x90000
+#define ARCHIVE_FORMAT_XAR 0xA0000
+
+/*-
+ * Basic outline for reading an archive:
+ * 1) Ask archive_read_new for an archive reader object.
+ * 2) Update any global properties as appropriate.
+ * In particular, you'll certainly want to call appropriate
+ * archive_read_support_XXX functions.
+ * 3) Call archive_read_open_XXX to open the archive
+ * 4) Repeatedly call archive_read_next_header to get information about
+ * successive archive entries. Call archive_read_data to extract
+ * data for entries of interest.
+ * 5) Call archive_read_finish to end processing.
+ */
+__LA_DECL struct archive *archive_read_new(void);
+
+/*
+ * The archive_read_support_XXX calls enable auto-detect for this
+ * archive handle. They also link in the necessary support code.
+ * For example, if you don't want bzlib linked in, don't invoke
+ * support_compression_bzip2(). The "all" functions provide the
+ * obvious shorthand.
+ */
+__LA_DECL int archive_read_support_compression_all(struct archive *);
+__LA_DECL int archive_read_support_compression_bzip2(struct archive *);
+__LA_DECL int archive_read_support_compression_compress(struct archive *);
+__LA_DECL int archive_read_support_compression_gzip(struct archive *);
+__LA_DECL int archive_read_support_compression_lzma(struct archive *);
+__LA_DECL int archive_read_support_compression_none(struct archive *);
+__LA_DECL int archive_read_support_compression_program(struct archive *,
+ const char *command);
+__LA_DECL int archive_read_support_compression_program_signature
+ (struct archive *, const char *,
+ const void * /* match */, size_t);
+
+__LA_DECL int archive_read_support_compression_rpm(struct archive *);
+__LA_DECL int archive_read_support_compression_uu(struct archive *);
+__LA_DECL int archive_read_support_compression_xz(struct archive *);
+
+__LA_DECL int archive_read_support_format_all(struct archive *);
+__LA_DECL int archive_read_support_format_ar(struct archive *);
+__LA_DECL int archive_read_support_format_cpio(struct archive *);
+__LA_DECL int archive_read_support_format_empty(struct archive *);
+__LA_DECL int archive_read_support_format_gnutar(struct archive *);
+__LA_DECL int archive_read_support_format_iso9660(struct archive *);
+__LA_DECL int archive_read_support_format_mtree(struct archive *);
+__LA_DECL int archive_read_support_format_raw(struct archive *);
+__LA_DECL int archive_read_support_format_tar(struct archive *);
+__LA_DECL int archive_read_support_format_xar(struct archive *);
+__LA_DECL int archive_read_support_format_zip(struct archive *);
+
+
+/* Open the archive using callbacks for archive I/O. */
+__LA_DECL int archive_read_open(struct archive *, void *_client_data,
+ archive_open_callback *, archive_read_callback *,
+ archive_close_callback *);
+__LA_DECL int archive_read_open2(struct archive *, void *_client_data,
+ archive_open_callback *, archive_read_callback *,
+ archive_skip_callback *, archive_close_callback *);
+
+/*
+ * A variety of shortcuts that invoke archive_read_open() with
+ * canned callbacks suitable for common situations. The ones that
+ * accept a block size handle tape blocking correctly.
+ */
+/* Use this if you know the filename. Note: NULL indicates stdin. */
+__LA_DECL int archive_read_open_filename(struct archive *,
+ const char *_filename, size_t _block_size);
+/* archive_read_open_file() is a deprecated synonym for ..._open_filename(). */
+__LA_DECL int archive_read_open_file(struct archive *,
+ const char *_filename, size_t _block_size);
+/* Read an archive that's stored in memory. */
+__LA_DECL int archive_read_open_memory(struct archive *,
+ void * buff, size_t size);
+/* A more involved version that is only used for internal testing. */
+__LA_DECL int archive_read_open_memory2(struct archive *a, void *buff,
+ size_t size, size_t read_size);
+/* Read an archive that's already open, using the file descriptor. */
+__LA_DECL int archive_read_open_fd(struct archive *, int _fd,
+ size_t _block_size);
+/* Read an archive that's already open, using a FILE *. */
+/* Note: DO NOT use this with tape drives. */
+__LA_DECL int archive_read_open_FILE(struct archive *, FILE *_file);
+
+/* Parses and returns next entry header. */
+__LA_DECL int archive_read_next_header(struct archive *,
+ struct archive_entry **);
+
+/* Parses and returns next entry header using the archive_entry passed in */
+__LA_DECL int archive_read_next_header2(struct archive *,
+ struct archive_entry *);
+
+/*
+ * Retrieve the byte offset in UNCOMPRESSED data where last-read
+ * header started.
+ */
+__LA_DECL __LA_INT64_T archive_read_header_position(struct archive *);
+
+/* Read data from the body of an entry. Similar to read(2). */
+__LA_DECL __LA_SSIZE_T archive_read_data(struct archive *,
+ void *, size_t);
+
+/*
+ * A zero-copy version of archive_read_data that also exposes the file offset
+ * of each returned block. Note that the client has no way to specify
+ * the desired size of the block. The API does guarantee that offsets will
+ * be strictly increasing and that returned blocks will not overlap.
+ */
+#if ARCHIVE_VERSION_NUMBER < 3000000
+__LA_DECL int archive_read_data_block(struct archive *a,
+ const void **buff, size_t *size, off_t *offset);
+#else
+__LA_DECL int archive_read_data_block(struct archive *a,
+ const void **buff, size_t *size,
+ __LA_INT64_T *offset);
+#endif
+
+/*-
+ * Some convenience functions that are built on archive_read_data:
+ * 'skip': skips entire entry
+ * 'into_buffer': writes data into memory buffer that you provide
+ * 'into_fd': writes data to specified filedes
+ */
+__LA_DECL int archive_read_data_skip(struct archive *);
+__LA_DECL int archive_read_data_into_buffer(struct archive *,
+ void *buffer, __LA_SSIZE_T len);
+__LA_DECL int archive_read_data_into_fd(struct archive *, int fd);
+
+/*
+ * Set read options.
+ */
+/* Apply option string to the format only. */
+__LA_DECL int archive_read_set_format_options(struct archive *_a,
+ const char *s);
+/* Apply option string to the filter only. */
+__LA_DECL int archive_read_set_filter_options(struct archive *_a,
+ const char *s);
+/* Apply option string to both the format and the filter. */
+__LA_DECL int archive_read_set_options(struct archive *_a,
+ const char *s);
+
+/*-
+ * Convenience function to recreate the current entry (whose header
+ * has just been read) on disk.
+ *
+ * This does quite a bit more than just copy data to disk. It also:
+ * - Creates intermediate directories as required.
+ * - Manages directory permissions: non-writable directories will
+ * be initially created with write permission enabled; when the
+ * archive is closed, dir permissions are edited to the values specified
+ * in the archive.
+ * - Checks hardlinks: hardlinks will not be extracted unless the
+ * linked-to file was also extracted within the same session. (TODO)
+ */
+
+/* The "flags" argument selects optional behavior, 'OR' the flags you want. */
+
+/* Default: Do not try to set owner/group. */
+#define ARCHIVE_EXTRACT_OWNER (0x0001)
+/* Default: Do obey umask, do not restore SUID/SGID/SVTX bits. */
+#define ARCHIVE_EXTRACT_PERM (0x0002)
+/* Default: Do not restore mtime/atime. */
+#define ARCHIVE_EXTRACT_TIME (0x0004)
+/* Default: Replace existing files. */
+#define ARCHIVE_EXTRACT_NO_OVERWRITE (0x0008)
+/* Default: Try create first, unlink only if create fails with EEXIST. */
+#define ARCHIVE_EXTRACT_UNLINK (0x0010)
+/* Default: Do not restore ACLs. */
+#define ARCHIVE_EXTRACT_ACL (0x0020)
+/* Default: Do not restore fflags. */
+#define ARCHIVE_EXTRACT_FFLAGS (0x0040)
+/* Default: Do not restore xattrs. */
+#define ARCHIVE_EXTRACT_XATTR (0x0080)
+/* Default: Do not try to guard against extracts redirected by symlinks. */
+/* Note: With ARCHIVE_EXTRACT_UNLINK, will remove any intermediate symlink. */
+#define ARCHIVE_EXTRACT_SECURE_SYMLINKS (0x0100)
+/* Default: Do not reject entries with '..' as path elements. */
+#define ARCHIVE_EXTRACT_SECURE_NODOTDOT (0x0200)
+/* Default: Create parent directories as needed. */
+#define ARCHIVE_EXTRACT_NO_AUTODIR (0x0400)
+/* Default: Overwrite files, even if one on disk is newer. */
+#define ARCHIVE_EXTRACT_NO_OVERWRITE_NEWER (0x0800)
+/* Detect blocks of 0 and write holes instead. */
+#define ARCHIVE_EXTRACT_SPARSE (0x1000)
+
+__LA_DECL int archive_read_extract(struct archive *, struct archive_entry *,
+ int flags);
+__LA_DECL int archive_read_extract2(struct archive *, struct archive_entry *,
+ struct archive * /* dest */);
+__LA_DECL void archive_read_extract_set_progress_callback(struct archive *,
+ void (*_progress_func)(void *), void *_user_data);
+
+/* Record the dev/ino of a file that will not be written. This is
+ * generally set to the dev/ino of the archive being read. */
+__LA_DECL void archive_read_extract_set_skip_file(struct archive *,
+ dev_t, ino_t);
+
+/* Close the file and release most resources. */
+__LA_DECL int archive_read_close(struct archive *);
+/* Release all resources and destroy the object. */
+/* Note that archive_read_finish will call archive_read_close for you. */
+#if ARCHIVE_VERSION_NUMBER < 2000000
+/* Erroneously declared to return void in libarchive 1.x */
+__LA_DECL void archive_read_finish(struct archive *);
+#else
+__LA_DECL int archive_read_finish(struct archive *);
+#endif
+
+/*-
+ * To create an archive:
+ * 1) Ask archive_write_new for a archive writer object.
+ * 2) Set any global properties. In particular, you should set
+ * the compression and format to use.
+ * 3) Call archive_write_open to open the file (most people
+ * will use archive_write_open_file or archive_write_open_fd,
+ * which provide convenient canned I/O callbacks for you).
+ * 4) For each entry:
+ * - construct an appropriate struct archive_entry structure
+ * - archive_write_header to write the header
+ * - archive_write_data to write the entry data
+ * 5) archive_write_close to close the output
+ * 6) archive_write_finish to cleanup the writer and release resources
+ */
+__LA_DECL struct archive *archive_write_new(void);
+__LA_DECL int archive_write_set_bytes_per_block(struct archive *,
+ int bytes_per_block);
+__LA_DECL int archive_write_get_bytes_per_block(struct archive *);
+/* XXX This is badly misnamed; suggestions appreciated. XXX */
+__LA_DECL int archive_write_set_bytes_in_last_block(struct archive *,
+ int bytes_in_last_block);
+__LA_DECL int archive_write_get_bytes_in_last_block(struct archive *);
+
+/* The dev/ino of a file that won't be archived. This is used
+ * to avoid recursively adding an archive to itself. */
+__LA_DECL int archive_write_set_skip_file(struct archive *, dev_t, ino_t);
+
+__LA_DECL int archive_write_set_compression_bzip2(struct archive *);
+__LA_DECL int archive_write_set_compression_compress(struct archive *);
+__LA_DECL int archive_write_set_compression_gzip(struct archive *);
+__LA_DECL int archive_write_set_compression_lzma(struct archive *);
+__LA_DECL int archive_write_set_compression_none(struct archive *);
+__LA_DECL int archive_write_set_compression_program(struct archive *,
+ const char *cmd);
+__LA_DECL int archive_write_set_compression_xz(struct archive *);
+/* A convenience function to set the format based on the code or name. */
+__LA_DECL int archive_write_set_format(struct archive *, int format_code);
+__LA_DECL int archive_write_set_format_by_name(struct archive *,
+ const char *name);
+/* To minimize link pollution, use one or more of the following. */
+__LA_DECL int archive_write_set_format_ar_bsd(struct archive *);
+__LA_DECL int archive_write_set_format_ar_svr4(struct archive *);
+__LA_DECL int archive_write_set_format_cpio(struct archive *);
+__LA_DECL int archive_write_set_format_cpio_newc(struct archive *);
+__LA_DECL int archive_write_set_format_mtree(struct archive *);
+/* TODO: int archive_write_set_format_old_tar(struct archive *); */
+__LA_DECL int archive_write_set_format_pax(struct archive *);
+__LA_DECL int archive_write_set_format_pax_restricted(struct archive *);
+__LA_DECL int archive_write_set_format_shar(struct archive *);
+__LA_DECL int archive_write_set_format_shar_dump(struct archive *);
+__LA_DECL int archive_write_set_format_ustar(struct archive *);
+__LA_DECL int archive_write_set_format_zip(struct archive *);
+__LA_DECL int archive_write_open(struct archive *, void *,
+ archive_open_callback *, archive_write_callback *,
+ archive_close_callback *);
+__LA_DECL int archive_write_open_fd(struct archive *, int _fd);
+__LA_DECL int archive_write_open_filename(struct archive *, const char *_file);
+/* A deprecated synonym for archive_write_open_filename() */
+__LA_DECL int archive_write_open_file(struct archive *, const char *_file);
+__LA_DECL int archive_write_open_FILE(struct archive *, FILE *);
+/* _buffSize is the size of the buffer, _used refers to a variable that
+ * will be updated after each write into the buffer. */
+__LA_DECL int archive_write_open_memory(struct archive *,
+ void *_buffer, size_t _buffSize, size_t *_used);
+
+/*
+ * Note that the library will truncate writes beyond the size provided
+ * to archive_write_header or pad if the provided data is short.
+ */
+__LA_DECL int archive_write_header(struct archive *,
+ struct archive_entry *);
+#if ARCHIVE_VERSION_NUMBER < 2000000
+/* This was erroneously declared to return "int" in libarchive 1.x. */
+__LA_DECL int archive_write_data(struct archive *,
+ const void *, size_t);
+#else
+/* Libarchive 2.0 and later return ssize_t here. */
+__LA_DECL __LA_SSIZE_T archive_write_data(struct archive *,
+ const void *, size_t);
+#endif
+
+#if ARCHIVE_VERSION_NUMBER < 3000000
+/* Libarchive 1.x and 2.x use off_t for the argument, but that's not
+ * stable on Linux. */
+__LA_DECL __LA_SSIZE_T archive_write_data_block(struct archive *,
+ const void *, size_t, off_t);
+#else
+/* Libarchive 3.0 uses explicit int64_t to ensure consistent 64-bit support. */
+__LA_DECL __LA_SSIZE_T archive_write_data_block(struct archive *,
+ const void *, size_t, __LA_INT64_T);
+#endif
+__LA_DECL int archive_write_finish_entry(struct archive *);
+__LA_DECL int archive_write_close(struct archive *);
+#if ARCHIVE_VERSION_NUMBER < 2000000
+/* Return value was incorrect in libarchive 1.x. */
+__LA_DECL void archive_write_finish(struct archive *);
+#else
+/* Libarchive 2.x and later returns an error if this fails. */
+/* It can fail if the archive wasn't already closed, in which case
+ * archive_write_finish() will implicitly call archive_write_close(). */
+__LA_DECL int archive_write_finish(struct archive *);
+#endif
+
+/*
+ * Set write options.
+ */
+/* Apply option string to the format only. */
+__LA_DECL int archive_write_set_format_options(struct archive *_a,
+ const char *s);
+/* Apply option string to the compressor only. */
+__LA_DECL int archive_write_set_compressor_options(struct archive *_a,
+ const char *s);
+/* Apply option string to both the format and the compressor. */
+__LA_DECL int archive_write_set_options(struct archive *_a,
+ const char *s);
+
+
+/*-
+ * ARCHIVE_WRITE_DISK API
+ *
+ * To create objects on disk:
+ * 1) Ask archive_write_disk_new for a new archive_write_disk object.
+ * 2) Set any global properties. In particular, you probably
+ * want to set the options.
+ * 3) For each entry:
+ * - construct an appropriate struct archive_entry structure
+ * - archive_write_header to create the file/dir/etc on disk
+ * - archive_write_data to write the entry data
+ * 4) archive_write_finish to cleanup the writer and release resources
+ *
+ * In particular, you can use this in conjunction with archive_read()
+ * to pull entries out of an archive and create them on disk.
+ */
+__LA_DECL struct archive *archive_write_disk_new(void);
+/* This file will not be overwritten. */
+__LA_DECL int archive_write_disk_set_skip_file(struct archive *,
+ dev_t, ino_t);
+/* Set flags to control how the next item gets created.
+ * This accepts a bitmask of ARCHIVE_EXTRACT_XXX flags defined above. */
+__LA_DECL int archive_write_disk_set_options(struct archive *,
+ int flags);
+/*
+ * The lookup functions are given uname/uid (or gname/gid) pairs and
+ * return a uid (gid) suitable for this system. These are used for
+ * restoring ownership and for setting ACLs. The default functions
+ * are naive, they just return the uid/gid. These are small, so reasonable
+ * for applications that don't need to preserve ownership; they
+ * are probably also appropriate for applications that are doing
+ * same-system backup and restore.
+ */
+/*
+ * The "standard" lookup functions use common system calls to lookup
+ * the uname/gname, falling back to the uid/gid if the names can't be
+ * found. They cache lookups and are reasonably fast, but can be very
+ * large, so they are not used unless you ask for them. In
+ * particular, these match the specifications of POSIX "pax" and old
+ * POSIX "tar".
+ */
+__LA_DECL int archive_write_disk_set_standard_lookup(struct archive *);
+/*
+ * If neither the default (naive) nor the standard (big) functions suit
+ * your needs, you can write your own and register them. Be sure to
+ * include a cleanup function if you have allocated private data.
+ */
+__LA_DECL int archive_write_disk_set_group_lookup(struct archive *,
+ void * /* private_data */,
+ __LA_GID_T (*)(void *, const char *, __LA_GID_T),
+ void (* /* cleanup */)(void *));
+__LA_DECL int archive_write_disk_set_user_lookup(struct archive *,
+ void * /* private_data */,
+ __LA_UID_T (*)(void *, const char *, __LA_UID_T),
+ void (* /* cleanup */)(void *));
+
+/*
+ * ARCHIVE_READ_DISK API
+ *
+ * This is still evolving and somewhat experimental.
+ */
+__LA_DECL struct archive *archive_read_disk_new(void);
+/* The names for symlink modes here correspond to an old BSD
+ * command-line argument convention: -L, -P, -H */
+/* Follow all symlinks. */
+__LA_DECL int archive_read_disk_set_symlink_logical(struct archive *);
+/* Follow no symlinks. */
+__LA_DECL int archive_read_disk_set_symlink_physical(struct archive *);
+/* Follow symlink initially, then not. */
+__LA_DECL int archive_read_disk_set_symlink_hybrid(struct archive *);
+/* TODO: Handle Linux stat32/stat64 ugliness. <sigh> */
+__LA_DECL int archive_read_disk_entry_from_file(struct archive *,
+ struct archive_entry *, int /* fd */, const struct stat *);
+/* Look up gname for gid or uname for uid. */
+/* Default implementations are very, very stupid. */
+__LA_DECL const char *archive_read_disk_gname(struct archive *, __LA_GID_T);
+__LA_DECL const char *archive_read_disk_uname(struct archive *, __LA_UID_T);
+/* "Standard" implementation uses getpwuid_r, getgrgid_r and caches the
+ * results for performance. */
+__LA_DECL int archive_read_disk_set_standard_lookup(struct archive *);
+/* You can install your own lookups if you like. */
+__LA_DECL int archive_read_disk_set_gname_lookup(struct archive *,
+ void * /* private_data */,
+ const char *(* /* lookup_fn */)(void *, __LA_GID_T),
+ void (* /* cleanup_fn */)(void *));
+__LA_DECL int archive_read_disk_set_uname_lookup(struct archive *,
+ void * /* private_data */,
+ const char *(* /* lookup_fn */)(void *, __LA_UID_T),
+ void (* /* cleanup_fn */)(void *));
+
+/*
+ * Accessor functions to read/set various information in
+ * the struct archive object:
+ */
+/* Bytes written after compression or read before decompression. */
+__LA_DECL __LA_INT64_T archive_position_compressed(struct archive *);
+/* Bytes written to compressor or read from decompressor. */
+__LA_DECL __LA_INT64_T archive_position_uncompressed(struct archive *);
+
+__LA_DECL const char *archive_compression_name(struct archive *);
+__LA_DECL int archive_compression(struct archive *);
+__LA_DECL int archive_errno(struct archive *);
+__LA_DECL const char *archive_error_string(struct archive *);
+__LA_DECL const char *archive_format_name(struct archive *);
+__LA_DECL int archive_format(struct archive *);
+__LA_DECL void archive_clear_error(struct archive *);
+__LA_DECL void archive_set_error(struct archive *, int _err,
+ const char *fmt, ...) __LA_PRINTF(3, 4);
+__LA_DECL void archive_copy_error(struct archive *dest,
+ struct archive *src);
+__LA_DECL int archive_file_count(struct archive *);
+
+#ifdef __cplusplus
+}
+#endif
+
+/* These are meaningless outside of this header. */
+#undef __LA_DECL
+#undef __LA_GID_T
+#undef __LA_UID_T
+
+/* These need to remain defined because they're used in the
+ * callback type definitions. XXX Fix this. This is ugly. XXX */
+/* #undef __LA_INT64_T */
+/* #undef __LA_SSIZE_T */
+
+#endif /* !ARCHIVE_H_INCLUDED */
diff --git a/libarchive/archive_check_magic.c b/libarchive/archive_check_magic.c
new file mode 100644
index 000000000000..e27e5d827089
--- /dev/null
+++ b/libarchive/archive_check_magic.c
@@ -0,0 +1,134 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_check_magic.c 201089 2009-12-28 02:20:23Z kientzle $");
+
+#ifdef HAVE_SYS_TYPES_H
+#include <sys/types.h>
+#endif
+
+#include <stdio.h>
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+#if defined(_WIN32) && !defined(__CYGWIN__)
+#include <windows.h>
+#include <winbase.h>
+#endif
+
+#include "archive_private.h"
+
+static void
+errmsg(const char *m)
+{
+ size_t s = strlen(m);
+ ssize_t written;
+
+ while (s > 0) {
+ written = write(2, m, strlen(m));
+ if (written <= 0)
+ return;
+ m += written;
+ s -= written;
+ }
+}
+
+static void
+diediedie(void)
+{
+#if defined(_WIN32) && !defined(__CYGWIN__) && defined(_DEBUG)
+ /* Cause a breakpoint exception */
+ DebugBreak();
+#endif
+ abort(); /* Terminate the program abnormally. */
+}
+
+static const char *
+state_name(unsigned s)
+{
+ switch (s) {
+ case ARCHIVE_STATE_NEW: return ("new");
+ case ARCHIVE_STATE_HEADER: return ("header");
+ case ARCHIVE_STATE_DATA: return ("data");
+ case ARCHIVE_STATE_EOF: return ("eof");
+ case ARCHIVE_STATE_CLOSED: return ("closed");
+ case ARCHIVE_STATE_FATAL: return ("fatal");
+ default: return ("??");
+ }
+}
+
+
+static void
+write_all_states(unsigned int states)
+{
+ unsigned int lowbit;
+
+ /* A trick for computing the lowest set bit. */
+ while ((lowbit = states & (1 + ~states)) != 0) {
+ states &= ~lowbit; /* Clear the low bit. */
+ errmsg(state_name(lowbit));
+ if (states != 0)
+ errmsg("/");
+ }
+}
+
+/*
+ * Check magic value and current state; bail if it isn't valid.
+ *
+ * This is designed to catch serious programming errors that violate
+ * the libarchive API.
+ */
+void
+__archive_check_magic(struct archive *a, unsigned int magic,
+ unsigned int state, const char *function)
+{
+ if (a->magic != magic) {
+ errmsg("INTERNAL ERROR: Function ");
+ errmsg(function);
+ errmsg(" invoked with invalid struct archive structure.\n");
+ diediedie();
+ }
+
+ if (state == ARCHIVE_STATE_ANY)
+ return;
+
+ if ((a->state & state) == 0) {
+ errmsg("INTERNAL ERROR: Function '");
+ errmsg(function);
+ errmsg("' invoked with archive structure in state '");
+ write_all_states(a->state);
+ errmsg("', should be in state '");
+ write_all_states(state);
+ errmsg("'\n");
+ diediedie();
+ }
+}
diff --git a/libarchive/archive_crc32.h b/libarchive/archive_crc32.h
new file mode 100644
index 000000000000..103e5df35c15
--- /dev/null
+++ b/libarchive/archive_crc32.h
@@ -0,0 +1,66 @@
+/*-
+ * Copyright (c) 2009 Joerg Sonnenberger
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD: head/lib/libarchive/archive_crc32.h 201102 2009-12-28 03:11:36Z kientzle $
+ */
+
+#ifndef __LIBARCHIVE_BUILD
+#error This header is only to be used internally to libarchive.
+#endif
+
+/*
+ * When zlib is unavailable, we should still be able to validate
+ * uncompressed zip archives. That requires us to be able to compute
+ * the CRC32 check value. This is a drop-in compatible replacement
+ * for crc32() from zlib. It's slower than the zlib implementation,
+ * but still pretty fast: This runs about 300MB/s on my 3GHz P4
+ * compared to about 800MB/s for the zlib implementation.
+ */
+static unsigned long
+crc32(unsigned long crc, const void *_p, size_t len)
+{
+ unsigned long crc2, b, i;
+ const unsigned char *p = _p;
+ static volatile int crc_tbl_inited = 0;
+ static unsigned long crc_tbl[256];
+
+ if (!crc_tbl_inited) {
+ for (b = 0; b < 256; ++b) {
+ crc2 = b;
+ for (i = 8; i > 0; --i) {
+ if (crc2 & 1)
+ crc2 = (crc2 >> 1) ^ 0xedb88320UL;
+ else
+ crc2 = (crc2 >> 1);
+ }
+ crc_tbl[b] = crc2;
+ }
+ crc_tbl_inited = 1;
+ }
+
+ crc = crc ^ 0xffffffffUL;
+ while (len--)
+ crc = crc_tbl[(crc ^ *p++) & 0xff] ^ (crc >> 8);
+ return (crc ^ 0xffffffffUL);
+}
diff --git a/libarchive/archive_endian.h b/libarchive/archive_endian.h
new file mode 100644
index 000000000000..edc90ee6a840
--- /dev/null
+++ b/libarchive/archive_endian.h
@@ -0,0 +1,162 @@
+/*-
+ * Copyright (c) 2002 Thomas Moestl <tmm@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD: head/lib/libarchive/archive_endian.h 201085 2009-12-28 02:17:15Z kientzle $
+ *
+ * Borrowed from FreeBSD's <sys/endian.h>
+ */
+
+#ifndef __LIBARCHIVE_BUILD
+#error This header is only to be used internally to libarchive.
+#endif
+
+/* Note: This is a purely internal header! */
+/* Do not use this outside of libarchive internal code! */
+
+#ifndef ARCHIVE_ENDIAN_H_INCLUDED
+#define ARCHIVE_ENDIAN_H_INCLUDED
+
+
+/*
+ * Disabling inline keyword for compilers known to choke on it:
+ * - Watcom C++ in C code. (For any version?)
+ * - SGI MIPSpro
+ * - Microsoft Visual C++ 6.0 (supposedly newer versions too)
+ */
+#if defined(__WATCOMC__) || defined(__sgi) || defined(__hpux) || defined(__BORLANDC__)
+#define inline
+#elif defined(_MSC_VER)
+#define inline __inline
+#endif
+
+/* Alignment-agnostic encode/decode bytestream to/from little/big endian. */
+
+static inline uint16_t
+archive_be16dec(const void *pp)
+{
+ unsigned char const *p = (unsigned char const *)pp;
+
+ return ((p[0] << 8) | p[1]);
+}
+
+static inline uint32_t
+archive_be32dec(const void *pp)
+{
+ unsigned char const *p = (unsigned char const *)pp;
+
+ return ((p[0] << 24) | (p[1] << 16) | (p[2] << 8) | p[3]);
+}
+
+static inline uint64_t
+archive_be64dec(const void *pp)
+{
+ unsigned char const *p = (unsigned char const *)pp;
+
+ return (((uint64_t)archive_be32dec(p) << 32) | archive_be32dec(p + 4));
+}
+
+static inline uint16_t
+archive_le16dec(const void *pp)
+{
+ unsigned char const *p = (unsigned char const *)pp;
+
+ return ((p[1] << 8) | p[0]);
+}
+
+static inline uint32_t
+archive_le32dec(const void *pp)
+{
+ unsigned char const *p = (unsigned char const *)pp;
+
+ return ((p[3] << 24) | (p[2] << 16) | (p[1] << 8) | p[0]);
+}
+
+static inline uint64_t
+archive_le64dec(const void *pp)
+{
+ unsigned char const *p = (unsigned char const *)pp;
+
+ return (((uint64_t)archive_le32dec(p + 4) << 32) | archive_le32dec(p));
+}
+
+static inline void
+archive_be16enc(void *pp, uint16_t u)
+{
+ unsigned char *p = (unsigned char *)pp;
+
+ p[0] = (u >> 8) & 0xff;
+ p[1] = u & 0xff;
+}
+
+static inline void
+archive_be32enc(void *pp, uint32_t u)
+{
+ unsigned char *p = (unsigned char *)pp;
+
+ p[0] = (u >> 24) & 0xff;
+ p[1] = (u >> 16) & 0xff;
+ p[2] = (u >> 8) & 0xff;
+ p[3] = u & 0xff;
+}
+
+static inline void
+archive_be64enc(void *pp, uint64_t u)
+{
+ unsigned char *p = (unsigned char *)pp;
+
+ archive_be32enc(p, u >> 32);
+ archive_be32enc(p + 4, u & 0xffffffff);
+}
+
+static inline void
+archive_le16enc(void *pp, uint16_t u)
+{
+ unsigned char *p = (unsigned char *)pp;
+
+ p[0] = u & 0xff;
+ p[1] = (u >> 8) & 0xff;
+}
+
+static inline void
+archive_le32enc(void *pp, uint32_t u)
+{
+ unsigned char *p = (unsigned char *)pp;
+
+ p[0] = u & 0xff;
+ p[1] = (u >> 8) & 0xff;
+ p[2] = (u >> 16) & 0xff;
+ p[3] = (u >> 24) & 0xff;
+}
+
+static inline void
+archive_le64enc(void *pp, uint64_t u)
+{
+ unsigned char *p = (unsigned char *)pp;
+
+ archive_le32enc(p, u & 0xffffffff);
+ archive_le32enc(p + 4, u >> 32);
+}
+
+#endif
diff --git a/libarchive/archive_entry.3 b/libarchive/archive_entry.3
new file mode 100644
index 000000000000..9ceb18b7aad6
--- /dev/null
+++ b/libarchive/archive_entry.3
@@ -0,0 +1,433 @@
+.\" Copyright (c) 2003-2007 Tim Kientzle
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\" notice, this list of conditions and the following disclaimer.
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\" notice, this list of conditions and the following disclaimer in the
+.\" documentation and/or other materials provided with the distribution.
+.\"
+.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+.\" SUCH DAMAGE.
+.\"
+.\" $FreeBSD: src/lib/libarchive/archive_entry.3,v 1.18 2008/05/26 17:00:22 kientzle Exp $
+.\"
+.Dd May 12, 2008
+.Dt archive_entry 3
+.Os
+.Sh NAME
+.Nm archive_entry_acl_add_entry ,
+.Nm archive_entry_acl_add_entry_w ,
+.Nm archive_entry_acl_clear ,
+.Nm archive_entry_acl_count ,
+.Nm archive_entry_acl_next ,
+.Nm archive_entry_acl_next_w ,
+.Nm archive_entry_acl_reset ,
+.Nm archive_entry_acl_text_w ,
+.Nm archive_entry_atime ,
+.Nm archive_entry_atime_nsec ,
+.Nm archive_entry_clear ,
+.Nm archive_entry_clone ,
+.Nm archive_entry_copy_fflags_text ,
+.Nm archive_entry_copy_fflags_text_w ,
+.Nm archive_entry_copy_gname ,
+.Nm archive_entry_copy_gname_w ,
+.Nm archive_entry_copy_hardlink ,
+.Nm archive_entry_copy_hardlink_w ,
+.Nm archive_entry_copy_link ,
+.Nm archive_entry_copy_link_w ,
+.Nm archive_entry_copy_pathname_w ,
+.Nm archive_entry_copy_sourcepath ,
+.Nm archive_entry_copy_stat ,
+.Nm archive_entry_copy_symlink ,
+.Nm archive_entry_copy_symlink_w ,
+.Nm archive_entry_copy_uname ,
+.Nm archive_entry_copy_uname_w ,
+.Nm archive_entry_dev ,
+.Nm archive_entry_devmajor ,
+.Nm archive_entry_devminor ,
+.Nm archive_entry_filetype ,
+.Nm archive_entry_fflags ,
+.Nm archive_entry_fflags_text ,
+.Nm archive_entry_free ,
+.Nm archive_entry_gid ,
+.Nm archive_entry_gname ,
+.Nm archive_entry_hardlink ,
+.Nm archive_entry_ino ,
+.Nm archive_entry_mode ,
+.Nm archive_entry_mtime ,
+.Nm archive_entry_mtime_nsec ,
+.Nm archive_entry_nlink ,
+.Nm archive_entry_new ,
+.Nm archive_entry_pathname ,
+.Nm archive_entry_pathname_w ,
+.Nm archive_entry_rdev ,
+.Nm archive_entry_rdevmajor ,
+.Nm archive_entry_rdevminor ,
+.Nm archive_entry_set_atime ,
+.Nm archive_entry_set_ctime ,
+.Nm archive_entry_set_dev ,
+.Nm archive_entry_set_devmajor ,
+.Nm archive_entry_set_devminor ,
+.Nm archive_entry_set_filetype ,
+.Nm archive_entry_set_fflags ,
+.Nm archive_entry_set_gid ,
+.Nm archive_entry_set_gname ,
+.Nm archive_entry_set_hardlink ,
+.Nm archive_entry_set_link ,
+.Nm archive_entry_set_mode ,
+.Nm archive_entry_set_mtime ,
+.Nm archive_entry_set_pathname ,
+.Nm archive_entry_set_rdevmajor ,
+.Nm archive_entry_set_rdevminor ,
+.Nm archive_entry_set_size ,
+.Nm archive_entry_set_symlink ,
+.Nm archive_entry_set_uid ,
+.Nm archive_entry_set_uname ,
+.Nm archive_entry_size ,
+.Nm archive_entry_sourcepath ,
+.Nm archive_entry_stat ,
+.Nm archive_entry_symlink ,
+.Nm archive_entry_uid ,
+.Nm archive_entry_uname
+.Nd functions for manipulating archive entry descriptions
+.Sh SYNOPSIS
+.In archive_entry.h
+.Ft void
+.Fo archive_entry_acl_add_entry
+.Fa "struct archive_entry *"
+.Fa "int type"
+.Fa "int permset"
+.Fa "int tag"
+.Fa "int qual"
+.Fa "const char *name"
+.Fc
+.Ft void
+.Fo archive_entry_acl_add_entry_w
+.Fa "struct archive_entry *"
+.Fa "int type"
+.Fa "int permset"
+.Fa "int tag"
+.Fa "int qual"
+.Fa "const wchar_t *name"
+.Fc
+.Ft void
+.Fn archive_entry_acl_clear "struct archive_entry *"
+.Ft int
+.Fn archive_entry_acl_count "struct archive_entry *" "int type"
+.Ft int
+.Fo archive_entry_acl_next
+.Fa "struct archive_entry *"
+.Fa "int want_type"
+.Fa "int *type"
+.Fa "int *permset"
+.Fa "int *tag"
+.Fa "int *qual"
+.Fa "const char **name"
+.Fc
+.Ft int
+.Fo archive_entry_acl_next_w
+.Fa "struct archive_entry *"
+.Fa "int want_type"
+.Fa "int *type"
+.Fa "int *permset"
+.Fa "int *tag"
+.Fa "int *qual"
+.Fa "const wchar_t **name"
+.Fc
+.Ft int
+.Fn archive_entry_acl_reset "struct archive_entry *" "int want_type"
+.Ft const wchar_t *
+.Fn archive_entry_acl_text_w "struct archive_entry *" "int flags"
+.Ft time_t
+.Fn archive_entry_atime "struct archive_entry *"
+.Ft long
+.Fn archive_entry_atime_nsec "struct archive_entry *"
+.Ft "struct archive_entry *"
+.Fn archive_entry_clear "struct archive_entry *"
+.Ft struct archive_entry *
+.Fn archive_entry_clone "struct archive_entry *"
+.Ft const char * *
+.Fn archive_entry_copy_fflags_text_w "struct archive_entry *" "const char *"
+.Ft const wchar_t *
+.Fn archive_entry_copy_fflags_text_w "struct archive_entry *" "const wchar_t *"
+.Ft void
+.Fn archive_entry_copy_gname "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_copy_gname_w "struct archive_entry *" "const wchar_t *"
+.Ft void
+.Fn archive_entry_copy_hardlink "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_copy_hardlink_w "struct archive_entry *" "const wchar_t *"
+.Ft void
+.Fn archive_entry_copy_sourcepath "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_copy_pathname_w "struct archive_entry *" "const wchar_t *"
+.Ft void
+.Fn archive_entry_copy_stat "struct archive_entry *" "const struct stat *"
+.Ft void
+.Fn archive_entry_copy_symlink "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_copy_symlink_w "struct archive_entry *" "const wchar_t *"
+.Ft void
+.Fn archive_entry_copy_uname "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_copy_uname_w "struct archive_entry *" "const wchar_t *"
+.Ft dev_t
+.Fn archive_entry_dev "struct archive_entry *"
+.Ft dev_t
+.Fn archive_entry_devmajor "struct archive_entry *"
+.Ft dev_t
+.Fn archive_entry_devminor "struct archive_entry *"
+.Ft mode_t
+.Fn archive_entry_filetype "struct archive_entry *"
+.Ft void
+.Fo archive_entry_fflags
+.Fa "struct archive_entry *"
+.Fa "unsigned long *set"
+.Fa "unsigned long *clear"
+.Fc
+.Ft const char *
+.Fn archive_entry_fflags_text "struct archive_entry *"
+.Ft void
+.Fn archive_entry_free "struct archive_entry *"
+.Ft const char *
+.Fn archive_entry_gname "struct archive_entry *"
+.Ft const char *
+.Fn archive_entry_hardlink "struct archive_entry *"
+.Ft ino_t
+.Fn archive_entry_ino "struct archive_entry *"
+.Ft mode_t
+.Fn archive_entry_mode "struct archive_entry *"
+.Ft time_t
+.Fn archive_entry_mtime "struct archive_entry *"
+.Ft long
+.Fn archive_entry_mtime_nsec "struct archive_entry *"
+.Ft unsigned int
+.Fn archive_entry_nlink "struct archive_entry *"
+.Ft struct archive_entry *
+.Fn archive_entry_new "void"
+.Ft const char *
+.Fn archive_entry_pathname "struct archive_entry *"
+.Ft const wchar_t *
+.Fn archive_entry_pathname_w "struct archive_entry *"
+.Ft dev_t
+.Fn archive_entry_rdev "struct archive_entry *"
+.Ft dev_t
+.Fn archive_entry_rdevmajor "struct archive_entry *"
+.Ft dev_t
+.Fn archive_entry_rdevminor "struct archive_entry *"
+.Ft void
+.Fn archive_entry_set_dev "struct archive_entry *" "dev_t"
+.Ft void
+.Fn archive_entry_set_devmajor "struct archive_entry *" "dev_t"
+.Ft void
+.Fn archive_entry_set_devminor "struct archive_entry *" "dev_t"
+.Ft void
+.Fn archive_entry_set_filetype "struct archive_entry *" "unsigned int"
+.Ft void
+.Fo archive_entry_set_fflags
+.Fa "struct archive_entry *"
+.Fa "unsigned long set"
+.Fa "unsigned long clear"
+.Fc
+.Ft void
+.Fn archive_entry_set_gid "struct archive_entry *" "gid_t"
+.Ft void
+.Fn archive_entry_set_gname "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_set_hardlink "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_set_ino "struct archive_entry *" "unsigned long"
+.Ft void
+.Fn archive_entry_set_link "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_set_mode "struct archive_entry *" "mode_t"
+.Ft void
+.Fn archive_entry_set_mtime "struct archive_entry *" "time_t" "long nanos"
+.Ft void
+.Fn archive_entry_set_nlink "struct archive_entry *" "unsigned int"
+.Ft void
+.Fn archive_entry_set_pathname "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_set_rdev "struct archive_entry *" "dev_t"
+.Ft void
+.Fn archive_entry_set_rdevmajor "struct archive_entry *" "dev_t"
+.Ft void
+.Fn archive_entry_set_rdevminor "struct archive_entry *" "dev_t"
+.Ft void
+.Fn archive_entry_set_size "struct archive_entry *" "int64_t"
+.Ft void
+.Fn archive_entry_set_symlink "struct archive_entry *" "const char *"
+.Ft void
+.Fn archive_entry_set_uid "struct archive_entry *" "uid_t"
+.Ft void
+.Fn archive_entry_set_uname "struct archive_entry *" "const char *"
+.Ft int64_t
+.Fn archive_entry_size "struct archive_entry *"
+.Ft const char *
+.Fn archive_entry_sourcepath "struct archive_entry *"
+.Ft const struct stat *
+.Fn archive_entry_stat "struct archive_entry *"
+.Ft const char *
+.Fn archive_entry_symlink "struct archive_entry *"
+.Ft const char *
+.Fn archive_entry_uname "struct archive_entry *"
+.Sh DESCRIPTION
+These functions create and manipulate data objects that
+represent entries within an archive.
+You can think of a
+.Tn struct archive_entry
+as a heavy-duty version of
+.Tn struct stat :
+it includes everything from
+.Tn struct stat
+plus associated pathname, textual group and user names, etc.
+These objects are used by
+.Xr libarchive 3
+to represent the metadata associated with a particular
+entry in an archive.
+.Ss Create and Destroy
+There are functions to allocate, destroy, clear, and copy
+.Va archive_entry
+objects:
+.Bl -tag -compact -width indent
+.It Fn archive_entry_clear
+Erases the object, resetting all internal fields to the
+same state as a newly-created object.
+This is provided to allow you to quickly recycle objects
+without thrashing the heap.
+.It Fn archive_entry_clone
+A deep copy operation; all text fields are duplicated.
+.It Fn archive_entry_free
+Releases the
+.Tn struct archive_entry
+object.
+.It Fn archive_entry_new
+Allocate and return a blank
+.Tn struct archive_entry
+object.
+.El
+.Ss Set and Get Functions
+Most of the functions here set or read entries in an object.
+Such functions have one of the following forms:
+.Bl -tag -compact -width indent
+.It Fn archive_entry_set_XXXX
+Stores the provided data in the object.
+In particular, for strings, the pointer is stored,
+not the referenced string.
+.It Fn archive_entry_copy_XXXX
+As above, except that the referenced data is copied
+into the object.
+.It Fn archive_entry_XXXX
+Returns the specified data.
+In the case of strings, a const-qualified pointer to
+the string is returned.
+.El
+String data can be set or accessed as wide character strings
+or normal
+.Va char
+strings.
+The functions that use wide character strings are suffixed with
+.Cm _w .
+Note that these are different representations of the same data:
+For example, if you store a narrow string and read the corresponding
+wide string, the object will transparently convert formats
+using the current locale.
+Similarly, if you store a wide string and then store a
+narrow string for the same data, the previously-set wide string will
+be discarded in favor of the new data.
+.Pp
+There are a few set/get functions that merit additional description:
+.Bl -tag -compact -width indent
+.It Fn archive_entry_set_link
+This function sets the symlink field if it is already set.
+Otherwise, it sets the hardlink field.
+.El
+.Ss File Flags
+File flags are transparently converted between a bitmap
+representation and a textual format.
+For example, if you set the bitmap and ask for text, the library
+will build a canonical text format.
+However, if you set a text format and request a text format,
+you will get back the same text, even if it is ill-formed.
+If you need to canonicalize a textual flags string, you should first set the
+text form, then request the bitmap form, then use that to set the bitmap form.
+Setting the bitmap format will clear the internal text representation
+and force it to be reconstructed when you next request the text form.
+.Pp
+The bitmap format consists of two integers, one containing bits
+that should be set, the other specifying bits that should be
+cleared.
+Bits not mentioned in either bitmap will be ignored.
+Usually, the bitmap of bits to be cleared will be set to zero.
+In unusual circumstances, you can force a fully-specified set
+of file flags by setting the bitmap of flags to clear to the complement
+of the bitmap of flags to set.
+(This differs from
+.Xr fflagstostr 3 ,
+which only includes names for set bits.)
+Converting a bitmap to a textual string is a platform-specific
+operation; bits that are not meaningful on the current platform
+will be ignored.
+.Pp
+The canonical text format is a comma-separated list of flag names.
+The
+.Fn archive_entry_copy_fflags_text
+and
+.Fn archive_entry_copy_fflags_text_w
+functions parse the provided text and sets the internal bitmap values.
+This is a platform-specific operation; names that are not meaningful
+on the current platform will be ignored.
+The function returns a pointer to the start of the first name that was not
+recognized, or NULL if every name was recognized.
+Note that every name--including names that follow an unrecognized name--will
+be evaluated, and the bitmaps will be set to reflect every name that is
+recognized.
+(In particular, this differs from
+.Xr strtofflags 3 ,
+which stops parsing at the first unrecognized name.)
+.Ss ACL Handling
+XXX This needs serious help.
+XXX
+.Pp
+An
+.Dq Access Control List
+(ACL) is a list of permissions that grant access to particular users or
+groups beyond what would normally be provided by standard POSIX mode bits.
+The ACL handling here addresses some deficiencies in the POSIX.1e draft 17 ACL
+specification.
+In particular, POSIX.1e draft 17 specifies several different formats, but
+none of those formats include both textual user/group names and numeric
+UIDs/GIDs.
+.Pp
+XXX explain ACL stuff XXX
+.\" .Sh EXAMPLE
+.\" .Sh RETURN VALUES
+.\" .Sh ERRORS
+.Sh SEE ALSO
+.Xr archive 3
+.Sh HISTORY
+The
+.Nm libarchive
+library first appeared in
+.Fx 5.3 .
+.Sh AUTHORS
+.An -nosplit
+The
+.Nm libarchive
+library was written by
+.An Tim Kientzle Aq kientzle@acm.org .
+.\" .Sh BUGS
diff --git a/libarchive/archive_entry.c b/libarchive/archive_entry.c
new file mode 100644
index 000000000000..f734b8c4f51d
--- /dev/null
+++ b/libarchive/archive_entry.c
@@ -0,0 +1,2202 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_entry.c 201096 2009-12-28 02:41:27Z kientzle $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_SYS_TYPES_H
+#include <sys/types.h>
+#endif
+#if MAJOR_IN_MKDEV
+#include <sys/mkdev.h>
+#define HAVE_MAJOR
+#elif MAJOR_IN_SYSMACROS
+#include <sys/sysmacros.h>
+#define HAVE_MAJOR
+#endif
+#ifdef HAVE_LIMITS_H
+#include <limits.h>
+#endif
+#ifdef HAVE_LINUX_FS_H
+#include <linux/fs.h> /* for Linux file flags */
+#endif
+/*
+ * Some Linux distributions have both linux/ext2_fs.h and ext2fs/ext2_fs.h.
+ * As the include guards don't agree, the order of include is important.
+ */
+#ifdef HAVE_LINUX_EXT2_FS_H
+#include <linux/ext2_fs.h> /* for Linux file flags */
+#endif
+#if defined(HAVE_EXT2FS_EXT2_FS_H) && !defined(__CYGWIN__)
+#include <ext2fs/ext2_fs.h> /* for Linux file flags */
+#endif
+#include <stddef.h>
+#include <stdio.h>
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_WCHAR_H
+#include <wchar.h>
+#endif
+
+#include "archive.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_entry_private.h"
+
+#undef max
+#define max(a, b) ((a)>(b)?(a):(b))
+
+#if !defined(HAVE_MAJOR) && !defined(major)
+/* Replacement for major/minor/makedev. */
+#define major(x) ((int)(0x00ff & ((x) >> 8)))
+#define minor(x) ((int)(0xffff00ff & (x)))
+#define makedev(maj,min) ((0xff00 & ((maj)<<8)) | (0xffff00ff & (min)))
+#endif
+
+/* Play games to come up with a suitable makedev() definition. */
+#ifdef __QNXNTO__
+/* QNX. <sigh> */
+#include <sys/netmgr.h>
+#define ae_makedev(maj, min) makedev(ND_LOCAL_NODE, (maj), (min))
+#elif defined makedev
+/* There's a "makedev" macro. */
+#define ae_makedev(maj, min) makedev((maj), (min))
+#elif defined mkdev || ((defined _WIN32 || defined __WIN32__) && !defined(__CYGWIN__))
+/* Windows. <sigh> */
+#define ae_makedev(maj, min) mkdev((maj), (min))
+#else
+/* There's a "makedev" function. */
+#define ae_makedev(maj, min) makedev((maj), (min))
+#endif
+
+static void aes_clean(struct aes *);
+static void aes_copy(struct aes *dest, struct aes *src);
+static const char * aes_get_mbs(struct aes *);
+static const wchar_t * aes_get_wcs(struct aes *);
+static int aes_set_mbs(struct aes *, const char *mbs);
+static int aes_copy_mbs(struct aes *, const char *mbs);
+/* static void aes_set_wcs(struct aes *, const wchar_t *wcs); */
+static int aes_copy_wcs(struct aes *, const wchar_t *wcs);
+static int aes_copy_wcs_len(struct aes *, const wchar_t *wcs, size_t);
+
+static char * ae_fflagstostr(unsigned long bitset, unsigned long bitclear);
+static const wchar_t *ae_wcstofflags(const wchar_t *stringp,
+ unsigned long *setp, unsigned long *clrp);
+static const char *ae_strtofflags(const char *stringp,
+ unsigned long *setp, unsigned long *clrp);
+static void append_entry_w(wchar_t **wp, const wchar_t *prefix, int tag,
+ const wchar_t *wname, int perm, int id);
+static void append_id_w(wchar_t **wp, int id);
+
+static int acl_special(struct archive_entry *entry,
+ int type, int permset, int tag);
+static struct ae_acl *acl_new_entry(struct archive_entry *entry,
+ int type, int permset, int tag, int id);
+static int isint_w(const wchar_t *start, const wchar_t *end, int *result);
+static int ismode_w(const wchar_t *start, const wchar_t *end, int *result);
+static void next_field_w(const wchar_t **wp, const wchar_t **start,
+ const wchar_t **end, wchar_t *sep);
+static int prefix_w(const wchar_t *start, const wchar_t *end,
+ const wchar_t *test);
+static void
+archive_entry_acl_add_entry_w_len(struct archive_entry *entry, int type,
+ int permset, int tag, int id, const wchar_t *name, size_t);
+
+
+#ifndef HAVE_WCSCPY
+static wchar_t * wcscpy(wchar_t *s1, const wchar_t *s2)
+{
+ wchar_t *dest = s1;
+ while ((*s1 = *s2) != L'\0')
+ ++s1, ++s2;
+ return dest;
+}
+#endif
+#ifndef HAVE_WCSLEN
+static size_t wcslen(const wchar_t *s)
+{
+ const wchar_t *p = s;
+ while (*p != L'\0')
+ ++p;
+ return p - s;
+}
+#endif
+#ifndef HAVE_WMEMCMP
+/* Good enough for simple equality testing, but not for sorting. */
+#define wmemcmp(a,b,i) memcmp((a), (b), (i) * sizeof(wchar_t))
+#endif
+#ifndef HAVE_WMEMCPY
+#define wmemcpy(a,b,i) (wchar_t *)memcpy((a), (b), (i) * sizeof(wchar_t))
+#endif
+
+static void
+aes_clean(struct aes *aes)
+{
+ if (aes->aes_wcs) {
+ free((wchar_t *)(uintptr_t)aes->aes_wcs);
+ aes->aes_wcs = NULL;
+ }
+ archive_string_free(&(aes->aes_mbs));
+ archive_string_free(&(aes->aes_utf8));
+ aes->aes_set = 0;
+}
+
+static void
+aes_copy(struct aes *dest, struct aes *src)
+{
+ wchar_t *wp;
+
+ dest->aes_set = src->aes_set;
+ archive_string_copy(&(dest->aes_mbs), &(src->aes_mbs));
+ archive_string_copy(&(dest->aes_utf8), &(src->aes_utf8));
+
+ if (src->aes_wcs != NULL) {
+ wp = (wchar_t *)malloc((wcslen(src->aes_wcs) + 1)
+ * sizeof(wchar_t));
+ if (wp == NULL)
+ __archive_errx(1, "No memory for aes_copy()");
+ wcscpy(wp, src->aes_wcs);
+ dest->aes_wcs = wp;
+ }
+}
+
+static const char *
+aes_get_utf8(struct aes *aes)
+{
+ if (aes->aes_set & AES_SET_UTF8)
+ return (aes->aes_utf8.s);
+ if ((aes->aes_set & AES_SET_WCS)
+ && archive_strappend_w_utf8(&(aes->aes_utf8), aes->aes_wcs) != NULL) {
+ aes->aes_set |= AES_SET_UTF8;
+ return (aes->aes_utf8.s);
+ }
+ return (NULL);
+}
+
+static const char *
+aes_get_mbs(struct aes *aes)
+{
+ /* If we already have an MBS form, return that immediately. */
+ if (aes->aes_set & AES_SET_MBS)
+ return (aes->aes_mbs.s);
+ /* If there's a WCS form, try converting with the native locale. */
+ if ((aes->aes_set & AES_SET_WCS)
+ && archive_strappend_w_mbs(&(aes->aes_mbs), aes->aes_wcs) != NULL) {
+ aes->aes_set |= AES_SET_MBS;
+ return (aes->aes_mbs.s);
+ }
+ /* We'll use UTF-8 for MBS if all else fails. */
+ return (aes_get_utf8(aes));
+}
+
+static const wchar_t *
+aes_get_wcs(struct aes *aes)
+{
+ wchar_t *w;
+ size_t r;
+
+ /* Return WCS form if we already have it. */
+ if (aes->aes_set & AES_SET_WCS)
+ return (aes->aes_wcs);
+
+ if (aes->aes_set & AES_SET_MBS) {
+ /* Try converting MBS to WCS using native locale. */
+ /*
+ * No single byte will be more than one wide character,
+ * so this length estimate will always be big enough.
+ */
+ size_t wcs_length = aes->aes_mbs.length;
+
+ w = (wchar_t *)malloc((wcs_length + 1) * sizeof(wchar_t));
+ if (w == NULL)
+ __archive_errx(1, "No memory for aes_get_wcs()");
+ r = mbstowcs(w, aes->aes_mbs.s, wcs_length);
+ if (r != (size_t)-1 && r != 0) {
+ w[r] = 0;
+ aes->aes_set |= AES_SET_WCS;
+ return (aes->aes_wcs = w);
+ }
+ free(w);
+ }
+
+ if (aes->aes_set & AES_SET_UTF8) {
+ /* Try converting UTF8 to WCS. */
+ aes->aes_wcs = __archive_string_utf8_w(&(aes->aes_utf8));
+ if (aes->aes_wcs != NULL)
+ aes->aes_set |= AES_SET_WCS;
+ return (aes->aes_wcs);
+ }
+ return (NULL);
+}
+
+static int
+aes_set_mbs(struct aes *aes, const char *mbs)
+{
+ return (aes_copy_mbs(aes, mbs));
+}
+
+static int
+aes_copy_mbs(struct aes *aes, const char *mbs)
+{
+ if (mbs == NULL) {
+ aes->aes_set = 0;
+ return (0);
+ }
+ aes->aes_set = AES_SET_MBS; /* Only MBS form is set now. */
+ archive_strcpy(&(aes->aes_mbs), mbs);
+ archive_string_empty(&(aes->aes_utf8));
+ if (aes->aes_wcs) {
+ free((wchar_t *)(uintptr_t)aes->aes_wcs);
+ aes->aes_wcs = NULL;
+ }
+ return (0);
+}
+
+/*
+ * The 'update' form tries to proactively update all forms of
+ * this string (WCS and MBS) and returns an error if any of
+ * them fail. This is used by the 'pax' handler, for instance,
+ * to detect and report character-conversion failures early while
+ * still allowing clients to get potentially useful values from
+ * the more tolerant lazy conversions. (get_mbs and get_wcs will
+ * strive to give the user something useful, so you can get hopefully
+ * usable values even if some of the character conversions are failing.)
+ */
+static int
+aes_update_utf8(struct aes *aes, const char *utf8)
+{
+ if (utf8 == NULL) {
+ aes->aes_set = 0;
+ return (1); /* Succeeded in clearing everything. */
+ }
+
+ /* Save the UTF8 string. */
+ archive_strcpy(&(aes->aes_utf8), utf8);
+
+ /* Empty the mbs and wcs strings. */
+ archive_string_empty(&(aes->aes_mbs));
+ if (aes->aes_wcs) {
+ free((wchar_t *)(uintptr_t)aes->aes_wcs);
+ aes->aes_wcs = NULL;
+ }
+
+ aes->aes_set = AES_SET_UTF8; /* Only UTF8 is set now. */
+
+ /* TODO: We should just do a direct UTF-8 to MBS conversion
+ * here. That would be faster, use less space, and give the
+ * same information. (If a UTF-8 to MBS conversion succeeds,
+ * then UTF-8->WCS and Unicode->MBS conversions will both
+ * succeed.) */
+
+ /* Try converting UTF8 to WCS, return false on failure. */
+ aes->aes_wcs = __archive_string_utf8_w(&(aes->aes_utf8));
+ if (aes->aes_wcs == NULL)
+ return (0);
+ aes->aes_set = AES_SET_UTF8 | AES_SET_WCS; /* Both UTF8 and WCS set. */
+
+ /* Try converting WCS to MBS, return false on failure. */
+ if (archive_strappend_w_mbs(&(aes->aes_mbs), aes->aes_wcs) == NULL)
+ return (0);
+ aes->aes_set = AES_SET_UTF8 | AES_SET_WCS | AES_SET_MBS;
+
+ /* All conversions succeeded. */
+ return (1);
+}
+
+static int
+aes_copy_wcs(struct aes *aes, const wchar_t *wcs)
+{
+ return aes_copy_wcs_len(aes, wcs, wcs == NULL ? 0 : wcslen(wcs));
+}
+
+static int
+aes_copy_wcs_len(struct aes *aes, const wchar_t *wcs, size_t len)
+{
+ wchar_t *w;
+
+ if (wcs == NULL) {
+ aes->aes_set = 0;
+ return (0);
+ }
+ aes->aes_set = AES_SET_WCS; /* Only WCS form set. */
+ archive_string_empty(&(aes->aes_mbs));
+ archive_string_empty(&(aes->aes_utf8));
+ if (aes->aes_wcs) {
+ free((wchar_t *)(uintptr_t)aes->aes_wcs);
+ aes->aes_wcs = NULL;
+ }
+ w = (wchar_t *)malloc((len + 1) * sizeof(wchar_t));
+ if (w == NULL)
+ __archive_errx(1, "No memory for aes_copy_wcs()");
+ wmemcpy(w, wcs, len);
+ w[len] = L'\0';
+ aes->aes_wcs = w;
+ return (0);
+}
+
+/****************************************************************************
+ *
+ * Public Interface
+ *
+ ****************************************************************************/
+
+struct archive_entry *
+archive_entry_clear(struct archive_entry *entry)
+{
+ if (entry == NULL)
+ return (NULL);
+ aes_clean(&entry->ae_fflags_text);
+ aes_clean(&entry->ae_gname);
+ aes_clean(&entry->ae_hardlink);
+ aes_clean(&entry->ae_pathname);
+ aes_clean(&entry->ae_sourcepath);
+ aes_clean(&entry->ae_symlink);
+ aes_clean(&entry->ae_uname);
+ archive_entry_acl_clear(entry);
+ archive_entry_xattr_clear(entry);
+ free(entry->stat);
+ memset(entry, 0, sizeof(*entry));
+ return entry;
+}
+
+struct archive_entry *
+archive_entry_clone(struct archive_entry *entry)
+{
+ struct archive_entry *entry2;
+ struct ae_acl *ap, *ap2;
+ struct ae_xattr *xp;
+
+ /* Allocate new structure and copy over all of the fields. */
+ entry2 = (struct archive_entry *)malloc(sizeof(*entry2));
+ if (entry2 == NULL)
+ return (NULL);
+ memset(entry2, 0, sizeof(*entry2));
+ entry2->ae_stat = entry->ae_stat;
+ entry2->ae_fflags_set = entry->ae_fflags_set;
+ entry2->ae_fflags_clear = entry->ae_fflags_clear;
+
+ aes_copy(&entry2->ae_fflags_text, &entry->ae_fflags_text);
+ aes_copy(&entry2->ae_gname, &entry->ae_gname);
+ aes_copy(&entry2->ae_hardlink, &entry->ae_hardlink);
+ aes_copy(&entry2->ae_pathname, &entry->ae_pathname);
+ aes_copy(&entry2->ae_sourcepath, &entry->ae_sourcepath);
+ aes_copy(&entry2->ae_symlink, &entry->ae_symlink);
+ entry2->ae_set = entry->ae_set;
+ aes_copy(&entry2->ae_uname, &entry->ae_uname);
+
+ /* Copy ACL data over. */
+ ap = entry->acl_head;
+ while (ap != NULL) {
+ ap2 = acl_new_entry(entry2,
+ ap->type, ap->permset, ap->tag, ap->id);
+ if (ap2 != NULL)
+ aes_copy(&ap2->name, &ap->name);
+ ap = ap->next;
+ }
+
+ /* Copy xattr data over. */
+ xp = entry->xattr_head;
+ while (xp != NULL) {
+ archive_entry_xattr_add_entry(entry2,
+ xp->name, xp->value, xp->size);
+ xp = xp->next;
+ }
+
+ return (entry2);
+}
+
+void
+archive_entry_free(struct archive_entry *entry)
+{
+ archive_entry_clear(entry);
+ free(entry);
+}
+
+struct archive_entry *
+archive_entry_new(void)
+{
+ struct archive_entry *entry;
+
+ entry = (struct archive_entry *)malloc(sizeof(*entry));
+ if (entry == NULL)
+ return (NULL);
+ memset(entry, 0, sizeof(*entry));
+ return (entry);
+}
+
+/*
+ * Functions for reading fields from an archive_entry.
+ */
+
+time_t
+archive_entry_atime(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_atime);
+}
+
+long
+archive_entry_atime_nsec(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_atime_nsec);
+}
+
+int
+archive_entry_atime_is_set(struct archive_entry *entry)
+{
+ return (entry->ae_set & AE_SET_ATIME);
+}
+
+time_t
+archive_entry_birthtime(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_birthtime);
+}
+
+long
+archive_entry_birthtime_nsec(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_birthtime_nsec);
+}
+
+int
+archive_entry_birthtime_is_set(struct archive_entry *entry)
+{
+ return (entry->ae_set & AE_SET_BIRTHTIME);
+}
+
+time_t
+archive_entry_ctime(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_ctime);
+}
+
+int
+archive_entry_ctime_is_set(struct archive_entry *entry)
+{
+ return (entry->ae_set & AE_SET_CTIME);
+}
+
+long
+archive_entry_ctime_nsec(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_ctime_nsec);
+}
+
+dev_t
+archive_entry_dev(struct archive_entry *entry)
+{
+ if (entry->ae_stat.aest_dev_is_broken_down)
+ return ae_makedev(entry->ae_stat.aest_devmajor,
+ entry->ae_stat.aest_devminor);
+ else
+ return (entry->ae_stat.aest_dev);
+}
+
+dev_t
+archive_entry_devmajor(struct archive_entry *entry)
+{
+ if (entry->ae_stat.aest_dev_is_broken_down)
+ return (entry->ae_stat.aest_devmajor);
+ else
+ return major(entry->ae_stat.aest_dev);
+}
+
+dev_t
+archive_entry_devminor(struct archive_entry *entry)
+{
+ if (entry->ae_stat.aest_dev_is_broken_down)
+ return (entry->ae_stat.aest_devminor);
+ else
+ return minor(entry->ae_stat.aest_dev);
+}
+
+mode_t
+archive_entry_filetype(struct archive_entry *entry)
+{
+ return (AE_IFMT & entry->ae_stat.aest_mode);
+}
+
+void
+archive_entry_fflags(struct archive_entry *entry,
+ unsigned long *set, unsigned long *clear)
+{
+ *set = entry->ae_fflags_set;
+ *clear = entry->ae_fflags_clear;
+}
+
+/*
+ * Note: if text was provided, this just returns that text. If you
+ * really need the text to be rebuilt in a canonical form, set the
+ * text, ask for the bitmaps, then set the bitmaps. (Setting the
+ * bitmaps clears any stored text.) This design is deliberate: if
+ * we're editing archives, we don't want to discard flags just because
+ * they aren't supported on the current system. The bitmap<->text
+ * conversions are platform-specific (see below).
+ */
+const char *
+archive_entry_fflags_text(struct archive_entry *entry)
+{
+ const char *f;
+ char *p;
+
+ f = aes_get_mbs(&entry->ae_fflags_text);
+ if (f != NULL)
+ return (f);
+
+ if (entry->ae_fflags_set == 0 && entry->ae_fflags_clear == 0)
+ return (NULL);
+
+ p = ae_fflagstostr(entry->ae_fflags_set, entry->ae_fflags_clear);
+ if (p == NULL)
+ return (NULL);
+
+ aes_copy_mbs(&entry->ae_fflags_text, p);
+ free(p);
+ f = aes_get_mbs(&entry->ae_fflags_text);
+ return (f);
+}
+
+gid_t
+archive_entry_gid(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_gid);
+}
+
+const char *
+archive_entry_gname(struct archive_entry *entry)
+{
+ return (aes_get_mbs(&entry->ae_gname));
+}
+
+const wchar_t *
+archive_entry_gname_w(struct archive_entry *entry)
+{
+ return (aes_get_wcs(&entry->ae_gname));
+}
+
+const char *
+archive_entry_hardlink(struct archive_entry *entry)
+{
+ if (entry->ae_set & AE_SET_HARDLINK)
+ return (aes_get_mbs(&entry->ae_hardlink));
+ return (NULL);
+}
+
+const wchar_t *
+archive_entry_hardlink_w(struct archive_entry *entry)
+{
+ if (entry->ae_set & AE_SET_HARDLINK)
+ return (aes_get_wcs(&entry->ae_hardlink));
+ return (NULL);
+}
+
+ino_t
+archive_entry_ino(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_ino);
+}
+
+int64_t
+archive_entry_ino64(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_ino);
+}
+
+mode_t
+archive_entry_mode(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_mode);
+}
+
+time_t
+archive_entry_mtime(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_mtime);
+}
+
+long
+archive_entry_mtime_nsec(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_mtime_nsec);
+}
+
+int
+archive_entry_mtime_is_set(struct archive_entry *entry)
+{
+ return (entry->ae_set & AE_SET_MTIME);
+}
+
+unsigned int
+archive_entry_nlink(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_nlink);
+}
+
+const char *
+archive_entry_pathname(struct archive_entry *entry)
+{
+ return (aes_get_mbs(&entry->ae_pathname));
+}
+
+const wchar_t *
+archive_entry_pathname_w(struct archive_entry *entry)
+{
+ return (aes_get_wcs(&entry->ae_pathname));
+}
+
+dev_t
+archive_entry_rdev(struct archive_entry *entry)
+{
+ if (entry->ae_stat.aest_rdev_is_broken_down)
+ return ae_makedev(entry->ae_stat.aest_rdevmajor,
+ entry->ae_stat.aest_rdevminor);
+ else
+ return (entry->ae_stat.aest_rdev);
+}
+
+dev_t
+archive_entry_rdevmajor(struct archive_entry *entry)
+{
+ if (entry->ae_stat.aest_rdev_is_broken_down)
+ return (entry->ae_stat.aest_rdevmajor);
+ else
+ return major(entry->ae_stat.aest_rdev);
+}
+
+dev_t
+archive_entry_rdevminor(struct archive_entry *entry)
+{
+ if (entry->ae_stat.aest_rdev_is_broken_down)
+ return (entry->ae_stat.aest_rdevminor);
+ else
+ return minor(entry->ae_stat.aest_rdev);
+}
+
+int64_t
+archive_entry_size(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_size);
+}
+
+int
+archive_entry_size_is_set(struct archive_entry *entry)
+{
+ return (entry->ae_set & AE_SET_SIZE);
+}
+
+const char *
+archive_entry_sourcepath(struct archive_entry *entry)
+{
+ return (aes_get_mbs(&entry->ae_sourcepath));
+}
+
+const char *
+archive_entry_symlink(struct archive_entry *entry)
+{
+ if (entry->ae_set & AE_SET_SYMLINK)
+ return (aes_get_mbs(&entry->ae_symlink));
+ return (NULL);
+}
+
+const wchar_t *
+archive_entry_symlink_w(struct archive_entry *entry)
+{
+ if (entry->ae_set & AE_SET_SYMLINK)
+ return (aes_get_wcs(&entry->ae_symlink));
+ return (NULL);
+}
+
+uid_t
+archive_entry_uid(struct archive_entry *entry)
+{
+ return (entry->ae_stat.aest_uid);
+}
+
+const char *
+archive_entry_uname(struct archive_entry *entry)
+{
+ return (aes_get_mbs(&entry->ae_uname));
+}
+
+const wchar_t *
+archive_entry_uname_w(struct archive_entry *entry)
+{
+ return (aes_get_wcs(&entry->ae_uname));
+}
+
+/*
+ * Functions to set archive_entry properties.
+ */
+
+void
+archive_entry_set_filetype(struct archive_entry *entry, unsigned int type)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_mode &= ~AE_IFMT;
+ entry->ae_stat.aest_mode |= AE_IFMT & type;
+}
+
+void
+archive_entry_set_fflags(struct archive_entry *entry,
+ unsigned long set, unsigned long clear)
+{
+ aes_clean(&entry->ae_fflags_text);
+ entry->ae_fflags_set = set;
+ entry->ae_fflags_clear = clear;
+}
+
+const char *
+archive_entry_copy_fflags_text(struct archive_entry *entry,
+ const char *flags)
+{
+ aes_copy_mbs(&entry->ae_fflags_text, flags);
+ return (ae_strtofflags(flags,
+ &entry->ae_fflags_set, &entry->ae_fflags_clear));
+}
+
+const wchar_t *
+archive_entry_copy_fflags_text_w(struct archive_entry *entry,
+ const wchar_t *flags)
+{
+ aes_copy_wcs(&entry->ae_fflags_text, flags);
+ return (ae_wcstofflags(flags,
+ &entry->ae_fflags_set, &entry->ae_fflags_clear));
+}
+
+void
+archive_entry_set_gid(struct archive_entry *entry, gid_t g)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_gid = g;
+}
+
+void
+archive_entry_set_gname(struct archive_entry *entry, const char *name)
+{
+ aes_set_mbs(&entry->ae_gname, name);
+}
+
+void
+archive_entry_copy_gname(struct archive_entry *entry, const char *name)
+{
+ aes_copy_mbs(&entry->ae_gname, name);
+}
+
+void
+archive_entry_copy_gname_w(struct archive_entry *entry, const wchar_t *name)
+{
+ aes_copy_wcs(&entry->ae_gname, name);
+}
+
+int
+archive_entry_update_gname_utf8(struct archive_entry *entry, const char *name)
+{
+ return (aes_update_utf8(&entry->ae_gname, name));
+}
+
+void
+archive_entry_set_ino(struct archive_entry *entry, unsigned long ino)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_ino = ino;
+}
+
+void
+archive_entry_set_ino64(struct archive_entry *entry, int64_t ino)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_ino = ino;
+}
+
+void
+archive_entry_set_hardlink(struct archive_entry *entry, const char *target)
+{
+ aes_set_mbs(&entry->ae_hardlink, target);
+ if (target != NULL)
+ entry->ae_set |= AE_SET_HARDLINK;
+ else
+ entry->ae_set &= ~AE_SET_HARDLINK;
+}
+
+void
+archive_entry_copy_hardlink(struct archive_entry *entry, const char *target)
+{
+ aes_copy_mbs(&entry->ae_hardlink, target);
+ if (target != NULL)
+ entry->ae_set |= AE_SET_HARDLINK;
+ else
+ entry->ae_set &= ~AE_SET_HARDLINK;
+}
+
+void
+archive_entry_copy_hardlink_w(struct archive_entry *entry, const wchar_t *target)
+{
+ aes_copy_wcs(&entry->ae_hardlink, target);
+ if (target != NULL)
+ entry->ae_set |= AE_SET_HARDLINK;
+ else
+ entry->ae_set &= ~AE_SET_HARDLINK;
+}
+
+int
+archive_entry_update_hardlink_utf8(struct archive_entry *entry, const char *target)
+{
+ if (target != NULL)
+ entry->ae_set |= AE_SET_HARDLINK;
+ else
+ entry->ae_set &= ~AE_SET_HARDLINK;
+ return (aes_update_utf8(&entry->ae_hardlink, target));
+}
+
+void
+archive_entry_set_atime(struct archive_entry *entry, time_t t, long ns)
+{
+ entry->stat_valid = 0;
+ entry->ae_set |= AE_SET_ATIME;
+ entry->ae_stat.aest_atime = t;
+ entry->ae_stat.aest_atime_nsec = ns;
+}
+
+void
+archive_entry_unset_atime(struct archive_entry *entry)
+{
+ archive_entry_set_atime(entry, 0, 0);
+ entry->ae_set &= ~AE_SET_ATIME;
+}
+
+void
+archive_entry_set_birthtime(struct archive_entry *entry, time_t m, long ns)
+{
+ entry->stat_valid = 0;
+ entry->ae_set |= AE_SET_BIRTHTIME;
+ entry->ae_stat.aest_birthtime = m;
+ entry->ae_stat.aest_birthtime_nsec = ns;
+}
+
+void
+archive_entry_unset_birthtime(struct archive_entry *entry)
+{
+ archive_entry_set_birthtime(entry, 0, 0);
+ entry->ae_set &= ~AE_SET_BIRTHTIME;
+}
+
+void
+archive_entry_set_ctime(struct archive_entry *entry, time_t t, long ns)
+{
+ entry->stat_valid = 0;
+ entry->ae_set |= AE_SET_CTIME;
+ entry->ae_stat.aest_ctime = t;
+ entry->ae_stat.aest_ctime_nsec = ns;
+}
+
+void
+archive_entry_unset_ctime(struct archive_entry *entry)
+{
+ archive_entry_set_ctime(entry, 0, 0);
+ entry->ae_set &= ~AE_SET_CTIME;
+}
+
+void
+archive_entry_set_dev(struct archive_entry *entry, dev_t d)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_dev_is_broken_down = 0;
+ entry->ae_stat.aest_dev = d;
+}
+
+void
+archive_entry_set_devmajor(struct archive_entry *entry, dev_t m)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_dev_is_broken_down = 1;
+ entry->ae_stat.aest_devmajor = m;
+}
+
+void
+archive_entry_set_devminor(struct archive_entry *entry, dev_t m)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_dev_is_broken_down = 1;
+ entry->ae_stat.aest_devminor = m;
+}
+
+/* Set symlink if symlink is already set, else set hardlink. */
+void
+archive_entry_set_link(struct archive_entry *entry, const char *target)
+{
+ if (entry->ae_set & AE_SET_SYMLINK)
+ aes_set_mbs(&entry->ae_symlink, target);
+ else
+ aes_set_mbs(&entry->ae_hardlink, target);
+}
+
+/* Set symlink if symlink is already set, else set hardlink. */
+void
+archive_entry_copy_link(struct archive_entry *entry, const char *target)
+{
+ if (entry->ae_set & AE_SET_SYMLINK)
+ aes_copy_mbs(&entry->ae_symlink, target);
+ else
+ aes_copy_mbs(&entry->ae_hardlink, target);
+}
+
+/* Set symlink if symlink is already set, else set hardlink. */
+void
+archive_entry_copy_link_w(struct archive_entry *entry, const wchar_t *target)
+{
+ if (entry->ae_set & AE_SET_SYMLINK)
+ aes_copy_wcs(&entry->ae_symlink, target);
+ else
+ aes_copy_wcs(&entry->ae_hardlink, target);
+}
+
+int
+archive_entry_update_link_utf8(struct archive_entry *entry, const char *target)
+{
+ if (entry->ae_set & AE_SET_SYMLINK)
+ return (aes_update_utf8(&entry->ae_symlink, target));
+ else
+ return (aes_update_utf8(&entry->ae_hardlink, target));
+}
+
+void
+archive_entry_set_mode(struct archive_entry *entry, mode_t m)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_mode = m;
+}
+
+void
+archive_entry_set_mtime(struct archive_entry *entry, time_t m, long ns)
+{
+ entry->stat_valid = 0;
+ entry->ae_set |= AE_SET_MTIME;
+ entry->ae_stat.aest_mtime = m;
+ entry->ae_stat.aest_mtime_nsec = ns;
+}
+
+void
+archive_entry_unset_mtime(struct archive_entry *entry)
+{
+ archive_entry_set_mtime(entry, 0, 0);
+ entry->ae_set &= ~AE_SET_MTIME;
+}
+
+void
+archive_entry_set_nlink(struct archive_entry *entry, unsigned int nlink)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_nlink = nlink;
+}
+
+void
+archive_entry_set_pathname(struct archive_entry *entry, const char *name)
+{
+ aes_set_mbs(&entry->ae_pathname, name);
+}
+
+void
+archive_entry_copy_pathname(struct archive_entry *entry, const char *name)
+{
+ aes_copy_mbs(&entry->ae_pathname, name);
+}
+
+void
+archive_entry_copy_pathname_w(struct archive_entry *entry, const wchar_t *name)
+{
+ aes_copy_wcs(&entry->ae_pathname, name);
+}
+
+int
+archive_entry_update_pathname_utf8(struct archive_entry *entry, const char *name)
+{
+ return (aes_update_utf8(&entry->ae_pathname, name));
+}
+
+void
+archive_entry_set_perm(struct archive_entry *entry, mode_t p)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_mode &= AE_IFMT;
+ entry->ae_stat.aest_mode |= ~AE_IFMT & p;
+}
+
+void
+archive_entry_set_rdev(struct archive_entry *entry, dev_t m)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_rdev = m;
+ entry->ae_stat.aest_rdev_is_broken_down = 0;
+}
+
+void
+archive_entry_set_rdevmajor(struct archive_entry *entry, dev_t m)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_rdev_is_broken_down = 1;
+ entry->ae_stat.aest_rdevmajor = m;
+}
+
+void
+archive_entry_set_rdevminor(struct archive_entry *entry, dev_t m)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_rdev_is_broken_down = 1;
+ entry->ae_stat.aest_rdevminor = m;
+}
+
+void
+archive_entry_set_size(struct archive_entry *entry, int64_t s)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_size = s;
+ entry->ae_set |= AE_SET_SIZE;
+}
+
+void
+archive_entry_unset_size(struct archive_entry *entry)
+{
+ archive_entry_set_size(entry, 0);
+ entry->ae_set &= ~AE_SET_SIZE;
+}
+
+void
+archive_entry_copy_sourcepath(struct archive_entry *entry, const char *path)
+{
+ aes_set_mbs(&entry->ae_sourcepath, path);
+}
+
+void
+archive_entry_set_symlink(struct archive_entry *entry, const char *linkname)
+{
+ aes_set_mbs(&entry->ae_symlink, linkname);
+ if (linkname != NULL)
+ entry->ae_set |= AE_SET_SYMLINK;
+ else
+ entry->ae_set &= ~AE_SET_SYMLINK;
+}
+
+void
+archive_entry_copy_symlink(struct archive_entry *entry, const char *linkname)
+{
+ aes_copy_mbs(&entry->ae_symlink, linkname);
+ if (linkname != NULL)
+ entry->ae_set |= AE_SET_SYMLINK;
+ else
+ entry->ae_set &= ~AE_SET_SYMLINK;
+}
+
+void
+archive_entry_copy_symlink_w(struct archive_entry *entry, const wchar_t *linkname)
+{
+ aes_copy_wcs(&entry->ae_symlink, linkname);
+ if (linkname != NULL)
+ entry->ae_set |= AE_SET_SYMLINK;
+ else
+ entry->ae_set &= ~AE_SET_SYMLINK;
+}
+
+int
+archive_entry_update_symlink_utf8(struct archive_entry *entry, const char *linkname)
+{
+ if (linkname != NULL)
+ entry->ae_set |= AE_SET_SYMLINK;
+ else
+ entry->ae_set &= ~AE_SET_SYMLINK;
+ return (aes_update_utf8(&entry->ae_symlink, linkname));
+}
+
+void
+archive_entry_set_uid(struct archive_entry *entry, uid_t u)
+{
+ entry->stat_valid = 0;
+ entry->ae_stat.aest_uid = u;
+}
+
+void
+archive_entry_set_uname(struct archive_entry *entry, const char *name)
+{
+ aes_set_mbs(&entry->ae_uname, name);
+}
+
+void
+archive_entry_copy_uname(struct archive_entry *entry, const char *name)
+{
+ aes_copy_mbs(&entry->ae_uname, name);
+}
+
+void
+archive_entry_copy_uname_w(struct archive_entry *entry, const wchar_t *name)
+{
+ aes_copy_wcs(&entry->ae_uname, name);
+}
+
+int
+archive_entry_update_uname_utf8(struct archive_entry *entry, const char *name)
+{
+ return (aes_update_utf8(&entry->ae_uname, name));
+}
+
+/*
+ * ACL management. The following would, of course, be a lot simpler
+ * if: 1) the last draft of POSIX.1e were a really thorough and
+ * complete standard that addressed the needs of ACL archiving and 2)
+ * everyone followed it faithfully. Alas, neither is true, so the
+ * following is a lot more complex than might seem necessary to the
+ * uninitiated.
+ */
+
+void
+archive_entry_acl_clear(struct archive_entry *entry)
+{
+ struct ae_acl *ap;
+
+ while (entry->acl_head != NULL) {
+ ap = entry->acl_head->next;
+ aes_clean(&entry->acl_head->name);
+ free(entry->acl_head);
+ entry->acl_head = ap;
+ }
+ if (entry->acl_text_w != NULL) {
+ free(entry->acl_text_w);
+ entry->acl_text_w = NULL;
+ }
+ entry->acl_p = NULL;
+ entry->acl_state = 0; /* Not counting. */
+}
+
+/*
+ * Add a single ACL entry to the internal list of ACL data.
+ */
+void
+archive_entry_acl_add_entry(struct archive_entry *entry,
+ int type, int permset, int tag, int id, const char *name)
+{
+ struct ae_acl *ap;
+
+ if (acl_special(entry, type, permset, tag) == 0)
+ return;
+ ap = acl_new_entry(entry, type, permset, tag, id);
+ if (ap == NULL) {
+ /* XXX Error XXX */
+ return;
+ }
+ if (name != NULL && *name != '\0')
+ aes_copy_mbs(&ap->name, name);
+ else
+ aes_clean(&ap->name);
+}
+
+/*
+ * As above, but with a wide-character name.
+ */
+void
+archive_entry_acl_add_entry_w(struct archive_entry *entry,
+ int type, int permset, int tag, int id, const wchar_t *name)
+{
+ archive_entry_acl_add_entry_w_len(entry, type, permset, tag, id, name, wcslen(name));
+}
+
+static void
+archive_entry_acl_add_entry_w_len(struct archive_entry *entry,
+ int type, int permset, int tag, int id, const wchar_t *name, size_t len)
+{
+ struct ae_acl *ap;
+
+ if (acl_special(entry, type, permset, tag) == 0)
+ return;
+ ap = acl_new_entry(entry, type, permset, tag, id);
+ if (ap == NULL) {
+ /* XXX Error XXX */
+ return;
+ }
+ if (name != NULL && *name != L'\0' && len > 0)
+ aes_copy_wcs_len(&ap->name, name, len);
+ else
+ aes_clean(&ap->name);
+}
+
+/*
+ * If this ACL entry is part of the standard POSIX permissions set,
+ * store the permissions in the stat structure and return zero.
+ */
+static int
+acl_special(struct archive_entry *entry, int type, int permset, int tag)
+{
+ if (type == ARCHIVE_ENTRY_ACL_TYPE_ACCESS) {
+ switch (tag) {
+ case ARCHIVE_ENTRY_ACL_USER_OBJ:
+ entry->ae_stat.aest_mode &= ~0700;
+ entry->ae_stat.aest_mode |= (permset & 7) << 6;
+ return (0);
+ case ARCHIVE_ENTRY_ACL_GROUP_OBJ:
+ entry->ae_stat.aest_mode &= ~0070;
+ entry->ae_stat.aest_mode |= (permset & 7) << 3;
+ return (0);
+ case ARCHIVE_ENTRY_ACL_OTHER:
+ entry->ae_stat.aest_mode &= ~0007;
+ entry->ae_stat.aest_mode |= permset & 7;
+ return (0);
+ }
+ }
+ return (1);
+}
+
+/*
+ * Allocate and populate a new ACL entry with everything but the
+ * name.
+ */
+static struct ae_acl *
+acl_new_entry(struct archive_entry *entry,
+ int type, int permset, int tag, int id)
+{
+ struct ae_acl *ap, *aq;
+
+ if (type != ARCHIVE_ENTRY_ACL_TYPE_ACCESS &&
+ type != ARCHIVE_ENTRY_ACL_TYPE_DEFAULT)
+ return (NULL);
+ if (entry->acl_text_w != NULL) {
+ free(entry->acl_text_w);
+ entry->acl_text_w = NULL;
+ }
+
+ /* XXX TODO: More sanity-checks on the arguments XXX */
+
+ /* If there's a matching entry already in the list, overwrite it. */
+ ap = entry->acl_head;
+ aq = NULL;
+ while (ap != NULL) {
+ if (ap->type == type && ap->tag == tag && ap->id == id) {
+ ap->permset = permset;
+ return (ap);
+ }
+ aq = ap;
+ ap = ap->next;
+ }
+
+ /* Add a new entry to the end of the list. */
+ ap = (struct ae_acl *)malloc(sizeof(*ap));
+ if (ap == NULL)
+ return (NULL);
+ memset(ap, 0, sizeof(*ap));
+ if (aq == NULL)
+ entry->acl_head = ap;
+ else
+ aq->next = ap;
+ ap->type = type;
+ ap->tag = tag;
+ ap->id = id;
+ ap->permset = permset;
+ return (ap);
+}
+
+/*
+ * Return a count of entries matching "want_type".
+ */
+int
+archive_entry_acl_count(struct archive_entry *entry, int want_type)
+{
+ int count;
+ struct ae_acl *ap;
+
+ count = 0;
+ ap = entry->acl_head;
+ while (ap != NULL) {
+ if ((ap->type & want_type) != 0)
+ count++;
+ ap = ap->next;
+ }
+
+ if (count > 0 && ((want_type & ARCHIVE_ENTRY_ACL_TYPE_ACCESS) != 0))
+ count += 3;
+ return (count);
+}
+
+/*
+ * Prepare for reading entries from the ACL data. Returns a count
+ * of entries matching "want_type", or zero if there are no
+ * non-extended ACL entries of that type.
+ */
+int
+archive_entry_acl_reset(struct archive_entry *entry, int want_type)
+{
+ int count, cutoff;
+
+ count = archive_entry_acl_count(entry, want_type);
+
+ /*
+ * If the only entries are the three standard ones,
+ * then don't return any ACL data. (In this case,
+ * client can just use chmod(2) to set permissions.)
+ */
+ if ((want_type & ARCHIVE_ENTRY_ACL_TYPE_ACCESS) != 0)
+ cutoff = 3;
+ else
+ cutoff = 0;
+
+ if (count > cutoff)
+ entry->acl_state = ARCHIVE_ENTRY_ACL_USER_OBJ;
+ else
+ entry->acl_state = 0;
+ entry->acl_p = entry->acl_head;
+ return (count);
+}
+
+/*
+ * Return the next ACL entry in the list. Fake entries for the
+ * standard permissions and include them in the returned list.
+ */
+
+int
+archive_entry_acl_next(struct archive_entry *entry, int want_type, int *type,
+ int *permset, int *tag, int *id, const char **name)
+{
+ *name = NULL;
+ *id = -1;
+
+ /*
+ * The acl_state is either zero (no entries available), -1
+ * (reading from list), or an entry type (retrieve that type
+ * from ae_stat.aest_mode).
+ */
+ if (entry->acl_state == 0)
+ return (ARCHIVE_WARN);
+
+ /* The first three access entries are special. */
+ if ((want_type & ARCHIVE_ENTRY_ACL_TYPE_ACCESS) != 0) {
+ switch (entry->acl_state) {
+ case ARCHIVE_ENTRY_ACL_USER_OBJ:
+ *permset = (entry->ae_stat.aest_mode >> 6) & 7;
+ *type = ARCHIVE_ENTRY_ACL_TYPE_ACCESS;
+ *tag = ARCHIVE_ENTRY_ACL_USER_OBJ;
+ entry->acl_state = ARCHIVE_ENTRY_ACL_GROUP_OBJ;
+ return (ARCHIVE_OK);
+ case ARCHIVE_ENTRY_ACL_GROUP_OBJ:
+ *permset = (entry->ae_stat.aest_mode >> 3) & 7;
+ *type = ARCHIVE_ENTRY_ACL_TYPE_ACCESS;
+ *tag = ARCHIVE_ENTRY_ACL_GROUP_OBJ;
+ entry->acl_state = ARCHIVE_ENTRY_ACL_OTHER;
+ return (ARCHIVE_OK);
+ case ARCHIVE_ENTRY_ACL_OTHER:
+ *permset = entry->ae_stat.aest_mode & 7;
+ *type = ARCHIVE_ENTRY_ACL_TYPE_ACCESS;
+ *tag = ARCHIVE_ENTRY_ACL_OTHER;
+ entry->acl_state = -1;
+ entry->acl_p = entry->acl_head;
+ return (ARCHIVE_OK);
+ default:
+ break;
+ }
+ }
+
+ while (entry->acl_p != NULL && (entry->acl_p->type & want_type) == 0)
+ entry->acl_p = entry->acl_p->next;
+ if (entry->acl_p == NULL) {
+ entry->acl_state = 0;
+ *type = 0;
+ *permset = 0;
+ *tag = 0;
+ *id = -1;
+ *name = NULL;
+ return (ARCHIVE_EOF); /* End of ACL entries. */
+ }
+ *type = entry->acl_p->type;
+ *permset = entry->acl_p->permset;
+ *tag = entry->acl_p->tag;
+ *id = entry->acl_p->id;
+ *name = aes_get_mbs(&entry->acl_p->name);
+ entry->acl_p = entry->acl_p->next;
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Generate a text version of the ACL. The flags parameter controls
+ * the style of the generated ACL.
+ */
+const wchar_t *
+archive_entry_acl_text_w(struct archive_entry *entry, int flags)
+{
+ int count;
+ size_t length;
+ const wchar_t *wname;
+ const wchar_t *prefix;
+ wchar_t separator;
+ struct ae_acl *ap;
+ int id;
+ wchar_t *wp;
+
+ if (entry->acl_text_w != NULL) {
+ free (entry->acl_text_w);
+ entry->acl_text_w = NULL;
+ }
+
+ separator = L',';
+ count = 0;
+ length = 0;
+ ap = entry->acl_head;
+ while (ap != NULL) {
+ if ((ap->type & flags) != 0) {
+ count++;
+ if ((flags & ARCHIVE_ENTRY_ACL_STYLE_MARK_DEFAULT) &&
+ (ap->type & ARCHIVE_ENTRY_ACL_TYPE_DEFAULT))
+ length += 8; /* "default:" */
+ length += 5; /* tag name */
+ length += 1; /* colon */
+ wname = aes_get_wcs(&ap->name);
+ if (wname != NULL)
+ length += wcslen(wname);
+ else
+ length += sizeof(uid_t) * 3 + 1;
+ length ++; /* colon */
+ length += 3; /* rwx */
+ length += 1; /* colon */
+ length += max(sizeof(uid_t), sizeof(gid_t)) * 3 + 1;
+ length ++; /* newline */
+ }
+ ap = ap->next;
+ }
+
+ if (count > 0 && ((flags & ARCHIVE_ENTRY_ACL_TYPE_ACCESS) != 0)) {
+ length += 10; /* "user::rwx\n" */
+ length += 11; /* "group::rwx\n" */
+ length += 11; /* "other::rwx\n" */
+ }
+
+ if (count == 0)
+ return (NULL);
+
+ /* Now, allocate the string and actually populate it. */
+ wp = entry->acl_text_w = (wchar_t *)malloc(length * sizeof(wchar_t));
+ if (wp == NULL)
+ __archive_errx(1, "No memory to generate the text version of the ACL");
+ count = 0;
+ if ((flags & ARCHIVE_ENTRY_ACL_TYPE_ACCESS) != 0) {
+ append_entry_w(&wp, NULL, ARCHIVE_ENTRY_ACL_USER_OBJ, NULL,
+ entry->ae_stat.aest_mode & 0700, -1);
+ *wp++ = ',';
+ append_entry_w(&wp, NULL, ARCHIVE_ENTRY_ACL_GROUP_OBJ, NULL,
+ entry->ae_stat.aest_mode & 0070, -1);
+ *wp++ = ',';
+ append_entry_w(&wp, NULL, ARCHIVE_ENTRY_ACL_OTHER, NULL,
+ entry->ae_stat.aest_mode & 0007, -1);
+ count += 3;
+
+ ap = entry->acl_head;
+ while (ap != NULL) {
+ if ((ap->type & ARCHIVE_ENTRY_ACL_TYPE_ACCESS) != 0) {
+ wname = aes_get_wcs(&ap->name);
+ *wp++ = separator;
+ if (flags & ARCHIVE_ENTRY_ACL_STYLE_EXTRA_ID)
+ id = ap->id;
+ else
+ id = -1;
+ append_entry_w(&wp, NULL, ap->tag, wname,
+ ap->permset, id);
+ count++;
+ }
+ ap = ap->next;
+ }
+ }
+
+
+ if ((flags & ARCHIVE_ENTRY_ACL_TYPE_DEFAULT) != 0) {
+ if (flags & ARCHIVE_ENTRY_ACL_STYLE_MARK_DEFAULT)
+ prefix = L"default:";
+ else
+ prefix = NULL;
+ ap = entry->acl_head;
+ count = 0;
+ while (ap != NULL) {
+ if ((ap->type & ARCHIVE_ENTRY_ACL_TYPE_DEFAULT) != 0) {
+ wname = aes_get_wcs(&ap->name);
+ if (count > 0)
+ *wp++ = separator;
+ if (flags & ARCHIVE_ENTRY_ACL_STYLE_EXTRA_ID)
+ id = ap->id;
+ else
+ id = -1;
+ append_entry_w(&wp, prefix, ap->tag,
+ wname, ap->permset, id);
+ count ++;
+ }
+ ap = ap->next;
+ }
+ }
+
+ return (entry->acl_text_w);
+}
+
+static void
+append_id_w(wchar_t **wp, int id)
+{
+ if (id < 0)
+ id = 0;
+ if (id > 9)
+ append_id_w(wp, id / 10);
+ *(*wp)++ = L"0123456789"[id % 10];
+}
+
+static void
+append_entry_w(wchar_t **wp, const wchar_t *prefix, int tag,
+ const wchar_t *wname, int perm, int id)
+{
+ if (prefix != NULL) {
+ wcscpy(*wp, prefix);
+ *wp += wcslen(*wp);
+ }
+ switch (tag) {
+ case ARCHIVE_ENTRY_ACL_USER_OBJ:
+ wname = NULL;
+ id = -1;
+ /* FALLTHROUGH */
+ case ARCHIVE_ENTRY_ACL_USER:
+ wcscpy(*wp, L"user");
+ break;
+ case ARCHIVE_ENTRY_ACL_GROUP_OBJ:
+ wname = NULL;
+ id = -1;
+ /* FALLTHROUGH */
+ case ARCHIVE_ENTRY_ACL_GROUP:
+ wcscpy(*wp, L"group");
+ break;
+ case ARCHIVE_ENTRY_ACL_MASK:
+ wcscpy(*wp, L"mask");
+ wname = NULL;
+ id = -1;
+ break;
+ case ARCHIVE_ENTRY_ACL_OTHER:
+ wcscpy(*wp, L"other");
+ wname = NULL;
+ id = -1;
+ break;
+ }
+ *wp += wcslen(*wp);
+ *(*wp)++ = L':';
+ if (wname != NULL) {
+ wcscpy(*wp, wname);
+ *wp += wcslen(*wp);
+ } else if (tag == ARCHIVE_ENTRY_ACL_USER
+ || tag == ARCHIVE_ENTRY_ACL_GROUP) {
+ append_id_w(wp, id);
+ id = -1;
+ }
+ *(*wp)++ = L':';
+ *(*wp)++ = (perm & 0444) ? L'r' : L'-';
+ *(*wp)++ = (perm & 0222) ? L'w' : L'-';
+ *(*wp)++ = (perm & 0111) ? L'x' : L'-';
+ if (id != -1) {
+ *(*wp)++ = L':';
+ append_id_w(wp, id);
+ }
+ **wp = L'\0';
+}
+
+/*
+ * Parse a textual ACL. This automatically recognizes and supports
+ * extensions described above. The 'type' argument is used to
+ * indicate the type that should be used for any entries not
+ * explicitly marked as "default:".
+ */
+int
+__archive_entry_acl_parse_w(struct archive_entry *entry,
+ const wchar_t *text, int default_type)
+{
+ struct {
+ const wchar_t *start;
+ const wchar_t *end;
+ } field[4], name;
+
+ int fields, n;
+ int type, tag, permset, id;
+ wchar_t sep;
+
+ while (text != NULL && *text != L'\0') {
+ /*
+ * Parse the fields out of the next entry,
+ * advance 'text' to start of next entry.
+ */
+ fields = 0;
+ do {
+ const wchar_t *start, *end;
+ next_field_w(&text, &start, &end, &sep);
+ if (fields < 4) {
+ field[fields].start = start;
+ field[fields].end = end;
+ }
+ ++fields;
+ } while (sep == L':');
+
+ /* Set remaining fields to blank. */
+ for (n = fields; n < 4; ++n)
+ field[n].start = field[n].end = NULL;
+
+ /* Check for a numeric ID in field 1 or 3. */
+ id = -1;
+ isint_w(field[1].start, field[1].end, &id);
+ /* Field 3 is optional. */
+ if (id == -1 && fields > 3)
+ isint_w(field[3].start, field[3].end, &id);
+
+ /*
+ * Solaris extension: "defaultuser::rwx" is the
+ * default ACL corresponding to "user::rwx", etc.
+ */
+ if (field[0].end - field[0].start > 7
+ && wmemcmp(field[0].start, L"default", 7) == 0) {
+ type = ARCHIVE_ENTRY_ACL_TYPE_DEFAULT;
+ field[0].start += 7;
+ } else
+ type = default_type;
+
+ name.start = name.end = NULL;
+ if (prefix_w(field[0].start, field[0].end, L"user")) {
+ if (!ismode_w(field[2].start, field[2].end, &permset))
+ return (ARCHIVE_WARN);
+ if (id != -1 || field[1].start < field[1].end) {
+ tag = ARCHIVE_ENTRY_ACL_USER;
+ name = field[1];
+ } else
+ tag = ARCHIVE_ENTRY_ACL_USER_OBJ;
+ } else if (prefix_w(field[0].start, field[0].end, L"group")) {
+ if (!ismode_w(field[2].start, field[2].end, &permset))
+ return (ARCHIVE_WARN);
+ if (id != -1 || field[1].start < field[1].end) {
+ tag = ARCHIVE_ENTRY_ACL_GROUP;
+ name = field[1];
+ } else
+ tag = ARCHIVE_ENTRY_ACL_GROUP_OBJ;
+ } else if (prefix_w(field[0].start, field[0].end, L"other")) {
+ if (fields == 2
+ && field[1].start < field[1].end
+ && ismode_w(field[1].start, field[1].end, &permset)) {
+ /* This is Solaris-style "other:rwx" */
+ } else if (fields == 3
+ && field[1].start == field[1].end
+ && field[2].start < field[2].end
+ && ismode_w(field[2].start, field[2].end, &permset)) {
+ /* This is FreeBSD-style "other::rwx" */
+ } else
+ return (ARCHIVE_WARN);
+ tag = ARCHIVE_ENTRY_ACL_OTHER;
+ } else if (prefix_w(field[0].start, field[0].end, L"mask")) {
+ if (fields == 2
+ && field[1].start < field[1].end
+ && ismode_w(field[1].start, field[1].end, &permset)) {
+ /* This is Solaris-style "mask:rwx" */
+ } else if (fields == 3
+ && field[1].start == field[1].end
+ && field[2].start < field[2].end
+ && ismode_w(field[2].start, field[2].end, &permset)) {
+ /* This is FreeBSD-style "mask::rwx" */
+ } else
+ return (ARCHIVE_WARN);
+ tag = ARCHIVE_ENTRY_ACL_MASK;
+ } else
+ return (ARCHIVE_WARN);
+
+ /* Add entry to the internal list. */
+ archive_entry_acl_add_entry_w_len(entry, type, permset,
+ tag, id, name.start, name.end - name.start);
+ }
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Parse a string to a positive decimal integer. Returns true if
+ * the string is non-empty and consists only of decimal digits,
+ * false otherwise.
+ */
+static int
+isint_w(const wchar_t *start, const wchar_t *end, int *result)
+{
+ int n = 0;
+ if (start >= end)
+ return (0);
+ while (start < end) {
+ if (*start < '0' || *start > '9')
+ return (0);
+ if (n > (INT_MAX / 10))
+ n = INT_MAX;
+ else {
+ n *= 10;
+ n += *start - '0';
+ }
+ start++;
+ }
+ *result = n;
+ return (1);
+}
+
+/*
+ * Parse a string as a mode field. Returns true if
+ * the string is non-empty and consists only of mode characters,
+ * false otherwise.
+ */
+static int
+ismode_w(const wchar_t *start, const wchar_t *end, int *permset)
+{
+ const wchar_t *p;
+
+ if (start >= end)
+ return (0);
+ p = start;
+ *permset = 0;
+ while (p < end) {
+ switch (*p++) {
+ case 'r': case 'R':
+ *permset |= ARCHIVE_ENTRY_ACL_READ;
+ break;
+ case 'w': case 'W':
+ *permset |= ARCHIVE_ENTRY_ACL_WRITE;
+ break;
+ case 'x': case 'X':
+ *permset |= ARCHIVE_ENTRY_ACL_EXECUTE;
+ break;
+ case '-':
+ break;
+ default:
+ return (0);
+ }
+ }
+ return (1);
+}
+
+/*
+ * Match "[:whitespace:]*(.*)[:whitespace:]*[:,\n]". *wp is updated
+ * to point to just after the separator. *start points to the first
+ * character of the matched text and *end just after the last
+ * character of the matched identifier. In particular *end - *start
+ * is the length of the field body, not including leading or trailing
+ * whitespace.
+ */
+static void
+next_field_w(const wchar_t **wp, const wchar_t **start,
+ const wchar_t **end, wchar_t *sep)
+{
+ /* Skip leading whitespace to find start of field. */
+ while (**wp == L' ' || **wp == L'\t' || **wp == L'\n') {
+ (*wp)++;
+ }
+ *start = *wp;
+
+ /* Scan for the separator. */
+ while (**wp != L'\0' && **wp != L',' && **wp != L':' &&
+ **wp != L'\n') {
+ (*wp)++;
+ }
+ *sep = **wp;
+
+ /* Trim trailing whitespace to locate end of field. */
+ *end = *wp - 1;
+ while (**end == L' ' || **end == L'\t' || **end == L'\n') {
+ (*end)--;
+ }
+ (*end)++;
+
+ /* Adjust scanner location. */
+ if (**wp != L'\0')
+ (*wp)++;
+}
+
+/*
+ * Return true if the characters [start...end) are a prefix of 'test'.
+ * This makes it easy to handle the obvious abbreviations: 'u' for 'user', etc.
+ */
+static int
+prefix_w(const wchar_t *start, const wchar_t *end, const wchar_t *test)
+{
+ if (start == end)
+ return (0);
+
+ if (*start++ != *test++)
+ return (0);
+
+ while (start < end && *start++ == *test++)
+ ;
+
+ if (start < end)
+ return (0);
+
+ return (1);
+}
+
+
+/*
+ * Following code is modified from UC Berkeley sources, and
+ * is subject to the following copyright notice.
+ */
+
+/*-
+ * Copyright (c) 1993
+ * The Regents of the University of California. All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 4. Neither the name of the University nor the names of its contributors
+ * may be used to endorse or promote products derived from this software
+ * without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+static struct flag {
+ const char *name;
+ const wchar_t *wname;
+ unsigned long set;
+ unsigned long clear;
+} flags[] = {
+ /* Preferred (shorter) names per flag first, all prefixed by "no" */
+#ifdef SF_APPEND
+ { "nosappnd", L"nosappnd", SF_APPEND, 0 },
+ { "nosappend", L"nosappend", SF_APPEND, 0 },
+#endif
+#ifdef EXT2_APPEND_FL /* 'a' */
+ { "nosappnd", L"nosappnd", EXT2_APPEND_FL, 0 },
+ { "nosappend", L"nosappend", EXT2_APPEND_FL, 0 },
+#endif
+#ifdef SF_ARCHIVED
+ { "noarch", L"noarch", SF_ARCHIVED, 0 },
+ { "noarchived", L"noarchived", SF_ARCHIVED, 0 },
+#endif
+#ifdef SF_IMMUTABLE
+ { "noschg", L"noschg", SF_IMMUTABLE, 0 },
+ { "noschange", L"noschange", SF_IMMUTABLE, 0 },
+ { "nosimmutable", L"nosimmutable", SF_IMMUTABLE, 0 },
+#endif
+#ifdef EXT2_IMMUTABLE_FL /* 'i' */
+ { "noschg", L"noschg", EXT2_IMMUTABLE_FL, 0 },
+ { "noschange", L"noschange", EXT2_IMMUTABLE_FL, 0 },
+ { "nosimmutable", L"nosimmutable", EXT2_IMMUTABLE_FL, 0 },
+#endif
+#ifdef SF_NOUNLINK
+ { "nosunlnk", L"nosunlnk", SF_NOUNLINK, 0 },
+ { "nosunlink", L"nosunlink", SF_NOUNLINK, 0 },
+#endif
+#ifdef SF_SNAPSHOT
+ { "nosnapshot", L"nosnapshot", SF_SNAPSHOT, 0 },
+#endif
+#ifdef UF_APPEND
+ { "nouappnd", L"nouappnd", UF_APPEND, 0 },
+ { "nouappend", L"nouappend", UF_APPEND, 0 },
+#endif
+#ifdef UF_IMMUTABLE
+ { "nouchg", L"nouchg", UF_IMMUTABLE, 0 },
+ { "nouchange", L"nouchange", UF_IMMUTABLE, 0 },
+ { "nouimmutable", L"nouimmutable", UF_IMMUTABLE, 0 },
+#endif
+#ifdef UF_NODUMP
+ { "nodump", L"nodump", 0, UF_NODUMP},
+#endif
+#ifdef EXT2_NODUMP_FL /* 'd' */
+ { "nodump", L"nodump", 0, EXT2_NODUMP_FL},
+#endif
+#ifdef UF_OPAQUE
+ { "noopaque", L"noopaque", UF_OPAQUE, 0 },
+#endif
+#ifdef UF_NOUNLINK
+ { "nouunlnk", L"nouunlnk", UF_NOUNLINK, 0 },
+ { "nouunlink", L"nouunlink", UF_NOUNLINK, 0 },
+#endif
+#ifdef EXT2_UNRM_FL
+ { "nouunlink", L"nouunlink", EXT2_UNRM_FL, 0},
+#endif
+
+#ifdef EXT2_BTREE_FL
+ { "nobtree", L"nobtree", EXT2_BTREE_FL, 0 },
+#endif
+
+#ifdef EXT2_ECOMPR_FL
+ { "nocomperr", L"nocomperr", EXT2_ECOMPR_FL, 0 },
+#endif
+
+#ifdef EXT2_COMPR_FL /* 'c' */
+ { "nocompress", L"nocompress", EXT2_COMPR_FL, 0 },
+#endif
+
+#ifdef EXT2_NOATIME_FL /* 'A' */
+ { "noatime", L"noatime", 0, EXT2_NOATIME_FL},
+#endif
+
+#ifdef EXT2_DIRTY_FL
+ { "nocompdirty",L"nocompdirty", EXT2_DIRTY_FL, 0},
+#endif
+
+#ifdef EXT2_COMPRBLK_FL
+#ifdef EXT2_NOCOMPR_FL
+ { "nocomprblk", L"nocomprblk", EXT2_COMPRBLK_FL, EXT2_NOCOMPR_FL},
+#else
+ { "nocomprblk", L"nocomprblk", EXT2_COMPRBLK_FL, 0},
+#endif
+#endif
+#ifdef EXT2_DIRSYNC_FL
+ { "nodirsync", L"nodirsync", EXT2_DIRSYNC_FL, 0},
+#endif
+#ifdef EXT2_INDEX_FL
+ { "nohashidx", L"nohashidx", EXT2_INDEX_FL, 0},
+#endif
+#ifdef EXT2_IMAGIC_FL
+ { "noimagic", L"noimagic", EXT2_IMAGIC_FL, 0},
+#endif
+#ifdef EXT3_JOURNAL_DATA_FL
+ { "nojournal", L"nojournal", EXT3_JOURNAL_DATA_FL, 0},
+#endif
+#ifdef EXT2_SECRM_FL
+ { "nosecuredeletion",L"nosecuredeletion",EXT2_SECRM_FL, 0},
+#endif
+#ifdef EXT2_SYNC_FL
+ { "nosync", L"nosync", EXT2_SYNC_FL, 0},
+#endif
+#ifdef EXT2_NOTAIL_FL
+ { "notail", L"notail", 0, EXT2_NOTAIL_FL},
+#endif
+#ifdef EXT2_TOPDIR_FL
+ { "notopdir", L"notopdir", EXT2_TOPDIR_FL, 0},
+#endif
+#ifdef EXT2_RESERVED_FL
+ { "noreserved", L"noreserved", EXT2_RESERVED_FL, 0},
+#endif
+
+ { NULL, NULL, 0, 0 }
+};
+
+/*
+ * fflagstostr --
+ * Convert file flags to a comma-separated string. If no flags
+ * are set, return the empty string.
+ */
+static char *
+ae_fflagstostr(unsigned long bitset, unsigned long bitclear)
+{
+ char *string, *dp;
+ const char *sp;
+ unsigned long bits;
+ struct flag *flag;
+ size_t length;
+
+ bits = bitset | bitclear;
+ length = 0;
+ for (flag = flags; flag->name != NULL; flag++)
+ if (bits & (flag->set | flag->clear)) {
+ length += strlen(flag->name) + 1;
+ bits &= ~(flag->set | flag->clear);
+ }
+
+ if (length == 0)
+ return (NULL);
+ string = (char *)malloc(length);
+ if (string == NULL)
+ return (NULL);
+
+ dp = string;
+ for (flag = flags; flag->name != NULL; flag++) {
+ if (bitset & flag->set || bitclear & flag->clear) {
+ sp = flag->name + 2;
+ } else if (bitset & flag->clear || bitclear & flag->set) {
+ sp = flag->name;
+ } else
+ continue;
+ bitset &= ~(flag->set | flag->clear);
+ bitclear &= ~(flag->set | flag->clear);
+ if (dp > string)
+ *dp++ = ',';
+ while ((*dp++ = *sp++) != '\0')
+ ;
+ dp--;
+ }
+
+ *dp = '\0';
+ return (string);
+}
+
+/*
+ * strtofflags --
+ * Take string of arguments and return file flags. This
+ * version works a little differently than strtofflags(3).
+ * In particular, it always tests every token, skipping any
+ * unrecognized tokens. It returns a pointer to the first
+ * unrecognized token, or NULL if every token was recognized.
+ * This version is also const-correct and does not modify the
+ * provided string.
+ */
+static const char *
+ae_strtofflags(const char *s, unsigned long *setp, unsigned long *clrp)
+{
+ const char *start, *end;
+ struct flag *flag;
+ unsigned long set, clear;
+ const char *failed;
+
+ set = clear = 0;
+ start = s;
+ failed = NULL;
+ /* Find start of first token. */
+ while (*start == '\t' || *start == ' ' || *start == ',')
+ start++;
+ while (*start != '\0') {
+ /* Locate end of token. */
+ end = start;
+ while (*end != '\0' && *end != '\t' &&
+ *end != ' ' && *end != ',')
+ end++;
+ for (flag = flags; flag->name != NULL; flag++) {
+ if (memcmp(start, flag->name, end - start) == 0) {
+ /* Matched "noXXXX", so reverse the sense. */
+ clear |= flag->set;
+ set |= flag->clear;
+ break;
+ } else if (memcmp(start, flag->name + 2, end - start)
+ == 0) {
+ /* Matched "XXXX", so don't reverse. */
+ set |= flag->set;
+ clear |= flag->clear;
+ break;
+ }
+ }
+ /* Ignore unknown flag names. */
+ if (flag->name == NULL && failed == NULL)
+ failed = start;
+
+ /* Find start of next token. */
+ start = end;
+ while (*start == '\t' || *start == ' ' || *start == ',')
+ start++;
+
+ }
+
+ if (setp)
+ *setp = set;
+ if (clrp)
+ *clrp = clear;
+
+ /* Return location of first failure. */
+ return (failed);
+}
+
+/*
+ * wcstofflags --
+ * Take string of arguments and return file flags. This
+ * version works a little differently than strtofflags(3).
+ * In particular, it always tests every token, skipping any
+ * unrecognized tokens. It returns a pointer to the first
+ * unrecognized token, or NULL if every token was recognized.
+ * This version is also const-correct and does not modify the
+ * provided string.
+ */
+static const wchar_t *
+ae_wcstofflags(const wchar_t *s, unsigned long *setp, unsigned long *clrp)
+{
+ const wchar_t *start, *end;
+ struct flag *flag;
+ unsigned long set, clear;
+ const wchar_t *failed;
+
+ set = clear = 0;
+ start = s;
+ failed = NULL;
+ /* Find start of first token. */
+ while (*start == L'\t' || *start == L' ' || *start == L',')
+ start++;
+ while (*start != L'\0') {
+ /* Locate end of token. */
+ end = start;
+ while (*end != L'\0' && *end != L'\t' &&
+ *end != L' ' && *end != L',')
+ end++;
+ for (flag = flags; flag->wname != NULL; flag++) {
+ if (wmemcmp(start, flag->wname, end - start) == 0) {
+ /* Matched "noXXXX", so reverse the sense. */
+ clear |= flag->set;
+ set |= flag->clear;
+ break;
+ } else if (wmemcmp(start, flag->wname + 2, end - start)
+ == 0) {
+ /* Matched "XXXX", so don't reverse. */
+ set |= flag->set;
+ clear |= flag->clear;
+ break;
+ }
+ }
+ /* Ignore unknown flag names. */
+ if (flag->wname == NULL && failed == NULL)
+ failed = start;
+
+ /* Find start of next token. */
+ start = end;
+ while (*start == L'\t' || *start == L' ' || *start == L',')
+ start++;
+
+ }
+
+ if (setp)
+ *setp = set;
+ if (clrp)
+ *clrp = clear;
+
+ /* Return location of first failure. */
+ return (failed);
+}
+
+
+#ifdef TEST
+#include <stdio.h>
+int
+main(int argc, char **argv)
+{
+ struct archive_entry *entry = archive_entry_new();
+ unsigned long set, clear;
+ const wchar_t *remainder;
+
+ remainder = archive_entry_copy_fflags_text_w(entry, L"nosappnd dump archive,,,,,,,");
+ archive_entry_fflags(entry, &set, &clear);
+
+ wprintf(L"set=0x%lX clear=0x%lX remainder='%ls'\n", set, clear, remainder);
+
+ wprintf(L"new flags='%s'\n", archive_entry_fflags_text(entry));
+ return (0);
+}
+#endif
diff --git a/libarchive/archive_entry.h b/libarchive/archive_entry.h
new file mode 100644
index 000000000000..d5728179eb7d
--- /dev/null
+++ b/libarchive/archive_entry.h
@@ -0,0 +1,524 @@
+/*-
+ * Copyright (c) 2003-2008 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD: head/lib/libarchive/archive_entry.h 201096 2009-12-28 02:41:27Z kientzle $
+ */
+
+#ifndef ARCHIVE_ENTRY_H_INCLUDED
+#define ARCHIVE_ENTRY_H_INCLUDED
+
+/*
+ * Note: archive_entry.h is for use outside of libarchive; the
+ * configuration headers (config.h, archive_platform.h, etc.) are
+ * purely internal. Do NOT use HAVE_XXX configuration macros to
+ * control the behavior of this header! If you must conditionalize,
+ * use predefined compiler and/or platform macros.
+ */
+
+#include <sys/types.h>
+#include <stddef.h> /* for wchar_t */
+#include <time.h>
+
+#if defined(_WIN32) && !defined(__CYGWIN__)
+#include <windows.h>
+#endif
+
+/* Get appropriate definitions of standard POSIX-style types. */
+/* These should match the types used in 'struct stat' */
+#if defined(_WIN32) && !defined(__CYGWIN__)
+#define __LA_INT64_T __int64
+# if defined(__BORLANDC__)
+# define __LA_UID_T uid_t
+# define __LA_GID_T gid_t
+# define __LA_DEV_T dev_t
+# define __LA_MODE_T mode_t
+# else
+# define __LA_UID_T short
+# define __LA_GID_T short
+# define __LA_DEV_T unsigned int
+# define __LA_MODE_T unsigned short
+# endif
+#else
+#include <unistd.h>
+#define __LA_INT64_T int64_t
+#define __LA_UID_T uid_t
+#define __LA_GID_T gid_t
+#define __LA_DEV_T dev_t
+#define __LA_MODE_T mode_t
+#endif
+
+/*
+ * XXX Is this defined for all Windows compilers? If so, in what
+ * header? It would be nice to remove the __LA_INO_T indirection and
+ * just use plain ino_t everywhere. Likewise for the other types just
+ * above.
+ */
+#define __LA_INO_T ino_t
+
+
+/*
+ * On Windows, define LIBARCHIVE_STATIC if you're building or using a
+ * .lib. The default here assumes you're building a DLL. Only
+ * libarchive source should ever define __LIBARCHIVE_BUILD.
+ */
+#if ((defined __WIN32__) || (defined _WIN32) || defined(__CYGWIN__)) && (!defined LIBARCHIVE_STATIC)
+# ifdef __LIBARCHIVE_BUILD
+# ifdef __GNUC__
+# define __LA_DECL __attribute__((dllexport)) extern
+# else
+# define __LA_DECL __declspec(dllexport)
+# endif
+# else
+# ifdef __GNUC__
+# define __LA_DECL __attribute__((dllimport)) extern
+# else
+# define __LA_DECL __declspec(dllimport)
+# endif
+# endif
+#else
+/* Static libraries on all platforms and shared libraries on non-Windows. */
+# define __LA_DECL
+#endif
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/*
+ * Description of an archive entry.
+ *
+ * You can think of this as "struct stat" with some text fields added in.
+ *
+ * TODO: Add "comment", "charset", and possibly other entries that are
+ * supported by "pax interchange" format. However, GNU, ustar, cpio,
+ * and other variants don't support these features, so they're not an
+ * excruciatingly high priority right now.
+ *
+ * TODO: "pax interchange" format allows essentially arbitrary
+ * key/value attributes to be attached to any entry. Supporting
+ * such extensions may make this library useful for special
+ * applications (e.g., a package manager could attach special
+ * package-management attributes to each entry).
+ */
+struct archive_entry;
+
+/*
+ * File-type constants. These are returned from archive_entry_filetype()
+ * and passed to archive_entry_set_filetype().
+ *
+ * These values match S_XXX defines on every platform I've checked,
+ * including Windows, AIX, Linux, Solaris, and BSD. They're
+ * (re)defined here because platforms generally don't define the ones
+ * they don't support. For example, Windows doesn't define S_IFLNK or
+ * S_IFBLK. Instead of having a mass of conditional logic and system
+ * checks to define any S_XXX values that aren't supported locally,
+ * I've just defined a new set of such constants so that
+ * libarchive-based applications can manipulate and identify archive
+ * entries properly even if the hosting platform can't store them on
+ * disk.
+ *
+ * These values are also used directly within some portable formats,
+ * such as cpio. If you find a platform that varies from these, the
+ * correct solution is to leave these alone and translate from these
+ * portable values to platform-native values when entries are read from
+ * or written to disk.
+ */
+#define AE_IFMT 0170000
+#define AE_IFREG 0100000
+#define AE_IFLNK 0120000
+#define AE_IFSOCK 0140000
+#define AE_IFCHR 0020000
+#define AE_IFBLK 0060000
+#define AE_IFDIR 0040000
+#define AE_IFIFO 0010000
+
+/*
+ * Basic object manipulation
+ */
+
+__LA_DECL struct archive_entry *archive_entry_clear(struct archive_entry *);
+/* The 'clone' function does a deep copy; all of the strings are copied too. */
+__LA_DECL struct archive_entry *archive_entry_clone(struct archive_entry *);
+__LA_DECL void archive_entry_free(struct archive_entry *);
+__LA_DECL struct archive_entry *archive_entry_new(void);
+
+/*
+ * Retrieve fields from an archive_entry.
+ *
+ * There are a number of implicit conversions among these fields. For
+ * example, if a regular string field is set and you read the _w wide
+ * character field, the entry will implicitly convert narrow-to-wide
+ * using the current locale. Similarly, dev values are automatically
+ * updated when you write devmajor or devminor and vice versa.
+ *
+ * In addition, fields can be "set" or "unset." Unset string fields
+ * return NULL, non-string fields have _is_set() functions to test
+ * whether they've been set. You can "unset" a string field by
+ * assigning NULL; non-string fields have _unset() functions to
+ * unset them.
+ *
+ * Note: There is one ambiguity in the above; string fields will
+ * also return NULL when implicit character set conversions fail.
+ * This is usually what you want.
+ */
+__LA_DECL time_t archive_entry_atime(struct archive_entry *);
+__LA_DECL long archive_entry_atime_nsec(struct archive_entry *);
+__LA_DECL int archive_entry_atime_is_set(struct archive_entry *);
+__LA_DECL time_t archive_entry_birthtime(struct archive_entry *);
+__LA_DECL long archive_entry_birthtime_nsec(struct archive_entry *);
+__LA_DECL int archive_entry_birthtime_is_set(struct archive_entry *);
+__LA_DECL time_t archive_entry_ctime(struct archive_entry *);
+__LA_DECL long archive_entry_ctime_nsec(struct archive_entry *);
+__LA_DECL int archive_entry_ctime_is_set(struct archive_entry *);
+__LA_DECL dev_t archive_entry_dev(struct archive_entry *);
+__LA_DECL dev_t archive_entry_devmajor(struct archive_entry *);
+__LA_DECL dev_t archive_entry_devminor(struct archive_entry *);
+__LA_DECL __LA_MODE_T archive_entry_filetype(struct archive_entry *);
+__LA_DECL void archive_entry_fflags(struct archive_entry *,
+ unsigned long * /* set */,
+ unsigned long * /* clear */);
+__LA_DECL const char *archive_entry_fflags_text(struct archive_entry *);
+__LA_DECL __LA_GID_T archive_entry_gid(struct archive_entry *);
+__LA_DECL const char *archive_entry_gname(struct archive_entry *);
+__LA_DECL const wchar_t *archive_entry_gname_w(struct archive_entry *);
+__LA_DECL const char *archive_entry_hardlink(struct archive_entry *);
+__LA_DECL const wchar_t *archive_entry_hardlink_w(struct archive_entry *);
+__LA_DECL __LA_INO_T archive_entry_ino(struct archive_entry *);
+__LA_DECL __LA_INT64_T archive_entry_ino64(struct archive_entry *);
+__LA_DECL __LA_MODE_T archive_entry_mode(struct archive_entry *);
+__LA_DECL time_t archive_entry_mtime(struct archive_entry *);
+__LA_DECL long archive_entry_mtime_nsec(struct archive_entry *);
+__LA_DECL int archive_entry_mtime_is_set(struct archive_entry *);
+__LA_DECL unsigned int archive_entry_nlink(struct archive_entry *);
+__LA_DECL const char *archive_entry_pathname(struct archive_entry *);
+__LA_DECL const wchar_t *archive_entry_pathname_w(struct archive_entry *);
+__LA_DECL dev_t archive_entry_rdev(struct archive_entry *);
+__LA_DECL dev_t archive_entry_rdevmajor(struct archive_entry *);
+__LA_DECL dev_t archive_entry_rdevminor(struct archive_entry *);
+__LA_DECL const char *archive_entry_sourcepath(struct archive_entry *);
+__LA_DECL __LA_INT64_T archive_entry_size(struct archive_entry *);
+__LA_DECL int archive_entry_size_is_set(struct archive_entry *);
+__LA_DECL const char *archive_entry_strmode(struct archive_entry *);
+__LA_DECL const char *archive_entry_symlink(struct archive_entry *);
+__LA_DECL const wchar_t *archive_entry_symlink_w(struct archive_entry *);
+__LA_DECL __LA_UID_T archive_entry_uid(struct archive_entry *);
+__LA_DECL const char *archive_entry_uname(struct archive_entry *);
+__LA_DECL const wchar_t *archive_entry_uname_w(struct archive_entry *);
+
+/*
+ * Set fields in an archive_entry.
+ *
+ * Note that string 'set' functions do not copy the string, only the pointer.
+ * In contrast, 'copy' functions do copy the object pointed to.
+ *
+ * Note: As of libarchive 2.4, 'set' functions do copy the string and
+ * are therefore exact synonyms for the 'copy' versions. The 'copy'
+ * names will be retired in libarchive 3.0.
+ */
+
+__LA_DECL void archive_entry_set_atime(struct archive_entry *, time_t, long);
+__LA_DECL void archive_entry_unset_atime(struct archive_entry *);
+#if defined(_WIN32) && !defined(__CYGWIN__)
+__LA_DECL void archive_entry_copy_bhfi(struct archive_entry *,
+ BY_HANDLE_FILE_INFORMATION *);
+#endif
+__LA_DECL void archive_entry_set_birthtime(struct archive_entry *, time_t, long);
+__LA_DECL void archive_entry_unset_birthtime(struct archive_entry *);
+__LA_DECL void archive_entry_set_ctime(struct archive_entry *, time_t, long);
+__LA_DECL void archive_entry_unset_ctime(struct archive_entry *);
+__LA_DECL void archive_entry_set_dev(struct archive_entry *, dev_t);
+__LA_DECL void archive_entry_set_devmajor(struct archive_entry *, dev_t);
+__LA_DECL void archive_entry_set_devminor(struct archive_entry *, dev_t);
+__LA_DECL void archive_entry_set_filetype(struct archive_entry *, unsigned int);
+__LA_DECL void archive_entry_set_fflags(struct archive_entry *,
+ unsigned long /* set */, unsigned long /* clear */);
+/* Returns pointer to start of first invalid token, or NULL if none. */
+/* Note that all recognized tokens are processed, regardless. */
+__LA_DECL const char *archive_entry_copy_fflags_text(struct archive_entry *,
+ const char *);
+__LA_DECL const wchar_t *archive_entry_copy_fflags_text_w(struct archive_entry *,
+ const wchar_t *);
+__LA_DECL void archive_entry_set_gid(struct archive_entry *, __LA_GID_T);
+__LA_DECL void archive_entry_set_gname(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_gname(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_gname_w(struct archive_entry *, const wchar_t *);
+__LA_DECL int archive_entry_update_gname_utf8(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_set_hardlink(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_hardlink(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_hardlink_w(struct archive_entry *, const wchar_t *);
+__LA_DECL int archive_entry_update_hardlink_utf8(struct archive_entry *, const char *);
+#if ARCHIVE_VERSION_NUMBER >= 3000000
+/* Starting with libarchive 3.0, this will be synonym for ino64. */
+__LA_DECL void archive_entry_set_ino(struct archive_entry *, __LA_INT64_T);
+#else
+__LA_DECL void archive_entry_set_ino(struct archive_entry *, unsigned long);
+#endif
+__LA_DECL void archive_entry_set_ino64(struct archive_entry *, __LA_INT64_T);
+__LA_DECL void archive_entry_set_link(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_link(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_link_w(struct archive_entry *, const wchar_t *);
+__LA_DECL int archive_entry_update_link_utf8(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_set_mode(struct archive_entry *, __LA_MODE_T);
+__LA_DECL void archive_entry_set_mtime(struct archive_entry *, time_t, long);
+__LA_DECL void archive_entry_unset_mtime(struct archive_entry *);
+__LA_DECL void archive_entry_set_nlink(struct archive_entry *, unsigned int);
+__LA_DECL void archive_entry_set_pathname(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_pathname(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_pathname_w(struct archive_entry *, const wchar_t *);
+__LA_DECL int archive_entry_update_pathname_utf8(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_set_perm(struct archive_entry *, __LA_MODE_T);
+__LA_DECL void archive_entry_set_rdev(struct archive_entry *, dev_t);
+__LA_DECL void archive_entry_set_rdevmajor(struct archive_entry *, dev_t);
+__LA_DECL void archive_entry_set_rdevminor(struct archive_entry *, dev_t);
+__LA_DECL void archive_entry_set_size(struct archive_entry *, __LA_INT64_T);
+__LA_DECL void archive_entry_unset_size(struct archive_entry *);
+__LA_DECL void archive_entry_copy_sourcepath(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_set_symlink(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_symlink(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_symlink_w(struct archive_entry *, const wchar_t *);
+__LA_DECL int archive_entry_update_symlink_utf8(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_set_uid(struct archive_entry *, __LA_UID_T);
+__LA_DECL void archive_entry_set_uname(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_uname(struct archive_entry *, const char *);
+__LA_DECL void archive_entry_copy_uname_w(struct archive_entry *, const wchar_t *);
+__LA_DECL int archive_entry_update_uname_utf8(struct archive_entry *, const char *);
+/*
+ * Routines to bulk copy fields to/from a platform-native "struct
+ * stat." Libarchive used to just store a struct stat inside of each
+ * archive_entry object, but this created issues when trying to
+ * manipulate archives on systems different than the ones they were
+ * created on.
+ *
+ * TODO: On Linux, provide both stat32 and stat64 versions of these functions.
+ */
+__LA_DECL const struct stat *archive_entry_stat(struct archive_entry *);
+__LA_DECL void archive_entry_copy_stat(struct archive_entry *, const struct stat *);
+
+
+/*
+ * ACL routines. This used to simply store and return text-format ACL
+ * strings, but that proved insufficient for a number of reasons:
+ * = clients need control over uname/uid and gname/gid mappings
+ * = there are many different ACL text formats
+ * = would like to be able to read/convert archives containing ACLs
+ * on platforms that lack ACL libraries
+ *
+ * This last point, in particular, forces me to implement a reasonably
+ * complete set of ACL support routines.
+ *
+ * TODO: Extend this to support NFSv4/NTFS permissions. That should
+ * allow full ACL support on Mac OS, in particular, which uses
+ * POSIX.1e-style interfaces to manipulate NFSv4/NTFS permissions.
+ */
+
+/*
+ * Permission bits mimic POSIX.1e. Note that I've not followed POSIX.1e's
+ * "permset"/"perm" abstract type nonsense. A permset is just a simple
+ * bitmap, following long-standing Unix tradition.
+ */
+#define ARCHIVE_ENTRY_ACL_EXECUTE 1
+#define ARCHIVE_ENTRY_ACL_WRITE 2
+#define ARCHIVE_ENTRY_ACL_READ 4
+
+/* We need to be able to specify either or both of these. */
+#define ARCHIVE_ENTRY_ACL_TYPE_ACCESS 256
+#define ARCHIVE_ENTRY_ACL_TYPE_DEFAULT 512
+
+/* Tag values mimic POSIX.1e */
+#define ARCHIVE_ENTRY_ACL_USER 10001 /* Specified user. */
+#define ARCHIVE_ENTRY_ACL_USER_OBJ 10002 /* User who owns the file. */
+#define ARCHIVE_ENTRY_ACL_GROUP 10003 /* Specified group. */
+#define ARCHIVE_ENTRY_ACL_GROUP_OBJ 10004 /* Group who owns the file. */
+#define ARCHIVE_ENTRY_ACL_MASK 10005 /* Modify group access. */
+#define ARCHIVE_ENTRY_ACL_OTHER 10006 /* Public. */
+
+/*
+ * Set the ACL by clearing it and adding entries one at a time.
+ * Unlike the POSIX.1e ACL routines, you must specify the type
+ * (access/default) for each entry. Internally, the ACL data is just
+ * a soup of entries. API calls here allow you to retrieve just the
+ * entries of interest. This design (which goes against the spirit of
+ * POSIX.1e) is useful for handling archive formats that combine
+ * default and access information in a single ACL list.
+ */
+__LA_DECL void archive_entry_acl_clear(struct archive_entry *);
+__LA_DECL void archive_entry_acl_add_entry(struct archive_entry *,
+ int /* type */, int /* permset */, int /* tag */,
+ int /* qual */, const char * /* name */);
+__LA_DECL void archive_entry_acl_add_entry_w(struct archive_entry *,
+ int /* type */, int /* permset */, int /* tag */,
+ int /* qual */, const wchar_t * /* name */);
+
+/*
+ * To retrieve the ACL, first "reset", then repeatedly ask for the
+ * "next" entry. The want_type parameter allows you to request only
+ * access entries or only default entries.
+ */
+__LA_DECL int archive_entry_acl_reset(struct archive_entry *, int /* want_type */);
+__LA_DECL int archive_entry_acl_next(struct archive_entry *, int /* want_type */,
+ int * /* type */, int * /* permset */, int * /* tag */,
+ int * /* qual */, const char ** /* name */);
+__LA_DECL int archive_entry_acl_next_w(struct archive_entry *, int /* want_type */,
+ int * /* type */, int * /* permset */, int * /* tag */,
+ int * /* qual */, const wchar_t ** /* name */);
+
+/*
+ * Construct a text-format ACL. The flags argument is a bitmask that
+ * can include any of the following:
+ *
+ * ARCHIVE_ENTRY_ACL_TYPE_ACCESS - Include access entries.
+ * ARCHIVE_ENTRY_ACL_TYPE_DEFAULT - Include default entries.
+ * ARCHIVE_ENTRY_ACL_STYLE_EXTRA_ID - Include extra numeric ID field in
+ * each ACL entry. (As used by 'star'.)
+ * ARCHIVE_ENTRY_ACL_STYLE_MARK_DEFAULT - Include "default:" before each
+ * default ACL entry.
+ */
+#define ARCHIVE_ENTRY_ACL_STYLE_EXTRA_ID 1024
+#define ARCHIVE_ENTRY_ACL_STYLE_MARK_DEFAULT 2048
+__LA_DECL const wchar_t *archive_entry_acl_text_w(struct archive_entry *,
+ int /* flags */);
+
+/* Return a count of entries matching 'want_type' */
+__LA_DECL int archive_entry_acl_count(struct archive_entry *, int /* want_type */);
+
+/*
+ * Private ACL parser. This is private because it handles some
+ * very weird formats that clients should not be messing with.
+ * Clients should only deal with their platform-native formats.
+ * Because of the need to support many formats cleanly, new arguments
+ * are likely to get added on a regular basis. Clients who try to use
+ * this interface are likely to be surprised when it changes.
+ *
+ * You were warned!
+ *
+ * TODO: Move this declaration out of the public header and into
+ * a private header. Warnings above are silly.
+ */
+__LA_DECL int __archive_entry_acl_parse_w(struct archive_entry *,
+ const wchar_t *, int /* type */);
+
+/*
+ * extended attributes
+ */
+
+__LA_DECL void archive_entry_xattr_clear(struct archive_entry *);
+__LA_DECL void archive_entry_xattr_add_entry(struct archive_entry *,
+ const char * /* name */, const void * /* value */,
+ size_t /* size */);
+
+/*
+ * To retrieve the xattr list, first "reset", then repeatedly ask for the
+ * "next" entry.
+ */
+
+__LA_DECL int archive_entry_xattr_count(struct archive_entry *);
+__LA_DECL int archive_entry_xattr_reset(struct archive_entry *);
+__LA_DECL int archive_entry_xattr_next(struct archive_entry *,
+ const char ** /* name */, const void ** /* value */, size_t *);
+
+/*
+ * Utility to match up hardlinks.
+ *
+ * The 'struct archive_entry_linkresolver' is a cache of archive entries
+ * for files with multiple links. Here's how to use it:
+ * 1. Create a lookup object with archive_entry_linkresolver_new()
+ * 2. Tell it the archive format you're using.
+ * 3. Hand each archive_entry to archive_entry_linkify().
+ * That function will return 0, 1, or 2 entries that should
+ * be written.
+ * 4. Call archive_entry_linkify(resolver, NULL) until
+ * no more entries are returned.
+ * 5. Call archive_entry_link_resolver_free(resolver) to free resources.
+ *
+ * The entries returned have their hardlink and size fields updated
+ * appropriately. If an entry is passed in that does not refer to
+ * a file with multiple links, it is returned unchanged. The intention
+ * is that you should be able to simply filter all entries through
+ * this machine.
+ *
+ * To make things more efficient, be sure that each entry has a valid
+ * nlinks value. The hardlink cache uses this to track when all links
+ * have been found. If the nlinks value is zero, it will keep every
+ * name in the cache indefinitely, which can use a lot of memory.
+ *
+ * Note that archive_entry_size() is reset to zero if the file
+ * body should not be written to the archive. Pay attention!
+ */
+struct archive_entry_linkresolver;
+
+/*
+ * There are three different strategies for marking hardlinks.
+ * The descriptions below name them after the best-known
+ * formats that rely on each strategy:
+ *
+ * "Old cpio" is the simplest, it always returns any entry unmodified.
+ * As far as I know, only cpio formats use this. Old cpio archives
+ * store every link with the full body; the onus is on the dearchiver
+ * to detect and properly link the files as they are restored.
+ * "tar" is also pretty simple; it caches a copy the first time it sees
+ * any link. Subsequent appearances are modified to be hardlink
+ * references to the first one without any body. Used by all tar
+ * formats, although the newest tar formats permit the "old cpio" strategy
+ * as well. This strategy is very simple for the dearchiver,
+ * and reasonably straightforward for the archiver.
+ * "new cpio" is trickier. It stores the body only with the last
+ * occurrence. The complication is that we might not
+ * see every link to a particular file in a single session, so
+ * there's no easy way to know when we've seen the last occurrence.
+ * The solution here is to queue one link until we see the next.
+ * At the end of the session, you can enumerate any remaining
+ * entries by calling archive_entry_linkify(NULL) and store those
+ * bodies. If you have a file with three links l1, l2, and l3,
+ * you'll get the following behavior if you see all three links:
+ * linkify(l1) => NULL (the resolver stores l1 internally)
+ * linkify(l2) => l1 (resolver stores l2, you write l1)
+ * linkify(l3) => l2, l3 (all links seen, you can write both).
+ * If you only see l1 and l2, you'll get this behavior:
+ * linkify(l1) => NULL
+ * linkify(l2) => l1
+ * linkify(NULL) => l2 (at end, you retrieve remaining links)
+ * As the name suggests, this strategy is used by newer cpio variants.
+ * It's noticably more complex for the archiver, slightly more complex
+ * for the dearchiver than the tar strategy, but makes it straightforward
+ * to restore a file using any link by simply continuing to scan until
+ * you see a link that is stored with a body. In contrast, the tar
+ * strategy requires you to rescan the archive from the beginning to
+ * correctly extract an arbitrary link.
+ */
+
+__LA_DECL struct archive_entry_linkresolver *archive_entry_linkresolver_new(void);
+__LA_DECL void archive_entry_linkresolver_set_strategy(
+ struct archive_entry_linkresolver *, int /* format_code */);
+__LA_DECL void archive_entry_linkresolver_free(struct archive_entry_linkresolver *);
+__LA_DECL void archive_entry_linkify(struct archive_entry_linkresolver *,
+ struct archive_entry **, struct archive_entry **);
+
+#ifdef __cplusplus
+}
+#endif
+
+/* This is meaningless outside of this header. */
+#undef __LA_DECL
+
+#endif /* !ARCHIVE_ENTRY_H_INCLUDED */
diff --git a/libarchive/archive_entry_copy_bhfi.c b/libarchive/archive_entry_copy_bhfi.c
new file mode 100644
index 000000000000..8339032c5c35
--- /dev/null
+++ b/libarchive/archive_entry_copy_bhfi.c
@@ -0,0 +1,74 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD$");
+
+#include "archive_private.h"
+#include "archive_entry.h"
+
+#if defined(_WIN32) && !defined(__CYGWIN__)
+
+#define EPOC_TIME ARCHIVE_LITERAL_ULL(116444736000000000)
+
+__inline static void
+fileTimeToUtc(const FILETIME *filetime, time_t *time, long *ns)
+{
+ ULARGE_INTEGER utc;
+
+ utc.HighPart = filetime->dwHighDateTime;
+ utc.LowPart = filetime->dwLowDateTime;
+ if (utc.QuadPart >= EPOC_TIME) {
+ utc.QuadPart -= EPOC_TIME;
+ *time = (time_t)(utc.QuadPart / 10000000); /* milli seconds base */
+ *ns = (long)(utc.QuadPart % 10000000) * 100;/* nano seconds base */
+ } else {
+ *time = 0;
+ *ns = 0;
+ }
+}
+
+void
+archive_entry_copy_bhfi(struct archive_entry *entry,
+ BY_HANDLE_FILE_INFORMATION *bhfi)
+{
+ time_t secs;
+ long nsecs;
+
+ fileTimeToUtc(&bhfi->ftLastAccessTime, &secs, &nsecs);
+ archive_entry_set_atime(entry, secs, nsecs);
+ fileTimeToUtc(&bhfi->ftLastWriteTime, &secs, &nsecs);
+ archive_entry_set_mtime(entry, secs, nsecs);
+ fileTimeToUtc(&bhfi->ftCreationTime, &secs, &nsecs);
+ archive_entry_set_birthtime(entry, secs, nsecs);
+ archive_entry_set_dev(entry, bhfi->dwVolumeSerialNumber);
+ archive_entry_set_ino64(entry, (((int64_t)bhfi->nFileIndexHigh) << 32)
+ + bhfi->nFileIndexLow);
+ archive_entry_set_nlink(entry, bhfi->nNumberOfLinks);
+ archive_entry_set_size(entry, (((int64_t)bhfi->nFileSizeHigh) << 32)
+ + bhfi->nFileSizeLow);
+// archive_entry_set_mode(entry, st->st_mode);
+}
+#endif
diff --git a/libarchive/archive_entry_copy_stat.c b/libarchive/archive_entry_copy_stat.c
new file mode 100644
index 000000000000..ef59a5e78a04
--- /dev/null
+++ b/libarchive/archive_entry_copy_stat.c
@@ -0,0 +1,77 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_entry_copy_stat.c 189466 2009-03-07 00:52:02Z kientzle $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+
+#include "archive_entry.h"
+
+void
+archive_entry_copy_stat(struct archive_entry *entry, const struct stat *st)
+{
+#if HAVE_STRUCT_STAT_ST_MTIMESPEC_TV_NSEC
+ archive_entry_set_atime(entry, st->st_atime, st->st_atimespec.tv_nsec);
+ archive_entry_set_ctime(entry, st->st_ctime, st->st_ctimespec.tv_nsec);
+ archive_entry_set_mtime(entry, st->st_mtime, st->st_mtimespec.tv_nsec);
+#elif HAVE_STRUCT_STAT_ST_MTIM_TV_NSEC
+ archive_entry_set_atime(entry, st->st_atime, st->st_atim.tv_nsec);
+ archive_entry_set_ctime(entry, st->st_ctime, st->st_ctim.tv_nsec);
+ archive_entry_set_mtime(entry, st->st_mtime, st->st_mtim.tv_nsec);
+#elif HAVE_STRUCT_STAT_ST_MTIME_N
+ archive_entry_set_atime(entry, st->st_atime, st->st_atime_n);
+ archive_entry_set_ctime(entry, st->st_ctime, st->st_ctime_n);
+ archive_entry_set_mtime(entry, st->st_mtime, st->st_mtime_n);
+#elif HAVE_STRUCT_STAT_ST_UMTIME
+ archive_entry_set_atime(entry, st->st_atime, st->st_uatime * 1000);
+ archive_entry_set_ctime(entry, st->st_ctime, st->st_uctime * 1000);
+ archive_entry_set_mtime(entry, st->st_mtime, st->st_umtime * 1000);
+#elif HAVE_STRUCT_STAT_ST_MTIME_USEC
+ archive_entry_set_atime(entry, st->st_atime, st->st_atime_usec * 1000);
+ archive_entry_set_ctime(entry, st->st_ctime, st->st_ctime_usec * 1000);
+ archive_entry_set_mtime(entry, st->st_mtime, st->st_mtime_usec * 1000);
+#else
+ archive_entry_set_atime(entry, st->st_atime, 0);
+ archive_entry_set_ctime(entry, st->st_ctime, 0);
+ archive_entry_set_mtime(entry, st->st_mtime, 0);
+#if HAVE_STRUCT_STAT_ST_BIRTHTIME
+ archive_entry_set_birthtime(entry, st->st_birthtime, 0);
+#endif
+#endif
+#if HAVE_STRUCT_STAT_ST_BIRTHTIMESPEC_TV_NSEC
+ archive_entry_set_birthtime(entry, st->st_birthtime, st->st_birthtimespec.tv_nsec);
+#endif
+ archive_entry_set_dev(entry, st->st_dev);
+ archive_entry_set_gid(entry, st->st_gid);
+ archive_entry_set_uid(entry, st->st_uid);
+ archive_entry_set_ino(entry, st->st_ino);
+ archive_entry_set_nlink(entry, st->st_nlink);
+ archive_entry_set_rdev(entry, st->st_rdev);
+ archive_entry_set_size(entry, st->st_size);
+ archive_entry_set_mode(entry, st->st_mode);
+}
diff --git a/libarchive/archive_entry_link_resolver.c b/libarchive/archive_entry_link_resolver.c
new file mode 100644
index 000000000000..3b13e19c8bf7
--- /dev/null
+++ b/libarchive/archive_entry_link_resolver.c
@@ -0,0 +1,405 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_entry_link_resolver.c 201100 2009-12-28 03:05:31Z kientzle $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#include <stdio.h>
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+
+#include "archive.h"
+#include "archive_entry.h"
+
+/*
+ * This is mostly a pretty straightforward hash table implementation.
+ * The only interesting bit is the different strategies used to
+ * match up links. These strategies match those used by various
+ * archiving formats:
+ * tar - content stored with first link, remainder refer back to it.
+ * This requires us to match each subsequent link up with the
+ * first appearance.
+ * cpio - Old cpio just stored body with each link, match-ups were
+ * implicit. This is trivial.
+ * new cpio - New cpio only stores body with last link, match-ups
+ * are implicit. This is actually quite tricky; see the notes
+ * below.
+ */
+
+/* Users pass us a format code, we translate that into a strategy here. */
+#define ARCHIVE_ENTRY_LINKIFY_LIKE_TAR 0
+#define ARCHIVE_ENTRY_LINKIFY_LIKE_MTREE 1
+#define ARCHIVE_ENTRY_LINKIFY_LIKE_OLD_CPIO 2
+#define ARCHIVE_ENTRY_LINKIFY_LIKE_NEW_CPIO 3
+
+/* Initial size of link cache. */
+#define links_cache_initial_size 1024
+
+struct links_entry {
+ struct links_entry *next;
+ struct links_entry *previous;
+ int links; /* # links not yet seen */
+ int hash;
+ struct archive_entry *canonical;
+ struct archive_entry *entry;
+};
+
+struct archive_entry_linkresolver {
+ struct links_entry **buckets;
+ struct links_entry *spare;
+ unsigned long number_entries;
+ size_t number_buckets;
+ int strategy;
+};
+
+static struct links_entry *find_entry(struct archive_entry_linkresolver *,
+ struct archive_entry *);
+static void grow_hash(struct archive_entry_linkresolver *);
+static struct links_entry *insert_entry(struct archive_entry_linkresolver *,
+ struct archive_entry *);
+static struct links_entry *next_entry(struct archive_entry_linkresolver *);
+
+struct archive_entry_linkresolver *
+archive_entry_linkresolver_new(void)
+{
+ struct archive_entry_linkresolver *res;
+ size_t i;
+
+ res = malloc(sizeof(struct archive_entry_linkresolver));
+ if (res == NULL)
+ return (NULL);
+ memset(res, 0, sizeof(struct archive_entry_linkresolver));
+ res->number_buckets = links_cache_initial_size;
+ res->buckets = malloc(res->number_buckets *
+ sizeof(res->buckets[0]));
+ if (res->buckets == NULL) {
+ free(res);
+ return (NULL);
+ }
+ for (i = 0; i < res->number_buckets; i++)
+ res->buckets[i] = NULL;
+ return (res);
+}
+
+void
+archive_entry_linkresolver_set_strategy(struct archive_entry_linkresolver *res,
+ int fmt)
+{
+ int fmtbase = fmt & ARCHIVE_FORMAT_BASE_MASK;
+
+ switch (fmtbase) {
+ case ARCHIVE_FORMAT_CPIO:
+ switch (fmt) {
+ case ARCHIVE_FORMAT_CPIO_SVR4_NOCRC:
+ case ARCHIVE_FORMAT_CPIO_SVR4_CRC:
+ res->strategy = ARCHIVE_ENTRY_LINKIFY_LIKE_NEW_CPIO;
+ break;
+ default:
+ res->strategy = ARCHIVE_ENTRY_LINKIFY_LIKE_OLD_CPIO;
+ break;
+ }
+ break;
+ case ARCHIVE_FORMAT_MTREE:
+ res->strategy = ARCHIVE_ENTRY_LINKIFY_LIKE_MTREE;
+ break;
+ case ARCHIVE_FORMAT_TAR:
+ res->strategy = ARCHIVE_ENTRY_LINKIFY_LIKE_TAR;
+ break;
+ default:
+ res->strategy = ARCHIVE_ENTRY_LINKIFY_LIKE_TAR;
+ break;
+ }
+}
+
+void
+archive_entry_linkresolver_free(struct archive_entry_linkresolver *res)
+{
+ struct links_entry *le;
+
+ if (res == NULL)
+ return;
+
+ if (res->buckets != NULL) {
+ while ((le = next_entry(res)) != NULL)
+ archive_entry_free(le->entry);
+ free(res->buckets);
+ res->buckets = NULL;
+ }
+ free(res);
+}
+
+void
+archive_entry_linkify(struct archive_entry_linkresolver *res,
+ struct archive_entry **e, struct archive_entry **f)
+{
+ struct links_entry *le;
+ struct archive_entry *t;
+
+ *f = NULL; /* Default: Don't return a second entry. */
+
+ if (*e == NULL) {
+ le = next_entry(res);
+ if (le != NULL) {
+ *e = le->entry;
+ le->entry = NULL;
+ }
+ return;
+ }
+
+ /* If it has only one link, then we're done. */
+ if (archive_entry_nlink(*e) == 1)
+ return;
+ /* Directories, devices never have hardlinks. */
+ if (archive_entry_filetype(*e) == AE_IFDIR
+ || archive_entry_filetype(*e) == AE_IFBLK
+ || archive_entry_filetype(*e) == AE_IFCHR)
+ return;
+
+ switch (res->strategy) {
+ case ARCHIVE_ENTRY_LINKIFY_LIKE_TAR:
+ le = find_entry(res, *e);
+ if (le != NULL) {
+ archive_entry_unset_size(*e);
+ archive_entry_copy_hardlink(*e,
+ archive_entry_pathname(le->canonical));
+ } else
+ insert_entry(res, *e);
+ return;
+ case ARCHIVE_ENTRY_LINKIFY_LIKE_MTREE:
+ le = find_entry(res, *e);
+ if (le != NULL) {
+ archive_entry_copy_hardlink(*e,
+ archive_entry_pathname(le->canonical));
+ } else
+ insert_entry(res, *e);
+ return;
+ case ARCHIVE_ENTRY_LINKIFY_LIKE_OLD_CPIO:
+ /* This one is trivial. */
+ return;
+ case ARCHIVE_ENTRY_LINKIFY_LIKE_NEW_CPIO:
+ le = find_entry(res, *e);
+ if (le != NULL) {
+ /*
+ * Put the new entry in le, return the
+ * old entry from le.
+ */
+ t = *e;
+ *e = le->entry;
+ le->entry = t;
+ /* Make the old entry into a hardlink. */
+ archive_entry_unset_size(*e);
+ archive_entry_copy_hardlink(*e,
+ archive_entry_pathname(le->canonical));
+ /* If we ran out of links, return the
+ * final entry as well. */
+ if (le->links == 0) {
+ *f = le->entry;
+ le->entry = NULL;
+ }
+ } else {
+ /*
+ * If we haven't seen it, tuck it away
+ * for future use.
+ */
+ le = insert_entry(res, *e);
+ le->entry = *e;
+ *e = NULL;
+ }
+ return;
+ default:
+ break;
+ }
+ return;
+}
+
+static struct links_entry *
+find_entry(struct archive_entry_linkresolver *res,
+ struct archive_entry *entry)
+{
+ struct links_entry *le;
+ int hash, bucket;
+ dev_t dev;
+ int64_t ino;
+
+ /* Free a held entry. */
+ if (res->spare != NULL) {
+ archive_entry_free(res->spare->canonical);
+ archive_entry_free(res->spare->entry);
+ free(res->spare);
+ res->spare = NULL;
+ }
+
+ /* If the links cache overflowed and got flushed, don't bother. */
+ if (res->buckets == NULL)
+ return (NULL);
+
+ dev = archive_entry_dev(entry);
+ ino = archive_entry_ino64(entry);
+ hash = (int)(dev ^ ino);
+
+ /* Try to locate this entry in the links cache. */
+ bucket = hash % res->number_buckets;
+ for (le = res->buckets[bucket]; le != NULL; le = le->next) {
+ if (le->hash == hash
+ && dev == archive_entry_dev(le->canonical)
+ && ino == archive_entry_ino64(le->canonical)) {
+ /*
+ * Decrement link count each time and release
+ * the entry if it hits zero. This saves
+ * memory and is necessary for detecting
+ * missed links.
+ */
+ --le->links;
+ if (le->links > 0)
+ return (le);
+ /* Remove it from this hash bucket. */
+ if (le->previous != NULL)
+ le->previous->next = le->next;
+ if (le->next != NULL)
+ le->next->previous = le->previous;
+ if (res->buckets[bucket] == le)
+ res->buckets[bucket] = le->next;
+ res->number_entries--;
+ /* Defer freeing this entry. */
+ res->spare = le;
+ return (le);
+ }
+ }
+ return (NULL);
+}
+
+static struct links_entry *
+next_entry(struct archive_entry_linkresolver *res)
+{
+ struct links_entry *le;
+ size_t bucket;
+
+ /* Free a held entry. */
+ if (res->spare != NULL) {
+ archive_entry_free(res->spare->canonical);
+ free(res->spare);
+ res->spare = NULL;
+ }
+
+ /* If the links cache overflowed and got flushed, don't bother. */
+ if (res->buckets == NULL)
+ return (NULL);
+
+ /* Look for next non-empty bucket in the links cache. */
+ for (bucket = 0; bucket < res->number_buckets; bucket++) {
+ le = res->buckets[bucket];
+ if (le != NULL) {
+ /* Remove it from this hash bucket. */
+ if (le->next != NULL)
+ le->next->previous = le->previous;
+ res->buckets[bucket] = le->next;
+ res->number_entries--;
+ /* Defer freeing this entry. */
+ res->spare = le;
+ return (le);
+ }
+ }
+ return (NULL);
+}
+
+static struct links_entry *
+insert_entry(struct archive_entry_linkresolver *res,
+ struct archive_entry *entry)
+{
+ struct links_entry *le;
+ int hash, bucket;
+
+ /* Add this entry to the links cache. */
+ le = malloc(sizeof(struct links_entry));
+ if (le == NULL)
+ return (NULL);
+ memset(le, 0, sizeof(*le));
+ le->canonical = archive_entry_clone(entry);
+
+ /* If the links cache is getting too full, enlarge the hash table. */
+ if (res->number_entries > res->number_buckets * 2)
+ grow_hash(res);
+
+ hash = archive_entry_dev(entry) ^ archive_entry_ino64(entry);
+ bucket = hash % res->number_buckets;
+
+ /* If we could allocate the entry, record it. */
+ if (res->buckets[bucket] != NULL)
+ res->buckets[bucket]->previous = le;
+ res->number_entries++;
+ le->next = res->buckets[bucket];
+ le->previous = NULL;
+ res->buckets[bucket] = le;
+ le->hash = hash;
+ le->links = archive_entry_nlink(entry) - 1;
+ return (le);
+}
+
+static void
+grow_hash(struct archive_entry_linkresolver *res)
+{
+ struct links_entry *le, **new_buckets;
+ size_t new_size;
+ size_t i, bucket;
+
+ /* Try to enlarge the bucket list. */
+ new_size = res->number_buckets * 2;
+ new_buckets = malloc(new_size * sizeof(struct links_entry *));
+
+ if (new_buckets != NULL) {
+ memset(new_buckets, 0,
+ new_size * sizeof(struct links_entry *));
+ for (i = 0; i < res->number_buckets; i++) {
+ while (res->buckets[i] != NULL) {
+ /* Remove entry from old bucket. */
+ le = res->buckets[i];
+ res->buckets[i] = le->next;
+
+ /* Add entry to new bucket. */
+ bucket = le->hash % new_size;
+
+ if (new_buckets[bucket] != NULL)
+ new_buckets[bucket]->previous =
+ le;
+ le->next = new_buckets[bucket];
+ le->previous = NULL;
+ new_buckets[bucket] = le;
+ }
+ }
+ free(res->buckets);
+ res->buckets = new_buckets;
+ res->number_buckets = new_size;
+ }
+}
diff --git a/libarchive/archive_entry_private.h b/libarchive/archive_entry_private.h
new file mode 100644
index 000000000000..5ab4f75f0bce
--- /dev/null
+++ b/libarchive/archive_entry_private.h
@@ -0,0 +1,184 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD: head/lib/libarchive/archive_entry_private.h 201096 2009-12-28 02:41:27Z kientzle $
+ */
+
+#ifndef __LIBARCHIVE_BUILD
+#error This header is only to be used internally to libarchive.
+#endif
+
+#ifndef ARCHIVE_ENTRY_PRIVATE_H_INCLUDED
+#define ARCHIVE_ENTRY_PRIVATE_H_INCLUDED
+
+#include "archive_string.h"
+
+/*
+ * Handle wide character (i.e., Unicode) and non-wide character
+ * strings transparently.
+ */
+
+struct aes {
+ struct archive_string aes_mbs;
+ struct archive_string aes_utf8;
+ const wchar_t *aes_wcs;
+ /* Bitmap of which of the above are valid. Because we're lazy
+ * about malloc-ing and reusing the underlying storage, we
+ * can't rely on NULL pointers to indicate whether a string
+ * has been set. */
+ int aes_set;
+#define AES_SET_MBS 1
+#define AES_SET_UTF8 2
+#define AES_SET_WCS 4
+};
+
+struct ae_acl {
+ struct ae_acl *next;
+ int type; /* E.g., access or default */
+ int tag; /* E.g., user/group/other/mask */
+ int permset; /* r/w/x bits */
+ int id; /* uid/gid for user/group */
+ struct aes name; /* uname/gname */
+};
+
+struct ae_xattr {
+ struct ae_xattr *next;
+
+ char *name;
+ void *value;
+ size_t size;
+};
+
+/*
+ * Description of an archive entry.
+ *
+ * Basically, this is a "struct stat" with a few text fields added in.
+ *
+ * TODO: Add "comment", "charset", and possibly other entries
+ * that are supported by "pax interchange" format. However, GNU, ustar,
+ * cpio, and other variants don't support these features, so they're not an
+ * excruciatingly high priority right now.
+ *
+ * TODO: "pax interchange" format allows essentially arbitrary
+ * key/value attributes to be attached to any entry. Supporting
+ * such extensions may make this library useful for special
+ * applications (e.g., a package manager could attach special
+ * package-management attributes to each entry). There are tricky
+ * API issues involved, so this is not going to happen until
+ * there's a real demand for it.
+ *
+ * TODO: Design a good API for handling sparse files.
+ */
+struct archive_entry {
+ /*
+ * Note that ae_stat.st_mode & AE_IFMT can be 0!
+ *
+ * This occurs when the actual file type of the object is not
+ * in the archive. For example, 'tar' archives store
+ * hardlinks without marking the type of the underlying
+ * object.
+ */
+
+ /*
+ * Read archive_entry_copy_stat.c for an explanation of why I
+ * don't just use "struct stat" instead of "struct aest" here
+ * and why I have this odd pointer to a separately-allocated
+ * struct stat.
+ */
+ void *stat;
+ int stat_valid; /* Set to 0 whenever a field in aest changes. */
+
+ struct aest {
+ int64_t aest_atime;
+ uint32_t aest_atime_nsec;
+ int64_t aest_ctime;
+ uint32_t aest_ctime_nsec;
+ int64_t aest_mtime;
+ uint32_t aest_mtime_nsec;
+ int64_t aest_birthtime;
+ uint32_t aest_birthtime_nsec;
+ gid_t aest_gid;
+ int64_t aest_ino;
+ mode_t aest_mode;
+ uint32_t aest_nlink;
+ uint64_t aest_size;
+ uid_t aest_uid;
+ /*
+ * Because converting between device codes and
+ * major/minor values is platform-specific and
+ * inherently a bit risky, we only do that conversion
+ * lazily. That way, we will do a better job of
+ * preserving information in those cases where no
+ * conversion is actually required.
+ */
+ int aest_dev_is_broken_down;
+ dev_t aest_dev;
+ dev_t aest_devmajor;
+ dev_t aest_devminor;
+ int aest_rdev_is_broken_down;
+ dev_t aest_rdev;
+ dev_t aest_rdevmajor;
+ dev_t aest_rdevminor;
+ } ae_stat;
+
+ int ae_set; /* bitmap of fields that are currently set */
+#define AE_SET_HARDLINK 1
+#define AE_SET_SYMLINK 2
+#define AE_SET_ATIME 4
+#define AE_SET_CTIME 8
+#define AE_SET_MTIME 16
+#define AE_SET_BIRTHTIME 32
+#define AE_SET_SIZE 64
+
+ /*
+ * Use aes here so that we get transparent mbs<->wcs conversions.
+ */
+ struct aes ae_fflags_text; /* Text fflags per fflagstostr(3) */
+ unsigned long ae_fflags_set; /* Bitmap fflags */
+ unsigned long ae_fflags_clear;
+ struct aes ae_gname; /* Name of owning group */
+ struct aes ae_hardlink; /* Name of target for hardlink */
+ struct aes ae_pathname; /* Name of entry */
+ struct aes ae_symlink; /* symlink contents */
+ struct aes ae_uname; /* Name of owner */
+
+ /* Not used within libarchive; useful for some clients. */
+ struct aes ae_sourcepath; /* Path this entry is sourced from. */
+
+ /* ACL support. */
+ struct ae_acl *acl_head;
+ struct ae_acl *acl_p;
+ int acl_state; /* See acl_next for details. */
+ wchar_t *acl_text_w;
+
+ /* extattr support. */
+ struct ae_xattr *xattr_head;
+ struct ae_xattr *xattr_p;
+
+ /* Miscellaneous. */
+ char strmode[12];
+};
+
+
+#endif /* ARCHIVE_ENTRY_PRIVATE_H_INCLUDED */
diff --git a/libarchive/archive_entry_stat.c b/libarchive/archive_entry_stat.c
new file mode 100644
index 000000000000..ad772c9564b6
--- /dev/null
+++ b/libarchive/archive_entry_stat.c
@@ -0,0 +1,118 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_entry_stat.c 201100 2009-12-28 03:05:31Z kientzle $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+
+#include "archive_entry.h"
+#include "archive_entry_private.h"
+
+const struct stat *
+archive_entry_stat(struct archive_entry *entry)
+{
+ struct stat *st;
+ if (entry->stat == NULL) {
+ entry->stat = malloc(sizeof(*st));
+ if (entry->stat == NULL)
+ return (NULL);
+ entry->stat_valid = 0;
+ }
+
+ /*
+ * If none of the underlying fields have been changed, we
+ * don't need to regenerate. In theory, we could use a bitmap
+ * here to flag only those items that have changed, but the
+ * extra complexity probably isn't worth it. It will be very
+ * rare for anyone to change just one field then request a new
+ * stat structure.
+ */
+ if (entry->stat_valid)
+ return (entry->stat);
+
+ st = entry->stat;
+ /*
+ * Use the public interfaces to extract items, so that
+ * the appropriate conversions get invoked.
+ */
+ st->st_atime = archive_entry_atime(entry);
+#if HAVE_STRUCT_STAT_ST_BIRTHTIME
+ st->st_birthtime = archive_entry_birthtime(entry);
+#endif
+ st->st_ctime = archive_entry_ctime(entry);
+ st->st_mtime = archive_entry_mtime(entry);
+ st->st_dev = archive_entry_dev(entry);
+ st->st_gid = archive_entry_gid(entry);
+ st->st_uid = archive_entry_uid(entry);
+ st->st_ino = archive_entry_ino64(entry);
+ st->st_nlink = archive_entry_nlink(entry);
+ st->st_rdev = archive_entry_rdev(entry);
+ st->st_size = archive_entry_size(entry);
+ st->st_mode = archive_entry_mode(entry);
+
+ /*
+ * On systems that support high-res timestamps, copy that
+ * information into struct stat.
+ */
+#if HAVE_STRUCT_STAT_ST_MTIMESPEC_TV_NSEC
+ st->st_atimespec.tv_nsec = archive_entry_atime_nsec(entry);
+ st->st_ctimespec.tv_nsec = archive_entry_ctime_nsec(entry);
+ st->st_mtimespec.tv_nsec = archive_entry_mtime_nsec(entry);
+#elif HAVE_STRUCT_STAT_ST_MTIM_TV_NSEC
+ st->st_atim.tv_nsec = archive_entry_atime_nsec(entry);
+ st->st_ctim.tv_nsec = archive_entry_ctime_nsec(entry);
+ st->st_mtim.tv_nsec = archive_entry_mtime_nsec(entry);
+#elif HAVE_STRUCT_STAT_ST_MTIME_N
+ st->st_atime_n = archive_entry_atime_nsec(entry);
+ st->st_ctime_n = archive_entry_ctime_nsec(entry);
+ st->st_mtime_n = archive_entry_mtime_nsec(entry);
+#elif HAVE_STRUCT_STAT_ST_UMTIME
+ st->st_uatime = archive_entry_atime_nsec(entry) / 1000;
+ st->st_uctime = archive_entry_ctime_nsec(entry) / 1000;
+ st->st_umtime = archive_entry_mtime_nsec(entry) / 1000;
+#elif HAVE_STRUCT_STAT_ST_MTIME_USEC
+ st->st_atime_usec = archive_entry_atime_nsec(entry) / 1000;
+ st->st_ctime_usec = archive_entry_ctime_nsec(entry) / 1000;
+ st->st_mtime_usec = archive_entry_mtime_nsec(entry) / 1000;
+#endif
+#if HAVE_STRUCT_STAT_ST_BIRTHTIMESPEC_TV_NSEC
+ st->st_birthtimespec.tv_nsec = archive_entry_birthtime_nsec(entry);
+#endif
+
+ /*
+ * TODO: On Linux, store 32 or 64 here depending on whether
+ * the cached stat structure is a stat32 or a stat64. This
+ * will allow us to support both variants interchangably.
+ */
+ entry->stat_valid = 1;
+
+ return (st);
+}
diff --git a/libarchive/archive_entry_strmode.c b/libarchive/archive_entry_strmode.c
new file mode 100644
index 000000000000..16cb3f7bb337
--- /dev/null
+++ b/libarchive/archive_entry_strmode.c
@@ -0,0 +1,87 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: src/lib/libarchive/archive_entry_strmode.c,v 1.4 2008/06/15 05:14:01 kientzle Exp $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+
+#include "archive_entry.h"
+#include "archive_entry_private.h"
+
+const char *
+archive_entry_strmode(struct archive_entry *entry)
+{
+ static const mode_t permbits[] =
+ { 0400, 0200, 0100, 0040, 0020, 0010, 0004, 0002, 0001 };
+ char *bp = entry->strmode;
+ mode_t mode;
+ int i;
+
+ /* Fill in a default string, then selectively override. */
+ strcpy(bp, "?rwxrwxrwx ");
+
+ mode = archive_entry_mode(entry);
+ switch (archive_entry_filetype(entry)) {
+ case AE_IFREG: bp[0] = '-'; break;
+ case AE_IFBLK: bp[0] = 'b'; break;
+ case AE_IFCHR: bp[0] = 'c'; break;
+ case AE_IFDIR: bp[0] = 'd'; break;
+ case AE_IFLNK: bp[0] = 'l'; break;
+ case AE_IFSOCK: bp[0] = 's'; break;
+ case AE_IFIFO: bp[0] = 'p'; break;
+ default:
+ if (archive_entry_hardlink(entry) != NULL) {
+ bp[0] = 'h';
+ break;
+ }
+ }
+
+ for (i = 0; i < 9; i++)
+ if (!(mode & permbits[i]))
+ bp[i+1] = '-';
+
+ if (mode & S_ISUID) {
+ if (mode & 0100) bp[3] = 's';
+ else bp[3] = 'S';
+ }
+ if (mode & S_ISGID) {
+ if (mode & 0010) bp[6] = 's';
+ else bp[6] = 'S';
+ }
+ if (mode & S_ISVTX) {
+ if (mode & 0001) bp[9] = 't';
+ else bp[9] = 'T';
+ }
+ if (archive_entry_acl_count(entry, ARCHIVE_ENTRY_ACL_TYPE_ACCESS))
+ bp[10] = '+';
+
+ return (bp);
+}
diff --git a/libarchive/archive_entry_xattr.c b/libarchive/archive_entry_xattr.c
new file mode 100644
index 000000000000..a3efe7ca8d9f
--- /dev/null
+++ b/libarchive/archive_entry_xattr.c
@@ -0,0 +1,158 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_entry_xattr.c 201096 2009-12-28 02:41:27Z kientzle $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_SYS_TYPES_H
+#include <sys/types.h>
+#endif
+#ifdef HAVE_LIMITS_H
+#include <limits.h>
+#endif
+#ifdef HAVE_LINUX_FS_H
+#include <linux/fs.h> /* for Linux file flags */
+#endif
+/*
+ * Some Linux distributions have both linux/ext2_fs.h and ext2fs/ext2_fs.h.
+ * As the include guards don't agree, the order of include is important.
+ */
+#ifdef HAVE_LINUX_EXT2_FS_H
+#include <linux/ext2_fs.h> /* for Linux file flags */
+#endif
+#if defined(HAVE_EXT2FS_EXT2_FS_H) && !defined(__CYGWIN__)
+#include <ext2fs/ext2_fs.h> /* for Linux file flags */
+#endif
+#include <stddef.h>
+#include <stdio.h>
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_WCHAR_H
+#include <wchar.h>
+#endif
+
+#include "archive.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_entry_private.h"
+
+/*
+ * extended attribute handling
+ */
+
+void
+archive_entry_xattr_clear(struct archive_entry *entry)
+{
+ struct ae_xattr *xp;
+
+ while (entry->xattr_head != NULL) {
+ xp = entry->xattr_head->next;
+ free(entry->xattr_head->name);
+ free(entry->xattr_head->value);
+ free(entry->xattr_head);
+ entry->xattr_head = xp;
+ }
+
+ entry->xattr_head = NULL;
+}
+
+void
+archive_entry_xattr_add_entry(struct archive_entry *entry,
+ const char *name, const void *value, size_t size)
+{
+ struct ae_xattr *xp;
+
+ for (xp = entry->xattr_head; xp != NULL; xp = xp->next)
+ ;
+
+ if ((xp = (struct ae_xattr *)malloc(sizeof(struct ae_xattr))) == NULL)
+ /* XXX Error XXX */
+ return;
+
+ xp->name = strdup(name);
+ if ((xp->value = malloc(size)) != NULL) {
+ memcpy(xp->value, value, size);
+ xp->size = size;
+ } else
+ xp->size = 0;
+
+ xp->next = entry->xattr_head;
+ entry->xattr_head = xp;
+}
+
+
+/*
+ * returns number of the extended attribute entries
+ */
+int
+archive_entry_xattr_count(struct archive_entry *entry)
+{
+ struct ae_xattr *xp;
+ int count = 0;
+
+ for (xp = entry->xattr_head; xp != NULL; xp = xp->next)
+ count++;
+
+ return count;
+}
+
+int
+archive_entry_xattr_reset(struct archive_entry * entry)
+{
+ entry->xattr_p = entry->xattr_head;
+
+ return archive_entry_xattr_count(entry);
+}
+
+int
+archive_entry_xattr_next(struct archive_entry * entry,
+ const char **name, const void **value, size_t *size)
+{
+ if (entry->xattr_p) {
+ *name = entry->xattr_p->name;
+ *value = entry->xattr_p->value;
+ *size = entry->xattr_p->size;
+
+ entry->xattr_p = entry->xattr_p->next;
+
+ return (ARCHIVE_OK);
+ } else {
+ *name = NULL;
+ *value = NULL;
+ *size = (size_t)0;
+ return (ARCHIVE_WARN);
+ }
+}
+
+/*
+ * end of xattr handling
+ */
diff --git a/libarchive/archive_hash.h b/libarchive/archive_hash.h
new file mode 100644
index 000000000000..da67a3026d6b
--- /dev/null
+++ b/libarchive/archive_hash.h
@@ -0,0 +1,309 @@
+/*-
+ * Copyright (c) 2009 Joerg Sonnenberger
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD: head/lib/libarchive/archive_hash.h 201171 2009-12-29 06:39:07Z kientzle $
+ */
+
+#ifndef __LIBARCHIVE_BUILD
+#error This header is only to be used internally to libarchive.
+#endif
+
+#ifdef HAVE_SYS_TYPES_H
+#include <sys/types.h>
+#endif
+
+/*
+ * Hash function support in various Operating Systems:
+ *
+ * NetBSD:
+ * - MD5 and SHA1 in libc: without _ after algorithm name
+ * - SHA2 in libc: with _ after algorithm name
+ *
+ * OpenBSD:
+ * - MD5, SHA1 and SHA2 in libc: without _ after algorithm name
+ * - OpenBSD 4.4 and earlier have SHA2 in libc with _ after algorithm name
+ *
+ * DragonFly and FreeBSD (XXX not used yet):
+ * - MD5 and SHA1 in libmd: without _ after algorithm name
+ * - SHA256: with _ after algorithm name
+ *
+ * Mac OS X (10.4 and later):
+ * - MD5, SHA1 and SHA2 in libSystem: with CC_ prefix and _ after algorithm name
+ *
+ * OpenSSL:
+ * - MD5, SHA1 and SHA2 in libcrypto: with _ after algorithm name
+ *
+ * Windows:
+ * - MD5, SHA1 and SHA2 in archive_windows.c: without algorithm name
+ * and with __la_ prefix.
+ */
+#if defined(ARCHIVE_HASH_MD5_WIN) ||\
+ defined(ARCHIVE_HASH_SHA1_WIN) || defined(ARCHIVE_HASH_SHA256_WIN) ||\
+ defined(ARCHIVE_HASH_SHA384_WIN) || defined(ARCHIVE_HASH_SHA512_WIN)
+#include <wincrypt.h>
+typedef struct {
+ int valid;
+ HCRYPTPROV cryptProv;
+ HCRYPTHASH hash;
+} Digest_CTX;
+extern void __la_hash_Init(Digest_CTX *, ALG_ID);
+extern void __la_hash_Final(unsigned char *, size_t, Digest_CTX *);
+extern void __la_hash_Update(Digest_CTX *, const unsigned char *, size_t);
+#endif
+
+#if defined(ARCHIVE_HASH_MD5_LIBC)
+# include <md5.h>
+# define ARCHIVE_HAS_MD5
+typedef MD5_CTX archive_md5_ctx;
+# define archive_md5_init(ctx) MD5Init(ctx)
+# define archive_md5_final(ctx, buf) MD5Final(buf, ctx)
+# define archive_md5_update(ctx, buf, n) MD5Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_MD5_LIBMD)
+# include <md5.h>
+# define ARCHIVE_HAS_MD5
+typedef MD5_CTX archive_md5_ctx;
+# define archive_md5_init(ctx) MD5Init(ctx)
+# define archive_md5_final(ctx, buf) MD5Final(buf, ctx)
+# define archive_md5_update(ctx, buf, n) MD5Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_MD5_LIBSYSTEM)
+# include <CommonCrypto/CommonDigest.h>
+# define ARCHIVE_HAS_MD5
+typedef CC_MD5_CTX archive_md5_ctx;
+# define archive_md5_init(ctx) CC_MD5_Init(ctx)
+# define archive_md5_final(ctx, buf) CC_MD5_Final(buf, ctx)
+# define archive_md5_update(ctx, buf, n) CC_MD5_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_MD5_OPENSSL)
+# include <openssl/md5.h>
+# define ARCHIVE_HAS_MD5
+typedef MD5_CTX archive_md5_ctx;
+# define archive_md5_init(ctx) MD5_Init(ctx)
+# define archive_md5_final(ctx, buf) MD5_Final(buf, ctx)
+# define archive_md5_update(ctx, buf, n) MD5_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_MD5_WIN)
+# define ARCHIVE_HAS_MD5
+# define MD5_DIGEST_LENGTH 16
+typedef Digest_CTX archive_md5_ctx;
+# define archive_md5_init(ctx) __la_hash_Init(ctx, CALG_MD5)
+# define archive_md5_final(ctx, buf) __la_hash_Final(buf, MD5_DIGEST_LENGTH, ctx)
+# define archive_md5_update(ctx, buf, n) __la_hash_Update(ctx, buf, n)
+#endif
+
+#if defined(ARCHIVE_HASH_RMD160_LIBC)
+# include <rmd160.h>
+# define ARCHIVE_HAS_RMD160
+typedef RMD160_CTX archive_rmd160_ctx;
+# define archive_rmd160_init(ctx) RMD160Init(ctx)
+# define archive_rmd160_final(ctx, buf) RMD160Final(buf, ctx)
+# define archive_rmd160_update(ctx, buf, n) RMD160Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_RMD160_OPENSSL)
+# include <openssl/ripemd.h>
+# define ARCHIVE_HAS_RMD160
+typedef RIPEMD160_CTX archive_rmd160_ctx;
+# define archive_rmd160_init(ctx) RIPEMD160_Init(ctx)
+# define archive_rmd160_final(ctx, buf) RIPEMD160_Final(buf, ctx)
+# define archive_rmd160_update(ctx, buf, n) RIPEMD160_Update(ctx, buf, n)
+#endif
+
+#if defined(ARCHIVE_HASH_SHA1_LIBC)
+# include <sha1.h>
+# define ARCHIVE_HAS_SHA1
+typedef SHA1_CTX archive_sha1_ctx;
+# define archive_sha1_init(ctx) SHA1Init(ctx)
+# define archive_sha1_final(ctx, buf) SHA1Final(buf, ctx)
+# define archive_sha1_update(ctx, buf, n) SHA1Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA1_LIBMD)
+# include <sha.h>
+# define ARCHIVE_HAS_SHA1
+typedef SHA1_CTX archive_sha1_ctx;
+# define archive_sha1_init(ctx) SHA1_Init(ctx)
+# define archive_sha1_final(ctx, buf) SHA1_Final(buf, ctx)
+# define archive_sha1_update(ctx, buf, n) SHA1_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA1_LIBSYSTEM)
+# include <CommonCrypto/CommonDigest.h>
+# define ARCHIVE_HAS_SHA1
+typedef CC_SHA1_CTX archive_sha1_ctx;
+# define archive_sha1_init(ctx) CC_SHA1_Init(ctx)
+# define archive_sha1_final(ctx, buf) CC_SHA1_Final(buf, ctx)
+# define archive_sha1_update(ctx, buf, n) CC_SHA1_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA1_OPENSSL)
+# include <openssl/sha.h>
+# define ARCHIVE_HAS_SHA1
+typedef SHA_CTX archive_sha1_ctx;
+# define archive_sha1_init(ctx) SHA1_Init(ctx)
+# define archive_sha1_final(ctx, buf) SHA1_Final(buf, ctx)
+# define archive_sha1_update(ctx, buf, n) SHA1_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA1_WIN)
+# define ARCHIVE_HAS_SHA1
+# define SHA1_DIGEST_LENGTH 20
+typedef Digest_CTX archive_sha1_ctx;
+# define archive_sha1_init(ctx) __la_hash_Init(ctx, CALG_SHA1)
+# define archive_sha1_final(ctx, buf) __la_hash_Final(buf, SHA1_DIGEST_LENGTH, ctx)
+# define archive_sha1_update(ctx, buf, n) __la_hash_Update(ctx, buf, n)
+#endif
+
+#if defined(ARCHIVE_HASH_SHA256_LIBC)
+# include <sha2.h>
+# define ARCHIVE_HAS_SHA256
+typedef SHA256_CTX archive_sha256_ctx;
+# define archive_sha256_init(ctx) SHA256_Init(ctx)
+# define archive_sha256_final(ctx, buf) SHA256_Final(buf, ctx)
+# define archive_sha256_update(ctx, buf, n) SHA256_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA256_LIBC2)
+# include <sha2.h>
+# define ARCHIVE_HAS_SHA256
+typedef SHA256_CTX archive_sha256_ctx;
+# define archive_sha256_init(ctx) SHA256Init(ctx)
+# define archive_sha256_final(ctx, buf) SHA256Final(buf, ctx)
+# define archive_sha256_update(ctx, buf, n) SHA256Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA256_LIBC3)
+# include <sha2.h>
+# define ARCHIVE_HAS_SHA256
+typedef SHA2_CTX archive_sha256_ctx;
+# define archive_sha256_init(ctx) SHA256Init(ctx)
+# define archive_sha256_final(ctx, buf) SHA256Final(buf, ctx)
+# define archive_sha256_update(ctx, buf, n) SHA256Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA256_LIBMD)
+# include <sha256.h>
+# define ARCHIVE_HAS_SHA256
+typedef SHA256_CTX archive_sha256_ctx;
+# define archive_sha256_init(ctx) SHA256_Init(ctx)
+# define archive_sha256_final(ctx, buf) SHA256_Final(buf, ctx)
+# define archive_sha256_update(ctx, buf, n) SHA256_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA256_LIBSYSTEM)
+# include <CommonCrypto/CommonDigest.h>
+# define ARCHIVE_HAS_SHA256
+typedef CC_SHA256_CTX archive_shs256_ctx;
+# define archive_shs256_init(ctx) CC_SHA256_Init(ctx)
+# define archive_shs256_final(ctx, buf) CC_SHA256_Final(buf, ctx)
+# define archive_shs256_update(ctx, buf, n) CC_SHA256_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA256_OPENSSL)
+# include <openssl/sha.h>
+# define ARCHIVE_HAS_SHA256
+typedef SHA256_CTX archive_sha256_ctx;
+# define archive_sha256_init(ctx) SHA256_Init(ctx)
+# define archive_sha256_final(ctx, buf) SHA256_Final(buf, ctx)
+# define archive_sha256_update(ctx, buf, n) SHA256_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA256_WIN)
+# define ARCHIVE_HAS_SHA256
+# define SHA256_DIGEST_LENGTH 32
+typedef Digest_CTX archive_sha256_ctx;
+# define archive_sha256_init(ctx) __la_hash_Init(ctx, CALG_SHA_256)
+# define archive_sha256_final(ctx, buf) __la_hash_Final(buf, SHA256_DIGEST_LENGTH, ctx)
+# define archive_sha256_update(ctx, buf, n) __la_hash_Update(ctx, buf, n)
+#endif
+
+#if defined(ARCHIVE_HASH_SHA384_LIBC)
+# include <sha2.h>
+# define ARCHIVE_HAS_SHA384
+typedef SHA384_CTX archive_sha384_ctx;
+# define archive_sha384_init(ctx) SHA384_Init(ctx)
+# define archive_sha384_final(ctx, buf) SHA384_Final(buf, ctx)
+# define archive_sha384_update(ctx, buf, n) SHA384_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA384_LIBC2)
+# include <sha2.h>
+# define ARCHIVE_HAS_SHA384
+typedef SHA384_CTX archive_sha384_ctx;
+# define archive_sha384_init(ctx) SHA384Init(ctx)
+# define archive_sha384_final(ctx, buf) SHA384Final(buf, ctx)
+# define archive_sha384_update(ctx, buf, n) SHA384Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA384_LIBC3)
+# include <sha2.h>
+# define ARCHIVE_HAS_SHA384
+typedef SHA2_CTX archive_sha384_ctx;
+# define archive_sha384_init(ctx) SHA384Init(ctx)
+# define archive_sha384_final(ctx, buf) SHA384Final(buf, ctx)
+# define archive_sha384_update(ctx, buf, n) SHA384Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA384_LIBSYSTEM)
+# include <CommonCrypto/CommonDigest.h>
+# define ARCHIVE_HAS_SHA384
+typedef CC_SHA512_CTX archive_shs384_ctx;
+# define archive_shs384_init(ctx) CC_SHA384_Init(ctx)
+# define archive_shs384_final(ctx, buf) CC_SHA384_Final(buf, ctx)
+# define archive_shs384_update(ctx, buf, n) CC_SHA384_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA384_OPENSSL)
+# include <openssl/sha.h>
+# define ARCHIVE_HAS_SHA384
+typedef SHA512_CTX archive_sha384_ctx;
+# define archive_sha384_init(ctx) SHA384_Init(ctx)
+# define archive_sha384_final(ctx, buf) SHA384_Final(buf, ctx)
+# define archive_sha384_update(ctx, buf, n) SHA384_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA384_WIN)
+# define ARCHIVE_HAS_SHA384
+# define SHA384_DIGEST_LENGTH 48
+typedef Digest_CTX archive_sha384_ctx;
+# define archive_sha384_init(ctx) __la_hash_Init(ctx, CALG_SHA_384)
+# define archive_sha384_final(ctx, buf) __la_hash_Final(buf, SHA384_DIGEST_LENGTH, ctx)
+# define archive_sha384_update(ctx, buf, n) __la_hash_Update(ctx, buf, n)
+#endif
+
+#if defined(ARCHIVE_HASH_SHA512_LIBC)
+# include <sha2.h>
+# define ARCHIVE_HAS_SHA512
+typedef SHA512_CTX archive_sha512_ctx;
+# define archive_sha512_init(ctx) SHA512_Init(ctx)
+# define archive_sha512_final(ctx, buf) SHA512_Final(buf, ctx)
+# define archive_sha512_update(ctx, buf, n) SHA512_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA512_LIBC2)
+# include <sha2.h>
+# define ARCHIVE_HAS_SHA512
+typedef SHA512_CTX archive_sha512_ctx;
+# define archive_sha512_init(ctx) SHA512Init(ctx)
+# define archive_sha512_final(ctx, buf) SHA512Final(buf, ctx)
+# define archive_sha512_update(ctx, buf, n) SHA512Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA512_LIBC3)
+# include <sha2.h>
+# define ARCHIVE_HAS_SHA512
+typedef SHA2_CTX archive_sha512_ctx;
+# define archive_sha512_init(ctx) SHA512Init(ctx)
+# define archive_sha512_final(ctx, buf) SHA512Final(buf, ctx)
+# define archive_sha512_update(ctx, buf, n) SHA512Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA512_LIBMD)
+# include <sha512.h>
+# define ARCHIVE_HAS_SHA512
+typedef SHA512_CTX archive_sha512_ctx;
+# define archive_sha512_init(ctx) SHA512_Init(ctx)
+# define archive_sha512_final(ctx, buf) SHA512_Final(buf, ctx)
+# define archive_sha512_update(ctx, buf, n) SHA512_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA512_LIBSYSTEM)
+# include <CommonCrypto/CommonDigest.h>
+# define ARCHIVE_HAS_SHA512
+typedef CC_SHA512_CTX archive_shs512_ctx;
+# define archive_shs512_init(ctx) CC_SHA512_Init(ctx)
+# define archive_shs512_final(ctx, buf) CC_SHA512_Final(buf, ctx)
+# define archive_shs512_update(ctx, buf, n) CC_SHA512_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA512_OPENSSL)
+# include <openssl/sha.h>
+# define ARCHIVE_HAS_SHA512
+typedef SHA512_CTX archive_sha512_ctx;
+# define archive_sha512_init(ctx) SHA512_Init(ctx)
+# define archive_sha512_final(ctx, buf) SHA512_Final(buf, ctx)
+# define archive_sha512_update(ctx, buf, n) SHA512_Update(ctx, buf, n)
+#elif defined(ARCHIVE_HASH_SHA512_WIN)
+# define ARCHIVE_HAS_SHA512
+# define SHA512_DIGEST_LENGTH 64
+typedef Digest_CTX archive_sha512_ctx;
+# define archive_sha512_init(ctx) __la_hash_Init(ctx, CALG_SHA_512)
+# define archive_sha512_final(ctx, buf) __la_hash_Final(buf, SHA512_DIGEST_LENGTH, ctx)
+# define archive_sha512_update(ctx, buf, n) __la_hash_Update(ctx, buf, n)
+#endif
diff --git a/libarchive/archive_platform.h b/libarchive/archive_platform.h
new file mode 100644
index 000000000000..ce2f482ba042
--- /dev/null
+++ b/libarchive/archive_platform.h
@@ -0,0 +1,165 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD: head/lib/libarchive/archive_platform.h 201090 2009-12-28 02:22:04Z kientzle $
+ */
+
+/* !!ONLY FOR USE INTERNALLY TO LIBARCHIVE!! */
+
+/*
+ * This header is the first thing included in any of the libarchive
+ * source files. As far as possible, platform-specific issues should
+ * be dealt with here and not within individual source files. I'm
+ * actively trying to minimize #if blocks within the main source,
+ * since they obfuscate the code.
+ */
+
+#ifndef ARCHIVE_PLATFORM_H_INCLUDED
+#define ARCHIVE_PLATFORM_H_INCLUDED
+
+/* archive.h and archive_entry.h require this. */
+#define __LIBARCHIVE_BUILD 1
+
+#if defined(PLATFORM_CONFIG_H)
+/* Use hand-built config.h in environments that need it. */
+#include PLATFORM_CONFIG_H
+#elif defined(HAVE_CONFIG_H)
+/* Most POSIX platforms use the 'configure' script to build config.h */
+#include "config.h"
+#else
+/* Warn if the library hasn't been (automatically or manually) configured. */
+#error Oops: No config.h and no pre-built configuration in archive_platform.h.
+#endif
+
+/* It should be possible to get rid of this by extending the feature-test
+ * macros to cover Windows API functions, probably along with non-trivial
+ * refactoring of code to find structures that sit more cleanly on top of
+ * either Windows or Posix APIs. */
+#if (defined(__WIN32__) || defined(_WIN32) || defined(__WIN32)) && !defined(__CYGWIN__)
+#include "archive_windows.h"
+#endif
+
+/*
+ * The config files define a lot of feature macros. The following
+ * uses those macros to select/define replacements and include key
+ * headers as required.
+ */
+
+/* Get a real definition for __FBSDID if we can */
+#if HAVE_SYS_CDEFS_H
+#include <sys/cdefs.h>
+#endif
+
+/* If not, define it so as to avoid dangling semicolons. */
+#ifndef __FBSDID
+#define __FBSDID(a) struct _undefined_hack
+#endif
+
+/* Try to get standard C99-style integer type definitions. */
+#if HAVE_INTTYPES_H
+#include <inttypes.h>
+#endif
+#if HAVE_STDINT_H
+#include <stdint.h>
+#endif
+
+/* Borland warns about its own constants! */
+#if defined(__BORLANDC__)
+# if HAVE_DECL_UINT64_MAX
+# undef UINT64_MAX
+# undef HAVE_DECL_UINT64_MAX
+# endif
+# if HAVE_DECL_UINT64_MIN
+# undef UINT64_MIN
+# undef HAVE_DECL_UINT64_MIN
+# endif
+# if HAVE_DECL_INT64_MAX
+# undef INT64_MAX
+# undef HAVE_DECL_INT64_MAX
+# endif
+# if HAVE_DECL_INT64_MIN
+# undef INT64_MIN
+# undef HAVE_DECL_INT64_MIN
+# endif
+#endif
+
+/* Some platforms lack the standard *_MAX definitions. */
+#if !HAVE_DECL_SIZE_MAX
+#define SIZE_MAX (~(size_t)0)
+#endif
+#if !HAVE_DECL_SSIZE_MAX
+#define SSIZE_MAX ((ssize_t)(SIZE_MAX >> 1))
+#endif
+#if !HAVE_DECL_UINT32_MAX
+#define UINT32_MAX (~(uint32_t)0)
+#endif
+#if !HAVE_DECL_UINT64_MAX
+#define UINT64_MAX (~(uint64_t)0)
+#endif
+#if !HAVE_DECL_INT64_MAX
+#define INT64_MAX ((int64_t)(UINT64_MAX >> 1))
+#endif
+#if !HAVE_DECL_INT64_MIN
+#define INT64_MIN ((int64_t)(~INT64_MAX))
+#endif
+
+/*
+ * If this platform has <sys/acl.h>, acl_create(), acl_init(),
+ * acl_set_file(), and ACL_USER, we assume it has the rest of the
+ * POSIX.1e draft functions used in archive_read_extract.c.
+ */
+#if HAVE_SYS_ACL_H && HAVE_ACL_CREATE_ENTRY && HAVE_ACL_INIT && HAVE_ACL_SET_FILE && HAVE_ACL_USER
+#define HAVE_POSIX_ACL 1
+#endif
+
+/*
+ * If we can't restore metadata using a file descriptor, then
+ * for compatibility's sake, close files before trying to restore metadata.
+ */
+#if defined(HAVE_FCHMOD) || defined(HAVE_FUTIMES) || defined(HAVE_ACL_SET_FD) || defined(HAVE_ACL_SET_FD_NP) || defined(HAVE_FCHOWN)
+#define CAN_RESTORE_METADATA_FD
+#endif
+
+/* Set up defaults for internal error codes. */
+#ifndef ARCHIVE_ERRNO_FILE_FORMAT
+#if HAVE_EFTYPE
+#define ARCHIVE_ERRNO_FILE_FORMAT EFTYPE
+#else
+#if HAVE_EILSEQ
+#define ARCHIVE_ERRNO_FILE_FORMAT EILSEQ
+#else
+#define ARCHIVE_ERRNO_FILE_FORMAT EINVAL
+#endif
+#endif
+#endif
+
+#ifndef ARCHIVE_ERRNO_PROGRAMMER
+#define ARCHIVE_ERRNO_PROGRAMMER EINVAL
+#endif
+
+#ifndef ARCHIVE_ERRNO_MISC
+#define ARCHIVE_ERRNO_MISC (-1)
+#endif
+
+#endif /* !ARCHIVE_PLATFORM_H_INCLUDED */
diff --git a/libarchive/archive_private.h b/libarchive/archive_private.h
new file mode 100644
index 000000000000..63384b82fed0
--- /dev/null
+++ b/libarchive/archive_private.h
@@ -0,0 +1,124 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD: head/lib/libarchive/archive_private.h 201098 2009-12-28 02:58:14Z kientzle $
+ */
+
+#ifndef __LIBARCHIVE_BUILD
+#error This header is only to be used internally to libarchive.
+#endif
+
+#ifndef ARCHIVE_PRIVATE_H_INCLUDED
+#define ARCHIVE_PRIVATE_H_INCLUDED
+
+#include "archive.h"
+#include "archive_string.h"
+
+#if defined(__GNUC__) && (__GNUC__ > 2 || \
+ (__GNUC__ == 2 && __GNUC_MINOR__ >= 5))
+#define __LA_DEAD __attribute__((__noreturn__))
+#else
+#define __LA_DEAD
+#endif
+
+#define ARCHIVE_WRITE_MAGIC (0xb0c5c0deU)
+#define ARCHIVE_READ_MAGIC (0xdeb0c5U)
+#define ARCHIVE_WRITE_DISK_MAGIC (0xc001b0c5U)
+#define ARCHIVE_READ_DISK_MAGIC (0xbadb0c5U)
+
+#define ARCHIVE_STATE_ANY 0xFFFFU
+#define ARCHIVE_STATE_NEW 1U
+#define ARCHIVE_STATE_HEADER 2U
+#define ARCHIVE_STATE_DATA 4U
+#define ARCHIVE_STATE_DATA_END 8U
+#define ARCHIVE_STATE_EOF 0x10U
+#define ARCHIVE_STATE_CLOSED 0x20U
+#define ARCHIVE_STATE_FATAL 0x8000U
+
+struct archive_vtable {
+ int (*archive_close)(struct archive *);
+ int (*archive_finish)(struct archive *);
+ int (*archive_write_header)(struct archive *,
+ struct archive_entry *);
+ int (*archive_write_finish_entry)(struct archive *);
+ ssize_t (*archive_write_data)(struct archive *,
+ const void *, size_t);
+ ssize_t (*archive_write_data_block)(struct archive *,
+ const void *, size_t, off_t);
+};
+
+struct archive {
+ /*
+ * The magic/state values are used to sanity-check the
+ * client's usage. If an API function is called at a
+ * ridiculous time, or the client passes us an invalid
+ * pointer, these values allow me to catch that.
+ */
+ unsigned int magic;
+ unsigned int state;
+
+ /*
+ * Some public API functions depend on the "real" type of the
+ * archive object.
+ */
+ struct archive_vtable *vtable;
+
+ int archive_format;
+ const char *archive_format_name;
+
+ int compression_code; /* Currently active compression. */
+ const char *compression_name;
+
+ /* Position in UNCOMPRESSED data stream. */
+ int64_t file_position;
+ /* Position in COMPRESSED data stream. */
+ int64_t raw_position;
+ /* Number of file entries processed. */
+ int file_count;
+
+ int archive_error_number;
+ const char *error;
+ struct archive_string error_string;
+};
+
+/* Check magic value and state; exit if it isn't valid. */
+void __archive_check_magic(struct archive *, unsigned int magic,
+ unsigned int state, const char *func);
+
+void __archive_errx(int retvalue, const char *msg) __LA_DEAD;
+
+int __archive_parse_options(const char *p, const char *fn,
+ int keysize, char *key, int valsize, char *val);
+
+#define err_combine(a,b) ((a) < (b) ? (a) : (b))
+
+#if defined(__BORLANDC__) || (defined(_MSC_VER) && _MSC_VER <= 1300)
+# define ARCHIVE_LITERAL_LL(x) x##i64
+# define ARCHIVE_LITERAL_ULL(x) x##ui64
+#else
+# define ARCHIVE_LITERAL_LL(x) x##ll
+# define ARCHIVE_LITERAL_ULL(x) x##ull
+#endif
+
+#endif
diff --git a/libarchive/archive_read.3 b/libarchive/archive_read.3
new file mode 100644
index 000000000000..5d5f539107ef
--- /dev/null
+++ b/libarchive/archive_read.3
@@ -0,0 +1,714 @@
+.\" Copyright (c) 2003-2007 Tim Kientzle
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\" notice, this list of conditions and the following disclaimer.
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\" notice, this list of conditions and the following disclaimer in the
+.\" documentation and/or other materials provided with the distribution.
+.\"
+.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+.\" SUCH DAMAGE.
+.\"
+.\" $FreeBSD: head/lib/libarchive/archive_read.3 191595 2009-04-27 20:13:13Z kientzle $
+.\"
+.Dd April 13, 2009
+.Dt archive_read 3
+.Os
+.Sh NAME
+.Nm archive_read_new ,
+.Nm archive_read_set_filter_options ,
+.Nm archive_read_set_format_options ,
+.Nm archive_read_set_options ,
+.Nm archive_read_support_compression_all ,
+.Nm archive_read_support_compression_bzip2 ,
+.Nm archive_read_support_compression_compress ,
+.Nm archive_read_support_compression_gzip ,
+.Nm archive_read_support_compression_lzma ,
+.Nm archive_read_support_compression_none ,
+.Nm archive_read_support_compression_xz ,
+.Nm archive_read_support_compression_program ,
+.Nm archive_read_support_compression_program_signature ,
+.Nm archive_read_support_format_all ,
+.Nm archive_read_support_format_ar ,
+.Nm archive_read_support_format_cpio ,
+.Nm archive_read_support_format_empty ,
+.Nm archive_read_support_format_iso9660 ,
+.Nm archive_read_support_format_mtree,
+.Nm archive_read_support_format_raw,
+.Nm archive_read_support_format_tar ,
+.Nm archive_read_support_format_zip ,
+.Nm archive_read_open ,
+.Nm archive_read_open2 ,
+.Nm archive_read_open_fd ,
+.Nm archive_read_open_FILE ,
+.Nm archive_read_open_filename ,
+.Nm archive_read_open_memory ,
+.Nm archive_read_next_header ,
+.Nm archive_read_next_header2 ,
+.Nm archive_read_data ,
+.Nm archive_read_data_block ,
+.Nm archive_read_data_skip ,
+.\" #if ARCHIVE_API_VERSION < 3
+.Nm archive_read_data_into_buffer ,
+.\" #endif
+.Nm archive_read_data_into_fd ,
+.Nm archive_read_extract ,
+.Nm archive_read_extract2 ,
+.Nm archive_read_extract_set_progress_callback ,
+.Nm archive_read_close ,
+.Nm archive_read_finish
+.Nd functions for reading streaming archives
+.Sh SYNOPSIS
+.In archive.h
+.Ft struct archive *
+.Fn archive_read_new "void"
+.Ft int
+.Fn archive_read_support_compression_all "struct archive *"
+.Ft int
+.Fn archive_read_support_compression_bzip2 "struct archive *"
+.Ft int
+.Fn archive_read_support_compression_compress "struct archive *"
+.Ft int
+.Fn archive_read_support_compression_gzip "struct archive *"
+.Ft int
+.Fn archive_read_support_compression_lzma "struct archive *"
+.Ft int
+.Fn archive_read_support_compression_none "struct archive *"
+.Ft int
+.Fn archive_read_support_compression_xz "struct archive *"
+.Ft int
+.Fo archive_read_support_compression_program
+.Fa "struct archive *"
+.Fa "const char *cmd"
+.Fc
+.Ft int
+.Fo archive_read_support_compression_program_signature
+.Fa "struct archive *"
+.Fa "const char *cmd"
+.Fa "const void *signature"
+.Fa "size_t signature_length"
+.Fc
+.Ft int
+.Fn archive_read_support_format_all "struct archive *"
+.Ft int
+.Fn archive_read_support_format_ar "struct archive *"
+.Ft int
+.Fn archive_read_support_format_cpio "struct archive *"
+.Ft int
+.Fn archive_read_support_format_empty "struct archive *"
+.Ft int
+.Fn archive_read_support_format_iso9660 "struct archive *"
+.Ft int
+.Fn archive_read_support_format_mtree "struct archive *"
+.Ft int
+.Fn archive_read_support_format_raw "struct archive *"
+.Ft int
+.Fn archive_read_support_format_tar "struct archive *"
+.Ft int
+.Fn archive_read_support_format_zip "struct archive *"
+.Ft int
+.Fn archive_read_set_filter_options "struct archive *" "const char *"
+.Ft int
+.Fn archive_read_set_format_options "struct archive *" "const char *"
+.Ft int
+.Fn archive_read_set_options "struct archive *" "const char *"
+.Ft int
+.Fo archive_read_open
+.Fa "struct archive *"
+.Fa "void *client_data"
+.Fa "archive_open_callback *"
+.Fa "archive_read_callback *"
+.Fa "archive_close_callback *"
+.Fc
+.Ft int
+.Fo archive_read_open2
+.Fa "struct archive *"
+.Fa "void *client_data"
+.Fa "archive_open_callback *"
+.Fa "archive_read_callback *"
+.Fa "archive_skip_callback *"
+.Fa "archive_close_callback *"
+.Fc
+.Ft int
+.Fn archive_read_open_FILE "struct archive *" "FILE *file"
+.Ft int
+.Fn archive_read_open_fd "struct archive *" "int fd" "size_t block_size"
+.Ft int
+.Fo archive_read_open_filename
+.Fa "struct archive *"
+.Fa "const char *filename"
+.Fa "size_t block_size"
+.Fc
+.Ft int
+.Fn archive_read_open_memory "struct archive *" "void *buff" "size_t size"
+.Ft int
+.Fn archive_read_next_header "struct archive *" "struct archive_entry **"
+.Ft int
+.Fn archive_read_next_header2 "struct archive *" "struct archive_entry *"
+.Ft ssize_t
+.Fn archive_read_data "struct archive *" "void *buff" "size_t len"
+.Ft int
+.Fo archive_read_data_block
+.Fa "struct archive *"
+.Fa "const void **buff"
+.Fa "size_t *len"
+.Fa "off_t *offset"
+.Fc
+.Ft int
+.Fn archive_read_data_skip "struct archive *"
+.\" #if ARCHIVE_API_VERSION < 3
+.Ft int
+.Fn archive_read_data_into_buffer "struct archive *" "void *" "ssize_t len"
+.\" #endif
+.Ft int
+.Fn archive_read_data_into_fd "struct archive *" "int fd"
+.Ft int
+.Fo archive_read_extract
+.Fa "struct archive *"
+.Fa "struct archive_entry *"
+.Fa "int flags"
+.Fc
+.Ft int
+.Fo archive_read_extract2
+.Fa "struct archive *src"
+.Fa "struct archive_entry *"
+.Fa "struct archive *dest"
+.Fc
+.Ft void
+.Fo archive_read_extract_set_progress_callback
+.Fa "struct archive *"
+.Fa "void (*func)(void *)"
+.Fa "void *user_data"
+.Fc
+.Ft int
+.Fn archive_read_close "struct archive *"
+.Ft int
+.Fn archive_read_finish "struct archive *"
+.Sh DESCRIPTION
+These functions provide a complete API for reading streaming archives.
+The general process is to first create the
+.Tn struct archive
+object, set options, initialize the reader, iterate over the archive
+headers and associated data, then close the archive and release all
+resources.
+The following summary describes the functions in approximately the
+order they would be used:
+.Bl -tag -compact -width indent
+.It Fn archive_read_new
+Allocates and initializes a
+.Tn struct archive
+object suitable for reading from an archive.
+.It Xo
+.Fn archive_read_support_compression_bzip2 ,
+.Fn archive_read_support_compression_compress ,
+.Fn archive_read_support_compression_gzip ,
+.Fn archive_read_support_compression_lzma ,
+.Fn archive_read_support_compression_none ,
+.Fn archive_read_support_compression_xz
+.Xc
+Enables auto-detection code and decompression support for the
+specified compression.
+Returns
+.Cm ARCHIVE_OK
+if the compression is fully supported, or
+.Cm ARCHIVE_WARN
+if the compression is supported only through an external program.
+Note that decompression using an external program is usually slower than
+decompression through built-in libraries.
+Note that
+.Dq none
+is always enabled by default.
+.It Fn archive_read_support_compression_all
+Enables all available decompression filters.
+.It Fn archive_read_support_compression_program
+Data is fed through the specified external program before being dearchived.
+Note that this disables automatic detection of the compression format,
+so it makes no sense to specify this in conjunction with any other
+decompression option.
+.It Fn archive_read_support_compression_program_signature
+This feeds data through the specified external program
+but only if the initial bytes of the data match the specified
+signature value.
+.It Xo
+.Fn archive_read_support_format_all ,
+.Fn archive_read_support_format_ar ,
+.Fn archive_read_support_format_cpio ,
+.Fn archive_read_support_format_empty ,
+.Fn archive_read_support_format_iso9660 ,
+.Fn archive_read_support_format_mtree ,
+.Fn archive_read_support_format_tar ,
+.Fn archive_read_support_format_zip
+.Xc
+Enables support---including auto-detection code---for the
+specified archive format.
+For example,
+.Fn archive_read_support_format_tar
+enables support for a variety of standard tar formats, old-style tar,
+ustar, pax interchange format, and many common variants.
+For convenience,
+.Fn archive_read_support_format_all
+enables support for all available formats.
+Only empty archives are supported by default.
+.It Fn archive_read_support_format_raw
+The
+.Dq raw
+format handler allows libarchive to be used to read arbitrary data.
+It treats any data stream as an archive with a single entry.
+The pathname of this entry is
+.Dq data ;
+all other entry fields are unset.
+This is not enabled by
+.Fn archive_read_support_format_all
+in order to avoid erroneous handling of damaged archives.
+.It Xo
+.Fn archive_read_set_filter_options ,
+.Fn archive_read_set_format_options ,
+.Fn archive_read_set_options
+.Xc
+Specifies options that will be passed to currently-registered
+filters (including decompression filters) and/or format readers.
+The argument is a comma-separated list of individual options.
+Individual options have one of the following forms:
+.Bl -tag -compact -width indent
+.It Ar option=value
+The option/value pair will be provided to every module.
+Modules that do not accept an option with this name will ignore it.
+.It Ar option
+The option will be provided to every module with a value of
+.Dq 1 .
+.It Ar !option
+The option will be provided to every module with a NULL value.
+.It Ar module:option=value , Ar module:option , Ar module:!option
+As above, but the corresponding option and value will be provided
+only to modules whose name matches
+.Ar module .
+.El
+The return value will be
+.Cm ARCHIVE_OK
+if any module accepts the option, or
+.Cm ARCHIVE_WARN
+if no module accepted the option, or
+.Cm ARCHIVE_FATAL
+if there was a fatal error while attempting to process the option.
+.Pp
+The currently supported options are:
+.Bl -tag -compact -width indent
+.It Format iso9660
+.Bl -tag -compact -width indent
+.It Cm joliet
+Support Joliet extensions.
+Defaults to enabled, use
+.Cm !joliet
+to disable.
+.El
+.El
+.It Fn archive_read_open
+The same as
+.Fn archive_read_open2 ,
+except that the skip callback is assumed to be
+.Dv NULL .
+.It Fn archive_read_open2
+Freeze the settings, open the archive, and prepare for reading entries.
+This is the most generic version of this call, which accepts
+four callback functions.
+Most clients will want to use
+.Fn archive_read_open_filename ,
+.Fn archive_read_open_FILE ,
+.Fn archive_read_open_fd ,
+or
+.Fn archive_read_open_memory
+instead.
+The library invokes the client-provided functions to obtain
+raw bytes from the archive.
+.It Fn archive_read_open_FILE
+Like
+.Fn archive_read_open ,
+except that it accepts a
+.Ft "FILE *"
+pointer.
+This function should not be used with tape drives or other devices
+that require strict I/O blocking.
+.It Fn archive_read_open_fd
+Like
+.Fn archive_read_open ,
+except that it accepts a file descriptor and block size rather than
+a set of function pointers.
+Note that the file descriptor will not be automatically closed at
+end-of-archive.
+This function is safe for use with tape drives or other blocked devices.
+.It Fn archive_read_open_file
+This is a deprecated synonym for
+.Fn archive_read_open_filename .
+.It Fn archive_read_open_filename
+Like
+.Fn archive_read_open ,
+except that it accepts a simple filename and a block size.
+A NULL filename represents standard input.
+This function is safe for use with tape drives or other blocked devices.
+.It Fn archive_read_open_memory
+Like
+.Fn archive_read_open ,
+except that it accepts a pointer and size of a block of
+memory containing the archive data.
+.It Fn archive_read_next_header
+Read the header for the next entry and return a pointer to
+a
+.Tn struct archive_entry .
+This is a convenience wrapper around
+.Fn archive_read_next_header2
+that reuses an internal
+.Tn struct archive_entry
+object for each request.
+.It Fn archive_read_next_header2
+Read the header for the next entry and populate the provided
+.Tn struct archive_entry .
+.It Fn archive_read_data
+Read data associated with the header just read.
+Internally, this is a convenience function that calls
+.Fn archive_read_data_block
+and fills any gaps with nulls so that callers see a single
+continuous stream of data.
+.It Fn archive_read_data_block
+Return the next available block of data for this entry.
+Unlike
+.Fn archive_read_data ,
+the
+.Fn archive_read_data_block
+function avoids copying data and allows you to correctly handle
+sparse files, as supported by some archive formats.
+The library guarantees that offsets will increase and that blocks
+will not overlap.
+Note that the blocks returned from this function can be much larger
+than the block size read from disk, due to compression
+and internal buffer optimizations.
+.It Fn archive_read_data_skip
+A convenience function that repeatedly calls
+.Fn archive_read_data_block
+to skip all of the data for this archive entry.
+.\" #if ARCHIVE_API_VERSION < 3
+.It Fn archive_read_data_into_buffer
+This function is deprecated and will be removed.
+Use
+.Fn archive_read_data
+instead.
+.\" #endif
+.It Fn archive_read_data_into_fd
+A convenience function that repeatedly calls
+.Fn archive_read_data_block
+to copy the entire entry to the provided file descriptor.
+.It Fn archive_read_extract , Fn archive_read_extract_set_skip_file
+A convenience function that wraps the corresponding
+.Xr archive_write_disk 3
+interfaces.
+The first call to
+.Fn archive_read_extract
+creates a restore object using
+.Xr archive_write_disk_new 3
+and
+.Xr archive_write_disk_set_standard_lookup 3 ,
+then transparently invokes
+.Xr archive_write_disk_set_options 3 ,
+.Xr archive_write_header 3 ,
+.Xr archive_write_data 3 ,
+and
+.Xr archive_write_finish_entry 3
+to create the entry on disk and copy data into it.
+The
+.Va flags
+argument is passed unmodified to
+.Xr archive_write_disk_set_options 3 .
+.It Fn archive_read_extract2
+This is another version of
+.Fn archive_read_extract
+that allows you to provide your own restore object.
+In particular, this allows you to override the standard lookup functions
+using
+.Xr archive_write_disk_set_group_lookup 3 ,
+and
+.Xr archive_write_disk_set_user_lookup 3 .
+Note that
+.Fn archive_read_extract2
+does not accept a
+.Va flags
+argument; you should use
+.Fn archive_write_disk_set_options
+to set the restore options yourself.
+.It Fn archive_read_extract_set_progress_callback
+Sets a pointer to a user-defined callback that can be used
+for updating progress displays during extraction.
+The progress function will be invoked during the extraction of large
+regular files.
+The progress function will be invoked with the pointer provided to this call.
+Generally, the data pointed to should include a reference to the archive
+object and the archive_entry object so that various statistics
+can be retrieved for the progress display.
+.It Fn archive_read_close
+Complete the archive and invoke the close callback.
+.It Fn archive_read_finish
+Invokes
+.Fn archive_read_close
+if it was not invoked manually, then release all resources.
+Note: In libarchive 1.x, this function was declared to return
+.Ft void ,
+which made it impossible to detect certain errors when
+.Fn archive_read_close
+was invoked implicitly from this function.
+The declaration is corrected beginning with libarchive 2.0.
+.El
+.Pp
+Note that the library determines most of the relevant information about
+the archive by inspection.
+In particular, it automatically detects
+.Xr gzip 1
+or
+.Xr bzip2 1
+compression and transparently performs the appropriate decompression.
+It also automatically detects the archive format.
+.Pp
+A complete description of the
+.Tn struct archive
+and
+.Tn struct archive_entry
+objects can be found in the overview manual page for
+.Xr libarchive 3 .
+.Sh CLIENT CALLBACKS
+The callback functions must match the following prototypes:
+.Bl -item -offset indent
+.It
+.Ft typedef ssize_t
+.Fo archive_read_callback
+.Fa "struct archive *"
+.Fa "void *client_data"
+.Fa "const void **buffer"
+.Fc
+.It
+.\" #if ARCHIVE_API_VERSION < 2
+.Ft typedef int
+.Fo archive_skip_callback
+.Fa "struct archive *"
+.Fa "void *client_data"
+.Fa "size_t request"
+.Fc
+.\" #else
+.\" .Ft typedef off_t
+.\" .Fo archive_skip_callback
+.\" .Fa "struct archive *"
+.\" .Fa "void *client_data"
+.\" .Fa "off_t request"
+.\" .Fc
+.\" #endif
+.It
+.Ft typedef int
+.Fn archive_open_callback "struct archive *" "void *client_data"
+.It
+.Ft typedef int
+.Fn archive_close_callback "struct archive *" "void *client_data"
+.El
+.Pp
+The open callback is invoked by
+.Fn archive_open .
+It should return
+.Cm ARCHIVE_OK
+if the underlying file or data source is successfully
+opened.
+If the open fails, it should call
+.Fn archive_set_error
+to register an error code and message and return
+.Cm ARCHIVE_FATAL .
+.Pp
+The read callback is invoked whenever the library
+requires raw bytes from the archive.
+The read callback should read data into a buffer,
+set the
+.Li const void **buffer
+argument to point to the available data, and
+return a count of the number of bytes available.
+The library will invoke the read callback again
+only after it has consumed this data.
+The library imposes no constraints on the size
+of the data blocks returned.
+On end-of-file, the read callback should
+return zero.
+On error, the read callback should invoke
+.Fn archive_set_error
+to register an error code and message and
+return -1.
+.Pp
+The skip callback is invoked when the
+library wants to ignore a block of data.
+The return value is the number of bytes actually
+skipped, which may differ from the request.
+If the callback cannot skip data, it should return
+zero.
+If the skip callback is not provided (the
+function pointer is
+.Dv NULL ),
+the library will invoke the read function
+instead and simply discard the result.
+A skip callback can provide significant
+performance gains when reading uncompressed
+archives from slow disk drives or other media
+that can skip quickly.
+.Pp
+The close callback is invoked by archive_close when
+the archive processing is complete.
+The callback should return
+.Cm ARCHIVE_OK
+on success.
+On failure, the callback should invoke
+.Fn archive_set_error
+to register an error code and message and
+return
+.Cm ARCHIVE_FATAL.
+.Sh EXAMPLE
+The following illustrates basic usage of the library.
+In this example,
+the callback functions are simply wrappers around the standard
+.Xr open 2 ,
+.Xr read 2 ,
+and
+.Xr close 2
+system calls.
+.Bd -literal -offset indent
+void
+list_archive(const char *name)
+{
+ struct mydata *mydata;
+ struct archive *a;
+ struct archive_entry *entry;
+
+ mydata = malloc(sizeof(struct mydata));
+ a = archive_read_new();
+ mydata->name = name;
+ archive_read_support_compression_all(a);
+ archive_read_support_format_all(a);
+ archive_read_open(a, mydata, myopen, myread, myclose);
+ while (archive_read_next_header(a, &entry) == ARCHIVE_OK) {
+ printf("%s\en",archive_entry_pathname(entry));
+ archive_read_data_skip(a);
+ }
+ archive_read_finish(a);
+ free(mydata);
+}
+
+ssize_t
+myread(struct archive *a, void *client_data, const void **buff)
+{
+ struct mydata *mydata = client_data;
+
+ *buff = mydata->buff;
+ return (read(mydata->fd, mydata->buff, 10240));
+}
+
+int
+myopen(struct archive *a, void *client_data)
+{
+ struct mydata *mydata = client_data;
+
+ mydata->fd = open(mydata->name, O_RDONLY);
+ return (mydata->fd >= 0 ? ARCHIVE_OK : ARCHIVE_FATAL);
+}
+
+int
+myclose(struct archive *a, void *client_data)
+{
+ struct mydata *mydata = client_data;
+
+ if (mydata->fd > 0)
+ close(mydata->fd);
+ return (ARCHIVE_OK);
+}
+.Ed
+.Sh RETURN VALUES
+Most functions return zero on success, non-zero on error.
+The possible return codes include:
+.Cm ARCHIVE_OK
+(the operation succeeded),
+.Cm ARCHIVE_WARN
+(the operation succeeded but a non-critical error was encountered),
+.Cm ARCHIVE_EOF
+(end-of-archive was encountered),
+.Cm ARCHIVE_RETRY
+(the operation failed but can be retried),
+and
+.Cm ARCHIVE_FATAL
+(there was a fatal error; the archive should be closed immediately).
+Detailed error codes and textual descriptions are available from the
+.Fn archive_errno
+and
+.Fn archive_error_string
+functions.
+.Pp
+.Fn archive_read_new
+returns a pointer to a freshly allocated
+.Tn struct archive
+object.
+It returns
+.Dv NULL
+on error.
+.Pp
+.Fn archive_read_data
+returns a count of bytes actually read or zero at the end of the entry.
+On error, a value of
+.Cm ARCHIVE_FATAL ,
+.Cm ARCHIVE_WARN ,
+or
+.Cm ARCHIVE_RETRY
+is returned and an error code and textual description can be retrieved from the
+.Fn archive_errno
+and
+.Fn archive_error_string
+functions.
+.Pp
+The library expects the client callbacks to behave similarly.
+If there is an error, you can use
+.Fn archive_set_error
+to set an appropriate error code and description,
+then return one of the non-zero values above.
+(Note that the value eventually returned to the client may
+not be the same; many errors that are not critical at the level
+of basic I/O can prevent the archive from being properly read,
+thus most I/O errors eventually cause
+.Cm ARCHIVE_FATAL
+to be returned.)
+.\" .Sh ERRORS
+.Sh SEE ALSO
+.Xr tar 1 ,
+.Xr archive 3 ,
+.Xr archive_util 3 ,
+.Xr tar 5
+.Sh HISTORY
+The
+.Nm libarchive
+library first appeared in
+.Fx 5.3 .
+.Sh AUTHORS
+.An -nosplit
+The
+.Nm libarchive
+library was written by
+.An Tim Kientzle Aq kientzle@acm.org .
+.Sh BUGS
+Many traditional archiver programs treat
+empty files as valid empty archives.
+For example, many implementations of
+.Xr tar 1
+allow you to append entries to an empty file.
+Of course, it is impossible to determine the format of an empty file
+by inspecting the contents, so this library treats empty files as
+having a special
+.Dq empty
+format.
diff --git a/libarchive/archive_read.c b/libarchive/archive_read.c
new file mode 100644
index 000000000000..b1558b666f10
--- /dev/null
+++ b/libarchive/archive_read.c
@@ -0,0 +1,1249 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/*
+ * This file contains the "essential" portions of the read API, that
+ * is, stuff that will probably always be used by any client that
+ * actually needs to read an archive. Optional pieces have been, as
+ * far as possible, separated out into separate files to avoid
+ * needlessly bloating statically-linked clients.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read.c 201157 2009-12-29 05:30:23Z kientzle $");
+
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#include <stdio.h>
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+
+#include "archive.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+#define minimum(a, b) (a < b ? a : b)
+
+static int build_stream(struct archive_read *);
+static int choose_format(struct archive_read *);
+static int cleanup_filters(struct archive_read *);
+static struct archive_vtable *archive_read_vtable(void);
+static int _archive_read_close(struct archive *);
+static int _archive_read_finish(struct archive *);
+
+static struct archive_vtable *
+archive_read_vtable(void)
+{
+ static struct archive_vtable av;
+ static int inited = 0;
+
+ if (!inited) {
+ av.archive_finish = _archive_read_finish;
+ av.archive_close = _archive_read_close;
+ }
+ return (&av);
+}
+
+/*
+ * Allocate, initialize and return a struct archive object.
+ */
+struct archive *
+archive_read_new(void)
+{
+ struct archive_read *a;
+
+ a = (struct archive_read *)malloc(sizeof(*a));
+ if (a == NULL)
+ return (NULL);
+ memset(a, 0, sizeof(*a));
+ a->archive.magic = ARCHIVE_READ_MAGIC;
+
+ a->archive.state = ARCHIVE_STATE_NEW;
+ a->entry = archive_entry_new();
+ a->archive.vtable = archive_read_vtable();
+
+ return (&a->archive);
+}
+
+/*
+ * Record the do-not-extract-to file. This belongs in archive_read_extract.c.
+ */
+void
+archive_read_extract_set_skip_file(struct archive *_a, dev_t d, ino_t i)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_ANY,
+ "archive_read_extract_set_skip_file");
+ a->skip_file_dev = d;
+ a->skip_file_ino = i;
+}
+
+/*
+ * Set read options for the format.
+ */
+int
+archive_read_set_format_options(struct archive *_a, const char *s)
+{
+ struct archive_read *a;
+ struct archive_format_descriptor *format;
+ char key[64], val[64];
+ char *valp;
+ size_t i;
+ int len, r;
+
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW,
+ "archive_read_set_format_options");
+
+ if (s == NULL || *s == '\0')
+ return (ARCHIVE_OK);
+ a = (struct archive_read *)_a;
+ __archive_check_magic(&a->archive, ARCHIVE_READ_MAGIC,
+ ARCHIVE_STATE_NEW, "archive_read_set_format_options");
+ len = 0;
+ for (i = 0; i < sizeof(a->formats)/sizeof(a->formats[0]); i++) {
+ format = &a->formats[i];
+ if (format == NULL || format->options == NULL ||
+ format->name == NULL)
+ /* This format does not support option. */
+ continue;
+
+ while ((len = __archive_parse_options(s, format->name,
+ sizeof(key), key, sizeof(val), val)) > 0) {
+ valp = val[0] == '\0' ? NULL : val;
+ a->format = format;
+ r = format->options(a, key, valp);
+ a->format = NULL;
+ if (r == ARCHIVE_FATAL)
+ return (r);
+ s += len;
+ }
+ }
+ if (len < 0) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
+ "Illegal format options.");
+ return (ARCHIVE_WARN);
+ }
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Set read options for the filter.
+ */
+int
+archive_read_set_filter_options(struct archive *_a, const char *s)
+{
+ struct archive_read *a;
+ struct archive_read_filter *filter;
+ struct archive_read_filter_bidder *bidder;
+ char key[64], val[64];
+ int len, r;
+
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW,
+ "archive_read_set_filter_options");
+
+ if (s == NULL || *s == '\0')
+ return (ARCHIVE_OK);
+ a = (struct archive_read *)_a;
+ __archive_check_magic(&a->archive, ARCHIVE_READ_MAGIC,
+ ARCHIVE_STATE_NEW, "archive_read_set_filter_options");
+ len = 0;
+ for (filter = a->filter; filter != NULL; filter = filter->upstream) {
+ bidder = filter->bidder;
+ if (bidder == NULL)
+ continue;
+ if (bidder->options == NULL)
+ /* This bidder does not support option */
+ continue;
+ while ((len = __archive_parse_options(s, filter->name,
+ sizeof(key), key, sizeof(val), val)) > 0) {
+ if (val[0] == '\0')
+ r = bidder->options(bidder, key, NULL);
+ else
+ r = bidder->options(bidder, key, val);
+ if (r == ARCHIVE_FATAL)
+ return (r);
+ s += len;
+ }
+ }
+ if (len < 0) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
+ "Illegal format options.");
+ return (ARCHIVE_WARN);
+ }
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Set read options for the format and the filter.
+ */
+int
+archive_read_set_options(struct archive *_a, const char *s)
+{
+ int r;
+
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW,
+ "archive_read_set_options");
+ archive_clear_error(_a);
+
+ r = archive_read_set_format_options(_a, s);
+ if (r != ARCHIVE_OK)
+ return (r);
+ r = archive_read_set_filter_options(_a, s);
+ if (r != ARCHIVE_OK)
+ return (r);
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Open the archive
+ */
+int
+archive_read_open(struct archive *a, void *client_data,
+ archive_open_callback *client_opener, archive_read_callback *client_reader,
+ archive_close_callback *client_closer)
+{
+ /* Old archive_read_open() is just a thin shell around
+ * archive_read_open2. */
+ return archive_read_open2(a, client_data, client_opener,
+ client_reader, NULL, client_closer);
+}
+
+static ssize_t
+client_read_proxy(struct archive_read_filter *self, const void **buff)
+{
+ ssize_t r;
+ r = (self->archive->client.reader)(&self->archive->archive,
+ self->data, buff);
+ self->archive->archive.raw_position += r;
+ return (r);
+}
+
+static int64_t
+client_skip_proxy(struct archive_read_filter *self, int64_t request)
+{
+ int64_t ask, get, total;
+ /* Limit our maximum seek request to 1GB on platforms
+ * with 32-bit off_t (such as Windows). */
+ int64_t skip_limit = ((int64_t)1) << (sizeof(off_t) * 8 - 2);
+
+ if (self->archive->client.skipper == NULL)
+ return (0);
+ total = 0;
+ for (;;) {
+ ask = request;
+ if (ask > skip_limit)
+ ask = skip_limit;
+ get = (self->archive->client.skipper)(&self->archive->archive,
+ self->data, ask);
+ if (get == 0)
+ return (total);
+ request -= get;
+ self->archive->archive.raw_position += get;
+ total += get;
+ }
+}
+
+static int
+client_close_proxy(struct archive_read_filter *self)
+{
+ int r = ARCHIVE_OK;
+
+ if (self->archive->client.closer != NULL)
+ r = (self->archive->client.closer)((struct archive *)self->archive,
+ self->data);
+ self->data = NULL;
+ return (r);
+}
+
+
+int
+archive_read_open2(struct archive *_a, void *client_data,
+ archive_open_callback *client_opener,
+ archive_read_callback *client_reader,
+ archive_skip_callback *client_skipper,
+ archive_close_callback *client_closer)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct archive_read_filter *filter;
+ int e;
+
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW,
+ "archive_read_open");
+ archive_clear_error(&a->archive);
+
+ if (client_reader == NULL)
+ __archive_errx(1,
+ "No reader function provided to archive_read_open");
+
+ /* Open data source. */
+ if (client_opener != NULL) {
+ e =(client_opener)(&a->archive, client_data);
+ if (e != 0) {
+ /* If the open failed, call the closer to clean up. */
+ if (client_closer)
+ (client_closer)(&a->archive, client_data);
+ return (e);
+ }
+ }
+
+ /* Save the client functions and mock up the initial source. */
+ a->client.reader = client_reader;
+ a->client.skipper = client_skipper;
+ a->client.closer = client_closer;
+
+ filter = calloc(1, sizeof(*filter));
+ if (filter == NULL)
+ return (ARCHIVE_FATAL);
+ filter->bidder = NULL;
+ filter->upstream = NULL;
+ filter->archive = a;
+ filter->data = client_data;
+ filter->read = client_read_proxy;
+ filter->skip = client_skip_proxy;
+ filter->close = client_close_proxy;
+ filter->name = "none";
+ filter->code = ARCHIVE_COMPRESSION_NONE;
+ a->filter = filter;
+
+ /* Build out the input pipeline. */
+ e = build_stream(a);
+ if (e == ARCHIVE_OK)
+ a->archive.state = ARCHIVE_STATE_HEADER;
+
+ return (e);
+}
+
+/*
+ * Allow each registered stream transform to bid on whether
+ * it wants to handle this stream. Repeat until we've finished
+ * building the pipeline.
+ */
+static int
+build_stream(struct archive_read *a)
+{
+ int number_bidders, i, bid, best_bid;
+ struct archive_read_filter_bidder *bidder, *best_bidder;
+ struct archive_read_filter *filter;
+ ssize_t avail;
+ int r;
+
+ for (;;) {
+ number_bidders = sizeof(a->bidders) / sizeof(a->bidders[0]);
+
+ best_bid = 0;
+ best_bidder = NULL;
+
+ bidder = a->bidders;
+ for (i = 0; i < number_bidders; i++, bidder++) {
+ if (bidder->bid != NULL) {
+ bid = (bidder->bid)(bidder, a->filter);
+ if (bid > best_bid) {
+ best_bid = bid;
+ best_bidder = bidder;
+ }
+ }
+ }
+
+ /* If no bidder, we're done. */
+ if (best_bidder == NULL) {
+ /* Verify the final pipelin by asking it for some data. */
+ __archive_read_filter_ahead(a->filter, 1, &avail);
+ if (avail < 0) {
+ cleanup_filters(a);
+ return (ARCHIVE_FATAL);
+ }
+ a->archive.compression_name = a->filter->name;
+ a->archive.compression_code = a->filter->code;
+ return (ARCHIVE_OK);
+ }
+
+ filter
+ = (struct archive_read_filter *)calloc(1, sizeof(*filter));
+ if (filter == NULL)
+ return (ARCHIVE_FATAL);
+ filter->bidder = best_bidder;
+ filter->archive = a;
+ filter->upstream = a->filter;
+ a->filter = filter;
+ r = (best_bidder->init)(a->filter);
+ if (r != ARCHIVE_OK) {
+ cleanup_filters(a);
+ return (r);
+ }
+ }
+}
+
+/*
+ * Read header of next entry.
+ */
+int
+archive_read_next_header2(struct archive *_a, struct archive_entry *entry)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ int slot, ret;
+
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC,
+ ARCHIVE_STATE_HEADER | ARCHIVE_STATE_DATA,
+ "archive_read_next_header");
+
+ ++_a->file_count;
+ archive_entry_clear(entry);
+ archive_clear_error(&a->archive);
+
+ /*
+ * If no format has yet been chosen, choose one.
+ */
+ if (a->format == NULL) {
+ slot = choose_format(a);
+ if (slot < 0) {
+ a->archive.state = ARCHIVE_STATE_FATAL;
+ return (ARCHIVE_FATAL);
+ }
+ a->format = &(a->formats[slot]);
+ }
+
+ /*
+ * If client didn't consume entire data, skip any remainder
+ * (This is especially important for GNU incremental directories.)
+ */
+ if (a->archive.state == ARCHIVE_STATE_DATA) {
+ ret = archive_read_data_skip(&a->archive);
+ if (ret == ARCHIVE_EOF) {
+ archive_set_error(&a->archive, EIO, "Premature end-of-file.");
+ a->archive.state = ARCHIVE_STATE_FATAL;
+ return (ARCHIVE_FATAL);
+ }
+ if (ret != ARCHIVE_OK)
+ return (ret);
+ }
+
+ /* Record start-of-header. */
+ a->header_position = a->archive.file_position;
+
+ ret = (a->format->read_header)(a, entry);
+
+ /*
+ * EOF and FATAL are persistent at this layer. By
+ * modifying the state, we guarantee that future calls to
+ * read a header or read data will fail.
+ */
+ switch (ret) {
+ case ARCHIVE_EOF:
+ a->archive.state = ARCHIVE_STATE_EOF;
+ break;
+ case ARCHIVE_OK:
+ a->archive.state = ARCHIVE_STATE_DATA;
+ break;
+ case ARCHIVE_WARN:
+ a->archive.state = ARCHIVE_STATE_DATA;
+ break;
+ case ARCHIVE_RETRY:
+ break;
+ case ARCHIVE_FATAL:
+ a->archive.state = ARCHIVE_STATE_FATAL;
+ break;
+ }
+
+ a->read_data_output_offset = 0;
+ a->read_data_remaining = 0;
+ return (ret);
+}
+
+int
+archive_read_next_header(struct archive *_a, struct archive_entry **entryp)
+{
+ int ret;
+ struct archive_read *a = (struct archive_read *)_a;
+ *entryp = NULL;
+ ret = archive_read_next_header2(_a, a->entry);
+ *entryp = a->entry;
+ return ret;
+}
+
+/*
+ * Allow each registered format to bid on whether it wants to handle
+ * the next entry. Return index of winning bidder.
+ */
+static int
+choose_format(struct archive_read *a)
+{
+ int slots;
+ int i;
+ int bid, best_bid;
+ int best_bid_slot;
+
+ slots = sizeof(a->formats) / sizeof(a->formats[0]);
+ best_bid = -1;
+ best_bid_slot = -1;
+
+ /* Set up a->format and a->pformat_data for convenience of bidders. */
+ a->format = &(a->formats[0]);
+ for (i = 0; i < slots; i++, a->format++) {
+ if (a->format->bid) {
+ bid = (a->format->bid)(a);
+ if (bid == ARCHIVE_FATAL)
+ return (ARCHIVE_FATAL);
+ if ((bid > best_bid) || (best_bid_slot < 0)) {
+ best_bid = bid;
+ best_bid_slot = i;
+ }
+ }
+ }
+
+ /*
+ * There were no bidders; this is a serious programmer error
+ * and demands a quick and definitive abort.
+ */
+ if (best_bid_slot < 0)
+ __archive_errx(1, "No formats were registered; you must "
+ "invoke at least one archive_read_support_format_XXX "
+ "function in order to successfully read an archive.");
+
+ /*
+ * There were bidders, but no non-zero bids; this means we
+ * can't support this stream.
+ */
+ if (best_bid < 1) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
+ "Unrecognized archive format");
+ return (ARCHIVE_FATAL);
+ }
+
+ return (best_bid_slot);
+}
+
+/*
+ * Return the file offset (within the uncompressed data stream) where
+ * the last header started.
+ */
+int64_t
+archive_read_header_position(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC,
+ ARCHIVE_STATE_ANY, "archive_read_header_position");
+ return (a->header_position);
+}
+
+/*
+ * Read data from an archive entry, using a read(2)-style interface.
+ * This is a convenience routine that just calls
+ * archive_read_data_block and copies the results into the client
+ * buffer, filling any gaps with zero bytes. Clients using this
+ * API can be completely ignorant of sparse-file issues; sparse files
+ * will simply be padded with nulls.
+ *
+ * DO NOT intermingle calls to this function and archive_read_data_block
+ * to read a single entry body.
+ */
+ssize_t
+archive_read_data(struct archive *_a, void *buff, size_t s)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ char *dest;
+ const void *read_buf;
+ size_t bytes_read;
+ size_t len;
+ int r;
+
+ bytes_read = 0;
+ dest = (char *)buff;
+
+ while (s > 0) {
+ if (a->read_data_remaining == 0) {
+ read_buf = a->read_data_block;
+ r = archive_read_data_block(&a->archive, &read_buf,
+ &a->read_data_remaining, &a->read_data_offset);
+ a->read_data_block = read_buf;
+ if (r == ARCHIVE_EOF)
+ return (bytes_read);
+ /*
+ * Error codes are all negative, so the status
+ * return here cannot be confused with a valid
+ * byte count. (ARCHIVE_OK is zero.)
+ */
+ if (r < ARCHIVE_OK)
+ return (r);
+ }
+
+ if (a->read_data_offset < a->read_data_output_offset) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT,
+ "Encountered out-of-order sparse blocks");
+ return (ARCHIVE_RETRY);
+ }
+
+ /* Compute the amount of zero padding needed. */
+ if (a->read_data_output_offset + (off_t)s <
+ a->read_data_offset) {
+ len = s;
+ } else if (a->read_data_output_offset <
+ a->read_data_offset) {
+ len = a->read_data_offset -
+ a->read_data_output_offset;
+ } else
+ len = 0;
+
+ /* Add zeroes. */
+ memset(dest, 0, len);
+ s -= len;
+ a->read_data_output_offset += len;
+ dest += len;
+ bytes_read += len;
+
+ /* Copy data if there is any space left. */
+ if (s > 0) {
+ len = a->read_data_remaining;
+ if (len > s)
+ len = s;
+ memcpy(dest, a->read_data_block, len);
+ s -= len;
+ a->read_data_block += len;
+ a->read_data_remaining -= len;
+ a->read_data_output_offset += len;
+ a->read_data_offset += len;
+ dest += len;
+ bytes_read += len;
+ }
+ }
+ return (bytes_read);
+}
+
+#if ARCHIVE_API_VERSION < 3
+/*
+ * Obsolete function provided for compatibility only. Note that the API
+ * of this function doesn't allow the caller to detect if the remaining
+ * data from the archive entry is shorter than the buffer provided, or
+ * even if an error occurred while reading data.
+ */
+int
+archive_read_data_into_buffer(struct archive *a, void *d, ssize_t len)
+{
+
+ archive_read_data(a, d, len);
+ return (ARCHIVE_OK);
+}
+#endif
+
+/*
+ * Skip over all remaining data in this entry.
+ */
+int
+archive_read_data_skip(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ int r;
+ const void *buff;
+ size_t size;
+ off_t offset;
+
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_DATA,
+ "archive_read_data_skip");
+
+ if (a->format->read_data_skip != NULL)
+ r = (a->format->read_data_skip)(a);
+ else {
+ while ((r = archive_read_data_block(&a->archive,
+ &buff, &size, &offset))
+ == ARCHIVE_OK)
+ ;
+ }
+
+ if (r == ARCHIVE_EOF)
+ r = ARCHIVE_OK;
+
+ a->archive.state = ARCHIVE_STATE_HEADER;
+ return (r);
+}
+
+/*
+ * Read the next block of entry data from the archive.
+ * This is a zero-copy interface; the client receives a pointer,
+ * size, and file offset of the next available block of data.
+ *
+ * Returns ARCHIVE_OK if the operation is successful, ARCHIVE_EOF if
+ * the end of entry is encountered.
+ */
+int
+archive_read_data_block(struct archive *_a,
+ const void **buff, size_t *size, off_t *offset)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_DATA,
+ "archive_read_data_block");
+
+ if (a->format->read_data == NULL) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_PROGRAMMER,
+ "Internal error: "
+ "No format_read_data_block function registered");
+ return (ARCHIVE_FATAL);
+ }
+
+ return (a->format->read_data)(a, buff, size, offset);
+}
+
+/*
+ * Close the file and release most resources.
+ *
+ * Be careful: client might just call read_new and then read_finish.
+ * Don't assume we actually read anything or performed any non-trivial
+ * initialization.
+ */
+static int
+_archive_read_close(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ int r = ARCHIVE_OK, r1 = ARCHIVE_OK;
+ size_t i, n;
+
+ __archive_check_magic(&a->archive, ARCHIVE_READ_MAGIC,
+ ARCHIVE_STATE_ANY, "archive_read_close");
+ archive_clear_error(&a->archive);
+ a->archive.state = ARCHIVE_STATE_CLOSED;
+
+
+ /* Call cleanup functions registered by optional components. */
+ if (a->cleanup_archive_extract != NULL)
+ r = (a->cleanup_archive_extract)(a);
+
+ /* TODO: Clean up the formatters. */
+
+ /* Release the filter objects. */
+ r1 = cleanup_filters(a);
+ if (r1 < r)
+ r = r1;
+
+ /* Release the bidder objects. */
+ n = sizeof(a->bidders)/sizeof(a->bidders[0]);
+ for (i = 0; i < n; i++) {
+ if (a->bidders[i].free != NULL) {
+ r1 = (a->bidders[i].free)(&a->bidders[i]);
+ if (r1 < r)
+ r = r1;
+ }
+ }
+
+ return (r);
+}
+
+static int
+cleanup_filters(struct archive_read *a)
+{
+ int r = ARCHIVE_OK;
+ /* Clean up the filter pipeline. */
+ while (a->filter != NULL) {
+ struct archive_read_filter *t = a->filter->upstream;
+ if (a->filter->close != NULL) {
+ int r1 = (a->filter->close)(a->filter);
+ if (r1 < r)
+ r = r1;
+ }
+ free(a->filter->buffer);
+ free(a->filter);
+ a->filter = t;
+ }
+ return r;
+}
+
+/*
+ * Release memory and other resources.
+ */
+static int
+_archive_read_finish(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ int i;
+ int slots;
+ int r = ARCHIVE_OK;
+
+ __archive_check_magic(_a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_ANY,
+ "archive_read_finish");
+ if (a->archive.state != ARCHIVE_STATE_CLOSED)
+ r = archive_read_close(&a->archive);
+
+ /* Cleanup format-specific data. */
+ slots = sizeof(a->formats) / sizeof(a->formats[0]);
+ for (i = 0; i < slots; i++) {
+ a->format = &(a->formats[i]);
+ if (a->formats[i].cleanup)
+ (a->formats[i].cleanup)(a);
+ }
+
+ archive_string_free(&a->archive.error_string);
+ if (a->entry)
+ archive_entry_free(a->entry);
+ a->archive.magic = 0;
+ free(a);
+#if ARCHIVE_API_VERSION > 1
+ return (r);
+#endif
+}
+
+/*
+ * Used internally by read format handlers to register their bid and
+ * initialization functions.
+ */
+int
+__archive_read_register_format(struct archive_read *a,
+ void *format_data,
+ const char *name,
+ int (*bid)(struct archive_read *),
+ int (*options)(struct archive_read *, const char *, const char *),
+ int (*read_header)(struct archive_read *, struct archive_entry *),
+ int (*read_data)(struct archive_read *, const void **, size_t *, off_t *),
+ int (*read_data_skip)(struct archive_read *),
+ int (*cleanup)(struct archive_read *))
+{
+ int i, number_slots;
+
+ __archive_check_magic(&a->archive,
+ ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW,
+ "__archive_read_register_format");
+
+ number_slots = sizeof(a->formats) / sizeof(a->formats[0]);
+
+ for (i = 0; i < number_slots; i++) {
+ if (a->formats[i].bid == bid)
+ return (ARCHIVE_WARN); /* We've already installed */
+ if (a->formats[i].bid == NULL) {
+ a->formats[i].bid = bid;
+ a->formats[i].options = options;
+ a->formats[i].read_header = read_header;
+ a->formats[i].read_data = read_data;
+ a->formats[i].read_data_skip = read_data_skip;
+ a->formats[i].cleanup = cleanup;
+ a->formats[i].data = format_data;
+ a->formats[i].name = name;
+ return (ARCHIVE_OK);
+ }
+ }
+
+ __archive_errx(1, "Not enough slots for format registration");
+ return (ARCHIVE_FATAL); /* Never actually called. */
+}
+
+/*
+ * Used internally by decompression routines to register their bid and
+ * initialization functions.
+ */
+struct archive_read_filter_bidder *
+__archive_read_get_bidder(struct archive_read *a)
+{
+ int i, number_slots;
+
+ __archive_check_magic(&a->archive,
+ ARCHIVE_READ_MAGIC, ARCHIVE_STATE_NEW,
+ "__archive_read_get_bidder");
+
+ number_slots = sizeof(a->bidders) / sizeof(a->bidders[0]);
+
+ for (i = 0; i < number_slots; i++) {
+ if (a->bidders[i].bid == NULL) {
+ memset(a->bidders + i, 0, sizeof(a->bidders[0]));
+ return (a->bidders + i);
+ }
+ }
+
+ __archive_errx(1, "Not enough slots for compression registration");
+ return (NULL); /* Never actually executed. */
+}
+
+/*
+ * The next three functions comprise the peek/consume internal I/O
+ * system used by archive format readers. This system allows fairly
+ * flexible read-ahead and allows the I/O code to operate in a
+ * zero-copy manner most of the time.
+ *
+ * In the ideal case, filters generate blocks of data
+ * and __archive_read_ahead() just returns pointers directly into
+ * those blocks. Then __archive_read_consume() just bumps those
+ * pointers. Only if your request would span blocks does the I/O
+ * layer use a copy buffer to provide you with a contiguous block of
+ * data. The __archive_read_skip() is an optimization; it scans ahead
+ * very quickly (it usually translates into a seek() operation if
+ * you're reading uncompressed disk files).
+ *
+ * A couple of useful idioms:
+ * * "I just want some data." Ask for 1 byte and pay attention to
+ * the "number of bytes available" from __archive_read_ahead().
+ * You can consume more than you asked for; you just can't consume
+ * more than is available. If you consume everything that's
+ * immediately available, the next read_ahead() call will pull
+ * the next block.
+ * * "I want to output a large block of data." As above, ask for 1 byte,
+ * emit all that's available (up to whatever limit you have), then
+ * repeat until you're done.
+ * * "I want to peek ahead by a large amount." Ask for 4k or so, then
+ * double and repeat until you get an error or have enough. Note
+ * that the I/O layer will likely end up expanding its copy buffer
+ * to fit your request, so use this technique cautiously. This
+ * technique is used, for example, by some of the format tasting
+ * code that has uncertain look-ahead needs.
+ *
+ * TODO: Someday, provide a more generic __archive_read_seek() for
+ * those cases where it's useful. This is tricky because there are lots
+ * of cases where seek() is not available (reading gzip data from a
+ * network socket, for instance), so there needs to be a good way to
+ * communicate whether seek() is available and users of that interface
+ * need to use non-seeking strategies whenever seek() is not available.
+ */
+
+/*
+ * Looks ahead in the input stream:
+ * * If 'avail' pointer is provided, that returns number of bytes available
+ * in the current buffer, which may be much larger than requested.
+ * * If end-of-file, *avail gets set to zero.
+ * * If error, *avail gets error code.
+ * * If request can be met, returns pointer to data, returns NULL
+ * if request is not met.
+ *
+ * Note: If you just want "some data", ask for 1 byte and pay attention
+ * to *avail, which will have the actual amount available. If you
+ * know exactly how many bytes you need, just ask for that and treat
+ * a NULL return as an error.
+ *
+ * Important: This does NOT move the file pointer. See
+ * __archive_read_consume() below.
+ */
+
+/*
+ * This is tricky. We need to provide our clients with pointers to
+ * contiguous blocks of memory but we want to avoid copying whenever
+ * possible.
+ *
+ * Mostly, this code returns pointers directly into the block of data
+ * provided by the client_read routine. It can do this unless the
+ * request would split across blocks. In that case, we have to copy
+ * into an internal buffer to combine reads.
+ */
+const void *
+__archive_read_ahead(struct archive_read *a, size_t min, ssize_t *avail)
+{
+ return (__archive_read_filter_ahead(a->filter, min, avail));
+}
+
+const void *
+__archive_read_filter_ahead(struct archive_read_filter *filter,
+ size_t min, ssize_t *avail)
+{
+ ssize_t bytes_read;
+ size_t tocopy;
+
+ if (filter->fatal) {
+ if (avail)
+ *avail = ARCHIVE_FATAL;
+ return (NULL);
+ }
+
+ /*
+ * Keep pulling more data until we can satisfy the request.
+ */
+ for (;;) {
+
+ /*
+ * If we can satisfy from the copy buffer (and the
+ * copy buffer isn't empty), we're done. In particular,
+ * note that min == 0 is a perfectly well-defined
+ * request.
+ */
+ if (filter->avail >= min && filter->avail > 0) {
+ if (avail != NULL)
+ *avail = filter->avail;
+ return (filter->next);
+ }
+
+ /*
+ * We can satisfy directly from client buffer if everything
+ * currently in the copy buffer is still in the client buffer.
+ */
+ if (filter->client_total >= filter->client_avail + filter->avail
+ && filter->client_avail + filter->avail >= min) {
+ /* "Roll back" to client buffer. */
+ filter->client_avail += filter->avail;
+ filter->client_next -= filter->avail;
+ /* Copy buffer is now empty. */
+ filter->avail = 0;
+ filter->next = filter->buffer;
+ /* Return data from client buffer. */
+ if (avail != NULL)
+ *avail = filter->client_avail;
+ return (filter->client_next);
+ }
+
+ /* Move data forward in copy buffer if necessary. */
+ if (filter->next > filter->buffer &&
+ filter->next + min > filter->buffer + filter->buffer_size) {
+ if (filter->avail > 0)
+ memmove(filter->buffer, filter->next, filter->avail);
+ filter->next = filter->buffer;
+ }
+
+ /* If we've used up the client data, get more. */
+ if (filter->client_avail <= 0) {
+ if (filter->end_of_file) {
+ if (avail != NULL)
+ *avail = 0;
+ return (NULL);
+ }
+ bytes_read = (filter->read)(filter,
+ &filter->client_buff);
+ if (bytes_read < 0) { /* Read error. */
+ filter->client_total = filter->client_avail = 0;
+ filter->client_next = filter->client_buff = NULL;
+ filter->fatal = 1;
+ if (avail != NULL)
+ *avail = ARCHIVE_FATAL;
+ return (NULL);
+ }
+ if (bytes_read == 0) { /* Premature end-of-file. */
+ filter->client_total = filter->client_avail = 0;
+ filter->client_next = filter->client_buff = NULL;
+ filter->end_of_file = 1;
+ /* Return whatever we do have. */
+ if (avail != NULL)
+ *avail = filter->avail;
+ return (NULL);
+ }
+ filter->position += bytes_read;
+ filter->client_total = bytes_read;
+ filter->client_avail = filter->client_total;
+ filter->client_next = filter->client_buff;
+ }
+ else
+ {
+ /*
+ * We can't satisfy the request from the copy
+ * buffer or the existing client data, so we
+ * need to copy more client data over to the
+ * copy buffer.
+ */
+
+ /* Ensure the buffer is big enough. */
+ if (min > filter->buffer_size) {
+ size_t s, t;
+ char *p;
+
+ /* Double the buffer; watch for overflow. */
+ s = t = filter->buffer_size;
+ if (s == 0)
+ s = min;
+ while (s < min) {
+ t *= 2;
+ if (t <= s) { /* Integer overflow! */
+ archive_set_error(
+ &filter->archive->archive,
+ ENOMEM,
+ "Unable to allocate copy buffer");
+ filter->fatal = 1;
+ if (avail != NULL)
+ *avail = ARCHIVE_FATAL;
+ return (NULL);
+ }
+ s = t;
+ }
+ /* Now s >= min, so allocate a new buffer. */
+ p = (char *)malloc(s);
+ if (p == NULL) {
+ archive_set_error(
+ &filter->archive->archive,
+ ENOMEM,
+ "Unable to allocate copy buffer");
+ filter->fatal = 1;
+ if (avail != NULL)
+ *avail = ARCHIVE_FATAL;
+ return (NULL);
+ }
+ /* Move data into newly-enlarged buffer. */
+ if (filter->avail > 0)
+ memmove(p, filter->next, filter->avail);
+ free(filter->buffer);
+ filter->next = filter->buffer = p;
+ filter->buffer_size = s;
+ }
+
+ /* We can add client data to copy buffer. */
+ /* First estimate: copy to fill rest of buffer. */
+ tocopy = (filter->buffer + filter->buffer_size)
+ - (filter->next + filter->avail);
+ /* Don't waste time buffering more than we need to. */
+ if (tocopy + filter->avail > min)
+ tocopy = min - filter->avail;
+ /* Don't copy more than is available. */
+ if (tocopy > filter->client_avail)
+ tocopy = filter->client_avail;
+
+ memcpy(filter->next + filter->avail, filter->client_next,
+ tocopy);
+ /* Remove this data from client buffer. */
+ filter->client_next += tocopy;
+ filter->client_avail -= tocopy;
+ /* add it to copy buffer. */
+ filter->avail += tocopy;
+ }
+ }
+}
+
+/*
+ * Move the file pointer forward. This should be called after
+ * __archive_read_ahead() returns data to you. Don't try to move
+ * ahead by more than the amount of data available according to
+ * __archive_read_ahead().
+ */
+/*
+ * Mark the appropriate data as used. Note that the request here will
+ * often be much smaller than the size of the previous read_ahead
+ * request.
+ */
+ssize_t
+__archive_read_consume(struct archive_read *a, size_t request)
+{
+ ssize_t r;
+ r = __archive_read_filter_consume(a->filter, request);
+ a->archive.file_position += r;
+ return (r);
+}
+
+ssize_t
+__archive_read_filter_consume(struct archive_read_filter * filter,
+ size_t request)
+{
+ if (filter->avail > 0) {
+ /* Read came from copy buffer. */
+ filter->next += request;
+ filter->avail -= request;
+ } else {
+ /* Read came from client buffer. */
+ filter->client_next += request;
+ filter->client_avail -= request;
+ }
+ return (request);
+}
+
+/*
+ * Move the file pointer ahead by an arbitrary amount. If you're
+ * reading uncompressed data from a disk file, this will actually
+ * translate into a seek() operation. Even in cases where seek()
+ * isn't feasible, this at least pushes the read-and-discard loop
+ * down closer to the data source.
+ */
+int64_t
+__archive_read_skip(struct archive_read *a, int64_t request)
+{
+ int64_t skipped = __archive_read_skip_lenient(a, request);
+ if (skipped == request)
+ return (skipped);
+ /* We hit EOF before we satisfied the skip request. */
+ if (skipped < 0) // Map error code to 0 for error message below.
+ skipped = 0;
+ archive_set_error(&a->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Truncated input file (needed %jd bytes, only %jd available)",
+ (intmax_t)request, (intmax_t)skipped);
+ return (ARCHIVE_FATAL);
+}
+
+int64_t
+__archive_read_skip_lenient(struct archive_read *a, int64_t request)
+{
+ int64_t skipped = __archive_read_filter_skip(a->filter, request);
+ if (skipped > 0)
+ a->archive.file_position += skipped;
+ return (skipped);
+}
+
+int64_t
+__archive_read_filter_skip(struct archive_read_filter *filter, int64_t request)
+{
+ int64_t bytes_skipped, total_bytes_skipped = 0;
+ size_t min;
+
+ if (filter->fatal)
+ return (-1);
+ /*
+ * If there is data in the buffers already, use that first.
+ */
+ if (filter->avail > 0) {
+ min = minimum(request, (off_t)filter->avail);
+ bytes_skipped = __archive_read_filter_consume(filter, min);
+ request -= bytes_skipped;
+ total_bytes_skipped += bytes_skipped;
+ }
+ if (filter->client_avail > 0) {
+ min = minimum(request, (int64_t)filter->client_avail);
+ bytes_skipped = __archive_read_filter_consume(filter, min);
+ request -= bytes_skipped;
+ total_bytes_skipped += bytes_skipped;
+ }
+ if (request == 0)
+ return (total_bytes_skipped);
+ /*
+ * If a client_skipper was provided, try that first.
+ */
+#if ARCHIVE_API_VERSION < 2
+ if ((filter->skip != NULL) && (request < SSIZE_MAX)) {
+#else
+ if (filter->skip != NULL) {
+#endif
+ bytes_skipped = (filter->skip)(filter, request);
+ if (bytes_skipped < 0) { /* error */
+ filter->client_total = filter->client_avail = 0;
+ filter->client_next = filter->client_buff = NULL;
+ filter->fatal = 1;
+ return (bytes_skipped);
+ }
+ total_bytes_skipped += bytes_skipped;
+ request -= bytes_skipped;
+ filter->client_next = filter->client_buff;
+ filter->client_avail = filter->client_total = 0;
+ }
+ /*
+ * Note that client_skipper will usually not satisfy the
+ * full request (due to low-level blocking concerns),
+ * so even if client_skipper is provided, we may still
+ * have to use ordinary reads to finish out the request.
+ */
+ while (request > 0) {
+ ssize_t bytes_read;
+ (void)__archive_read_filter_ahead(filter, 1, &bytes_read);
+ if (bytes_read < 0)
+ return (bytes_read);
+ if (bytes_read == 0) {
+ return (total_bytes_skipped);
+ }
+ min = (size_t)(minimum(bytes_read, request));
+ bytes_read = __archive_read_filter_consume(filter, min);
+ total_bytes_skipped += bytes_read;
+ request -= bytes_read;
+ }
+ return (total_bytes_skipped);
+}
diff --git a/libarchive/archive_read_data_into_fd.c b/libarchive/archive_read_data_into_fd.c
new file mode 100644
index 000000000000..3aeef3bc2da8
--- /dev/null
+++ b/libarchive/archive_read_data_into_fd.c
@@ -0,0 +1,93 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: src/lib/libarchive/archive_read_data_into_fd.c,v 1.16 2008/05/23 05:01:29 cperciva Exp $");
+
+#ifdef HAVE_SYS_TYPES_H
+#include <sys/types.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+
+#include "archive.h"
+#include "archive_private.h"
+
+/* Maximum amount of data to write at one time. */
+#define MAX_WRITE (1024 * 1024)
+
+/*
+ * This implementation minimizes copying of data and is sparse-file aware.
+ */
+int
+archive_read_data_into_fd(struct archive *a, int fd)
+{
+ int r;
+ const void *buff;
+ size_t size, bytes_to_write;
+ ssize_t bytes_written, total_written;
+ off_t offset;
+ off_t output_offset;
+
+ __archive_check_magic(a, ARCHIVE_READ_MAGIC, ARCHIVE_STATE_DATA, "archive_read_data_into_fd");
+
+ total_written = 0;
+ output_offset = 0;
+
+ while ((r = archive_read_data_block(a, &buff, &size, &offset)) ==
+ ARCHIVE_OK) {
+ const char *p = buff;
+ if (offset > output_offset) {
+ output_offset = lseek(fd,
+ offset - output_offset, SEEK_CUR);
+ if (output_offset != offset) {
+ archive_set_error(a, errno, "Seek error");
+ return (ARCHIVE_FATAL);
+ }
+ }
+ while (size > 0) {
+ bytes_to_write = size;
+ if (bytes_to_write > MAX_WRITE)
+ bytes_to_write = MAX_WRITE;
+ bytes_written = write(fd, p, bytes_to_write);
+ if (bytes_written < 0) {
+ archive_set_error(a, errno, "Write error");
+ return (ARCHIVE_FATAL);
+ }
+ output_offset += bytes_written;
+ total_written += bytes_written;
+ p += bytes_written;
+ size -= bytes_written;
+ }
+ }
+
+ if (r != ARCHIVE_EOF)
+ return (r);
+ return (ARCHIVE_OK);
+}
diff --git a/libarchive/archive_read_disk.3 b/libarchive/archive_read_disk.3
new file mode 100644
index 000000000000..98ec6482f3f0
--- /dev/null
+++ b/libarchive/archive_read_disk.3
@@ -0,0 +1,308 @@
+.\" Copyright (c) 2003-2009 Tim Kientzle
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\" notice, this list of conditions and the following disclaimer.
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\" notice, this list of conditions and the following disclaimer in the
+.\" documentation and/or other materials provided with the distribution.
+.\"
+.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+.\" SUCH DAMAGE.
+.\"
+.\" $FreeBSD: head/lib/libarchive/archive_read_disk.3 190957 2009-04-12 05:04:02Z kientzle $
+.\"
+.Dd March 10, 2009
+.Dt archive_read_disk 3
+.Os
+.Sh NAME
+.Nm archive_read_disk_new ,
+.Nm archive_read_disk_set_symlink_logical ,
+.Nm archive_read_disk_set_symlink_physical ,
+.Nm archive_read_disk_set_symlink_hybrid ,
+.Nm archive_read_disk_entry_from_file ,
+.Nm archive_read_disk_gname ,
+.Nm archive_read_disk_uname ,
+.Nm archive_read_disk_set_uname_lookup ,
+.Nm archive_read_disk_set_gname_lookup ,
+.Nm archive_read_disk_set_standard_lookup ,
+.Nm archive_read_close ,
+.Nm archive_read_finish
+.Nd functions for reading objects from disk
+.Sh SYNOPSIS
+.In archive.h
+.Ft struct archive *
+.Fn archive_read_disk_new "void"
+.Ft int
+.Fn archive_read_disk_set_symlink_logical "struct archive *"
+.Ft int
+.Fn archive_read_disk_set_symlink_physical "struct archive *"
+.Ft int
+.Fn archive_read_disk_set_symlink_hybrid "struct archive *"
+.Ft int
+.Fn archive_read_disk_gname "struct archive *" "gid_t"
+.Ft int
+.Fn archive_read_disk_uname "struct archive *" "uid_t"
+.Ft int
+.Fo archive_read_disk_set_gname_lookup
+.Fa "struct archive *"
+.Fa "void *"
+.Fa "const char *(*lookup)(void *, gid_t)"
+.Fa "void (*cleanup)(void *)"
+.Fc
+.Ft int
+.Fo archive_read_disk_set_uname_lookup
+.Fa "struct archive *"
+.Fa "void *"
+.Fa "const char *(*lookup)(void *, uid_t)"
+.Fa "void (*cleanup)(void *)"
+.Fc
+.Ft int
+.Fn archive_read_disk_set_standard_lookup "struct archive *"
+.Ft int
+.Fo archive_read_disk_entry_from_file
+.Fa "struct archive *"
+.Fa "struct archive_entry *"
+.Fa "int fd"
+.Fa "const struct stat *"
+.Fc
+.Ft int
+.Fn archive_read_close "struct archive *"
+.Ft int
+.Fn archive_read_finish "struct archive *"
+.Sh DESCRIPTION
+These functions provide an API for reading information about
+objects on disk.
+In particular, they provide an interface for populating
+.Tn struct archive_entry
+objects.
+.Bl -tag -width indent
+.It Fn archive_read_disk_new
+Allocates and initializes a
+.Tn struct archive
+object suitable for reading object information from disk.
+.It Xo
+.Fn archive_read_disk_set_symlink_logical ,
+.Fn archive_read_disk_set_symlink_physical ,
+.Fn archive_read_disk_set_symlink_hybrid
+.Xc
+This sets the mode used for handling symbolic links.
+The
+.Dq logical
+mode follows all symbolic links.
+The
+.Dq physical
+mode does not follow any symbolic links.
+The
+.Dq hybrid
+mode currently behaves identically to the
+.Dq logical
+mode.
+.It Xo
+.Fn archive_read_disk_gname ,
+.Fn archive_read_disk_uname
+.Xc
+Returns a user or group name given a gid or uid value.
+By default, these always return a NULL string.
+.It Xo
+.Fn archive_read_disk_set_gname_lookup ,
+.Fn archive_read_disk_set_uname_lookup
+.Xc
+These allow you to override the functions used for
+user and group name lookups.
+You may also provide a
+.Tn void *
+pointer to a private data structure and a cleanup function for
+that data.
+The cleanup function will be invoked when the
+.Tn struct archive
+object is destroyed or when new lookup functions are registered.
+.It Fn archive_read_disk_set_standard_lookup
+This convenience function installs a standard set of user
+and group name lookup functions.
+These functions use
+.Xr getpwuid 3
+and
+.Xr getgrgid 3
+to convert ids to names, defaulting to NULL if the names cannot
+be looked up.
+These functions also implement a simple memory cache to reduce
+the number of calls to
+.Xr getpwuid 3
+and
+.Xr getgrgid 3 .
+.It Fn archive_read_disk_entry_from_file
+Populates a
+.Tn struct archive_entry
+object with information about a particular file.
+The
+.Tn archive_entry
+object must have already been created with
+.Xr archive_entry_new 3
+and at least one of the source path or path fields must already be set.
+(If both are set, the source path will be used.)
+.Pp
+Information is read from disk using the path name from the
+.Tn struct archive_entry
+object.
+If a file descriptor is provided, some information will be obtained using
+that file descriptor, on platforms that support the appropriate
+system calls.
+.Pp
+If a pointer to a
+.Tn struct stat
+is provided, information from that structure will be used instead
+of reading from the disk where appropriate.
+This can provide performance benefits in scenarios where
+.Tn struct stat
+information has already been read from the disk as a side effect
+of some other operation.
+(For example, directory traversal libraries often provide this information.)
+.Pp
+Where necessary, user and group ids are converted to user and group names
+using the currently registered lookup functions above.
+This affects the file ownership fields and ACL values in the
+.Tn struct archive_entry
+object.
+.It Fn archive_read_close
+This currently does nothing.
+.It Fn archive_write_finish
+Invokes
+.Fn archive_write_close
+if it was not invoked manually, then releases all resources.
+.El
+More information about the
+.Va struct archive
+object and the overall design of the library can be found in the
+.Xr libarchive 3
+overview.
+.Sh EXAMPLE
+The following illustrates basic usage of the library by
+showing how to use it to copy an item on disk into an archive.
+.Bd -literal -offset indent
+void
+file_to_archive(struct archive *a, const char *name)
+{
+ char buff[8192];
+ size_t bytes_read;
+ struct archive *ard;
+ struct archive_entry *entry;
+ int fd;
+
+ ard = archive_read_disk_new();
+ archive_read_disk_set_standard_lookup(ard);
+ entry = archive_entry_new();
+ fd = open(name, O_RDONLY);
+ if (fd < 0)
+ return;
+ archive_entry_copy_sourcepath(entry, name);
+ archive_read_disk_entry_from_file(ard, entry, fd, NULL);
+ archive_write_header(a, entry);
+ while ((bytes_read = read(fd, buff, sizeof(buff))) > 0)
+ archive_write_data(a, buff, bytes_read);
+ archive_write_finish_entry(a);
+ archive_read_finish(ard);
+ archive_entry_free(entry);
+}
+.Ed
+.Sh RETURN VALUES
+Most functions return
+.Cm ARCHIVE_OK
+(zero) on success, or one of several negative
+error codes for errors.
+Specific error codes include:
+.Cm ARCHIVE_RETRY
+for operations that might succeed if retried,
+.Cm ARCHIVE_WARN
+for unusual conditions that do not prevent further operations, and
+.Cm ARCHIVE_FATAL
+for serious errors that make remaining operations impossible.
+The
+.Xr archive_errno 3
+and
+.Xr archive_error_string 3
+functions can be used to retrieve an appropriate error code and a
+textual error message.
+(See
+.Xr archive_util 3
+for details.)
+.Pp
+.Fn archive_read_disk_new
+returns a pointer to a newly-allocated
+.Tn struct archive
+object or NULL if the allocation failed for any reason.
+.Pp
+.Fn archive_read_disk_gname
+and
+.Fn archive_read_disk_uname
+return
+.Tn const char *
+pointers to the textual name or NULL if the lookup failed for any reason.
+The returned pointer points to internal storage that
+may be reused on the next call to either of these functions;
+callers should copy the string if they need to continue accessing it.
+.Pp
+.Sh SEE ALSO
+.Xr archive_read 3 ,
+.Xr archive_write 3 ,
+.Xr archive_write_disk 3 ,
+.Xr tar 1 ,
+.Xr libarchive 3
+.Sh HISTORY
+The
+.Nm libarchive
+library first appeared in
+.Fx 5.3 .
+The
+.Nm archive_read_disk
+interface was added to
+.Nm libarchive 2.6
+and first appeared in
+.Fx 8.0 .
+.Sh AUTHORS
+.An -nosplit
+The
+.Nm libarchive
+library was written by
+.An Tim Kientzle Aq kientzle@freebsd.org .
+.Sh BUGS
+The
+.Dq standard
+user name and group name lookup functions are not the defaults because
+.Xr getgrgid 3
+and
+.Xr getpwuid 3
+are sometimes too large for particular applications.
+The current design allows the application author to use a more
+compact implementation when appropriate.
+.Pp
+The full list of metadata read from disk by
+.Fn archive_read_disk_entry_from_file
+is necessarily system-dependent.
+.Pp
+The
+.Fn archive_read_disk_entry_from_file
+function reads as much information as it can from disk.
+Some method should be provided to limit this so that clients who
+do not need ACLs, for instance, can avoid the extra work needed
+to look up such information.
+.Pp
+This API should provide a set of methods for walking a directory tree.
+That would make it a direct parallel of the
+.Xr archive_read 3
+API.
+When such methods are implemented, the
+.Dq hybrid
+symbolic link mode will make sense.
diff --git a/libarchive/archive_read_disk.c b/libarchive/archive_read_disk.c
new file mode 100644
index 000000000000..8fad7f137334
--- /dev/null
+++ b/libarchive/archive_read_disk.c
@@ -0,0 +1,198 @@
+/*-
+ * Copyright (c) 2003-2009 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer
+ * in this position and unchanged.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_disk.c 189429 2009-03-06 04:35:31Z kientzle $");
+
+#include "archive.h"
+#include "archive_string.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_read_disk_private.h"
+
+static int _archive_read_finish(struct archive *);
+static int _archive_read_close(struct archive *);
+static const char *trivial_lookup_gname(void *, gid_t gid);
+static const char *trivial_lookup_uname(void *, uid_t uid);
+
+static struct archive_vtable *
+archive_read_disk_vtable(void)
+{
+ static struct archive_vtable av;
+ static int inited = 0;
+
+ if (!inited) {
+ av.archive_finish = _archive_read_finish;
+ av.archive_close = _archive_read_close;
+ }
+ return (&av);
+}
+
+const char *
+archive_read_disk_gname(struct archive *_a, gid_t gid)
+{
+ struct archive_read_disk *a = (struct archive_read_disk *)_a;
+ if (a->lookup_gname != NULL)
+ return ((*a->lookup_gname)(a->lookup_gname_data, gid));
+ return (NULL);
+}
+
+const char *
+archive_read_disk_uname(struct archive *_a, uid_t uid)
+{
+ struct archive_read_disk *a = (struct archive_read_disk *)_a;
+ if (a->lookup_uname != NULL)
+ return ((*a->lookup_uname)(a->lookup_uname_data, uid));
+ return (NULL);
+}
+
+int
+archive_read_disk_set_gname_lookup(struct archive *_a,
+ void *private_data,
+ const char * (*lookup_gname)(void *private, gid_t gid),
+ void (*cleanup_gname)(void *private))
+{
+ struct archive_read_disk *a = (struct archive_read_disk *)_a;
+ __archive_check_magic(&a->archive, ARCHIVE_READ_DISK_MAGIC,
+ ARCHIVE_STATE_ANY, "archive_read_disk_set_gname_lookup");
+
+ if (a->cleanup_gname != NULL && a->lookup_gname_data != NULL)
+ (a->cleanup_gname)(a->lookup_gname_data);
+
+ a->lookup_gname = lookup_gname;
+ a->cleanup_gname = cleanup_gname;
+ a->lookup_gname_data = private_data;
+ return (ARCHIVE_OK);
+}
+
+int
+archive_read_disk_set_uname_lookup(struct archive *_a,
+ void *private_data,
+ const char * (*lookup_uname)(void *private, uid_t uid),
+ void (*cleanup_uname)(void *private))
+{
+ struct archive_read_disk *a = (struct archive_read_disk *)_a;
+ __archive_check_magic(&a->archive, ARCHIVE_READ_DISK_MAGIC,
+ ARCHIVE_STATE_ANY, "archive_read_disk_set_uname_lookup");
+
+ if (a->cleanup_uname != NULL && a->lookup_uname_data != NULL)
+ (a->cleanup_uname)(a->lookup_uname_data);
+
+ a->lookup_uname = lookup_uname;
+ a->cleanup_uname = cleanup_uname;
+ a->lookup_uname_data = private_data;
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Create a new archive_read_disk object and initialize it with global state.
+ */
+struct archive *
+archive_read_disk_new(void)
+{
+ struct archive_read_disk *a;
+
+ a = (struct archive_read_disk *)malloc(sizeof(*a));
+ if (a == NULL)
+ return (NULL);
+ memset(a, 0, sizeof(*a));
+ a->archive.magic = ARCHIVE_READ_DISK_MAGIC;
+ /* We're ready to write a header immediately. */
+ a->archive.state = ARCHIVE_STATE_HEADER;
+ a->archive.vtable = archive_read_disk_vtable();
+ a->lookup_uname = trivial_lookup_uname;
+ a->lookup_gname = trivial_lookup_gname;
+ return (&a->archive);
+}
+
+static int
+_archive_read_finish(struct archive *_a)
+{
+ struct archive_read_disk *a = (struct archive_read_disk *)_a;
+
+ if (a->cleanup_gname != NULL && a->lookup_gname_data != NULL)
+ (a->cleanup_gname)(a->lookup_gname_data);
+ if (a->cleanup_uname != NULL && a->lookup_uname_data != NULL)
+ (a->cleanup_uname)(a->lookup_uname_data);
+ archive_string_free(&a->archive.error_string);
+ free(a);
+ return (ARCHIVE_OK);
+}
+
+static int
+_archive_read_close(struct archive *_a)
+{
+ (void)_a; /* UNUSED */
+ return (ARCHIVE_OK);
+}
+
+int
+archive_read_disk_set_symlink_logical(struct archive *_a)
+{
+ struct archive_read_disk *a = (struct archive_read_disk *)_a;
+ a->symlink_mode = 'L';
+ a->follow_symlinks = 1;
+ return (ARCHIVE_OK);
+}
+
+int
+archive_read_disk_set_symlink_physical(struct archive *_a)
+{
+ struct archive_read_disk *a = (struct archive_read_disk *)_a;
+ a->symlink_mode = 'P';
+ a->follow_symlinks = 0;
+ return (ARCHIVE_OK);
+}
+
+int
+archive_read_disk_set_symlink_hybrid(struct archive *_a)
+{
+ struct archive_read_disk *a = (struct archive_read_disk *)_a;
+ a->symlink_mode = 'H';
+ a->follow_symlinks = 1; /* Follow symlinks initially. */
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Trivial implementations of gname/uname lookup functions.
+ * These are normally overridden by the client, but these stub
+ * versions ensure that we always have something that works.
+ */
+static const char *
+trivial_lookup_gname(void *private_data, gid_t gid)
+{
+ (void)private_data; /* UNUSED */
+ (void)gid; /* UNUSED */
+ return (NULL);
+}
+
+static const char *
+trivial_lookup_uname(void *private_data, uid_t uid)
+{
+ (void)private_data; /* UNUSED */
+ (void)uid; /* UNUSED */
+ return (NULL);
+}
diff --git a/libarchive/archive_read_disk_entry_from_file.c b/libarchive/archive_read_disk_entry_from_file.c
new file mode 100644
index 000000000000..7473c506e832
--- /dev/null
+++ b/libarchive/archive_read_disk_entry_from_file.c
@@ -0,0 +1,570 @@
+/*-
+ * Copyright (c) 2003-2009 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_disk_entry_from_file.c 201084 2009-12-28 02:14:09Z kientzle $");
+
+#ifdef HAVE_SYS_TYPES_H
+/* Mac OSX requires sys/types.h before sys/acl.h. */
+#include <sys/types.h>
+#endif
+#ifdef HAVE_SYS_ACL_H
+#include <sys/acl.h>
+#endif
+#ifdef HAVE_SYS_EXTATTR_H
+#include <sys/extattr.h>
+#endif
+#ifdef HAVE_SYS_PARAM_H
+#include <sys/param.h>
+#endif
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_SYS_XATTR_H
+#include <sys/xattr.h>
+#endif
+#ifdef HAVE_ACL_LIBACL_H
+#include <acl/libacl.h>
+#endif
+#ifdef HAVE_ATTR_XATTR_H
+#include <attr/xattr.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_LIMITS_H
+#include <limits.h>
+#endif
+#ifdef HAVE_WINDOWS_H
+#include <windows.h>
+#endif
+
+#include "archive.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_read_disk_private.h"
+
+/*
+ * Linux and FreeBSD plug this obvious hole in POSIX.1e in
+ * different ways.
+ */
+#if HAVE_ACL_GET_PERM
+#define ACL_GET_PERM acl_get_perm
+#elif HAVE_ACL_GET_PERM_NP
+#define ACL_GET_PERM acl_get_perm_np
+#endif
+
+static int setup_acls_posix1e(struct archive_read_disk *,
+ struct archive_entry *, int fd);
+static int setup_xattrs(struct archive_read_disk *,
+ struct archive_entry *, int fd);
+
+int
+archive_read_disk_entry_from_file(struct archive *_a,
+ struct archive_entry *entry,
+ int fd, const struct stat *st)
+{
+ struct archive_read_disk *a = (struct archive_read_disk *)_a;
+ const char *path, *name;
+ struct stat s;
+ int initial_fd = fd;
+ int r, r1;
+
+ archive_clear_error(_a);
+ path = archive_entry_sourcepath(entry);
+ if (path == NULL)
+ path = archive_entry_pathname(entry);
+
+#ifdef EXT2_IOC_GETFLAGS
+ /* Linux requires an extra ioctl to pull the flags. Although
+ * this is an extra step, it has a nice side-effect: We get an
+ * open file descriptor which we can use in the subsequent lookups. */
+ if ((S_ISREG(st->st_mode) || S_ISDIR(st->st_mode))) {
+ if (fd < 0)
+ fd = open(pathname, O_RDONLY | O_NONBLOCK | O_BINARY);
+ if (fd >= 0) {
+ unsigned long stflags;
+ int r = ioctl(fd, EXT2_IOC_GETFLAGS, &stflags);
+ if (r == 0 && stflags != 0)
+ archive_entry_set_fflags(entry, stflags, 0);
+ }
+ }
+#endif
+
+ if (st == NULL) {
+ /* TODO: On Windows, use GetFileInfoByHandle() here.
+ * Using Windows stat() call is badly broken, but
+ * even the stat() wrapper has problems because
+ * 'struct stat' is broken on Windows.
+ */
+#if HAVE_FSTAT
+ if (fd >= 0) {
+ if (fstat(fd, &s) != 0) {
+ archive_set_error(&a->archive, errno,
+ "Can't fstat");
+ return (ARCHIVE_FAILED);
+ }
+ } else
+#endif
+#if HAVE_LSTAT
+ if (!a->follow_symlinks) {
+ if (lstat(path, &s) != 0) {
+ archive_set_error(&a->archive, errno,
+ "Can't lstat %s", path);
+ return (ARCHIVE_FAILED);
+ }
+ } else
+#endif
+ if (stat(path, &s) != 0) {
+ archive_set_error(&a->archive, errno,
+ "Can't lstat %s", path);
+ return (ARCHIVE_FAILED);
+ }
+ st = &s;
+ }
+ archive_entry_copy_stat(entry, st);
+
+ /* Lookup uname/gname */
+ name = archive_read_disk_uname(_a, archive_entry_uid(entry));
+ if (name != NULL)
+ archive_entry_copy_uname(entry, name);
+ name = archive_read_disk_gname(_a, archive_entry_gid(entry));
+ if (name != NULL)
+ archive_entry_copy_gname(entry, name);
+
+#ifdef HAVE_STRUCT_STAT_ST_FLAGS
+ /* On FreeBSD, we get flags for free with the stat. */
+ /* TODO: Does this belong in copy_stat()? */
+ if (st->st_flags != 0)
+ archive_entry_set_fflags(entry, st->st_flags, 0);
+#endif
+
+#ifdef HAVE_READLINK
+ if (S_ISLNK(st->st_mode)) {
+ char linkbuffer[PATH_MAX + 1];
+ int lnklen = readlink(path, linkbuffer, PATH_MAX);
+ if (lnklen < 0) {
+ archive_set_error(&a->archive, errno,
+ "Couldn't read link data");
+ return (ARCHIVE_FAILED);
+ }
+ linkbuffer[lnklen] = 0;
+ archive_entry_set_symlink(entry, linkbuffer);
+ }
+#endif
+
+ r = setup_acls_posix1e(a, entry, fd);
+ r1 = setup_xattrs(a, entry, fd);
+ if (r1 < r)
+ r = r1;
+ /* If we opened the file earlier in this function, close it. */
+ if (initial_fd != fd)
+ close(fd);
+ return (r);
+}
+
+#ifdef HAVE_POSIX_ACL
+static void setup_acl_posix1e(struct archive_read_disk *a,
+ struct archive_entry *entry, acl_t acl, int archive_entry_acl_type);
+
+static int
+setup_acls_posix1e(struct archive_read_disk *a,
+ struct archive_entry *entry, int fd)
+{
+ const char *accpath;
+ acl_t acl;
+
+ accpath = archive_entry_sourcepath(entry);
+ if (accpath == NULL)
+ accpath = archive_entry_pathname(entry);
+
+ archive_entry_acl_clear(entry);
+
+ /* Retrieve access ACL from file. */
+ if (fd >= 0)
+ acl = acl_get_fd(fd);
+#if HAVE_ACL_GET_LINK_NP
+ else if (!a->follow_symlinks)
+ acl = acl_get_link_np(accpath, ACL_TYPE_ACCESS);
+#else
+ else if ((!a->follow_symlinks)
+ && (archive_entry_filetype(entry) == AE_IFLNK))
+ /* We can't get the ACL of a symlink, so we assume it can't
+ have one. */
+ acl = NULL;
+#endif
+ else
+ acl = acl_get_file(accpath, ACL_TYPE_ACCESS);
+ if (acl != NULL) {
+ setup_acl_posix1e(a, entry, acl,
+ ARCHIVE_ENTRY_ACL_TYPE_ACCESS);
+ acl_free(acl);
+ }
+
+ /* Only directories can have default ACLs. */
+ if (S_ISDIR(archive_entry_mode(entry))) {
+ acl = acl_get_file(accpath, ACL_TYPE_DEFAULT);
+ if (acl != NULL) {
+ setup_acl_posix1e(a, entry, acl,
+ ARCHIVE_ENTRY_ACL_TYPE_DEFAULT);
+ acl_free(acl);
+ }
+ }
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Translate POSIX.1e ACL into libarchive internal structure.
+ */
+static void
+setup_acl_posix1e(struct archive_read_disk *a,
+ struct archive_entry *entry, acl_t acl, int archive_entry_acl_type)
+{
+ acl_tag_t acl_tag;
+ acl_entry_t acl_entry;
+ acl_permset_t acl_permset;
+ int s, ae_id, ae_tag, ae_perm;
+ const char *ae_name;
+
+ s = acl_get_entry(acl, ACL_FIRST_ENTRY, &acl_entry);
+ while (s == 1) {
+ ae_id = -1;
+ ae_name = NULL;
+
+ acl_get_tag_type(acl_entry, &acl_tag);
+ if (acl_tag == ACL_USER) {
+ ae_id = (int)*(uid_t *)acl_get_qualifier(acl_entry);
+ ae_name = archive_read_disk_uname(&a->archive, ae_id);
+ ae_tag = ARCHIVE_ENTRY_ACL_USER;
+ } else if (acl_tag == ACL_GROUP) {
+ ae_id = (int)*(gid_t *)acl_get_qualifier(acl_entry);
+ ae_name = archive_read_disk_gname(&a->archive, ae_id);
+ ae_tag = ARCHIVE_ENTRY_ACL_GROUP;
+ } else if (acl_tag == ACL_MASK) {
+ ae_tag = ARCHIVE_ENTRY_ACL_MASK;
+ } else if (acl_tag == ACL_USER_OBJ) {
+ ae_tag = ARCHIVE_ENTRY_ACL_USER_OBJ;
+ } else if (acl_tag == ACL_GROUP_OBJ) {
+ ae_tag = ARCHIVE_ENTRY_ACL_GROUP_OBJ;
+ } else if (acl_tag == ACL_OTHER) {
+ ae_tag = ARCHIVE_ENTRY_ACL_OTHER;
+ } else {
+ /* Skip types that libarchive can't support. */
+ continue;
+ }
+
+ acl_get_permset(acl_entry, &acl_permset);
+ ae_perm = 0;
+ /*
+ * acl_get_perm() is spelled differently on different
+ * platforms; see above.
+ */
+ if (ACL_GET_PERM(acl_permset, ACL_EXECUTE))
+ ae_perm |= ARCHIVE_ENTRY_ACL_EXECUTE;
+ if (ACL_GET_PERM(acl_permset, ACL_READ))
+ ae_perm |= ARCHIVE_ENTRY_ACL_READ;
+ if (ACL_GET_PERM(acl_permset, ACL_WRITE))
+ ae_perm |= ARCHIVE_ENTRY_ACL_WRITE;
+
+ archive_entry_acl_add_entry(entry,
+ archive_entry_acl_type, ae_perm, ae_tag,
+ ae_id, ae_name);
+
+ s = acl_get_entry(acl, ACL_NEXT_ENTRY, &acl_entry);
+ }
+}
+#else
+static int
+setup_acls_posix1e(struct archive_read_disk *a,
+ struct archive_entry *entry, int fd)
+{
+ (void)a; /* UNUSED */
+ (void)entry; /* UNUSED */
+ (void)fd; /* UNUSED */
+ return (ARCHIVE_OK);
+}
+#endif
+
+#if HAVE_LISTXATTR && HAVE_LLISTXATTR && HAVE_GETXATTR && HAVE_LGETXATTR
+
+/*
+ * Linux extended attribute support.
+ *
+ * TODO: By using a stack-allocated buffer for the first
+ * call to getxattr(), we might be able to avoid the second
+ * call entirely. We only need the second call if the
+ * stack-allocated buffer is too small. But a modest buffer
+ * of 1024 bytes or so will often be big enough. Same applies
+ * to listxattr().
+ */
+
+
+static int
+setup_xattr(struct archive_read_disk *a,
+ struct archive_entry *entry, const char *name, int fd)
+{
+ ssize_t size;
+ void *value = NULL;
+ const char *accpath;
+
+ (void)fd; /* UNUSED */
+
+ accpath = archive_entry_sourcepath(entry);
+ if (accpath == NULL)
+ accpath = archive_entry_pathname(entry);
+
+ if (!a->follow_symlinks)
+ size = lgetxattr(accpath, name, NULL, 0);
+ else
+ size = getxattr(accpath, name, NULL, 0);
+
+ if (size == -1) {
+ archive_set_error(&a->archive, errno,
+ "Couldn't query extended attribute");
+ return (ARCHIVE_WARN);
+ }
+
+ if (size > 0 && (value = malloc(size)) == NULL) {
+ archive_set_error(&a->archive, errno, "Out of memory");
+ return (ARCHIVE_FATAL);
+ }
+
+ if (!a->follow_symlinks)
+ size = lgetxattr(accpath, name, value, size);
+ else
+ size = getxattr(accpath, name, value, size);
+
+ if (size == -1) {
+ archive_set_error(&a->archive, errno,
+ "Couldn't read extended attribute");
+ return (ARCHIVE_WARN);
+ }
+
+ archive_entry_xattr_add_entry(entry, name, value, size);
+
+ free(value);
+ return (ARCHIVE_OK);
+}
+
+static int
+setup_xattrs(struct archive_read_disk *a,
+ struct archive_entry *entry, int fd)
+{
+ char *list, *p;
+ const char *path;
+ ssize_t list_size;
+
+
+ path = archive_entry_sourcepath(entry);
+ if (path == NULL)
+ path = archive_entry_pathname(entry);
+
+ if (!a->follow_symlinks)
+ list_size = llistxattr(path, NULL, 0);
+ else
+ list_size = listxattr(path, NULL, 0);
+
+ if (list_size == -1) {
+ if (errno == ENOTSUP)
+ return (ARCHIVE_OK);
+ archive_set_error(&a->archive, errno,
+ "Couldn't list extended attributes");
+ return (ARCHIVE_WARN);
+ }
+
+ if (list_size == 0)
+ return (ARCHIVE_OK);
+
+ if ((list = malloc(list_size)) == NULL) {
+ archive_set_error(&a->archive, errno, "Out of memory");
+ return (ARCHIVE_FATAL);
+ }
+
+ if (!a->follow_symlinks)
+ list_size = llistxattr(path, list, list_size);
+ else
+ list_size = listxattr(path, list, list_size);
+
+ if (list_size == -1) {
+ archive_set_error(&a->archive, errno,
+ "Couldn't retrieve extended attributes");
+ free(list);
+ return (ARCHIVE_WARN);
+ }
+
+ for (p = list; (p - list) < list_size; p += strlen(p) + 1) {
+ if (strncmp(p, "system.", 7) == 0 ||
+ strncmp(p, "xfsroot.", 8) == 0)
+ continue;
+ setup_xattr(a, entry, p, fd);
+ }
+
+ free(list);
+ return (ARCHIVE_OK);
+}
+
+#elif HAVE_EXTATTR_GET_FILE && HAVE_EXTATTR_LIST_FILE && \
+ HAVE_DECL_EXTATTR_NAMESPACE_USER
+
+/*
+ * FreeBSD extattr interface.
+ */
+
+/* TODO: Implement this. Follow the Linux model above, but
+ * with FreeBSD-specific system calls, of course. Be careful
+ * to not include the system extattrs that hold ACLs; we handle
+ * those separately.
+ */
+static int
+setup_xattr(struct archive_read_disk *a, struct archive_entry *entry,
+ int namespace, const char *name, const char *fullname, int fd);
+
+static int
+setup_xattr(struct archive_read_disk *a, struct archive_entry *entry,
+ int namespace, const char *name, const char *fullname, int fd)
+{
+ ssize_t size;
+ void *value = NULL;
+ const char *accpath;
+
+ (void)fd; /* UNUSED */
+
+ accpath = archive_entry_sourcepath(entry);
+ if (accpath == NULL)
+ accpath = archive_entry_pathname(entry);
+
+ if (!a->follow_symlinks)
+ size = extattr_get_link(accpath, namespace, name, NULL, 0);
+ else
+ size = extattr_get_file(accpath, namespace, name, NULL, 0);
+
+ if (size == -1) {
+ archive_set_error(&a->archive, errno,
+ "Couldn't query extended attribute");
+ return (ARCHIVE_WARN);
+ }
+
+ if (size > 0 && (value = malloc(size)) == NULL) {
+ archive_set_error(&a->archive, errno, "Out of memory");
+ return (ARCHIVE_FATAL);
+ }
+
+ if (!a->follow_symlinks)
+ size = extattr_get_link(accpath, namespace, name, value, size);
+ else
+ size = extattr_get_file(accpath, namespace, name, value, size);
+
+ if (size == -1) {
+ archive_set_error(&a->archive, errno,
+ "Couldn't read extended attribute");
+ return (ARCHIVE_WARN);
+ }
+
+ archive_entry_xattr_add_entry(entry, fullname, value, size);
+
+ free(value);
+ return (ARCHIVE_OK);
+}
+
+static int
+setup_xattrs(struct archive_read_disk *a,
+ struct archive_entry *entry, int fd)
+{
+ char buff[512];
+ char *list, *p;
+ ssize_t list_size;
+ const char *path;
+ int namespace = EXTATTR_NAMESPACE_USER;
+
+ path = archive_entry_sourcepath(entry);
+ if (path == NULL)
+ path = archive_entry_pathname(entry);
+
+ if (!a->follow_symlinks)
+ list_size = extattr_list_link(path, namespace, NULL, 0);
+ else
+ list_size = extattr_list_file(path, namespace, NULL, 0);
+
+ if (list_size == -1 && errno == EOPNOTSUPP)
+ return (ARCHIVE_OK);
+ if (list_size == -1) {
+ archive_set_error(&a->archive, errno,
+ "Couldn't list extended attributes");
+ return (ARCHIVE_WARN);
+ }
+
+ if (list_size == 0)
+ return (ARCHIVE_OK);
+
+ if ((list = malloc(list_size)) == NULL) {
+ archive_set_error(&a->archive, errno, "Out of memory");
+ return (ARCHIVE_FATAL);
+ }
+
+ if (!a->follow_symlinks)
+ list_size = extattr_list_link(path, namespace, list, list_size);
+ else
+ list_size = extattr_list_file(path, namespace, list, list_size);
+
+ if (list_size == -1) {
+ archive_set_error(&a->archive, errno,
+ "Couldn't retrieve extended attributes");
+ free(list);
+ return (ARCHIVE_WARN);
+ }
+
+ p = list;
+ while ((p - list) < list_size) {
+ size_t len = 255 & (int)*p;
+ char *name;
+
+ strcpy(buff, "user.");
+ name = buff + strlen(buff);
+ memcpy(name, p + 1, len);
+ name[len] = '\0';
+ setup_xattr(a, entry, namespace, name, buff, fd);
+ p += 1 + len;
+ }
+
+ free(list);
+ return (ARCHIVE_OK);
+}
+
+#else
+
+/*
+ * Generic (stub) extended attribute support.
+ */
+static int
+setup_xattrs(struct archive_read_disk *a,
+ struct archive_entry *entry, int fd)
+{
+ (void)a; /* UNUSED */
+ (void)entry; /* UNUSED */
+ (void)fd; /* UNUSED */
+ return (ARCHIVE_OK);
+}
+
+#endif
diff --git a/libarchive/archive_read_disk_private.h b/libarchive/archive_read_disk_private.h
new file mode 100644
index 000000000000..b674b7106d73
--- /dev/null
+++ b/libarchive/archive_read_disk_private.h
@@ -0,0 +1,62 @@
+/*-
+ * Copyright (c) 2003-2009 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer
+ * in this position and unchanged.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD: head/lib/libarchive/archive_read_disk_private.h 201105 2009-12-28 03:20:54Z kientzle $
+ */
+
+#ifndef __LIBARCHIVE_BUILD
+#error This header is only to be used internally to libarchive.
+#endif
+
+#ifndef ARCHIVE_READ_DISK_PRIVATE_H_INCLUDED
+#define ARCHIVE_READ_DISK_PRIVATE_H_INCLUDED
+
+struct archive_read_disk {
+ struct archive archive;
+
+ /*
+ * Symlink mode is one of 'L'ogical, 'P'hysical, or 'H'ybrid,
+ * following an old BSD convention. 'L' follows all symlinks,
+ * 'P' follows none, 'H' follows symlinks only for the first
+ * item.
+ */
+ char symlink_mode;
+
+ /*
+ * Since symlink interaction changes, we need to track whether
+ * we're following symlinks for the current item. 'L' mode above
+ * sets this true, 'P' sets it false, 'H' changes it as we traverse.
+ */
+ char follow_symlinks; /* Either 'L' or 'P'. */
+
+ const char * (*lookup_gname)(void *private, gid_t gid);
+ void (*cleanup_gname)(void *private);
+ void *lookup_gname_data;
+ const char * (*lookup_uname)(void *private, gid_t gid);
+ void (*cleanup_uname)(void *private);
+ void *lookup_uname_data;
+};
+
+#endif
diff --git a/libarchive/archive_read_disk_set_standard_lookup.c b/libarchive/archive_read_disk_set_standard_lookup.c
new file mode 100644
index 000000000000..97a568e057af
--- /dev/null
+++ b/libarchive/archive_read_disk_set_standard_lookup.c
@@ -0,0 +1,303 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_disk_set_standard_lookup.c 201109 2009-12-28 03:30:31Z kientzle $");
+
+#ifdef HAVE_SYS_TYPES_H
+#include <sys/types.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_GRP_H
+#include <grp.h>
+#endif
+#ifdef HAVE_PWD_H
+#include <pwd.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+
+#include "archive.h"
+
+#if defined(_WIN32) && !defined(__CYGWIN__)
+int
+archive_read_disk_set_standard_lookup(struct archive *a)
+{
+ archive_set_error(a, -1, "Standard lookups not available on Windows");
+ return (ARCHIVE_FATAL);
+}
+#else /* ! (_WIN32 && !__CYGWIN__) */
+#define name_cache_size 127
+
+static const char * const NO_NAME = "(noname)";
+
+struct name_cache {
+ struct archive *archive;
+ char *buff;
+ size_t buff_size;
+ int probes;
+ int hits;
+ size_t size;
+ struct {
+ id_t id;
+ const char *name;
+ } cache[name_cache_size];
+};
+
+static const char * lookup_gname(void *, gid_t);
+static const char * lookup_uname(void *, uid_t);
+static void cleanup(void *);
+static const char * lookup_gname_helper(struct name_cache *, id_t gid);
+static const char * lookup_uname_helper(struct name_cache *, id_t uid);
+
+/*
+ * Installs functions that use getpwuid()/getgrgid()---along with
+ * a simple cache to accelerate such lookups---into the archive_read_disk
+ * object. This is in a separate file because getpwuid()/getgrgid()
+ * can pull in a LOT of library code (including NIS/LDAP functions, which
+ * pull in DNS resolveers, etc). This can easily top 500kB, which makes
+ * it inappropriate for some space-constrained applications.
+ *
+ * Applications that are size-sensitive may want to just use the
+ * real default functions (defined in archive_read_disk.c) that just
+ * use the uid/gid without the lookup. Or define your own custom functions
+ * if you prefer.
+ */
+int
+archive_read_disk_set_standard_lookup(struct archive *a)
+{
+ struct name_cache *ucache = malloc(sizeof(struct name_cache));
+ struct name_cache *gcache = malloc(sizeof(struct name_cache));
+
+ if (ucache == NULL || gcache == NULL) {
+ archive_set_error(a, ENOMEM,
+ "Can't allocate uname/gname lookup cache");
+ free(ucache);
+ free(gcache);
+ return (ARCHIVE_FATAL);
+ }
+
+ memset(ucache, 0, sizeof(*ucache));
+ ucache->archive = a;
+ ucache->size = name_cache_size;
+ memset(gcache, 0, sizeof(*gcache));
+ gcache->archive = a;
+ gcache->size = name_cache_size;
+
+ archive_read_disk_set_gname_lookup(a, gcache, lookup_gname, cleanup);
+ archive_read_disk_set_uname_lookup(a, ucache, lookup_uname, cleanup);
+
+ return (ARCHIVE_OK);
+}
+
+static void
+cleanup(void *data)
+{
+ struct name_cache *cache = (struct name_cache *)data;
+ size_t i;
+
+ if (cache != NULL) {
+ for (i = 0; i < cache->size; i++) {
+ if (cache->cache[i].name != NULL &&
+ cache->cache[i].name != NO_NAME)
+ free((void *)(uintptr_t)cache->cache[i].name);
+ }
+ free(cache->buff);
+ free(cache);
+ }
+}
+
+/*
+ * Lookup uid/gid from uname/gname, return NULL if no match.
+ */
+static const char *
+lookup_name(struct name_cache *cache,
+ const char * (*lookup_fn)(struct name_cache *, id_t), id_t id)
+{
+ const char *name;
+ int slot;
+
+
+ cache->probes++;
+
+ slot = id % cache->size;
+ if (cache->cache[slot].name != NULL) {
+ if (cache->cache[slot].id == id) {
+ cache->hits++;
+ if (cache->cache[slot].name == NO_NAME)
+ return (NULL);
+ return (cache->cache[slot].name);
+ }
+ if (cache->cache[slot].name != NO_NAME)
+ free((void *)(uintptr_t)cache->cache[slot].name);
+ cache->cache[slot].name = NULL;
+ }
+
+ name = (lookup_fn)(cache, id);
+ if (name == NULL) {
+ /* Cache and return the negative response. */
+ cache->cache[slot].name = NO_NAME;
+ cache->cache[slot].id = id;
+ return (NULL);
+ }
+
+ cache->cache[slot].name = name;
+ cache->cache[slot].id = id;
+ return (cache->cache[slot].name);
+}
+
+static const char *
+lookup_uname(void *data, uid_t uid)
+{
+ struct name_cache *uname_cache = (struct name_cache *)data;
+ return (lookup_name(uname_cache,
+ &lookup_uname_helper, (id_t)uid));
+}
+
+#if HAVE_GETPWUID_R
+static const char *
+lookup_uname_helper(struct name_cache *cache, id_t id)
+{
+ struct passwd pwent, *result;
+ int r;
+
+ if (cache->buff_size == 0) {
+ cache->buff_size = 256;
+ cache->buff = malloc(cache->buff_size);
+ }
+ if (cache->buff == NULL)
+ return (NULL);
+ for (;;) {
+ result = &pwent; /* Old getpwuid_r ignores last arg. */
+ r = getpwuid_r((uid_t)id, &pwent,
+ cache->buff, cache->buff_size, &result);
+ if (r == 0)
+ break;
+ if (r != ERANGE)
+ break;
+ /* ERANGE means our buffer was too small, but POSIX
+ * doesn't tell us how big the buffer should be, so
+ * we just double it and try again. Because the buffer
+ * is kept around in the cache object, we shouldn't
+ * have to do this very often. */
+ cache->buff_size *= 2;
+ cache->buff = realloc(cache->buff, cache->buff_size);
+ if (cache->buff == NULL)
+ break;
+ }
+ if (r != 0) {
+ archive_set_error(cache->archive, errno,
+ "Can't lookup user for id %d", (int)id);
+ return (NULL);
+ }
+ if (result == NULL)
+ return (NULL);
+
+ return strdup(result->pw_name);
+}
+#else
+static const char *
+lookup_uname_helper(struct name_cache *cache, id_t id)
+{
+ struct passwd *result;
+
+ result = getpwuid((uid_t)id);
+
+ if (result == NULL)
+ return (NULL);
+
+ return strdup(result->pw_name);
+}
+#endif
+
+static const char *
+lookup_gname(void *data, gid_t gid)
+{
+ struct name_cache *gname_cache = (struct name_cache *)data;
+ return (lookup_name(gname_cache,
+ &lookup_gname_helper, (id_t)gid));
+}
+
+#if HAVE_GETGRGID_R
+static const char *
+lookup_gname_helper(struct name_cache *cache, id_t id)
+{
+ struct group grent, *result;
+ int r;
+
+ if (cache->buff_size == 0) {
+ cache->buff_size = 256;
+ cache->buff = malloc(cache->buff_size);
+ }
+ if (cache->buff == NULL)
+ return (NULL);
+ for (;;) {
+ result = &grent; /* Old getgrgid_r ignores last arg. */
+ r = getgrgid_r((gid_t)id, &grent,
+ cache->buff, cache->buff_size, &result);
+ if (r == 0)
+ break;
+ if (r != ERANGE)
+ break;
+ /* ERANGE means our buffer was too small, but POSIX
+ * doesn't tell us how big the buffer should be, so
+ * we just double it and try again. */
+ cache->buff_size *= 2;
+ cache->buff = realloc(cache->buff, cache->buff_size);
+ if (cache->buff == NULL)
+ break;
+ }
+ if (r != 0) {
+ archive_set_error(cache->archive, errno,
+ "Can't lookup group for id %d", (int)id);
+ return (NULL);
+ }
+ if (result == NULL)
+ return (NULL);
+
+ return strdup(result->gr_name);
+}
+#else
+static const char *
+lookup_gname_helper(struct name_cache *cache, id_t id)
+{
+ struct group *result;
+
+ result = getgrgid((gid_t)id);
+
+ if (result == NULL)
+ return (NULL);
+
+ return strdup(result->gr_name);
+}
+#endif
+
+#endif /* ! (_WIN32 && !__CYGWIN__) */
diff --git a/libarchive/archive_read_extract.c b/libarchive/archive_read_extract.c
new file mode 100644
index 000000000000..0096127702b9
--- /dev/null
+++ b/libarchive/archive_read_extract.c
@@ -0,0 +1,183 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: src/lib/libarchive/archive_read_extract.c,v 1.61 2008/05/26 17:00:22 kientzle Exp $");
+
+#ifdef HAVE_SYS_TYPES_H
+#include <sys/types.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+
+#include "archive.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+#include "archive_write_disk_private.h"
+
+struct extract {
+ struct archive *ad; /* archive_write_disk object */
+
+ /* Progress function invoked during extract. */
+ void (*extract_progress)(void *);
+ void *extract_progress_user_data;
+};
+
+static int archive_read_extract_cleanup(struct archive_read *);
+static int copy_data(struct archive *ar, struct archive *aw);
+static struct extract *get_extract(struct archive_read *);
+
+static struct extract *
+get_extract(struct archive_read *a)
+{
+ /* If we haven't initialized, do it now. */
+ /* This also sets up a lot of global state. */
+ if (a->extract == NULL) {
+ a->extract = (struct extract *)malloc(sizeof(*a->extract));
+ if (a->extract == NULL) {
+ archive_set_error(&a->archive, ENOMEM, "Can't extract");
+ return (NULL);
+ }
+ memset(a->extract, 0, sizeof(*a->extract));
+ a->extract->ad = archive_write_disk_new();
+ if (a->extract->ad == NULL) {
+ archive_set_error(&a->archive, ENOMEM, "Can't extract");
+ return (NULL);
+ }
+ archive_write_disk_set_standard_lookup(a->extract->ad);
+ a->cleanup_archive_extract = archive_read_extract_cleanup;
+ }
+ return (a->extract);
+}
+
+int
+archive_read_extract(struct archive *_a, struct archive_entry *entry, int flags)
+{
+ struct extract *extract;
+
+ extract = get_extract((struct archive_read *)_a);
+ if (extract == NULL)
+ return (ARCHIVE_FATAL);
+ archive_write_disk_set_options(extract->ad, flags);
+ return (archive_read_extract2(_a, entry, extract->ad));
+}
+
+int
+archive_read_extract2(struct archive *_a, struct archive_entry *entry,
+ struct archive *ad)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ int r, r2;
+
+ /* Set up for this particular entry. */
+ archive_write_disk_set_skip_file(ad,
+ a->skip_file_dev, a->skip_file_ino);
+ r = archive_write_header(ad, entry);
+ if (r < ARCHIVE_WARN)
+ r = ARCHIVE_WARN;
+ if (r != ARCHIVE_OK)
+ /* If _write_header failed, copy the error. */
+ archive_copy_error(&a->archive, ad);
+ else if (!archive_entry_size_is_set(entry) || archive_entry_size(entry) > 0)
+ /* Otherwise, pour data into the entry. */
+ r = copy_data(_a, ad);
+ r2 = archive_write_finish_entry(ad);
+ if (r2 < ARCHIVE_WARN)
+ r2 = ARCHIVE_WARN;
+ /* Use the first message. */
+ if (r2 != ARCHIVE_OK && r == ARCHIVE_OK)
+ archive_copy_error(&a->archive, ad);
+ /* Use the worst error return. */
+ if (r2 < r)
+ r = r2;
+ return (r);
+}
+
+void
+archive_read_extract_set_progress_callback(struct archive *_a,
+ void (*progress_func)(void *), void *user_data)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct extract *extract = get_extract(a);
+ if (extract != NULL) {
+ extract->extract_progress = progress_func;
+ extract->extract_progress_user_data = user_data;
+ }
+}
+
+static int
+copy_data(struct archive *ar, struct archive *aw)
+{
+ off_t offset;
+ const void *buff;
+ struct extract *extract;
+ size_t size;
+ int r;
+
+ extract = get_extract((struct archive_read *)ar);
+ for (;;) {
+ r = archive_read_data_block(ar, &buff, &size, &offset);
+ if (r == ARCHIVE_EOF)
+ return (ARCHIVE_OK);
+ if (r != ARCHIVE_OK)
+ return (r);
+ r = archive_write_data_block(aw, buff, size, offset);
+ if (r < ARCHIVE_WARN)
+ r = ARCHIVE_WARN;
+ if (r != ARCHIVE_OK) {
+ archive_set_error(ar, archive_errno(aw),
+ "%s", archive_error_string(aw));
+ return (r);
+ }
+ if (extract->extract_progress)
+ (extract->extract_progress)
+ (extract->extract_progress_user_data);
+ }
+}
+
+/*
+ * Cleanup function for archive_extract.
+ */
+static int
+archive_read_extract_cleanup(struct archive_read *a)
+{
+ int ret = ARCHIVE_OK;
+
+#if ARCHIVE_API_VERSION > 1
+ ret =
+#endif
+ archive_write_finish(a->extract->ad);
+ free(a->extract);
+ a->extract = NULL;
+ return (ret);
+}
diff --git a/libarchive/archive_read_open_fd.c b/libarchive/archive_read_open_fd.c
new file mode 100644
index 000000000000..9660da8f9600
--- /dev/null
+++ b/libarchive/archive_read_open_fd.c
@@ -0,0 +1,190 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_open_fd.c 201103 2009-12-28 03:13:49Z kientzle $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_FCNTL_H
+#include <fcntl.h>
+#endif
+#ifdef HAVE_IO_H
+#include <io.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+
+#include "archive.h"
+
+struct read_fd_data {
+ int fd;
+ size_t block_size;
+ char can_skip;
+ void *buffer;
+};
+
+static int file_close(struct archive *, void *);
+static ssize_t file_read(struct archive *, void *, const void **buff);
+#if ARCHIVE_API_VERSION < 2
+static ssize_t file_skip(struct archive *, void *, size_t request);
+#else
+static off_t file_skip(struct archive *, void *, off_t request);
+#endif
+
+int
+archive_read_open_fd(struct archive *a, int fd, size_t block_size)
+{
+ struct stat st;
+ struct read_fd_data *mine;
+ void *b;
+
+ archive_clear_error(a);
+ if (fstat(fd, &st) != 0) {
+ archive_set_error(a, errno, "Can't stat fd %d", fd);
+ return (ARCHIVE_FATAL);
+ }
+
+ mine = (struct read_fd_data *)malloc(sizeof(*mine));
+ b = malloc(block_size);
+ if (mine == NULL || b == NULL) {
+ archive_set_error(a, ENOMEM, "No memory");
+ free(mine);
+ free(b);
+ return (ARCHIVE_FATAL);
+ }
+ mine->block_size = block_size;
+ mine->buffer = b;
+ mine->fd = fd;
+ /*
+ * Skip support is a performance optimization for anything
+ * that supports lseek(). On FreeBSD, only regular files and
+ * raw disk devices support lseek() and there's no portable
+ * way to determine if a device is a raw disk device, so we
+ * only enable this optimization for regular files.
+ */
+ if (S_ISREG(st.st_mode)) {
+ archive_read_extract_set_skip_file(a, st.st_dev, st.st_ino);
+ mine->can_skip = 1;
+ } else
+ mine->can_skip = 0;
+#if defined(__CYGWIN__) || defined(_WIN32)
+ setmode(mine->fd, O_BINARY);
+#endif
+
+ return (archive_read_open2(a, mine,
+ NULL, file_read, file_skip, file_close));
+}
+
+static ssize_t
+file_read(struct archive *a, void *client_data, const void **buff)
+{
+ struct read_fd_data *mine = (struct read_fd_data *)client_data;
+ ssize_t bytes_read;
+
+ *buff = mine->buffer;
+ for (;;) {
+ bytes_read = read(mine->fd, mine->buffer, mine->block_size);
+ if (bytes_read < 0) {
+ if (errno == EINTR)
+ continue;
+ archive_set_error(a, errno, "Error reading fd %d", mine->fd);
+ }
+ return (bytes_read);
+ }
+}
+
+#if ARCHIVE_API_VERSION < 2
+static ssize_t
+file_skip(struct archive *a, void *client_data, size_t request)
+#else
+static off_t
+file_skip(struct archive *a, void *client_data, off_t request)
+#endif
+{
+ struct read_fd_data *mine = (struct read_fd_data *)client_data;
+ off_t old_offset, new_offset;
+
+ if (!mine->can_skip)
+ return (0);
+
+ /* Reduce request to the next smallest multiple of block_size */
+ request = (request / mine->block_size) * mine->block_size;
+ if (request == 0)
+ return (0);
+
+ /*
+ * Hurray for lazy evaluation: if the first lseek fails, the second
+ * one will not be executed.
+ */
+ if (((old_offset = lseek(mine->fd, 0, SEEK_CUR)) < 0) ||
+ ((new_offset = lseek(mine->fd, request, SEEK_CUR)) < 0))
+ {
+ /* If seek failed once, it will probably fail again. */
+ mine->can_skip = 0;
+
+ if (errno == ESPIPE)
+ {
+ /*
+ * Failure to lseek() can be caused by the file
+ * descriptor pointing to a pipe, socket or FIFO.
+ * Return 0 here, so the compression layer will use
+ * read()s instead to advance the file descriptor.
+ * It's slower of course, but works as well.
+ */
+ return (0);
+ }
+ /*
+ * There's been an error other than ESPIPE. This is most
+ * likely caused by a programmer error (too large request)
+ * or a corrupted archive file.
+ */
+ archive_set_error(a, errno, "Error seeking");
+ return (-1);
+ }
+ return (new_offset - old_offset);
+}
+
+static int
+file_close(struct archive *a, void *client_data)
+{
+ struct read_fd_data *mine = (struct read_fd_data *)client_data;
+
+ (void)a; /* UNUSED */
+ free(mine->buffer);
+ free(mine);
+ return (ARCHIVE_OK);
+}
diff --git a/libarchive/archive_read_open_file.c b/libarchive/archive_read_open_file.c
new file mode 100644
index 000000000000..095ae6eb5ce0
--- /dev/null
+++ b/libarchive/archive_read_open_file.c
@@ -0,0 +1,165 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_open_file.c 201093 2009-12-28 02:28:44Z kientzle $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_FCNTL_H
+#include <fcntl.h>
+#endif
+#ifdef HAVE_IO_H
+#include <io.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+
+#include "archive.h"
+
+struct read_FILE_data {
+ FILE *f;
+ size_t block_size;
+ void *buffer;
+ char can_skip;
+};
+
+static int file_close(struct archive *, void *);
+static ssize_t file_read(struct archive *, void *, const void **buff);
+#if ARCHIVE_API_VERSION < 2
+static ssize_t file_skip(struct archive *, void *, size_t request);
+#else
+static off_t file_skip(struct archive *, void *, off_t request);
+#endif
+
+int
+archive_read_open_FILE(struct archive *a, FILE *f)
+{
+ struct stat st;
+ struct read_FILE_data *mine;
+ size_t block_size = 128 * 1024;
+ void *b;
+
+ archive_clear_error(a);
+ mine = (struct read_FILE_data *)malloc(sizeof(*mine));
+ b = malloc(block_size);
+ if (mine == NULL || b == NULL) {
+ archive_set_error(a, ENOMEM, "No memory");
+ free(mine);
+ free(b);
+ return (ARCHIVE_FATAL);
+ }
+ mine->block_size = block_size;
+ mine->buffer = b;
+ mine->f = f;
+ /*
+ * If we can't fstat() the file, it may just be that it's not
+ * a file. (FILE * objects can wrap many kinds of I/O
+ * streams, some of which don't support fileno()).)
+ */
+ if (fstat(fileno(mine->f), &st) == 0 && S_ISREG(st.st_mode)) {
+ archive_read_extract_set_skip_file(a, st.st_dev, st.st_ino);
+ /* Enable the seek optimization only for regular files. */
+ mine->can_skip = 1;
+ } else
+ mine->can_skip = 0;
+
+#if defined(__CYGWIN__) || defined(_WIN32)
+ setmode(fileno(mine->f), O_BINARY);
+#endif
+
+ return (archive_read_open2(a, mine, NULL, file_read,
+ file_skip, file_close));
+}
+
+static ssize_t
+file_read(struct archive *a, void *client_data, const void **buff)
+{
+ struct read_FILE_data *mine = (struct read_FILE_data *)client_data;
+ ssize_t bytes_read;
+
+ *buff = mine->buffer;
+ bytes_read = fread(mine->buffer, 1, mine->block_size, mine->f);
+ if (bytes_read < 0) {
+ archive_set_error(a, errno, "Error reading file");
+ }
+ return (bytes_read);
+}
+
+#if ARCHIVE_API_VERSION < 2
+static ssize_t
+file_skip(struct archive *a, void *client_data, size_t request)
+#else
+static off_t
+file_skip(struct archive *a, void *client_data, off_t request)
+#endif
+{
+ struct read_FILE_data *mine = (struct read_FILE_data *)client_data;
+
+ (void)a; /* UNUSED */
+
+ /*
+ * If we can't skip, return 0 as the amount we did step and
+ * the caller will work around by reading and discarding.
+ */
+ if (!mine->can_skip)
+ return (0);
+ if (request == 0)
+ return (0);
+
+#if HAVE_FSEEKO
+ if (fseeko(mine->f, request, SEEK_CUR) != 0)
+#else
+ if (fseek(mine->f, request, SEEK_CUR) != 0)
+#endif
+ {
+ mine->can_skip = 0;
+ return (0);
+ }
+ return (request);
+}
+
+static int
+file_close(struct archive *a, void *client_data)
+{
+ struct read_FILE_data *mine = (struct read_FILE_data *)client_data;
+
+ (void)a; /* UNUSED */
+ if (mine->buffer != NULL)
+ free(mine->buffer);
+ free(mine);
+ return (ARCHIVE_OK);
+}
diff --git a/libarchive/archive_read_open_filename.c b/libarchive/archive_read_open_filename.c
new file mode 100644
index 000000000000..74f3e60c43ed
--- /dev/null
+++ b/libarchive/archive_read_open_filename.c
@@ -0,0 +1,272 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_open_filename.c 201093 2009-12-28 02:28:44Z kientzle $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_FCNTL_H
+#include <fcntl.h>
+#endif
+#ifdef HAVE_IO_H
+#include <io.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+
+#include "archive.h"
+
+#ifndef O_BINARY
+#define O_BINARY 0
+#endif
+
+struct read_file_data {
+ int fd;
+ size_t block_size;
+ void *buffer;
+ mode_t st_mode; /* Mode bits for opened file. */
+ char can_skip; /* This file supports skipping. */
+ char filename[1]; /* Must be last! */
+};
+
+static int file_close(struct archive *, void *);
+static ssize_t file_read(struct archive *, void *, const void **buff);
+#if ARCHIVE_API_VERSION < 2
+static ssize_t file_skip(struct archive *, void *, size_t request);
+#else
+static off_t file_skip(struct archive *, void *, off_t request);
+#endif
+
+int
+archive_read_open_file(struct archive *a, const char *filename,
+ size_t block_size)
+{
+ return (archive_read_open_filename(a, filename, block_size));
+}
+
+int
+archive_read_open_filename(struct archive *a, const char *filename,
+ size_t block_size)
+{
+ struct stat st;
+ struct read_file_data *mine;
+ void *b;
+ int fd;
+
+ archive_clear_error(a);
+ if (filename == NULL || filename[0] == '\0') {
+ /* We used to invoke archive_read_open_fd(a,0,block_size)
+ * here, but that doesn't (and shouldn't) handle the
+ * end-of-file flush when reading stdout from a pipe.
+ * Basically, read_open_fd() is intended for folks who
+ * are willing to handle such details themselves. This
+ * API is intended to be a little smarter for folks who
+ * want easy handling of the common case.
+ */
+ filename = ""; /* Normalize NULL to "" */
+ fd = 0;
+#if defined(__CYGWIN__) || defined(_WIN32)
+ setmode(0, O_BINARY);
+#endif
+ } else {
+ fd = open(filename, O_RDONLY | O_BINARY);
+ if (fd < 0) {
+ archive_set_error(a, errno,
+ "Failed to open '%s'", filename);
+ return (ARCHIVE_FATAL);
+ }
+ }
+ if (fstat(fd, &st) != 0) {
+ archive_set_error(a, errno, "Can't stat '%s'", filename);
+ return (ARCHIVE_FATAL);
+ }
+
+ mine = (struct read_file_data *)calloc(1,
+ sizeof(*mine) + strlen(filename));
+ b = malloc(block_size);
+ if (mine == NULL || b == NULL) {
+ archive_set_error(a, ENOMEM, "No memory");
+ free(mine);
+ free(b);
+ return (ARCHIVE_FATAL);
+ }
+ strcpy(mine->filename, filename);
+ mine->block_size = block_size;
+ mine->buffer = b;
+ mine->fd = fd;
+ /* Remember mode so close can decide whether to flush. */
+ mine->st_mode = st.st_mode;
+ /* If we're reading a file from disk, ensure that we don't
+ overwrite it with an extracted file. */
+ if (S_ISREG(st.st_mode)) {
+ archive_read_extract_set_skip_file(a, st.st_dev, st.st_ino);
+ /*
+ * Enabling skip here is a performance optimization
+ * for anything that supports lseek(). On FreeBSD
+ * (and probably many other systems), only regular
+ * files and raw disk devices support lseek() (on
+ * other input types, lseek() returns success but
+ * doesn't actually change the file pointer, which
+ * just completely screws up the position-tracking
+ * logic). In addition, I've yet to find a portable
+ * way to determine if a device is a raw disk device.
+ * So I don't see a way to do much better than to only
+ * enable this optimization for regular files.
+ */
+ mine->can_skip = 1;
+ }
+ return (archive_read_open2(a, mine,
+ NULL, file_read, file_skip, file_close));
+}
+
+static ssize_t
+file_read(struct archive *a, void *client_data, const void **buff)
+{
+ struct read_file_data *mine = (struct read_file_data *)client_data;
+ ssize_t bytes_read;
+
+ *buff = mine->buffer;
+ for (;;) {
+ bytes_read = read(mine->fd, mine->buffer, mine->block_size);
+ if (bytes_read < 0) {
+ if (errno == EINTR)
+ continue;
+ else if (mine->filename[0] == '\0')
+ archive_set_error(a, errno, "Error reading stdin");
+ else
+ archive_set_error(a, errno, "Error reading '%s'",
+ mine->filename);
+ }
+ return (bytes_read);
+ }
+}
+
+#if ARCHIVE_API_VERSION < 2
+static ssize_t
+file_skip(struct archive *a, void *client_data, size_t request)
+#else
+static off_t
+file_skip(struct archive *a, void *client_data, off_t request)
+#endif
+{
+ struct read_file_data *mine = (struct read_file_data *)client_data;
+ off_t old_offset, new_offset;
+
+ if (!mine->can_skip) /* We can't skip, so ... */
+ return (0); /* ... skip zero bytes. */
+
+ /* Reduce request to the next smallest multiple of block_size */
+ request = (request / mine->block_size) * mine->block_size;
+ if (request == 0)
+ return (0);
+
+ /*
+ * Hurray for lazy evaluation: if the first lseek fails, the second
+ * one will not be executed.
+ */
+ if (((old_offset = lseek(mine->fd, 0, SEEK_CUR)) < 0) ||
+ ((new_offset = lseek(mine->fd, request, SEEK_CUR)) < 0))
+ {
+ /* If skip failed once, it will probably fail again. */
+ mine->can_skip = 0;
+
+ if (errno == ESPIPE)
+ {
+ /*
+ * Failure to lseek() can be caused by the file
+ * descriptor pointing to a pipe, socket or FIFO.
+ * Return 0 here, so the compression layer will use
+ * read()s instead to advance the file descriptor.
+ * It's slower of course, but works as well.
+ */
+ return (0);
+ }
+ /*
+ * There's been an error other than ESPIPE. This is most
+ * likely caused by a programmer error (too large request)
+ * or a corrupted archive file.
+ */
+ if (mine->filename[0] == '\0')
+ /*
+ * Should never get here, since lseek() on stdin ought
+ * to return an ESPIPE error.
+ */
+ archive_set_error(a, errno, "Error seeking in stdin");
+ else
+ archive_set_error(a, errno, "Error seeking in '%s'",
+ mine->filename);
+ return (-1);
+ }
+ return (new_offset - old_offset);
+}
+
+static int
+file_close(struct archive *a, void *client_data)
+{
+ struct read_file_data *mine = (struct read_file_data *)client_data;
+
+ (void)a; /* UNUSED */
+
+ /* Only flush and close if open succeeded. */
+ if (mine->fd >= 0) {
+ /*
+ * Sometimes, we should flush the input before closing.
+ * Regular files: faster to just close without flush.
+ * Devices: must not flush (user might need to
+ * read the "next" item on a non-rewind device).
+ * Pipes and sockets: must flush (otherwise, the
+ * program feeding the pipe or socket may complain).
+ * Here, I flush everything except for regular files and
+ * device nodes.
+ */
+ if (!S_ISREG(mine->st_mode)
+ && !S_ISCHR(mine->st_mode)
+ && !S_ISBLK(mine->st_mode)) {
+ ssize_t bytesRead;
+ do {
+ bytesRead = read(mine->fd, mine->buffer,
+ mine->block_size);
+ } while (bytesRead > 0);
+ }
+ /* If a named file was opened, then it needs to be closed. */
+ if (mine->filename[0] != '\0')
+ close(mine->fd);
+ }
+ free(mine->buffer);
+ free(mine);
+ return (ARCHIVE_OK);
+}
diff --git a/libarchive/archive_read_open_memory.c b/libarchive/archive_read_open_memory.c
new file mode 100644
index 000000000000..61f574fa7705
--- /dev/null
+++ b/libarchive/archive_read_open_memory.c
@@ -0,0 +1,156 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: src/lib/libarchive/archive_read_open_memory.c,v 1.6 2007/07/06 15:51:59 kientzle Exp $");
+
+#include <errno.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "archive.h"
+
+/*
+ * Glue to read an archive from a block of memory.
+ *
+ * This is mostly a huge help in building test harnesses;
+ * test programs can build archives in memory and read them
+ * back again without having to mess with files on disk.
+ */
+
+struct read_memory_data {
+ unsigned char *buffer;
+ unsigned char *end;
+ ssize_t read_size;
+};
+
+static int memory_read_close(struct archive *, void *);
+static int memory_read_open(struct archive *, void *);
+#if ARCHIVE_API_VERSION < 2
+static ssize_t memory_read_skip(struct archive *, void *, size_t request);
+#else
+static off_t memory_read_skip(struct archive *, void *, off_t request);
+#endif
+static ssize_t memory_read(struct archive *, void *, const void **buff);
+
+int
+archive_read_open_memory(struct archive *a, void *buff, size_t size)
+{
+ return archive_read_open_memory2(a, buff, size, size);
+}
+
+/*
+ * Don't use _open_memory2() in production code; the archive_read_open_memory()
+ * version is the one you really want. This is just here so that
+ * test harnesses can exercise block operations inside the library.
+ */
+int
+archive_read_open_memory2(struct archive *a, void *buff,
+ size_t size, size_t read_size)
+{
+ struct read_memory_data *mine;
+
+ mine = (struct read_memory_data *)malloc(sizeof(*mine));
+ if (mine == NULL) {
+ archive_set_error(a, ENOMEM, "No memory");
+ return (ARCHIVE_FATAL);
+ }
+ memset(mine, 0, sizeof(*mine));
+ mine->buffer = (unsigned char *)buff;
+ mine->end = mine->buffer + size;
+ mine->read_size = read_size;
+ return (archive_read_open2(a, mine, memory_read_open,
+ memory_read, memory_read_skip, memory_read_close));
+}
+
+/*
+ * There's nothing to open.
+ */
+static int
+memory_read_open(struct archive *a, void *client_data)
+{
+ (void)a; /* UNUSED */
+ (void)client_data; /* UNUSED */
+ return (ARCHIVE_OK);
+}
+
+/*
+ * This is scary simple: Just advance a pointer. Limiting
+ * to read_size is not technically necessary, but it exercises
+ * more of the internal logic when used with a small block size
+ * in a test harness. Production use should not specify a block
+ * size; then this is much faster.
+ */
+static ssize_t
+memory_read(struct archive *a, void *client_data, const void **buff)
+{
+ struct read_memory_data *mine = (struct read_memory_data *)client_data;
+ ssize_t size;
+
+ (void)a; /* UNUSED */
+ *buff = mine->buffer;
+ size = mine->end - mine->buffer;
+ if (size > mine->read_size)
+ size = mine->read_size;
+ mine->buffer += size;
+ return (size);
+}
+
+/*
+ * Advancing is just as simple. Again, this is doing more than
+ * necessary in order to better exercise internal code when used
+ * as a test harness.
+ */
+#if ARCHIVE_API_VERSION < 2
+static ssize_t
+memory_read_skip(struct archive *a, void *client_data, size_t skip)
+#else
+static off_t
+memory_read_skip(struct archive *a, void *client_data, off_t skip)
+#endif
+{
+ struct read_memory_data *mine = (struct read_memory_data *)client_data;
+
+ (void)a; /* UNUSED */
+ if ((off_t)skip > (off_t)(mine->end - mine->buffer))
+ skip = mine->end - mine->buffer;
+ /* Round down to block size. */
+ skip /= mine->read_size;
+ skip *= mine->read_size;
+ mine->buffer += skip;
+ return (skip);
+}
+
+/*
+ * Close is just cleaning up our one small bit of data.
+ */
+static int
+memory_read_close(struct archive *a, void *client_data)
+{
+ struct read_memory_data *mine = (struct read_memory_data *)client_data;
+ (void)a; /* UNUSED */
+ free(mine);
+ return (ARCHIVE_OK);
+}
diff --git a/libarchive/archive_read_private.h b/libarchive/archive_read_private.h
new file mode 100644
index 000000000000..5a850189042a
--- /dev/null
+++ b/libarchive/archive_read_private.h
@@ -0,0 +1,199 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD: head/lib/libarchive/archive_read_private.h 201088 2009-12-28 02:18:55Z kientzle $
+ */
+
+#ifndef __LIBARCHIVE_BUILD
+#error This header is only to be used internally to libarchive.
+#endif
+
+#ifndef ARCHIVE_READ_PRIVATE_H_INCLUDED
+#define ARCHIVE_READ_PRIVATE_H_INCLUDED
+
+#include "archive.h"
+#include "archive_string.h"
+#include "archive_private.h"
+
+struct archive_read;
+struct archive_read_filter_bidder;
+struct archive_read_filter;
+
+/*
+ * How bidding works for filters:
+ * * The bid manager reads the first block from the current source.
+ * * It shows that block to each registered bidder.
+ * * The bid manager creates a new filter structure for the winning
+ * bidder and gives the winning bidder a chance to initialize it.
+ * * The new filter becomes the top filter in the archive_read structure
+ * and we repeat the process.
+ * This ends only when no bidder provides a non-zero bid.
+ */
+struct archive_read_filter_bidder {
+ /* Configuration data for the bidder. */
+ void *data;
+ /* Taste the upstream filter to see if we handle this. */
+ int (*bid)(struct archive_read_filter_bidder *,
+ struct archive_read_filter *);
+ /* Initialize a newly-created filter. */
+ int (*init)(struct archive_read_filter *);
+ /* Set an option for the filter bidder. */
+ int (*options)(struct archive_read_filter_bidder *,
+ const char *key, const char *value);
+ /* Release the bidder's configuration data. */
+ int (*free)(struct archive_read_filter_bidder *);
+};
+
+/*
+ * This structure is allocated within the archive_read core
+ * and initialized by archive_read and the init() method of the
+ * corresponding bidder above.
+ */
+struct archive_read_filter {
+ /* Essentially all filters will need these values, so
+ * just declare them here. */
+ struct archive_read_filter_bidder *bidder; /* My bidder. */
+ struct archive_read_filter *upstream; /* Who I read from. */
+ struct archive_read *archive; /* Associated archive. */
+ /* Return next block. */
+ ssize_t (*read)(struct archive_read_filter *, const void **);
+ /* Skip forward this many bytes. */
+ int64_t (*skip)(struct archive_read_filter *self, int64_t request);
+ /* Close (just this filter) and free(self). */
+ int (*close)(struct archive_read_filter *self);
+ /* My private data. */
+ void *data;
+
+ const char *name;
+ int code;
+
+ /* Used by reblocking logic. */
+ char *buffer;
+ size_t buffer_size;
+ char *next; /* Current read location. */
+ size_t avail; /* Bytes in my buffer. */
+ const void *client_buff; /* Client buffer information. */
+ size_t client_total;
+ const char *client_next;
+ size_t client_avail;
+ int64_t position;
+ char end_of_file;
+ char fatal;
+};
+
+/*
+ * The client looks a lot like a filter, so we just wrap it here.
+ *
+ * TODO: Make archive_read_filter and archive_read_client identical so
+ * that users of the library can easily register their own
+ * transformation filters. This will probably break the API/ABI and
+ * so should be deferred at least until libarchive 3.0.
+ */
+struct archive_read_client {
+ archive_read_callback *reader;
+ archive_skip_callback *skipper;
+ archive_close_callback *closer;
+};
+
+struct archive_read {
+ struct archive archive;
+
+ struct archive_entry *entry;
+
+ /* Dev/ino of the archive being read/written. */
+ dev_t skip_file_dev;
+ ino_t skip_file_ino;
+
+ /*
+ * Used by archive_read_data() to track blocks and copy
+ * data to client buffers, filling gaps with zero bytes.
+ */
+ const char *read_data_block;
+ off_t read_data_offset;
+ off_t read_data_output_offset;
+ size_t read_data_remaining;
+
+ /* Callbacks to open/read/write/close client archive stream. */
+ struct archive_read_client client;
+
+ /* Registered filter bidders. */
+ struct archive_read_filter_bidder bidders[8];
+
+ /* Last filter in chain */
+ struct archive_read_filter *filter;
+
+ /* File offset of beginning of most recently-read header. */
+ off_t header_position;
+
+ /*
+ * Format detection is mostly the same as compression
+ * detection, with one significant difference: The bidders
+ * use the read_ahead calls above to examine the stream rather
+ * than having the supervisor hand them a block of data to
+ * examine.
+ */
+
+ struct archive_format_descriptor {
+ void *data;
+ const char *name;
+ int (*bid)(struct archive_read *);
+ int (*options)(struct archive_read *, const char *key,
+ const char *value);
+ int (*read_header)(struct archive_read *, struct archive_entry *);
+ int (*read_data)(struct archive_read *, const void **, size_t *, off_t *);
+ int (*read_data_skip)(struct archive_read *);
+ int (*cleanup)(struct archive_read *);
+ } formats[9];
+ struct archive_format_descriptor *format; /* Active format. */
+
+ /*
+ * Various information needed by archive_extract.
+ */
+ struct extract *extract;
+ int (*cleanup_archive_extract)(struct archive_read *);
+};
+
+int __archive_read_register_format(struct archive_read *a,
+ void *format_data,
+ const char *name,
+ int (*bid)(struct archive_read *),
+ int (*options)(struct archive_read *, const char *, const char *),
+ int (*read_header)(struct archive_read *, struct archive_entry *),
+ int (*read_data)(struct archive_read *, const void **, size_t *, off_t *),
+ int (*read_data_skip)(struct archive_read *),
+ int (*cleanup)(struct archive_read *));
+
+struct archive_read_filter_bidder
+ *__archive_read_get_bidder(struct archive_read *a);
+
+const void *__archive_read_ahead(struct archive_read *, size_t, ssize_t *);
+const void *__archive_read_filter_ahead(struct archive_read_filter *,
+ size_t, ssize_t *);
+ssize_t __archive_read_consume(struct archive_read *, size_t);
+ssize_t __archive_read_filter_consume(struct archive_read_filter *, size_t);
+int64_t __archive_read_skip(struct archive_read *, int64_t);
+int64_t __archive_read_skip_lenient(struct archive_read *, int64_t);
+int64_t __archive_read_filter_skip(struct archive_read_filter *, int64_t);
+int __archive_read_program(struct archive_read_filter *, const char *);
+#endif
diff --git a/libarchive/archive_read_support_compression_all.c b/libarchive/archive_read_support_compression_all.c
new file mode 100644
index 000000000000..a6db7364d098
--- /dev/null
+++ b/libarchive/archive_read_support_compression_all.c
@@ -0,0 +1,60 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_compression_all.c 201248 2009-12-30 06:12:03Z kientzle $");
+
+#include "archive.h"
+
+int
+archive_read_support_compression_all(struct archive *a)
+{
+ /* Bzip falls back to "bunzip2" command-line */
+ archive_read_support_compression_bzip2(a);
+ /* The decompress code doesn't use an outside library. */
+ archive_read_support_compression_compress(a);
+ /* Gzip decompress falls back to "gunzip" command-line. */
+ archive_read_support_compression_gzip(a);
+ /* The LZMA file format has a very weak signature, so it
+ * may not be feasible to keep this here, but we'll try.
+ * This will come back out if there are problems. */
+ /* Lzma falls back to "unlzma" command-line program. */
+ archive_read_support_compression_lzma(a);
+ /* Xz falls back to "unxz" command-line program. */
+ archive_read_support_compression_xz(a);
+ /* The decode code doesn't use an outside library. */
+ archive_read_support_compression_uu(a);
+ /* The decode code doesn't use an outside library. */
+ archive_read_support_compression_rpm(a);
+
+ /* Note: We always return ARCHIVE_OK here, even if some of the
+ * above return ARCHIVE_WARN. The intent here is to enable
+ * "as much as possible." Clients who need specific
+ * compression should enable those individually so they can
+ * verify the level of support. */
+ /* Clear any warning messages set by the above functions. */
+ archive_clear_error(a);
+ return (ARCHIVE_OK);
+}
diff --git a/libarchive/archive_read_support_compression_bzip2.c b/libarchive/archive_read_support_compression_bzip2.c
new file mode 100644
index 000000000000..8381c9a578c6
--- /dev/null
+++ b/libarchive/archive_read_support_compression_bzip2.c
@@ -0,0 +1,353 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_compression_bzip2.c 201108 2009-12-28 03:28:21Z kientzle $");
+
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#include <stdio.h>
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+#ifdef HAVE_BZLIB_H
+#include <bzlib.h>
+#endif
+
+#include "archive.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+#if defined(HAVE_BZLIB_H) && defined(BZ_CONFIG_ERROR)
+struct private_data {
+ bz_stream stream;
+ char *out_block;
+ size_t out_block_size;
+ char valid; /* True = decompressor is initialized */
+ char eof; /* True = found end of compressed data. */
+};
+
+/* Bzip2 filter */
+static ssize_t bzip2_filter_read(struct archive_read_filter *, const void **);
+static int bzip2_filter_close(struct archive_read_filter *);
+#endif
+
+/*
+ * Note that we can detect bzip2 archives even if we can't decompress
+ * them. (In fact, we like detecting them because we can give better
+ * error messages.) So the bid framework here gets compiled even
+ * if bzlib is unavailable.
+ */
+static int bzip2_reader_bid(struct archive_read_filter_bidder *, struct archive_read_filter *);
+static int bzip2_reader_init(struct archive_read_filter *);
+static int bzip2_reader_free(struct archive_read_filter_bidder *);
+
+int
+archive_read_support_compression_bzip2(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct archive_read_filter_bidder *reader = __archive_read_get_bidder(a);
+
+ if (reader == NULL)
+ return (ARCHIVE_FATAL);
+
+ reader->data = NULL;
+ reader->bid = bzip2_reader_bid;
+ reader->init = bzip2_reader_init;
+ reader->options = NULL;
+ reader->free = bzip2_reader_free;
+#if defined(HAVE_BZLIB_H) && defined(BZ_CONFIG_ERROR)
+ return (ARCHIVE_OK);
+#else
+ archive_set_error(_a, ARCHIVE_ERRNO_MISC,
+ "Using external bunzip2 program");
+ return (ARCHIVE_WARN);
+#endif
+}
+
+static int
+bzip2_reader_free(struct archive_read_filter_bidder *self){
+ (void)self; /* UNUSED */
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Test whether we can handle this data.
+ *
+ * This logic returns zero if any part of the signature fails. It
+ * also tries to Do The Right Thing if a very short buffer prevents us
+ * from verifying as much as we would like.
+ */
+static int
+bzip2_reader_bid(struct archive_read_filter_bidder *self, struct archive_read_filter *filter)
+{
+ const unsigned char *buffer;
+ ssize_t avail;
+ int bits_checked;
+
+ (void)self; /* UNUSED */
+
+ /* Minimal bzip2 archive is 14 bytes. */
+ buffer = __archive_read_filter_ahead(filter, 14, &avail);
+ if (buffer == NULL)
+ return (0);
+
+ /* First three bytes must be "BZh" */
+ bits_checked = 0;
+ if (buffer[0] != 'B' || buffer[1] != 'Z' || buffer[2] != 'h')
+ return (0);
+ bits_checked += 24;
+
+ /* Next follows a compression flag which must be an ASCII digit. */
+ if (buffer[3] < '1' || buffer[3] > '9')
+ return (0);
+ bits_checked += 5;
+
+ /* After BZh[1-9], there must be either a data block
+ * which begins with 0x314159265359 or an end-of-data
+ * marker of 0x177245385090. */
+ if (memcmp(buffer + 4, "\x31\x41\x59\x26\x53\x59", 6) == 0)
+ bits_checked += 48;
+ else if (memcmp(buffer + 4, "\x17\x72\x45\x38\x50\x90", 6) == 0)
+ bits_checked += 48;
+ else
+ return (0);
+
+ return (bits_checked);
+}
+
+#if !defined(HAVE_BZLIB_H) || !defined(BZ_CONFIG_ERROR)
+
+/*
+ * If we don't have the library on this system, we can't actually do the
+ * decompression. We can, however, still detect compressed archives
+ * and emit a useful message.
+ */
+static int
+bzip2_reader_init(struct archive_read_filter *self)
+{
+ int r;
+
+ r = __archive_read_program(self, "bunzip2");
+ /* Note: We set the format here even if __archive_read_program()
+ * above fails. We do, after all, know what the format is
+ * even if we weren't able to read it. */
+ self->code = ARCHIVE_COMPRESSION_BZIP2;
+ self->name = "bzip2";
+ return (r);
+}
+
+
+#else
+
+/*
+ * Setup the callbacks.
+ */
+static int
+bzip2_reader_init(struct archive_read_filter *self)
+{
+ static const size_t out_block_size = 64 * 1024;
+ void *out_block;
+ struct private_data *state;
+
+ self->code = ARCHIVE_COMPRESSION_BZIP2;
+ self->name = "bzip2";
+
+ state = (struct private_data *)calloc(sizeof(*state), 1);
+ out_block = (unsigned char *)malloc(out_block_size);
+ if (self == NULL || state == NULL || out_block == NULL) {
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Can't allocate data for bzip2 decompression");
+ free(out_block);
+ free(state);
+ return (ARCHIVE_FATAL);
+ }
+
+ self->data = state;
+ state->out_block_size = out_block_size;
+ state->out_block = out_block;
+ self->read = bzip2_filter_read;
+ self->skip = NULL; /* not supported */
+ self->close = bzip2_filter_close;
+
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Return the next block of decompressed data.
+ */
+static ssize_t
+bzip2_filter_read(struct archive_read_filter *self, const void **p)
+{
+ struct private_data *state;
+ size_t decompressed;
+ const char *read_buf;
+ ssize_t ret;
+
+ state = (struct private_data *)self->data;
+
+ if (state->eof) {
+ *p = NULL;
+ return (0);
+ }
+
+ /* Empty our output buffer. */
+ state->stream.next_out = state->out_block;
+ state->stream.avail_out = state->out_block_size;
+
+ /* Try to fill the output buffer. */
+ for (;;) {
+ if (!state->valid) {
+ if (bzip2_reader_bid(self->bidder, self->upstream) == 0) {
+ state->eof = 1;
+ *p = state->out_block;
+ decompressed = state->stream.next_out
+ - state->out_block;
+ return (decompressed);
+ }
+ /* Initialize compression library. */
+ ret = BZ2_bzDecompressInit(&(state->stream),
+ 0 /* library verbosity */,
+ 0 /* don't use low-mem algorithm */);
+
+ /* If init fails, try low-memory algorithm instead. */
+ if (ret == BZ_MEM_ERROR)
+ ret = BZ2_bzDecompressInit(&(state->stream),
+ 0 /* library verbosity */,
+ 1 /* do use low-mem algo */);
+
+ if (ret != BZ_OK) {
+ const char *detail = NULL;
+ int err = ARCHIVE_ERRNO_MISC;
+ switch (ret) {
+ case BZ_PARAM_ERROR:
+ detail = "invalid setup parameter";
+ break;
+ case BZ_MEM_ERROR:
+ err = ENOMEM;
+ detail = "out of memory";
+ break;
+ case BZ_CONFIG_ERROR:
+ detail = "mis-compiled library";
+ break;
+ }
+ archive_set_error(&self->archive->archive, err,
+ "Internal error initializing decompressor%s%s",
+ detail == NULL ? "" : ": ",
+ detail);
+ return (ARCHIVE_FATAL);
+ }
+ state->valid = 1;
+ }
+
+ /* stream.next_in is really const, but bzlib
+ * doesn't declare it so. <sigh> */
+ read_buf =
+ __archive_read_filter_ahead(self->upstream, 1, &ret);
+ if (read_buf == NULL)
+ return (ARCHIVE_FATAL);
+ state->stream.next_in = (char *)(uintptr_t)read_buf;
+ state->stream.avail_in = ret;
+ /* There is no more data, return whatever we have. */
+ if (ret == 0) {
+ state->eof = 1;
+ *p = state->out_block;
+ decompressed = state->stream.next_out
+ - state->out_block;
+ return (decompressed);
+ }
+
+ /* Decompress as much as we can in one pass. */
+ ret = BZ2_bzDecompress(&(state->stream));
+ __archive_read_filter_consume(self->upstream,
+ state->stream.next_in - read_buf);
+
+ switch (ret) {
+ case BZ_STREAM_END: /* Found end of stream. */
+ switch (BZ2_bzDecompressEnd(&(state->stream))) {
+ case BZ_OK:
+ break;
+ default:
+ archive_set_error(&(self->archive->archive),
+ ARCHIVE_ERRNO_MISC,
+ "Failed to clean up decompressor");
+ return (ARCHIVE_FATAL);
+ }
+ state->valid = 0;
+ /* FALLTHROUGH */
+ case BZ_OK: /* Decompressor made some progress. */
+ /* If we filled our buffer, update stats and return. */
+ if (state->stream.avail_out == 0) {
+ *p = state->out_block;
+ decompressed = state->stream.next_out
+ - state->out_block;
+ return (decompressed);
+ }
+ break;
+ default: /* Return an error. */
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC, "bzip decompression failed");
+ return (ARCHIVE_FATAL);
+ }
+ }
+}
+
+/*
+ * Clean up the decompressor.
+ */
+static int
+bzip2_filter_close(struct archive_read_filter *self)
+{
+ struct private_data *state;
+ int ret = ARCHIVE_OK;
+
+ state = (struct private_data *)self->data;
+
+ if (state->valid) {
+ switch (BZ2_bzDecompressEnd(&state->stream)) {
+ case BZ_OK:
+ break;
+ default:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Failed to clean up decompressor");
+ ret = ARCHIVE_FATAL;
+ }
+ }
+
+ free(state->out_block);
+ free(state);
+ return (ret);
+}
+
+#endif /* HAVE_BZLIB_H && BZ_CONFIG_ERROR */
diff --git a/libarchive/archive_read_support_compression_compress.c b/libarchive/archive_read_support_compression_compress.c
new file mode 100644
index 000000000000..2461975e59bc
--- /dev/null
+++ b/libarchive/archive_read_support_compression_compress.c
@@ -0,0 +1,444 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+/*
+ * This code borrows heavily from "compress" source code, which is
+ * protected by the following copyright. (Clause 3 dropped by request
+ * of the Regents.)
+ */
+
+/*-
+ * Copyright (c) 1985, 1986, 1992, 1993
+ * The Regents of the University of California. All rights reserved.
+ *
+ * This code is derived from software contributed to Berkeley by
+ * Diomidis Spinellis and James A. Woods, derived from original
+ * work by Spencer Thomas and Joseph Orost.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 4. Neither the name of the University nor the names of its contributors
+ * may be used to endorse or promote products derived from this software
+ * without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_compression_compress.c 201094 2009-12-28 02:29:21Z kientzle $");
+
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+
+#include "archive.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+/*
+ * Because LZW decompression is pretty simple, I've just implemented
+ * the whole decompressor here (cribbing from "compress" source code,
+ * of course), rather than relying on an external library. I have
+ * made an effort to clarify and simplify the algorithm, so the
+ * names and structure here don't exactly match those used by compress.
+ */
+
+struct private_data {
+ /* Input variables. */
+ const unsigned char *next_in;
+ size_t avail_in;
+ int bit_buffer;
+ int bits_avail;
+ size_t bytes_in_section;
+
+ /* Output variables. */
+ size_t out_block_size;
+ void *out_block;
+
+ /* Decompression status variables. */
+ int use_reset_code;
+ int end_of_stream; /* EOF status. */
+ int maxcode; /* Largest code. */
+ int maxcode_bits; /* Length of largest code. */
+ int section_end_code; /* When to increase bits. */
+ int bits; /* Current code length. */
+ int oldcode; /* Previous code. */
+ int finbyte; /* Last byte of prev code. */
+
+ /* Dictionary. */
+ int free_ent; /* Next dictionary entry. */
+ unsigned char suffix[65536];
+ uint16_t prefix[65536];
+
+ /*
+ * Scratch area for expanding dictionary entries. Note:
+ * "worst" case here comes from compressing /dev/zero: the
+ * last code in the dictionary will code a sequence of
+ * 65536-256 zero bytes. Thus, we need stack space to expand
+ * a 65280-byte dictionary entry. (Of course, 32640:1
+ * compression could also be considered the "best" case. ;-)
+ */
+ unsigned char *stackp;
+ unsigned char stack[65300];
+};
+
+static int compress_bidder_bid(struct archive_read_filter_bidder *, struct archive_read_filter *);
+static int compress_bidder_init(struct archive_read_filter *);
+static int compress_bidder_free(struct archive_read_filter_bidder *);
+
+static ssize_t compress_filter_read(struct archive_read_filter *, const void **);
+static int compress_filter_close(struct archive_read_filter *);
+
+static int getbits(struct archive_read_filter *, int n);
+static int next_code(struct archive_read_filter *);
+
+int
+archive_read_support_compression_compress(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct archive_read_filter_bidder *bidder = __archive_read_get_bidder(a);
+
+ if (bidder == NULL)
+ return (ARCHIVE_FATAL);
+
+ bidder->data = NULL;
+ bidder->bid = compress_bidder_bid;
+ bidder->init = compress_bidder_init;
+ bidder->options = NULL;
+ bidder->free = compress_bidder_free;
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Test whether we can handle this data.
+ *
+ * This logic returns zero if any part of the signature fails. It
+ * also tries to Do The Right Thing if a very short buffer prevents us
+ * from verifying as much as we would like.
+ */
+static int
+compress_bidder_bid(struct archive_read_filter_bidder *self,
+ struct archive_read_filter *filter)
+{
+ const unsigned char *buffer;
+ ssize_t avail;
+ int bits_checked;
+
+ (void)self; /* UNUSED */
+
+ buffer = __archive_read_filter_ahead(filter, 2, &avail);
+
+ if (buffer == NULL)
+ return (0);
+
+ bits_checked = 0;
+ if (buffer[0] != 037) /* Verify first ID byte. */
+ return (0);
+ bits_checked += 8;
+
+ if (buffer[1] != 0235) /* Verify second ID byte. */
+ return (0);
+ bits_checked += 8;
+
+ /*
+ * TODO: Verify more.
+ */
+
+ return (bits_checked);
+}
+
+/*
+ * Setup the callbacks.
+ */
+static int
+compress_bidder_init(struct archive_read_filter *self)
+{
+ struct private_data *state;
+ static const size_t out_block_size = 64 * 1024;
+ void *out_block;
+ int code;
+
+ self->code = ARCHIVE_COMPRESSION_COMPRESS;
+ self->name = "compress (.Z)";
+
+ state = (struct private_data *)calloc(sizeof(*state), 1);
+ out_block = malloc(out_block_size);
+ if (state == NULL || out_block == NULL) {
+ free(out_block);
+ free(state);
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Can't allocate data for %s decompression",
+ self->name);
+ return (ARCHIVE_FATAL);
+ }
+
+ self->data = state;
+ state->out_block_size = out_block_size;
+ state->out_block = out_block;
+ self->read = compress_filter_read;
+ self->skip = NULL; /* not supported */
+ self->close = compress_filter_close;
+
+ /* XXX MOVE THE FOLLOWING OUT OF INIT() XXX */
+
+ (void)getbits(self, 8); /* Skip first signature byte. */
+ (void)getbits(self, 8); /* Skip second signature byte. */
+
+ code = getbits(self, 8);
+ state->maxcode_bits = code & 0x1f;
+ state->maxcode = (1 << state->maxcode_bits);
+ state->use_reset_code = code & 0x80;
+
+ /* Initialize decompressor. */
+ state->free_ent = 256;
+ state->stackp = state->stack;
+ if (state->use_reset_code)
+ state->free_ent++;
+ state->bits = 9;
+ state->section_end_code = (1<<state->bits) - 1;
+ state->oldcode = -1;
+ for (code = 255; code >= 0; code--) {
+ state->prefix[code] = 0;
+ state->suffix[code] = code;
+ }
+ next_code(self);
+
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Return a block of data from the decompression buffer. Decompress more
+ * as necessary.
+ */
+static ssize_t
+compress_filter_read(struct archive_read_filter *self, const void **pblock)
+{
+ struct private_data *state;
+ unsigned char *p, *start, *end;
+ int ret;
+
+ state = (struct private_data *)self->data;
+ if (state->end_of_stream) {
+ *pblock = NULL;
+ return (0);
+ }
+ p = start = (unsigned char *)state->out_block;
+ end = start + state->out_block_size;
+
+ while (p < end && !state->end_of_stream) {
+ if (state->stackp > state->stack) {
+ *p++ = *--state->stackp;
+ } else {
+ ret = next_code(self);
+ if (ret == -1)
+ state->end_of_stream = ret;
+ else if (ret != ARCHIVE_OK)
+ return (ret);
+ }
+ }
+
+ *pblock = start;
+ return (p - start);
+}
+
+/*
+ * Clean up the reader.
+ */
+static int
+compress_bidder_free(struct archive_read_filter_bidder *self)
+{
+ self->data = NULL;
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Close and release the filter.
+ */
+static int
+compress_filter_close(struct archive_read_filter *self)
+{
+ struct private_data *state = (struct private_data *)self->data;
+
+ free(state->out_block);
+ free(state);
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Process the next code and fill the stack with the expansion
+ * of the code. Returns ARCHIVE_FATAL if there is a fatal I/O or
+ * format error, ARCHIVE_EOF if we hit end of data, ARCHIVE_OK otherwise.
+ */
+static int
+next_code(struct archive_read_filter *self)
+{
+ struct private_data *state = (struct private_data *)self->data;
+ int code, newcode;
+
+ static int debug_buff[1024];
+ static unsigned debug_index;
+
+ code = newcode = getbits(self, state->bits);
+ if (code < 0)
+ return (code);
+
+ debug_buff[debug_index++] = code;
+ if (debug_index >= sizeof(debug_buff)/sizeof(debug_buff[0]))
+ debug_index = 0;
+
+ /* If it's a reset code, reset the dictionary. */
+ if ((code == 256) && state->use_reset_code) {
+ /*
+ * The original 'compress' implementation blocked its
+ * I/O in a manner that resulted in junk bytes being
+ * inserted after every reset. The next section skips
+ * this junk. (Yes, the number of *bytes* to skip is
+ * a function of the current *bit* length.)
+ */
+ int skip_bytes = state->bits -
+ (state->bytes_in_section % state->bits);
+ skip_bytes %= state->bits;
+ state->bits_avail = 0; /* Discard rest of this byte. */
+ while (skip_bytes-- > 0) {
+ code = getbits(self, 8);
+ if (code < 0)
+ return (code);
+ }
+ /* Now, actually do the reset. */
+ state->bytes_in_section = 0;
+ state->bits = 9;
+ state->section_end_code = (1 << state->bits) - 1;
+ state->free_ent = 257;
+ state->oldcode = -1;
+ return (next_code(self));
+ }
+
+ if (code > state->free_ent) {
+ /* An invalid code is a fatal error. */
+ archive_set_error(&(self->archive->archive), -1,
+ "Invalid compressed data");
+ return (ARCHIVE_FATAL);
+ }
+
+ /* Special case for KwKwK string. */
+ if (code >= state->free_ent) {
+ *state->stackp++ = state->finbyte;
+ code = state->oldcode;
+ }
+
+ /* Generate output characters in reverse order. */
+ while (code >= 256) {
+ *state->stackp++ = state->suffix[code];
+ code = state->prefix[code];
+ }
+ *state->stackp++ = state->finbyte = code;
+
+ /* Generate the new entry. */
+ code = state->free_ent;
+ if (code < state->maxcode && state->oldcode >= 0) {
+ state->prefix[code] = state->oldcode;
+ state->suffix[code] = state->finbyte;
+ ++state->free_ent;
+ }
+ if (state->free_ent > state->section_end_code) {
+ state->bits++;
+ state->bytes_in_section = 0;
+ if (state->bits == state->maxcode_bits)
+ state->section_end_code = state->maxcode;
+ else
+ state->section_end_code = (1 << state->bits) - 1;
+ }
+
+ /* Remember previous code. */
+ state->oldcode = newcode;
+ return (ARCHIVE_OK);
+}
+
+/*
+ * Return next 'n' bits from stream.
+ *
+ * -1 indicates end of available data.
+ */
+static int
+getbits(struct archive_read_filter *self, int n)
+{
+ struct private_data *state = (struct private_data *)self->data;
+ int code;
+ ssize_t ret;
+ static const int mask[] = {
+ 0x00, 0x01, 0x03, 0x07, 0x0f, 0x1f, 0x3f, 0x7f, 0xff,
+ 0x1ff, 0x3ff, 0x7ff, 0xfff, 0x1fff, 0x3fff, 0x7fff, 0xffff
+ };
+
+ while (state->bits_avail < n) {
+ if (state->avail_in <= 0) {
+ state->next_in
+ = __archive_read_filter_ahead(self->upstream,
+ 1, &ret);
+ if (ret == 0)
+ return (-1);
+ if (ret < 0 || state->next_in == NULL)
+ return (ARCHIVE_FATAL);
+ state->avail_in = ret;
+ __archive_read_filter_consume(self->upstream, ret);
+ }
+ state->bit_buffer |= *state->next_in++ << state->bits_avail;
+ state->avail_in--;
+ state->bits_avail += 8;
+ state->bytes_in_section++;
+ }
+
+ code = state->bit_buffer;
+ state->bit_buffer >>= n;
+ state->bits_avail -= n;
+
+ return (code & mask[n]);
+}
diff --git a/libarchive/archive_read_support_compression_gzip.c b/libarchive/archive_read_support_compression_gzip.c
new file mode 100644
index 000000000000..8cc924c9fc13
--- /dev/null
+++ b/libarchive/archive_read_support_compression_gzip.c
@@ -0,0 +1,465 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_compression_gzip.c 201082 2009-12-28 02:05:28Z kientzle $");
+
+
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+#ifdef HAVE_ZLIB_H
+#include <zlib.h>
+#endif
+
+#include "archive.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+#ifdef HAVE_ZLIB_H
+struct private_data {
+ z_stream stream;
+ char in_stream;
+ unsigned char *out_block;
+ size_t out_block_size;
+ int64_t total_out;
+ unsigned long crc;
+ char eof; /* True = found end of compressed data. */
+};
+
+/* Gzip Filter. */
+static ssize_t gzip_filter_read(struct archive_read_filter *, const void **);
+static int gzip_filter_close(struct archive_read_filter *);
+#endif
+
+/*
+ * Note that we can detect gzip archives even if we can't decompress
+ * them. (In fact, we like detecting them because we can give better
+ * error messages.) So the bid framework here gets compiled even
+ * if zlib is unavailable.
+ *
+ * TODO: If zlib is unavailable, gzip_bidder_init() should
+ * use the compress_program framework to try to fire up an external
+ * gunzip program.
+ */
+static int gzip_bidder_bid(struct archive_read_filter_bidder *,
+ struct archive_read_filter *);
+static int gzip_bidder_init(struct archive_read_filter *);
+
+int
+archive_read_support_compression_gzip(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct archive_read_filter_bidder *bidder = __archive_read_get_bidder(a);
+
+ if (bidder == NULL)
+ return (ARCHIVE_FATAL);
+
+ bidder->data = NULL;
+ bidder->bid = gzip_bidder_bid;
+ bidder->init = gzip_bidder_init;
+ bidder->options = NULL;
+ bidder->free = NULL; /* No data, so no cleanup necessary. */
+ /* Signal the extent of gzip support with the return value here. */
+#if HAVE_ZLIB_H
+ return (ARCHIVE_OK);
+#else
+ archive_set_error(_a, ARCHIVE_ERRNO_MISC,
+ "Using external gunzip program");
+ return (ARCHIVE_WARN);
+#endif
+}
+
+/*
+ * Read and verify the header.
+ *
+ * Returns zero if the header couldn't be validated, else returns
+ * number of bytes in header. If pbits is non-NULL, it receives a
+ * count of bits verified, suitable for use by bidder.
+ */
+static int
+peek_at_header(struct archive_read_filter *filter, int *pbits)
+{
+ const unsigned char *p;
+ ssize_t avail, len;
+ int bits = 0;
+ int header_flags;
+
+ /* Start by looking at the first ten bytes of the header, which
+ * is all fixed layout. */
+ len = 10;
+ p = __archive_read_filter_ahead(filter, len, &avail);
+ if (p == NULL || avail == 0)
+ return (0);
+ if (p[0] != 037)
+ return (0);
+ bits += 8;
+ if (p[1] != 0213)
+ return (0);
+ bits += 8;
+ if (p[2] != 8) /* We only support deflation. */
+ return (0);
+ bits += 8;
+ if ((p[3] & 0xE0)!= 0) /* No reserved flags set. */
+ return (0);
+ bits += 3;
+ header_flags = p[3];
+ /* Bytes 4-7 are mod time. */
+ /* Byte 8 is deflate flags. */
+ /* XXXX TODO: return deflate flags back to consume_header for use
+ in initializing the decompressor. */
+ /* Byte 9 is OS. */
+
+ /* Optional extra data: 2 byte length plus variable body. */
+ if (header_flags & 4) {
+ p = __archive_read_filter_ahead(filter, len + 2, &avail);
+ if (p == NULL)
+ return (0);
+ len += ((int)p[len + 1] << 8) | (int)p[len];
+ len += 2;
+ }
+
+ /* Null-terminated optional filename. */
+ if (header_flags & 8) {
+ do {
+ ++len;
+ if (avail < len)
+ p = __archive_read_filter_ahead(filter,
+ len, &avail);
+ if (p == NULL)
+ return (0);
+ } while (p[len - 1] != 0);
+ }
+
+ /* Null-terminated optional comment. */
+ if (header_flags & 16) {
+ do {
+ ++len;
+ if (avail < len)
+ p = __archive_read_filter_ahead(filter,
+ len, &avail);
+ if (p == NULL)
+ return (0);
+ } while (p[len - 1] != 0);
+ }
+
+ /* Optional header CRC */
+ if ((header_flags & 2)) {
+ p = __archive_read_filter_ahead(filter, len + 2, &avail);
+ if (p == NULL)
+ return (0);
+#if 0
+ int hcrc = ((int)p[len + 1] << 8) | (int)p[len];
+ int crc = /* XXX TODO: Compute header CRC. */;
+ if (crc != hcrc)
+ return (0);
+ bits += 16;
+#endif
+ len += 2;
+ }
+
+ if (pbits != NULL)
+ *pbits = bits;
+ return (len);
+}
+
+/*
+ * Bidder just verifies the header and returns the number of verified bits.
+ */
+static int
+gzip_bidder_bid(struct archive_read_filter_bidder *self,
+ struct archive_read_filter *filter)
+{
+ int bits_checked;
+
+ (void)self; /* UNUSED */
+
+ if (peek_at_header(filter, &bits_checked))
+ return (bits_checked);
+ return (0);
+}
+
+
+#ifndef HAVE_ZLIB_H
+
+/*
+ * If we don't have the library on this system, we can't do the
+ * decompression directly. We can, however, try to run gunzip
+ * in case that's available.
+ */
+static int
+gzip_bidder_init(struct archive_read_filter *self)
+{
+ int r;
+
+ r = __archive_read_program(self, "gunzip");
+ /* Note: We set the format here even if __archive_read_program()
+ * above fails. We do, after all, know what the format is
+ * even if we weren't able to read it. */
+ self->code = ARCHIVE_COMPRESSION_GZIP;
+ self->name = "gzip";
+ return (r);
+}
+
+#else
+
+/*
+ * Initialize the filter object.
+ */
+static int
+gzip_bidder_init(struct archive_read_filter *self)
+{
+ struct private_data *state;
+ static const size_t out_block_size = 64 * 1024;
+ void *out_block;
+
+ self->code = ARCHIVE_COMPRESSION_GZIP;
+ self->name = "gzip";
+
+ state = (struct private_data *)calloc(sizeof(*state), 1);
+ out_block = (unsigned char *)malloc(out_block_size);
+ if (state == NULL || out_block == NULL) {
+ free(out_block);
+ free(state);
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Can't allocate data for gzip decompression");
+ return (ARCHIVE_FATAL);
+ }
+
+ self->data = state;
+ state->out_block_size = out_block_size;
+ state->out_block = out_block;
+ self->read = gzip_filter_read;
+ self->skip = NULL; /* not supported */
+ self->close = gzip_filter_close;
+
+ state->in_stream = 0; /* We're not actually within a stream yet. */
+
+ return (ARCHIVE_OK);
+}
+
+static int
+consume_header(struct archive_read_filter *self)
+{
+ struct private_data *state;
+ ssize_t avail;
+ size_t len;
+ int ret;
+
+ state = (struct private_data *)self->data;
+
+ /* If this is a real header, consume it. */
+ len = peek_at_header(self->upstream, NULL);
+ if (len == 0)
+ return (ARCHIVE_EOF);
+ __archive_read_filter_consume(self->upstream, len);
+
+ /* Initialize CRC accumulator. */
+ state->crc = crc32(0L, NULL, 0);
+
+ /* Initialize compression library. */
+ state->stream.next_in = (unsigned char *)(uintptr_t)
+ __archive_read_filter_ahead(self->upstream, 1, &avail);
+ state->stream.avail_in = avail;
+ ret = inflateInit2(&(state->stream),
+ -15 /* Don't check for zlib header */);
+
+ /* Decipher the error code. */
+ switch (ret) {
+ case Z_OK:
+ state->in_stream = 1;
+ return (ARCHIVE_OK);
+ case Z_STREAM_ERROR:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Internal error initializing compression library: "
+ "invalid setup parameter");
+ break;
+ case Z_MEM_ERROR:
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Internal error initializing compression library: "
+ "out of memory");
+ break;
+ case Z_VERSION_ERROR:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Internal error initializing compression library: "
+ "invalid library version");
+ break;
+ default:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Internal error initializing compression library: "
+ " Zlib error %d", ret);
+ break;
+ }
+ return (ARCHIVE_FATAL);
+}
+
+static int
+consume_trailer(struct archive_read_filter *self)
+{
+ struct private_data *state;
+ const unsigned char *p;
+ ssize_t avail;
+
+ state = (struct private_data *)self->data;
+
+ state->in_stream = 0;
+ switch (inflateEnd(&(state->stream))) {
+ case Z_OK:
+ break;
+ default:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Failed to clean up gzip decompressor");
+ return (ARCHIVE_FATAL);
+ }
+
+ /* GZip trailer is a fixed 8 byte structure. */
+ p = __archive_read_filter_ahead(self->upstream, 8, &avail);
+ if (p == NULL || avail == 0)
+ return (ARCHIVE_FATAL);
+
+ /* XXX TODO: Verify the length and CRC. */
+
+ /* We've verified the trailer, so consume it now. */
+ __archive_read_filter_consume(self->upstream, 8);
+
+ return (ARCHIVE_OK);
+}
+
+static ssize_t
+gzip_filter_read(struct archive_read_filter *self, const void **p)
+{
+ struct private_data *state;
+ size_t decompressed;
+ ssize_t avail_in;
+ int ret;
+
+ state = (struct private_data *)self->data;
+
+ /* Empty our output buffer. */
+ state->stream.next_out = state->out_block;
+ state->stream.avail_out = state->out_block_size;
+
+ /* Try to fill the output buffer. */
+ while (state->stream.avail_out > 0 && !state->eof) {
+ /* If we're not in a stream, read a header
+ * and initialize the decompression library. */
+ if (!state->in_stream) {
+ ret = consume_header(self);
+ if (ret == ARCHIVE_EOF) {
+ state->eof = 1;
+ break;
+ }
+ if (ret < ARCHIVE_OK)
+ return (ret);
+ }
+
+ /* Peek at the next available data. */
+ /* ZLib treats stream.next_in as const but doesn't declare
+ * it so, hence this ugly cast. */
+ state->stream.next_in = (unsigned char *)(uintptr_t)
+ __archive_read_filter_ahead(self->upstream, 1, &avail_in);
+ if (state->stream.next_in == NULL)
+ return (ARCHIVE_FATAL);
+ state->stream.avail_in = avail_in;
+
+ /* Decompress and consume some of that data. */
+ ret = inflate(&(state->stream), 0);
+ switch (ret) {
+ case Z_OK: /* Decompressor made some progress. */
+ __archive_read_filter_consume(self->upstream,
+ avail_in - state->stream.avail_in);
+ break;
+ case Z_STREAM_END: /* Found end of stream. */
+ __archive_read_filter_consume(self->upstream,
+ avail_in - state->stream.avail_in);
+ /* Consume the stream trailer; release the
+ * decompression library. */
+ ret = consume_trailer(self);
+ if (ret < ARCHIVE_OK)
+ return (ret);
+ break;
+ default:
+ /* Return an error. */
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "gzip decompression failed");
+ return (ARCHIVE_FATAL);
+ }
+ }
+
+ /* We've read as much as we can. */
+ decompressed = state->stream.next_out - state->out_block;
+ state->total_out += decompressed;
+ if (decompressed == 0)
+ *p = NULL;
+ else
+ *p = state->out_block;
+ return (decompressed);
+}
+
+/*
+ * Clean up the decompressor.
+ */
+static int
+gzip_filter_close(struct archive_read_filter *self)
+{
+ struct private_data *state;
+ int ret;
+
+ state = (struct private_data *)self->data;
+ ret = ARCHIVE_OK;
+
+ if (state->in_stream) {
+ switch (inflateEnd(&(state->stream))) {
+ case Z_OK:
+ break;
+ default:
+ archive_set_error(&(self->archive->archive),
+ ARCHIVE_ERRNO_MISC,
+ "Failed to clean up gzip compressor");
+ ret = ARCHIVE_FATAL;
+ }
+ }
+
+ free(state->out_block);
+ free(state);
+ return (ret);
+}
+
+#endif /* HAVE_ZLIB_H */
diff --git a/libarchive/archive_read_support_compression_none.c b/libarchive/archive_read_support_compression_none.c
new file mode 100644
index 000000000000..955d06d9aa18
--- /dev/null
+++ b/libarchive/archive_read_support_compression_none.c
@@ -0,0 +1,40 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_compression_none.c 185679 2008-12-06 06:45:15Z kientzle $");
+
+#include "archive.h"
+
+/*
+ * Uncompressed streams are handled implicitly by the read core,
+ * so this is now a no-op.
+ */
+int
+archive_read_support_compression_none(struct archive *a)
+{
+ (void)a; /* UNUSED */
+ return (ARCHIVE_OK);
+}
diff --git a/libarchive/archive_read_support_compression_program.c b/libarchive/archive_read_support_compression_program.c
new file mode 100644
index 000000000000..0c63f2e83129
--- /dev/null
+++ b/libarchive/archive_read_support_compression_program.c
@@ -0,0 +1,459 @@
+/*-
+ * Copyright (c) 2007 Joerg Sonnenberger
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_compression_program.c 201112 2009-12-28 06:59:35Z kientzle $");
+
+#ifdef HAVE_SYS_WAIT_H
+# include <sys/wait.h>
+#endif
+#ifdef HAVE_ERRNO_H
+# include <errno.h>
+#endif
+#ifdef HAVE_FCNTL_H
+# include <fcntl.h>
+#endif
+#ifdef HAVE_LIMITS_H
+# include <limits.h>
+#endif
+#ifdef HAVE_SIGNAL_H
+# include <signal.h>
+#endif
+#ifdef HAVE_STDLIB_H
+# include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+# include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+# include <unistd.h>
+#endif
+
+#include "archive.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+int
+archive_read_support_compression_program(struct archive *a, const char *cmd)
+{
+ return (archive_read_support_compression_program_signature(a, cmd, NULL, 0));
+}
+
+
+/* This capability is only available on POSIX systems. */
+#if (!defined(HAVE_PIPE) || !defined(HAVE_FCNTL) || \
+ !(defined(HAVE_FORK) || defined(HAVE_VFORK))) && (!defined(_WIN32) || defined(__CYGWIN__))
+
+/*
+ * On non-Posix systems, allow the program to build, but choke if
+ * this function is actually invoked.
+ */
+int
+archive_read_support_compression_program_signature(struct archive *_a,
+ const char *cmd, void *signature, size_t signature_len)
+{
+ (void)_a; /* UNUSED */
+ (void)cmd; /* UNUSED */
+ (void)signature; /* UNUSED */
+ (void)signature_len; /* UNUSED */
+
+ archive_set_error(_a, -1,
+ "External compression programs not supported on this platform");
+ return (ARCHIVE_FATAL);
+}
+
+int
+__archive_read_program(struct archive_read_filter *self, const char *cmd)
+{
+ (void)self; /* UNUSED */
+ (void)cmd; /* UNUSED */
+
+ archive_set_error(&self->archive->archive, -1,
+ "External compression programs not supported on this platform");
+ return (ARCHIVE_FATAL);
+}
+
+#else
+
+#include "filter_fork.h"
+
+/*
+ * The bidder object stores the command and the signature to watch for.
+ * The 'inhibit' entry here is used to ensure that unchecked filters never
+ * bid twice in the same pipeline.
+ */
+struct program_bidder {
+ char *cmd;
+ void *signature;
+ size_t signature_len;
+ int inhibit;
+};
+
+static int program_bidder_bid(struct archive_read_filter_bidder *,
+ struct archive_read_filter *upstream);
+static int program_bidder_init(struct archive_read_filter *);
+static int program_bidder_free(struct archive_read_filter_bidder *);
+
+/*
+ * The actual filter needs to track input and output data.
+ */
+struct program_filter {
+ char *description;
+ pid_t child;
+ int exit_status;
+ int waitpid_return;
+ int child_stdin, child_stdout;
+
+ char *out_buf;
+ size_t out_buf_len;
+};
+
+static ssize_t program_filter_read(struct archive_read_filter *,
+ const void **);
+static int program_filter_close(struct archive_read_filter *);
+
+int
+archive_read_support_compression_program_signature(struct archive *_a,
+ const char *cmd, const void *signature, size_t signature_len)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct archive_read_filter_bidder *bidder;
+ struct program_bidder *state;
+
+ /*
+ * Get a bidder object from the read core.
+ */
+ bidder = __archive_read_get_bidder(a);
+ if (bidder == NULL)
+ return (ARCHIVE_FATAL);
+
+ /*
+ * Allocate our private state.
+ */
+ state = (struct program_bidder *)calloc(sizeof (*state), 1);
+ if (state == NULL)
+ return (ARCHIVE_FATAL);
+ state->cmd = strdup(cmd);
+ if (signature != NULL && signature_len > 0) {
+ state->signature_len = signature_len;
+ state->signature = malloc(signature_len);
+ memcpy(state->signature, signature, signature_len);
+ }
+
+ /*
+ * Fill in the bidder object.
+ */
+ bidder->data = state;
+ bidder->bid = program_bidder_bid;
+ bidder->init = program_bidder_init;
+ bidder->options = NULL;
+ bidder->free = program_bidder_free;
+ return (ARCHIVE_OK);
+}
+
+static int
+program_bidder_free(struct archive_read_filter_bidder *self)
+{
+ struct program_bidder *state = (struct program_bidder *)self->data;
+ free(state->cmd);
+ free(state->signature);
+ free(self->data);
+ return (ARCHIVE_OK);
+}
+
+/*
+ * If we do have a signature, bid only if that matches.
+ *
+ * If there's no signature, we bid INT_MAX the first time
+ * we're called, then never bid again.
+ */
+static int
+program_bidder_bid(struct archive_read_filter_bidder *self,
+ struct archive_read_filter *upstream)
+{
+ struct program_bidder *state = self->data;
+ const char *p;
+
+ /* If we have a signature, use that to match. */
+ if (state->signature_len > 0) {
+ p = __archive_read_filter_ahead(upstream,
+ state->signature_len, NULL);
+ if (p == NULL)
+ return (0);
+ /* No match, so don't bid. */
+ if (memcmp(p, state->signature, state->signature_len) != 0)
+ return (0);
+ return ((int)state->signature_len * 8);
+ }
+
+ /* Otherwise, bid once and then never bid again. */
+ if (state->inhibit)
+ return (0);
+ state->inhibit = 1;
+ return (INT_MAX);
+}
+
+/*
+ * Shut down the child, return ARCHIVE_OK if it exited normally.
+ *
+ * Note that the return value is sticky; if we're called again,
+ * we won't reap the child again, but we will return the same status
+ * (including error message if the child came to a bad end).
+ */
+static int
+child_stop(struct archive_read_filter *self, struct program_filter *state)
+{
+ /* Close our side of the I/O with the child. */
+ if (state->child_stdin != -1) {
+ close(state->child_stdin);
+ state->child_stdin = -1;
+ }
+ if (state->child_stdout != -1) {
+ close(state->child_stdout);
+ state->child_stdout = -1;
+ }
+
+ if (state->child != 0) {
+ /* Reap the child. */
+ do {
+ state->waitpid_return
+ = waitpid(state->child, &state->exit_status, 0);
+ } while (state->waitpid_return == -1 && errno == EINTR);
+ state->child = 0;
+ }
+
+ if (state->waitpid_return < 0) {
+ /* waitpid() failed? This is ugly. */
+ archive_set_error(&self->archive->archive, ARCHIVE_ERRNO_MISC,
+ "Child process exited badly");
+ return (ARCHIVE_WARN);
+ }
+
+#if !defined(_WIN32) || defined(__CYGWIN__)
+ if (WIFSIGNALED(state->exit_status)) {
+#ifdef SIGPIPE
+ /* If the child died because we stopped reading before
+ * it was done, that's okay. Some archive formats
+ * have padding at the end that we routinely ignore. */
+ /* The alternative to this would be to add a step
+ * before close(child_stdout) above to read from the
+ * child until the child has no more to write. */
+ if (WTERMSIG(state->exit_status) == SIGPIPE)
+ return (ARCHIVE_OK);
+#endif
+ archive_set_error(&self->archive->archive, ARCHIVE_ERRNO_MISC,
+ "Child process exited with signal %d",
+ WTERMSIG(state->exit_status));
+ return (ARCHIVE_WARN);
+ }
+#endif /* !_WIN32 || __CYGWIN__ */
+
+ if (WIFEXITED(state->exit_status)) {
+ if (WEXITSTATUS(state->exit_status) == 0)
+ return (ARCHIVE_OK);
+
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Child process exited with status %d",
+ WEXITSTATUS(state->exit_status));
+ return (ARCHIVE_WARN);
+ }
+
+ return (ARCHIVE_WARN);
+}
+
+/*
+ * Use select() to decide whether the child is ready for read or write.
+ */
+static ssize_t
+child_read(struct archive_read_filter *self, char *buf, size_t buf_len)
+{
+ struct program_filter *state = self->data;
+ ssize_t ret, requested, avail;
+ const char *p;
+
+ requested = buf_len > SSIZE_MAX ? SSIZE_MAX : buf_len;
+
+ for (;;) {
+ do {
+ ret = read(state->child_stdout, buf, requested);
+ } while (ret == -1 && errno == EINTR);
+
+ if (ret > 0)
+ return (ret);
+ if (ret == 0 || (ret == -1 && errno == EPIPE))
+ /* Child has closed its output; reap the child
+ * and return the status. */
+ return (child_stop(self, state));
+ if (ret == -1 && errno != EAGAIN)
+ return (-1);
+
+ if (state->child_stdin == -1) {
+ /* Block until child has some I/O ready. */
+ __archive_check_child(state->child_stdin,
+ state->child_stdout);
+ continue;
+ }
+
+ /* Get some more data from upstream. */
+ p = __archive_read_filter_ahead(self->upstream, 1, &avail);
+ if (p == NULL) {
+ close(state->child_stdin);
+ state->child_stdin = -1;
+ fcntl(state->child_stdout, F_SETFL, 0);
+ if (avail < 0)
+ return (avail);
+ continue;
+ }
+
+ do {
+ ret = write(state->child_stdin, p, avail);
+ } while (ret == -1 && errno == EINTR);
+
+ if (ret > 0) {
+ /* Consume whatever we managed to write. */
+ __archive_read_filter_consume(self->upstream, ret);
+ } else if (ret == -1 && errno == EAGAIN) {
+ /* Block until child has some I/O ready. */
+ __archive_check_child(state->child_stdin,
+ state->child_stdout);
+ } else {
+ /* Write failed. */
+ close(state->child_stdin);
+ state->child_stdin = -1;
+ fcntl(state->child_stdout, F_SETFL, 0);
+ /* If it was a bad error, we're done; otherwise
+ * it was EPIPE or EOF, and we can still read
+ * from the child. */
+ if (ret == -1 && errno != EPIPE)
+ return (-1);
+ }
+ }
+}
+
+int
+__archive_read_program(struct archive_read_filter *self, const char *cmd)
+{
+ struct program_filter *state;
+ static const size_t out_buf_len = 65536;
+ char *out_buf;
+ char *description;
+ const char *prefix = "Program: ";
+
+ state = (struct program_filter *)calloc(1, sizeof(*state));
+ out_buf = (char *)malloc(out_buf_len);
+ description = (char *)malloc(strlen(prefix) + strlen(cmd) + 1);
+ if (state == NULL || out_buf == NULL || description == NULL) {
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Can't allocate input data");
+ free(state);
+ free(out_buf);
+ free(description);
+ return (ARCHIVE_FATAL);
+ }
+
+ self->code = ARCHIVE_COMPRESSION_PROGRAM;
+ state->description = description;
+ strcpy(state->description, prefix);
+ strcat(state->description, cmd);
+ self->name = state->description;
+
+ state->out_buf = out_buf;
+ state->out_buf_len = out_buf_len;
+
+ if ((state->child = __archive_create_child(cmd,
+ &state->child_stdin, &state->child_stdout)) == -1) {
+ free(state->out_buf);
+ free(state);
+ archive_set_error(&self->archive->archive, EINVAL,
+ "Can't initialise filter");
+ return (ARCHIVE_FATAL);
+ }
+
+ self->data = state;
+ self->read = program_filter_read;
+ self->skip = NULL;
+ self->close = program_filter_close;
+
+ /* XXX Check that we can read at least one byte? */
+ return (ARCHIVE_OK);
+}
+
+static int
+program_bidder_init(struct archive_read_filter *self)
+{
+ struct program_bidder *bidder_state;
+
+ bidder_state = (struct program_bidder *)self->bidder->data;
+ return (__archive_read_program(self, bidder_state->cmd));
+}
+
+static ssize_t
+program_filter_read(struct archive_read_filter *self, const void **buff)
+{
+ struct program_filter *state;
+ ssize_t bytes;
+ size_t total;
+ char *p;
+
+ state = (struct program_filter *)self->data;
+
+ total = 0;
+ p = state->out_buf;
+ while (state->child_stdout != -1 && total < state->out_buf_len) {
+ bytes = child_read(self, p, state->out_buf_len - total);
+ if (bytes < 0)
+ /* No recovery is possible if we can no longer
+ * read from the child. */
+ return (ARCHIVE_FATAL);
+ if (bytes == 0)
+ /* We got EOF from the child. */
+ break;
+ total += bytes;
+ p += bytes;
+ }
+
+ *buff = state->out_buf;
+ return (total);
+}
+
+static int
+program_filter_close(struct archive_read_filter *self)
+{
+ struct program_filter *state;
+ int e;
+
+ state = (struct program_filter *)self->data;
+ e = child_stop(self, state);
+
+ /* Release our private data. */
+ free(state->out_buf);
+ free(state->description);
+ free(state);
+
+ return (e);
+}
+
+#endif /* !defined(HAVE_PIPE) || !defined(HAVE_VFORK) || !defined(HAVE_FCNTL) */
diff --git a/libarchive/archive_read_support_compression_rpm.c b/libarchive/archive_read_support_compression_rpm.c
new file mode 100644
index 000000000000..051baa5e1229
--- /dev/null
+++ b/libarchive/archive_read_support_compression_rpm.c
@@ -0,0 +1,287 @@
+/*-
+ * Copyright (c) 2009 Michihiro NAKAJIMA
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+
+#include "archive.h"
+#include "archive_endian.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+struct rpm {
+ int64_t total_in;
+ size_t hpos;
+ size_t hlen;
+ unsigned char header[16];
+ enum {
+ ST_LEAD, /* Skipping 'Lead' section. */
+ ST_HEADER, /* Reading 'Header' section;
+ * first 16 bytes. */
+ ST_HEADER_DATA, /* Skipping 'Header' section. */
+ ST_PADDING, /* Skipping padding data after the
+ * 'Header' section. */
+ ST_ARCHIVE /* Reading 'Archive' section. */
+ } state;
+ int first_header;
+};
+#define RPM_LEAD_SIZE 96 /* Size of 'Lead' section. */
+
+static int rpm_bidder_bid(struct archive_read_filter_bidder *,
+ struct archive_read_filter *);
+static int rpm_bidder_init(struct archive_read_filter *);
+
+static ssize_t rpm_filter_read(struct archive_read_filter *,
+ const void **);
+static int rpm_filter_close(struct archive_read_filter *);
+
+int
+archive_read_support_compression_rpm(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct archive_read_filter_bidder *bidder;
+
+ bidder = __archive_read_get_bidder(a);
+ archive_clear_error(_a);
+ if (bidder == NULL)
+ return (ARCHIVE_FATAL);
+
+ bidder->data = NULL;
+ bidder->bid = rpm_bidder_bid;
+ bidder->init = rpm_bidder_init;
+ bidder->options = NULL;
+ bidder->free = NULL;
+ return (ARCHIVE_OK);
+}
+
+static int
+rpm_bidder_bid(struct archive_read_filter_bidder *self,
+ struct archive_read_filter *filter)
+{
+ const unsigned char *b;
+ ssize_t avail;
+ int bits_checked;
+
+ (void)self; /* UNUSED */
+
+ b = __archive_read_filter_ahead(filter, 8, &avail);
+ if (b == NULL)
+ return (0);
+
+ bits_checked = 0;
+ /*
+ * Verify Header Magic Bytes : 0xed 0xab 0xee 0xdb
+ */
+ if (b[0] != 0xed)
+ return (0);
+ bits_checked += 8;
+ if (b[1] != 0xab)
+ return (0);
+ bits_checked += 8;
+ if (b[2] != 0xee)
+ return (0);
+ bits_checked += 8;
+ if (b[3] != 0xdb)
+ return (0);
+ bits_checked += 8;
+ /*
+ * Check major version.
+ */
+ if (b[4] != 3 && b[4] != 4)
+ return (0);
+ bits_checked += 8;
+ /*
+ * Check package type; binary or source.
+ */
+ if (b[6] != 0)
+ return (0);
+ bits_checked += 8;
+ if (b[7] != 0 && b[7] != 1)
+ return (0);
+ bits_checked += 8;
+
+ return (bits_checked);
+}
+
+static int
+rpm_bidder_init(struct archive_read_filter *self)
+{
+ struct rpm *rpm;
+
+ self->code = ARCHIVE_COMPRESSION_RPM;
+ self->name = "rpm";
+ self->read = rpm_filter_read;
+ self->skip = NULL; /* not supported */
+ self->close = rpm_filter_close;
+
+ rpm = (struct rpm *)calloc(sizeof(*rpm), 1);
+ if (rpm == NULL) {
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Can't allocate data for rpm");
+ return (ARCHIVE_FATAL);
+ }
+
+ self->data = rpm;
+ rpm->state = ST_LEAD;
+
+ return (ARCHIVE_OK);
+}
+
+static ssize_t
+rpm_filter_read(struct archive_read_filter *self, const void **buff)
+{
+ struct rpm *rpm;
+ const unsigned char *b;
+ ssize_t avail_in, total;
+ size_t used, n;
+ uint32_t section;
+ uint32_t bytes;
+
+ rpm = (struct rpm *)self->data;
+ *buff = NULL;
+ total = avail_in = 0;
+ b = NULL;
+ used = 0;
+ do {
+ if (b == NULL) {
+ b = __archive_read_filter_ahead(self->upstream, 1,
+ &avail_in);
+ if (b == NULL) {
+ if (avail_in < 0)
+ return (ARCHIVE_FATAL);
+ else
+ break;
+ }
+ }
+
+ switch (rpm->state) {
+ case ST_LEAD:
+ if (rpm->total_in + avail_in < RPM_LEAD_SIZE)
+ used += avail_in;
+ else {
+ n = RPM_LEAD_SIZE - rpm->total_in;
+ used += n;
+ b += n;
+ rpm->state = ST_HEADER;
+ rpm->hpos = 0;
+ rpm->hlen = 0;
+ rpm->first_header = 1;
+ }
+ break;
+ case ST_HEADER:
+ n = 16 - rpm->hpos;
+ if (n > avail_in - used)
+ n = avail_in - used;
+ memcpy(rpm->header+rpm->hpos, b, n);
+ b += n;
+ used += n;
+ rpm->hpos += n;
+
+ if (rpm->hpos == 16) {
+ if (rpm->header[0] != 0x8e ||
+ rpm->header[1] != 0xad ||
+ rpm->header[2] != 0xe8 ||
+ rpm->header[3] != 0x01) {
+ if (rpm->first_header) {
+ archive_set_error(
+ &self->archive->archive,
+ ARCHIVE_ERRNO_FILE_FORMAT,
+ "Unrecoginized rpm header");
+ return (ARCHIVE_FATAL);
+ }
+ rpm->state = ST_ARCHIVE;
+ *buff = rpm->header;
+ total = rpm->hpos;
+ break;
+ }
+ /* Calculate 'Header' length. */
+ section = archive_be32dec(rpm->header+8);
+ bytes = archive_be32dec(rpm->header+12);
+ rpm->hlen = 16 + section * 16 + bytes;
+ rpm->state = ST_HEADER_DATA;
+ rpm->first_header = 0;
+ }
+ break;
+ case ST_HEADER_DATA:
+ n = rpm->hlen - rpm->hpos;
+ if (n > avail_in - used)
+ n = avail_in - used;
+ b += n;
+ used += n;
+ rpm->hpos += n;
+ if (rpm->hpos == rpm->hlen)
+ rpm->state = ST_PADDING;
+ break;
+ case ST_PADDING:
+ while (used < (size_t)avail_in) {
+ if (*b != 0) {
+ /* Read next header. */
+ rpm->state = ST_HEADER;
+ rpm->hpos = 0;
+ rpm->hlen = 0;
+ break;
+ }
+ b++;
+ used++;
+ }
+ break;
+ case ST_ARCHIVE:
+ *buff = b;
+ total = avail_in;
+ used = avail_in;
+ break;
+ }
+ if (used == (size_t)avail_in) {
+ rpm->total_in += used;
+ __archive_read_filter_consume(self->upstream, used);
+ b = NULL;
+ used = 0;
+ }
+ } while (total == 0 && avail_in > 0);
+
+ if (used > 0 && b != NULL) {
+ rpm->total_in += used;
+ __archive_read_filter_consume(self->upstream, used);
+ }
+ return (total);
+}
+
+static int
+rpm_filter_close(struct archive_read_filter *self)
+{
+ struct rpm *rpm;
+
+ rpm = (struct rpm *)self->data;
+ free(rpm);
+
+ return (ARCHIVE_OK);
+}
+
diff --git a/libarchive/archive_read_support_compression_uu.c b/libarchive/archive_read_support_compression_uu.c
new file mode 100644
index 000000000000..f7bea828fccd
--- /dev/null
+++ b/libarchive/archive_read_support_compression_uu.c
@@ -0,0 +1,637 @@
+/*-
+ * Copyright (c) 2009 Michihiro NAKAJIMA
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_compression_uu.c 201248 2009-12-30 06:12:03Z kientzle $");
+
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+
+#include "archive.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+struct uudecode {
+ int64_t total;
+ unsigned char *in_buff;
+#define IN_BUFF_SIZE (1024)
+ int in_cnt;
+ size_t in_allocated;
+ unsigned char *out_buff;
+#define OUT_BUFF_SIZE (64 * 1024)
+ int state;
+#define ST_FIND_HEAD 0
+#define ST_READ_UU 1
+#define ST_UUEND 2
+#define ST_READ_BASE64 3
+};
+
+static int uudecode_bidder_bid(struct archive_read_filter_bidder *,
+ struct archive_read_filter *filter);
+static int uudecode_bidder_init(struct archive_read_filter *);
+
+static ssize_t uudecode_filter_read(struct archive_read_filter *,
+ const void **);
+static int uudecode_filter_close(struct archive_read_filter *);
+
+int
+archive_read_support_compression_uu(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct archive_read_filter_bidder *bidder;
+
+ bidder = __archive_read_get_bidder(a);
+ archive_clear_error(_a);
+ if (bidder == NULL)
+ return (ARCHIVE_FATAL);
+
+ bidder->data = NULL;
+ bidder->bid = uudecode_bidder_bid;
+ bidder->init = uudecode_bidder_init;
+ bidder->options = NULL;
+ bidder->free = NULL;
+ return (ARCHIVE_OK);
+}
+
+static const unsigned char ascii[256] = {
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, '\n', 0, 0, '\r', 0, 0, /* 00 - 0F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 10 - 1F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 20 - 2F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 30 - 3F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 40 - 4F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 50 - 5F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 60 - 6F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, /* 70 - 7F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 80 - 8F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 90 - 9F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* A0 - AF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* B0 - BF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* C0 - CF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* D0 - DF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* E0 - EF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* F0 - FF */
+};
+
+static const unsigned char uuchar[256] = {
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 00 - 0F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 10 - 1F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 20 - 2F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 30 - 3F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 40 - 4F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 50 - 5F */
+ 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 60 - 6F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 70 - 7F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 80 - 8F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 90 - 9F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* A0 - AF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* B0 - BF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* C0 - CF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* D0 - DF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* E0 - EF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* F0 - FF */
+};
+
+static const unsigned char base64[256] = {
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 00 - 0F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 10 - 1F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, /* 20 - 2F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0, /* 30 - 3F */
+ 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 40 - 4F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, /* 50 - 5F */
+ 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, /* 60 - 6F */
+ 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, /* 70 - 7F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 80 - 8F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* 90 - 9F */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* A0 - AF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* B0 - BF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* C0 - CF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* D0 - DF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* E0 - EF */
+ 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, /* F0 - FF */
+};
+
+static const int base64num[128] = {
+ 0, 0, 0, 0, 0, 0, 0, 0,
+ 0, 0, 0, 0, 0, 0, 0, 0, /* 00 - 0F */
+ 0, 0, 0, 0, 0, 0, 0, 0,
+ 0, 0, 0, 0, 0, 0, 0, 0, /* 10 - 1F */
+ 0, 0, 0, 0, 0, 0, 0, 0,
+ 0, 0, 0, 62, 0, 0, 0, 63, /* 20 - 2F */
+ 52, 53, 54, 55, 56, 57, 58, 59,
+ 60, 61, 0, 0, 0, 0, 0, 0, /* 30 - 3F */
+ 0, 0, 1, 2, 3, 4, 5, 6,
+ 7, 8, 9, 10, 11, 12, 13, 14, /* 40 - 4F */
+ 15, 16, 17, 18, 19, 20, 21, 22,
+ 23, 24, 25, 0, 0, 0, 0, 0, /* 50 - 5F */
+ 0, 26, 27, 28, 29, 30, 31, 32,
+ 33, 34, 35, 36, 37, 38, 39, 40, /* 60 - 6F */
+ 41, 42, 43, 44, 45, 46, 47, 48,
+ 49, 50, 51, 0, 0, 0, 0, 0, /* 70 - 7F */
+};
+
+static ssize_t
+get_line(const unsigned char *b, ssize_t avail, ssize_t *nlsize)
+{
+ ssize_t len;
+
+ len = 0;
+ while (len < avail) {
+ switch (ascii[*b]) {
+ case 0: /* Non-ascii character or control character. */
+ if (nlsize != NULL)
+ *nlsize = 0;
+ return (-1);
+ case '\r':
+ if (avail-len > 1 && b[1] == '\n') {
+ if (nlsize != NULL)
+ *nlsize = 2;
+ return (len+2);
+ }
+ /* FALL THROUGH */
+ case '\n':
+ if (nlsize != NULL)
+ *nlsize = 1;
+ return (len+1);
+ case 1:
+ b++;
+ len++;
+ break;
+ }
+ }
+ if (nlsize != NULL)
+ *nlsize = 0;
+ return (avail);
+}
+
+static ssize_t
+bid_get_line(struct archive_read_filter *filter,
+ const unsigned char **b, ssize_t *avail, ssize_t *ravail, ssize_t *nl)
+{
+ ssize_t len;
+ int quit;
+
+ quit = 0;
+ if (*avail == 0) {
+ *nl = 0;
+ len = 0;
+ } else
+ len = get_line(*b, *avail, nl);
+ /*
+ * Read bytes more while it does not reach the end of line.
+ */
+ while (*nl == 0 && len == *avail && !quit) {
+ ssize_t diff = *ravail - *avail;
+
+ *b = __archive_read_filter_ahead(filter, 160 + *ravail, avail);
+ if (*b == NULL) {
+ if (*ravail >= *avail)
+ return (0);
+ /* Reading bytes reaches the end of file. */
+ *b = __archive_read_filter_ahead(filter, *avail, avail);
+ quit = 1;
+ }
+ *ravail = *avail;
+ *b += diff;
+ *avail -= diff;
+ len = get_line(*b, *avail, nl);
+ }
+ return (len);
+}
+
+#define UUDECODE(c) (((c) - 0x20) & 0x3f)
+
+static int
+uudecode_bidder_bid(struct archive_read_filter_bidder *self,
+ struct archive_read_filter *filter)
+{
+ const unsigned char *b;
+ ssize_t avail, ravail;
+ ssize_t len, nl;
+ int l;
+ int firstline;
+
+ (void)self; /* UNUSED */
+
+ b = __archive_read_filter_ahead(filter, 1, &avail);
+ if (b == NULL)
+ return (0);
+
+ firstline = 20;
+ ravail = avail;
+ for (;;) {
+ len = bid_get_line(filter, &b, &avail, &ravail, &nl);
+ if (len < 0 || nl == 0)
+ return (0);/* Binary data. */
+ if (memcmp(b, "begin ", 6) == 0 && len - nl >= 11)
+ l = 6;
+ else if (memcmp(b, "begin-base64 ", 13) == 0 && len - nl >= 18)
+ l = 13;
+ else
+ l = 0;
+
+ if (l > 0 && (b[l] < '0' || b[l] > '7' ||
+ b[l+1] < '0' || b[l+1] > '7' ||
+ b[l+2] < '0' || b[l+2] > '7' || b[l+3] != ' '))
+ l = 0;
+
+ b += len;
+ avail -= len;
+ if (l)
+ break;
+ firstline = 0;
+ }
+ if (!avail)
+ return (0);
+ len = bid_get_line(filter, &b, &avail, &ravail, &nl);
+ if (len < 0 || nl == 0)
+ return (0);/* There are non-ascii characters. */
+ avail -= len;
+
+ if (l == 6) {
+ if (!uuchar[*b])
+ return (0);
+ /* Get a length of decoded bytes. */
+ l = UUDECODE(*b++); len--;
+ if (l > 45)
+ /* Normally, maximum length is 45(character 'M'). */
+ return (0);
+ while (l && len-nl > 0) {
+ if (l > 0) {
+ if (!uuchar[*b++])
+ return (0);
+ if (!uuchar[*b++])
+ return (0);
+ len -= 2;
+ --l;
+ }
+ if (l > 0) {
+ if (!uuchar[*b++])
+ return (0);
+ --len;
+ --l;
+ }
+ if (l > 0) {
+ if (!uuchar[*b++])
+ return (0);
+ --len;
+ --l;
+ }
+ }
+ if (len-nl < 0)
+ return (0);
+ if (len-nl == 1 &&
+ (uuchar[*b] || /* Check sum. */
+ (*b >= 'a' && *b <= 'z'))) {/* Padding data(MINIX). */
+ ++b;
+ --len;
+ }
+ b += nl;
+ if (avail && uuchar[*b])
+ return (firstline+30);
+ }
+ if (l == 13) {
+ while (len-nl > 0) {
+ if (!base64[*b++])
+ return (0);
+ --len;
+ }
+ b += nl;
+
+ if (avail >= 5 && memcmp(b, "====\n", 5) == 0)
+ return (firstline+40);
+ if (avail >= 6 && memcmp(b, "====\r\n", 6) == 0)
+ return (firstline+40);
+ if (avail > 0 && base64[*b])
+ return (firstline+30);
+ }
+
+ return (0);
+}
+
+static int
+uudecode_bidder_init(struct archive_read_filter *self)
+{
+ struct uudecode *uudecode;
+ void *out_buff;
+ void *in_buff;
+
+ self->code = ARCHIVE_COMPRESSION_UU;
+ self->name = "uu";
+ self->read = uudecode_filter_read;
+ self->skip = NULL; /* not supported */
+ self->close = uudecode_filter_close;
+
+ uudecode = (struct uudecode *)calloc(sizeof(*uudecode), 1);
+ out_buff = malloc(OUT_BUFF_SIZE);
+ in_buff = malloc(IN_BUFF_SIZE);
+ if (uudecode == NULL || out_buff == NULL || in_buff == NULL) {
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Can't allocate data for uudecode");
+ free(uudecode);
+ free(out_buff);
+ free(in_buff);
+ return (ARCHIVE_FATAL);
+ }
+
+ self->data = uudecode;
+ uudecode->in_buff = in_buff;
+ uudecode->in_cnt = 0;
+ uudecode->in_allocated = IN_BUFF_SIZE;
+ uudecode->out_buff = out_buff;
+ uudecode->state = ST_FIND_HEAD;
+
+ return (ARCHIVE_OK);
+}
+
+static int
+ensure_in_buff_size(struct archive_read_filter *self,
+ struct uudecode *uudecode, size_t size)
+{
+
+ if (size > uudecode->in_allocated) {
+ unsigned char *ptr;
+ size_t newsize;
+
+ /*
+ * Calculate a new buffer size for in_buff.
+ * Increase its value until it has enough size we need.
+ */
+ newsize = uudecode->in_allocated;
+ do {
+ if (newsize < IN_BUFF_SIZE*32)
+ newsize <<= 1;
+ else
+ newsize += IN_BUFF_SIZE;
+ } while (size > newsize);
+ ptr = malloc(newsize);
+ if (ptr == NULL ||
+ newsize < uudecode->in_allocated) {
+ free(ptr);
+ archive_set_error(&self->archive->archive,
+ ENOMEM,
+ "Can't allocate data for uudecode");
+ return (ARCHIVE_FATAL);
+ }
+ if (uudecode->in_cnt)
+ memmove(ptr, uudecode->in_buff,
+ uudecode->in_cnt);
+ free(uudecode->in_buff);
+ uudecode->in_buff = ptr;
+ uudecode->in_allocated = newsize;
+ }
+ return (ARCHIVE_OK);
+}
+
+static ssize_t
+uudecode_filter_read(struct archive_read_filter *self, const void **buff)
+{
+ struct uudecode *uudecode;
+ const unsigned char *b, *d;
+ unsigned char *out;
+ ssize_t avail_in, ravail;
+ ssize_t used;
+ ssize_t total;
+ ssize_t len, llen, nl;
+
+ uudecode = (struct uudecode *)self->data;
+
+read_more:
+ d = __archive_read_filter_ahead(self->upstream, 1, &avail_in);
+ if (d == NULL && avail_in < 0)
+ return (ARCHIVE_FATAL);
+ /* Quiet a code analyzer; make sure avail_in must be zero
+ * when d is NULL. */
+ if (d == NULL)
+ avail_in = 0;
+ used = 0;
+ total = 0;
+ out = uudecode->out_buff;
+ ravail = avail_in;
+ if (uudecode->in_cnt) {
+ /*
+ * If there is remaining data which is saved by
+ * previous calling, use it first.
+ */
+ if (ensure_in_buff_size(self, uudecode,
+ avail_in + uudecode->in_cnt) != ARCHIVE_OK)
+ return (ARCHIVE_FATAL);
+ memcpy(uudecode->in_buff + uudecode->in_cnt,
+ d, avail_in);
+ d = uudecode->in_buff;
+ avail_in += uudecode->in_cnt;
+ uudecode->in_cnt = 0;
+ }
+ for (;used < avail_in; d += llen, used += llen) {
+ int l, body;
+
+ b = d;
+ len = get_line(b, avail_in - used, &nl);
+ if (len < 0) {
+ /* Non-ascii character is found. */
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Insufficient compressed data");
+ return (ARCHIVE_FATAL);
+ }
+ llen = len;
+ if (nl == 0) {
+ /*
+ * Save remaining data which does not contain
+ * NL('\n','\r').
+ */
+ if (ensure_in_buff_size(self, uudecode, len)
+ != ARCHIVE_OK)
+ return (ARCHIVE_FATAL);
+ if (uudecode->in_buff != b)
+ memmove(uudecode->in_buff, b, len);
+ uudecode->in_cnt = len;
+ if (total == 0) {
+ /* Do not return 0; it means end-of-file.
+ * We should try to read bytes more. */
+ __archive_read_filter_consume(
+ self->upstream, ravail);
+ goto read_more;
+ }
+ break;
+ }
+ if (total + len * 2 > OUT_BUFF_SIZE)
+ break;
+ switch (uudecode->state) {
+ default:
+ case ST_FIND_HEAD:
+ if (len - nl >= 11 && memcmp(b, "begin ", 6) == 0)
+ l = 6;
+ else if (len - nl >= 18 &&
+ memcmp(b, "begin-base64 ", 13) == 0)
+ l = 13;
+ else
+ l = 0;
+ if (l != 0 && b[l] >= '0' && b[l] <= '7' &&
+ b[l+1] >= '0' && b[l+1] <= '7' &&
+ b[l+2] >= '0' && b[l+2] <= '7' && b[l+3] == ' ') {
+ if (l == 6)
+ uudecode->state = ST_READ_UU;
+ else
+ uudecode->state = ST_READ_BASE64;
+ }
+ break;
+ case ST_READ_UU:
+ body = len - nl;
+ if (!uuchar[*b] || body <= 0) {
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Insufficient compressed data");
+ return (ARCHIVE_FATAL);
+ }
+ /* Get length of undecoded bytes of curent line. */
+ l = UUDECODE(*b++);
+ body--;
+ if (l > body) {
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Insufficient compressed data");
+ return (ARCHIVE_FATAL);
+ }
+ if (l == 0) {
+ uudecode->state = ST_UUEND;
+ break;
+ }
+ while (l > 0) {
+ int n = 0;
+
+ if (l > 0) {
+ if (!uuchar[b[0]] || !uuchar[b[1]])
+ break;
+ n = UUDECODE(*b++) << 18;
+ n |= UUDECODE(*b++) << 12;
+ *out++ = n >> 16; total++;
+ --l;
+ }
+ if (l > 0) {
+ if (!uuchar[b[0]])
+ break;
+ n |= UUDECODE(*b++) << 6;
+ *out++ = (n >> 8) & 0xFF; total++;
+ --l;
+ }
+ if (l > 0) {
+ if (!uuchar[b[0]])
+ break;
+ n |= UUDECODE(*b++);
+ *out++ = n & 0xFF; total++;
+ --l;
+ }
+ }
+ if (l) {
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Insufficient compressed data");
+ return (ARCHIVE_FATAL);
+ }
+ break;
+ case ST_UUEND:
+ if (len - nl == 3 && memcmp(b, "end ", 3) == 0)
+ uudecode->state = ST_FIND_HEAD;
+ else {
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Insufficient compressed data");
+ return (ARCHIVE_FATAL);
+ }
+ break;
+ case ST_READ_BASE64:
+ l = len - nl;
+ if (l >= 3 && b[0] == '=' && b[1] == '=' &&
+ b[2] == '=') {
+ uudecode->state = ST_FIND_HEAD;
+ break;
+ }
+ while (l > 0) {
+ int n = 0;
+
+ if (l > 0) {
+ if (!base64[b[0]] || !base64[b[1]])
+ break;
+ n = base64num[*b++] << 18;
+ n |= base64num[*b++] << 12;
+ *out++ = n >> 16; total++;
+ l -= 2;
+ }
+ if (l > 0) {
+ if (*b == '=')
+ break;
+ if (!base64[*b])
+ break;
+ n |= base64num[*b++] << 6;
+ *out++ = (n >> 8) & 0xFF; total++;
+ --l;
+ }
+ if (l > 0) {
+ if (*b == '=')
+ break;
+ if (!base64[*b])
+ break;
+ n |= base64num[*b++];
+ *out++ = n & 0xFF; total++;
+ --l;
+ }
+ }
+ if (l && *b != '=') {
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Insufficient compressed data");
+ return (ARCHIVE_FATAL);
+ }
+ break;
+ }
+ }
+
+ __archive_read_filter_consume(self->upstream, ravail);
+
+ *buff = uudecode->out_buff;
+ uudecode->total += total;
+ return (total);
+}
+
+static int
+uudecode_filter_close(struct archive_read_filter *self)
+{
+ struct uudecode *uudecode;
+
+ uudecode = (struct uudecode *)self->data;
+ free(uudecode->in_buff);
+ free(uudecode->out_buff);
+ free(uudecode);
+
+ return (ARCHIVE_OK);
+}
+
diff --git a/libarchive/archive_read_support_compression_xz.c b/libarchive/archive_read_support_compression_xz.c
new file mode 100644
index 000000000000..28c4e2d9fc51
--- /dev/null
+++ b/libarchive/archive_read_support_compression_xz.c
@@ -0,0 +1,708 @@
+/*-
+ * Copyright (c) 2009 Michihiro NAKAJIMA
+ * Copyright (c) 2003-2008 Tim Kientzle and Miklos Vajna
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_compression_xz.c 201167 2009-12-29 06:06:20Z kientzle $");
+
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#include <stdio.h>
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif
+#if HAVE_LZMA_H
+#include <lzma.h>
+#elif HAVE_LZMADEC_H
+#include <lzmadec.h>
+#endif
+
+#include "archive.h"
+#include "archive_endian.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+#if HAVE_LZMA_H && HAVE_LIBLZMA
+
+struct private_data {
+ lzma_stream stream;
+ unsigned char *out_block;
+ size_t out_block_size;
+ int64_t total_out;
+ char eof; /* True = found end of compressed data. */
+};
+
+/* Combined lzma/xz filter */
+static ssize_t xz_filter_read(struct archive_read_filter *, const void **);
+static int xz_filter_close(struct archive_read_filter *);
+static int xz_lzma_bidder_init(struct archive_read_filter *);
+
+#elif HAVE_LZMADEC_H && HAVE_LIBLZMADEC
+
+struct private_data {
+ lzmadec_stream stream;
+ unsigned char *out_block;
+ size_t out_block_size;
+ int64_t total_out;
+ char eof; /* True = found end of compressed data. */
+};
+
+/* Lzma-only filter */
+static ssize_t lzma_filter_read(struct archive_read_filter *, const void **);
+static int lzma_filter_close(struct archive_read_filter *);
+#endif
+
+/*
+ * Note that we can detect xz and lzma compressed files even if we
+ * can't decompress them. (In fact, we like detecting them because we
+ * can give better error messages.) So the bid framework here gets
+ * compiled even if no lzma library is available.
+ */
+static int xz_bidder_bid(struct archive_read_filter_bidder *,
+ struct archive_read_filter *);
+static int xz_bidder_init(struct archive_read_filter *);
+static int lzma_bidder_bid(struct archive_read_filter_bidder *,
+ struct archive_read_filter *);
+static int lzma_bidder_init(struct archive_read_filter *);
+
+int
+archive_read_support_compression_xz(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct archive_read_filter_bidder *bidder = __archive_read_get_bidder(a);
+
+ archive_clear_error(_a);
+ if (bidder == NULL)
+ return (ARCHIVE_FATAL);
+
+ bidder->data = NULL;
+ bidder->bid = xz_bidder_bid;
+ bidder->init = xz_bidder_init;
+ bidder->options = NULL;
+ bidder->free = NULL;
+#if HAVE_LZMA_H && HAVE_LIBLZMA
+ return (ARCHIVE_OK);
+#else
+ archive_set_error(_a, ARCHIVE_ERRNO_MISC,
+ "Using external unxz program for xz decompression");
+ return (ARCHIVE_WARN);
+#endif
+}
+
+int
+archive_read_support_compression_lzma(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct archive_read_filter_bidder *bidder = __archive_read_get_bidder(a);
+
+ archive_clear_error(_a);
+ if (bidder == NULL)
+ return (ARCHIVE_FATAL);
+
+ bidder->data = NULL;
+ bidder->bid = lzma_bidder_bid;
+ bidder->init = lzma_bidder_init;
+ bidder->options = NULL;
+ bidder->free = NULL;
+#if HAVE_LZMA_H && HAVE_LIBLZMA
+ return (ARCHIVE_OK);
+#elif HAVE_LZMADEC_H && HAVE_LIBLZMADEC
+ return (ARCHIVE_OK);
+#else
+ archive_set_error(_a, ARCHIVE_ERRNO_MISC,
+ "Using external unlzma program for lzma decompression");
+ return (ARCHIVE_WARN);
+#endif
+}
+
+/*
+ * Test whether we can handle this data.
+ */
+static int
+xz_bidder_bid(struct archive_read_filter_bidder *self,
+ struct archive_read_filter *filter)
+{
+ const unsigned char *buffer;
+ ssize_t avail;
+ int bits_checked;
+
+ (void)self; /* UNUSED */
+
+ buffer = __archive_read_filter_ahead(filter, 6, &avail);
+ if (buffer == NULL)
+ return (0);
+
+ /*
+ * Verify Header Magic Bytes : FD 37 7A 58 5A 00
+ */
+ bits_checked = 0;
+ if (buffer[0] != 0xFD)
+ return (0);
+ bits_checked += 8;
+ if (buffer[1] != 0x37)
+ return (0);
+ bits_checked += 8;
+ if (buffer[2] != 0x7A)
+ return (0);
+ bits_checked += 8;
+ if (buffer[3] != 0x58)
+ return (0);
+ bits_checked += 8;
+ if (buffer[4] != 0x5A)
+ return (0);
+ bits_checked += 8;
+ if (buffer[5] != 0x00)
+ return (0);
+ bits_checked += 8;
+
+ return (bits_checked);
+}
+
+/*
+ * Test whether we can handle this data.
+ *
+ * <sigh> LZMA has a rather poor file signature. Zeros do not
+ * make good signature bytes as a rule, and the only non-zero byte
+ * here is an ASCII character. For example, an uncompressed tar
+ * archive whose first file is ']' would satisfy this check. It may
+ * be necessary to exclude LZMA from compression_all() because of
+ * this. Clients of libarchive would then have to explicitly enable
+ * LZMA checking instead of (or in addition to) compression_all() when
+ * they have other evidence (file name, command-line option) to go on.
+ */
+static int
+lzma_bidder_bid(struct archive_read_filter_bidder *self,
+ struct archive_read_filter *filter)
+{
+ const unsigned char *buffer;
+ ssize_t avail;
+ uint32_t dicsize;
+ uint64_t uncompressed_size;
+ int bits_checked;
+
+ (void)self; /* UNUSED */
+
+ buffer = __archive_read_filter_ahead(filter, 14, &avail);
+ if (buffer == NULL)
+ return (0);
+
+ /* First byte of raw LZMA stream is commonly 0x5d.
+ * The first byte is a special number, which consists of
+ * three parameters of LZMA compression, a number of literal
+ * context bits(which is from 0 to 8, default is 3), a number
+ * of literal pos bits(which is from 0 to 4, default is 0),
+ * a number of pos bits(which is from 0 to 4, default is 2).
+ * The first byte is made by
+ * (pos bits * 5 + literal pos bit) * 9 + * literal contest bit,
+ * and so the default value in this field is
+ * (2 * 5 + 0) * 9 + 3 = 0x5d.
+ * lzma of LZMA SDK has options to change those parameters.
+ * It means a range of this field is from 0 to 224. And lzma of
+ * XZ Utils with option -e records 0x5e in this field. */
+ /* NOTE: If this checking of the first byte increases false
+ * recognition, we should allow only 0x5d and 0x5e for the first
+ * byte of LZMA stream. */
+ bits_checked = 0;
+ if (buffer[0] > (4 * 5 + 4) * 9 + 8)
+ return (0);
+ /* Most likely value in the first byte of LZMA stream. */
+ if (buffer[0] == 0x5d || buffer[0] == 0x5e)
+ bits_checked += 8;
+
+ /* Sixth through fourteenth bytes are uncompressed size,
+ * stored in little-endian order. `-1' means uncompressed
+ * size is unknown and lzma of XZ Utils always records `-1'
+ * in this field. */
+ uncompressed_size = archive_le64dec(buffer+5);
+ if (uncompressed_size == (uint64_t)ARCHIVE_LITERAL_LL(-1))
+ bits_checked += 64;
+
+ /* Second through fifth bytes are dictionary size, stored in
+ * little-endian order. The minimum dictionary size is
+ * 1 << 12(4KiB) which the lzma of LZMA SDK uses with option
+ * -d12 and the maxinam dictionary size is 1 << 27(128MiB)
+ * which the one uses with option -d27.
+ * NOTE: A comment of LZMA SDK source code says this dictionary
+ * range is from 1 << 12 to 1 << 30. */
+ dicsize = archive_le32dec(buffer+1);
+ switch (dicsize) {
+ case 0x00001000:/* lzma of LZMA SDK option -d12. */
+ case 0x00002000:/* lzma of LZMA SDK option -d13. */
+ case 0x00004000:/* lzma of LZMA SDK option -d14. */
+ case 0x00008000:/* lzma of LZMA SDK option -d15. */
+ case 0x00010000:/* lzma of XZ Utils option -0 and -1.
+ * lzma of LZMA SDK option -d16. */
+ case 0x00020000:/* lzma of LZMA SDK option -d17. */
+ case 0x00040000:/* lzma of LZMA SDK option -d18. */
+ case 0x00080000:/* lzma of XZ Utils option -2.
+ * lzma of LZMA SDK option -d19. */
+ case 0x00100000:/* lzma of XZ Utils option -3.
+ * lzma of LZMA SDK option -d20. */
+ case 0x00200000:/* lzma of XZ Utils option -4.
+ * lzma of LZMA SDK option -d21. */
+ case 0x00400000:/* lzma of XZ Utils option -5.
+ * lzma of LZMA SDK option -d22. */
+ case 0x00800000:/* lzma of XZ Utils option -6.
+ * lzma of LZMA SDK option -d23. */
+ case 0x01000000:/* lzma of XZ Utils option -7.
+ * lzma of LZMA SDK option -d24. */
+ case 0x02000000:/* lzma of XZ Utils option -8.
+ * lzma of LZMA SDK option -d25. */
+ case 0x04000000:/* lzma of XZ Utils option -9.
+ * lzma of LZMA SDK option -d26. */
+ case 0x08000000:/* lzma of LZMA SDK option -d27. */
+ bits_checked += 32;
+ break;
+ default:
+ /* If a memory usage for encoding was not enough on
+ * the platform where LZMA stream was made, lzma of
+ * XZ Utils automatically decreased the dictionary
+ * size to enough memory for encoding by 1Mi bytes
+ * (1 << 20).*/
+ if (dicsize <= 0x03F00000 && dicsize >= 0x00300000 &&
+ (dicsize & ((1 << 20)-1)) == 0 &&
+ bits_checked == 8 + 64) {
+ bits_checked += 32;
+ break;
+ }
+ /* Otherwise dictionary size is unlikely. But it is
+ * possible that someone makes lzma stream with
+ * liblzma/LZMA SDK in one's dictionary size. */
+ return (0);
+ }
+
+ /* TODO: The above test is still very weak. It would be
+ * good to do better. */
+
+ return (bits_checked);
+}
+
+#if HAVE_LZMA_H && HAVE_LIBLZMA
+
+/*
+ * liblzma 4.999.7 and later support both lzma and xz streams.
+ */
+static int
+xz_bidder_init(struct archive_read_filter *self)
+{
+ self->code = ARCHIVE_COMPRESSION_XZ;
+ self->name = "xz";
+ return (xz_lzma_bidder_init(self));
+}
+
+static int
+lzma_bidder_init(struct archive_read_filter *self)
+{
+ self->code = ARCHIVE_COMPRESSION_LZMA;
+ self->name = "lzma";
+ return (xz_lzma_bidder_init(self));
+}
+
+/*
+ * Setup the callbacks.
+ */
+static int
+xz_lzma_bidder_init(struct archive_read_filter *self)
+{
+ static const size_t out_block_size = 64 * 1024;
+ void *out_block;
+ struct private_data *state;
+ int ret;
+
+ state = (struct private_data *)calloc(sizeof(*state), 1);
+ out_block = (unsigned char *)malloc(out_block_size);
+ if (state == NULL || out_block == NULL) {
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Can't allocate data for xz decompression");
+ free(out_block);
+ free(state);
+ return (ARCHIVE_FATAL);
+ }
+
+ self->data = state;
+ state->out_block_size = out_block_size;
+ state->out_block = out_block;
+ self->read = xz_filter_read;
+ self->skip = NULL; /* not supported */
+ self->close = xz_filter_close;
+
+ state->stream.avail_in = 0;
+
+ state->stream.next_out = state->out_block;
+ state->stream.avail_out = state->out_block_size;
+
+ /* Initialize compression library.
+ * TODO: I don't know what value is best for memlimit.
+ * maybe, it needs to check memory size which
+ * running system has.
+ */
+ if (self->code == ARCHIVE_COMPRESSION_XZ)
+ ret = lzma_stream_decoder(&(state->stream),
+ (1U << 30),/* memlimit */
+ LZMA_CONCATENATED);
+ else
+ ret = lzma_alone_decoder(&(state->stream),
+ (1U << 30));/* memlimit */
+
+ if (ret == LZMA_OK)
+ return (ARCHIVE_OK);
+
+ /* Library setup failed: Choose an error message and clean up. */
+ switch (ret) {
+ case LZMA_MEM_ERROR:
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Internal error initializing compression library: "
+ "Cannot allocate memory");
+ break;
+ case LZMA_OPTIONS_ERROR:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Internal error initializing compression library: "
+ "Invalid or unsupported options");
+ break;
+ default:
+ archive_set_error(&self->archive->archive, ARCHIVE_ERRNO_MISC,
+ "Internal error initializing lzma library");
+ break;
+ }
+
+ free(state->out_block);
+ free(state);
+ self->data = NULL;
+ return (ARCHIVE_FATAL);
+}
+
+/*
+ * Return the next block of decompressed data.
+ */
+static ssize_t
+xz_filter_read(struct archive_read_filter *self, const void **p)
+{
+ struct private_data *state;
+ size_t decompressed;
+ ssize_t avail_in;
+ int ret;
+
+ state = (struct private_data *)self->data;
+
+ /* Empty our output buffer. */
+ state->stream.next_out = state->out_block;
+ state->stream.avail_out = state->out_block_size;
+
+ /* Try to fill the output buffer. */
+ while (state->stream.avail_out > 0 && !state->eof) {
+ state->stream.next_in =
+ __archive_read_filter_ahead(self->upstream, 1, &avail_in);
+ if (state->stream.next_in == NULL && avail_in < 0)
+ return (ARCHIVE_FATAL);
+ state->stream.avail_in = avail_in;
+
+ /* Decompress as much as we can in one pass. */
+ ret = lzma_code(&(state->stream),
+ (state->stream.avail_in == 0)? LZMA_FINISH: LZMA_RUN);
+ switch (ret) {
+ case LZMA_STREAM_END: /* Found end of stream. */
+ state->eof = 1;
+ /* FALL THROUGH */
+ case LZMA_OK: /* Decompressor made some progress. */
+ __archive_read_filter_consume(self->upstream,
+ avail_in - state->stream.avail_in);
+ break;
+ case LZMA_MEM_ERROR:
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Lzma library error: Cannot allocate memory");
+ return (ARCHIVE_FATAL);
+ case LZMA_MEMLIMIT_ERROR:
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Lzma library error: Out of memory");
+ return (ARCHIVE_FATAL);
+ case LZMA_FORMAT_ERROR:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Lzma library error: format not recognized");
+ return (ARCHIVE_FATAL);
+ case LZMA_OPTIONS_ERROR:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Lzma library error: Invalid options");
+ return (ARCHIVE_FATAL);
+ case LZMA_DATA_ERROR:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Lzma library error: Corrupted input data");
+ return (ARCHIVE_FATAL);
+ case LZMA_BUF_ERROR:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Lzma library error: No progress is possible");
+ return (ARCHIVE_FATAL);
+ default:
+ /* Return an error. */
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Lzma decompression failed: Unknown error");
+ return (ARCHIVE_FATAL);
+ }
+ }
+
+ decompressed = state->stream.next_out - state->out_block;
+ state->total_out += decompressed;
+ if (decompressed == 0)
+ *p = NULL;
+ else
+ *p = state->out_block;
+ return (decompressed);
+}
+
+/*
+ * Clean up the decompressor.
+ */
+static int
+xz_filter_close(struct archive_read_filter *self)
+{
+ struct private_data *state;
+
+ state = (struct private_data *)self->data;
+ lzma_end(&(state->stream));
+ free(state->out_block);
+ free(state);
+ return (ARCHIVE_OK);
+}
+
+#else
+
+#if HAVE_LZMADEC_H && HAVE_LIBLZMADEC
+
+/*
+ * If we have the older liblzmadec library, then we can handle
+ * LZMA streams but not XZ streams.
+ */
+
+/*
+ * Setup the callbacks.
+ */
+static int
+lzma_bidder_init(struct archive_read_filter *self)
+{
+ static const size_t out_block_size = 64 * 1024;
+ void *out_block;
+ struct private_data *state;
+ ssize_t ret, avail_in;
+
+ self->code = ARCHIVE_COMPRESSION_LZMA;
+ self->name = "lzma";
+
+ state = (struct private_data *)calloc(sizeof(*state), 1);
+ out_block = (unsigned char *)malloc(out_block_size);
+ if (state == NULL || out_block == NULL) {
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Can't allocate data for lzma decompression");
+ free(out_block);
+ free(state);
+ return (ARCHIVE_FATAL);
+ }
+
+ self->data = state;
+ state->out_block_size = out_block_size;
+ state->out_block = out_block;
+ self->read = lzma_filter_read;
+ self->skip = NULL; /* not supported */
+ self->close = lzma_filter_close;
+
+ /* Prime the lzma library with 18 bytes of input. */
+ state->stream.next_in = (unsigned char *)(uintptr_t)
+ __archive_read_filter_ahead(self->upstream, 18, &avail_in);
+ if (state->stream.next_in == NULL)
+ return (ARCHIVE_FATAL);
+ state->stream.avail_in = avail_in;
+ state->stream.next_out = state->out_block;
+ state->stream.avail_out = state->out_block_size;
+
+ /* Initialize compression library. */
+ ret = lzmadec_init(&(state->stream));
+ __archive_read_filter_consume(self->upstream,
+ avail_in - state->stream.avail_in);
+ if (ret == LZMADEC_OK)
+ return (ARCHIVE_OK);
+
+ /* Library setup failed: Clean up. */
+ archive_set_error(&self->archive->archive, ARCHIVE_ERRNO_MISC,
+ "Internal error initializing lzma library");
+
+ /* Override the error message if we know what really went wrong. */
+ switch (ret) {
+ case LZMADEC_HEADER_ERROR:
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Internal error initializing compression library: "
+ "invalid header");
+ break;
+ case LZMADEC_MEM_ERROR:
+ archive_set_error(&self->archive->archive, ENOMEM,
+ "Internal error initializing compression library: "
+ "out of memory");
+ break;
+ }
+
+ free(state->out_block);
+ free(state);
+ self->data = NULL;
+ return (ARCHIVE_FATAL);
+}
+
+/*
+ * Return the next block of decompressed data.
+ */
+static ssize_t
+lzma_filter_read(struct archive_read_filter *self, const void **p)
+{
+ struct private_data *state;
+ size_t decompressed;
+ ssize_t avail_in, ret;
+
+ state = (struct private_data *)self->data;
+
+ /* Empty our output buffer. */
+ state->stream.next_out = state->out_block;
+ state->stream.avail_out = state->out_block_size;
+
+ /* Try to fill the output buffer. */
+ while (state->stream.avail_out > 0 && !state->eof) {
+ state->stream.next_in = (unsigned char *)(uintptr_t)
+ __archive_read_filter_ahead(self->upstream, 1, &avail_in);
+ if (state->stream.next_in == NULL && avail_in < 0)
+ return (ARCHIVE_FATAL);
+ state->stream.avail_in = avail_in;
+
+ /* Decompress as much as we can in one pass. */
+ ret = lzmadec_decode(&(state->stream), avail_in == 0);
+ switch (ret) {
+ case LZMADEC_STREAM_END: /* Found end of stream. */
+ state->eof = 1;
+ /* FALL THROUGH */
+ case LZMADEC_OK: /* Decompressor made some progress. */
+ __archive_read_filter_consume(self->upstream,
+ avail_in - state->stream.avail_in);
+ break;
+ case LZMADEC_BUF_ERROR: /* Insufficient input data? */
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Insufficient compressed data");
+ return (ARCHIVE_FATAL);
+ default:
+ /* Return an error. */
+ archive_set_error(&self->archive->archive,
+ ARCHIVE_ERRNO_MISC,
+ "Lzma decompression failed");
+ return (ARCHIVE_FATAL);
+ }
+ }
+
+ decompressed = state->stream.next_out - state->out_block;
+ state->total_out += decompressed;
+ if (decompressed == 0)
+ *p = NULL;
+ else
+ *p = state->out_block;
+ return (decompressed);
+}
+
+/*
+ * Clean up the decompressor.
+ */
+static int
+lzma_filter_close(struct archive_read_filter *self)
+{
+ struct private_data *state;
+ int ret;
+
+ state = (struct private_data *)self->data;
+ ret = ARCHIVE_OK;
+ switch (lzmadec_end(&(state->stream))) {
+ case LZMADEC_OK:
+ break;
+ default:
+ archive_set_error(&(self->archive->archive),
+ ARCHIVE_ERRNO_MISC,
+ "Failed to clean up %s compressor",
+ self->archive->archive.compression_name);
+ ret = ARCHIVE_FATAL;
+ }
+
+ free(state->out_block);
+ free(state);
+ return (ret);
+}
+
+#else
+
+/*
+ *
+ * If we have no suitable library on this system, we can't actually do
+ * the decompression. We can, however, still detect compressed
+ * archives and emit a useful message.
+ *
+ */
+static int
+lzma_bidder_init(struct archive_read_filter *self)
+{
+ int r;
+
+ r = __archive_read_program(self, "unlzma");
+ /* Note: We set the format here even if __archive_read_program()
+ * above fails. We do, after all, know what the format is
+ * even if we weren't able to read it. */
+ self->code = ARCHIVE_COMPRESSION_LZMA;
+ self->name = "lzma";
+ return (r);
+}
+
+#endif /* HAVE_LZMADEC_H */
+
+
+static int
+xz_bidder_init(struct archive_read_filter *self)
+{
+ int r;
+
+ r = __archive_read_program(self, "unxz");
+ /* Note: We set the format here even if __archive_read_program()
+ * above fails. We do, after all, know what the format is
+ * even if we weren't able to read it. */
+ self->code = ARCHIVE_COMPRESSION_XZ;
+ self->name = "xz";
+ return (r);
+}
+
+
+#endif /* HAVE_LZMA_H */
diff --git a/libarchive/archive_read_support_format_all.c b/libarchive/archive_read_support_format_all.c
new file mode 100644
index 000000000000..573750811d2c
--- /dev/null
+++ b/libarchive/archive_read_support_format_all.c
@@ -0,0 +1,51 @@
+/*-
+ * Copyright (c) 2003-2011 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_format_all.c 174991 2007-12-30 04:58:22Z kientzle $");
+
+#include "archive.h"
+
+int
+archive_read_support_format_all(struct archive *a)
+{
+ archive_read_support_format_ar(a);
+ archive_read_support_format_cpio(a);
+ archive_read_support_format_empty(a);
+ archive_read_support_format_iso9660(a);
+ archive_read_support_format_mtree(a);
+ archive_read_support_format_tar(a);
+ archive_read_support_format_xar(a);
+ archive_read_support_format_zip(a);
+
+ /* Note: We always return ARCHIVE_OK here, even if some of the
+ * above return ARCHIVE_WARN. The intent here is to enable
+ * "as much as possible." Clients who need specific
+ * compression should enable those individually so they can
+ * verify the level of support. */
+ /* Clear any warning messages set by the above functions. */
+ archive_clear_error(a);
+ return (ARCHIVE_OK);
+}
diff --git a/libarchive/archive_read_support_format_ar.c b/libarchive/archive_read_support_format_ar.c
new file mode 100644
index 000000000000..7c1ca3a0c876
--- /dev/null
+++ b/libarchive/archive_read_support_format_ar.c
@@ -0,0 +1,584 @@
+/*-
+ * Copyright (c) 2007 Kai Wang
+ * Copyright (c) 2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer
+ * in this position and unchanged.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_format_ar.c 201101 2009-12-28 03:06:27Z kientzle $");
+
+#ifdef HAVE_SYS_STAT_H
+#include <sys/stat.h>
+#endif
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#ifdef HAVE_LIMITS_H
+#include <limits.h>
+#endif
+
+#include "archive.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+struct ar {
+ off_t entry_bytes_remaining;
+ off_t entry_offset;
+ off_t entry_padding;
+ char *strtab;
+ size_t strtab_size;
+};
+
+/*
+ * Define structure of the "ar" header.
+ */
+#define AR_name_offset 0
+#define AR_name_size 16
+#define AR_date_offset 16
+#define AR_date_size 12
+#define AR_uid_offset 28
+#define AR_uid_size 6
+#define AR_gid_offset 34
+#define AR_gid_size 6
+#define AR_mode_offset 40
+#define AR_mode_size 8
+#define AR_size_offset 48
+#define AR_size_size 10
+#define AR_fmag_offset 58
+#define AR_fmag_size 2
+
+static int archive_read_format_ar_bid(struct archive_read *a);
+static int archive_read_format_ar_cleanup(struct archive_read *a);
+static int archive_read_format_ar_read_data(struct archive_read *a,
+ const void **buff, size_t *size, off_t *offset);
+static int archive_read_format_ar_skip(struct archive_read *a);
+static int archive_read_format_ar_read_header(struct archive_read *a,
+ struct archive_entry *e);
+static uint64_t ar_atol8(const char *p, unsigned char_cnt);
+static uint64_t ar_atol10(const char *p, unsigned char_cnt);
+static int ar_parse_gnu_filename_table(struct archive_read *a);
+static int ar_parse_common_header(struct ar *ar, struct archive_entry *,
+ const char *h);
+
+int
+archive_read_support_format_ar(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct ar *ar;
+ int r;
+
+ ar = (struct ar *)malloc(sizeof(*ar));
+ if (ar == NULL) {
+ archive_set_error(&a->archive, ENOMEM,
+ "Can't allocate ar data");
+ return (ARCHIVE_FATAL);
+ }
+ memset(ar, 0, sizeof(*ar));
+ ar->strtab = NULL;
+
+ r = __archive_read_register_format(a,
+ ar,
+ "ar",
+ archive_read_format_ar_bid,
+ NULL,
+ archive_read_format_ar_read_header,
+ archive_read_format_ar_read_data,
+ archive_read_format_ar_skip,
+ archive_read_format_ar_cleanup);
+
+ if (r != ARCHIVE_OK) {
+ free(ar);
+ return (r);
+ }
+ return (ARCHIVE_OK);
+}
+
+static int
+archive_read_format_ar_cleanup(struct archive_read *a)
+{
+ struct ar *ar;
+
+ ar = (struct ar *)(a->format->data);
+ if (ar->strtab)
+ free(ar->strtab);
+ free(ar);
+ (a->format->data) = NULL;
+ return (ARCHIVE_OK);
+}
+
+static int
+archive_read_format_ar_bid(struct archive_read *a)
+{
+ const void *h;
+
+ if (a->archive.archive_format != 0 &&
+ (a->archive.archive_format & ARCHIVE_FORMAT_BASE_MASK) !=
+ ARCHIVE_FORMAT_AR)
+ return(0);
+
+ /*
+ * Verify the 8-byte file signature.
+ * TODO: Do we need to check more than this?
+ */
+ if ((h = __archive_read_ahead(a, 8, NULL)) == NULL)
+ return (-1);
+ if (strncmp((const char*)h, "!<arch>\n", 8) == 0) {
+ return (64);
+ }
+ return (-1);
+}
+
+static int
+archive_read_format_ar_read_header(struct archive_read *a,
+ struct archive_entry *entry)
+{
+ char filename[AR_name_size + 1];
+ struct ar *ar;
+ uint64_t number; /* Used to hold parsed numbers before validation. */
+ ssize_t bytes_read;
+ size_t bsd_name_length, entry_size;
+ char *p, *st;
+ const void *b;
+ const char *h;
+ int r;
+
+ ar = (struct ar*)(a->format->data);
+
+ if (a->archive.file_position == 0) {
+ /*
+ * We are now at the beginning of the archive,
+ * so we need first consume the ar global header.
+ */
+ __archive_read_consume(a, 8);
+ /* Set a default format code for now. */
+ a->archive.archive_format = ARCHIVE_FORMAT_AR;
+ }
+
+ /* Read the header for the next file entry. */
+ if ((b = __archive_read_ahead(a, 60, &bytes_read)) == NULL)
+ /* Broken header. */
+ return (ARCHIVE_EOF);
+ __archive_read_consume(a, 60);
+ h = (const char *)b;
+
+ /* Verify the magic signature on the file header. */
+ if (strncmp(h + AR_fmag_offset, "`\n", 2) != 0) {
+ archive_set_error(&a->archive, EINVAL,
+ "Incorrect file header signature");
+ return (ARCHIVE_WARN);
+ }
+
+ /* Copy filename into work buffer. */
+ strncpy(filename, h + AR_name_offset, AR_name_size);
+ filename[AR_name_size] = '\0';
+
+ /*
+ * Guess the format variant based on the filename.
+ */
+ if (a->archive.archive_format == ARCHIVE_FORMAT_AR) {
+ /* We don't already know the variant, so let's guess. */
+ /*
+ * Biggest clue is presence of '/': GNU starts special
+ * filenames with '/', appends '/' as terminator to
+ * non-special names, so anything with '/' should be
+ * GNU except for BSD long filenames.
+ */
+ if (strncmp(filename, "#1/", 3) == 0)
+ a->archive.archive_format = ARCHIVE_FORMAT_AR_BSD;
+ else if (strchr(filename, '/') != NULL)
+ a->archive.archive_format = ARCHIVE_FORMAT_AR_GNU;
+ else if (strncmp(filename, "__.SYMDEF", 9) == 0)
+ a->archive.archive_format = ARCHIVE_FORMAT_AR_BSD;
+ /*
+ * XXX Do GNU/SVR4 'ar' programs ever omit trailing '/'
+ * if name exactly fills 16-byte field? If so, we
+ * can't assume entries without '/' are BSD. XXX
+ */
+ }
+
+ /* Update format name from the code. */
+ if (a->archive.archive_format == ARCHIVE_FORMAT_AR_GNU)
+ a->archive.archive_format_name = "ar (GNU/SVR4)";
+ else if (a->archive.archive_format == ARCHIVE_FORMAT_AR_BSD)
+ a->archive.archive_format_name = "ar (BSD)";
+ else
+ a->archive.archive_format_name = "ar";
+
+ /*
+ * Remove trailing spaces from the filename. GNU and BSD
+ * variants both pad filename area out with spaces.
+ * This will only be wrong if GNU/SVR4 'ar' implementations
+ * omit trailing '/' for 16-char filenames and we have
+ * a 16-char filename that ends in ' '.
+ */
+ p = filename + AR_name_size - 1;
+ while (p >= filename && *p == ' ') {
+ *p = '\0';
+ p--;
+ }
+
+ /*
+ * Remove trailing slash unless first character is '/'.
+ * (BSD entries never end in '/', so this will only trim
+ * GNU-format entries. GNU special entries start with '/'
+ * and are not terminated in '/', so we don't trim anything
+ * that starts with '/'.)
+ */
+ if (filename[0] != '/' && *p == '/')
+ *p = '\0';
+
+ /*
+ * '//' is the GNU filename table.
+ * Later entries can refer to names in this table.
+ */
+ if (strcmp(filename, "//") == 0) {
+ /* This must come before any call to _read_ahead. */
+ ar_parse_common_header(ar, entry, h);
+ archive_entry_copy_pathname(entry, filename);
+ archive_entry_set_filetype(entry, AE_IFREG);
+ /* Get the size of the filename table. */
+ number = ar_atol10(h + AR_size_offset, AR_size_size);
+ if (number > SIZE_MAX) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
+ "Filename table too large");
+ return (ARCHIVE_FATAL);
+ }
+ entry_size = (size_t)number;
+ if (entry_size == 0) {
+ archive_set_error(&a->archive, EINVAL,
+ "Invalid string table");
+ return (ARCHIVE_WARN);
+ }
+ if (ar->strtab != NULL) {
+ archive_set_error(&a->archive, EINVAL,
+ "More than one string tables exist");
+ return (ARCHIVE_WARN);
+ }
+
+ /* Read the filename table into memory. */
+ st = malloc(entry_size);
+ if (st == NULL) {
+ archive_set_error(&a->archive, ENOMEM,
+ "Can't allocate filename table buffer");
+ return (ARCHIVE_FATAL);
+ }
+ ar->strtab = st;
+ ar->strtab_size = entry_size;
+ if ((b = __archive_read_ahead(a, entry_size, NULL)) == NULL)
+ return (ARCHIVE_FATAL);
+ memcpy(st, b, entry_size);
+ __archive_read_consume(a, entry_size);
+ /* All contents are consumed. */
+ ar->entry_bytes_remaining = 0;
+ archive_entry_set_size(entry, ar->entry_bytes_remaining);
+
+ /* Parse the filename table. */
+ return (ar_parse_gnu_filename_table(a));
+ }
+
+ /*
+ * GNU variant handles long filenames by storing /<number>
+ * to indicate a name stored in the filename table.
+ * XXX TODO: Verify that it's all digits... Don't be fooled
+ * by "/9xyz" XXX
+ */
+ if (filename[0] == '/' && filename[1] >= '0' && filename[1] <= '9') {
+ number = ar_atol10(h + AR_name_offset + 1, AR_name_size - 1);
+ /*
+ * If we can't look up the real name, warn and return
+ * the entry with the wrong name.
+ */
+ if (ar->strtab == NULL || number > ar->strtab_size) {
+ archive_set_error(&a->archive, EINVAL,
+ "Can't find long filename for entry");
+ archive_entry_copy_pathname(entry, filename);
+ /* Parse the time, owner, mode, size fields. */
+ ar_parse_common_header(ar, entry, h);
+ return (ARCHIVE_WARN);
+ }
+
+ archive_entry_copy_pathname(entry, &ar->strtab[(size_t)number]);
+ /* Parse the time, owner, mode, size fields. */
+ return (ar_parse_common_header(ar, entry, h));
+ }
+
+ /*
+ * BSD handles long filenames by storing "#1/" followed by the
+ * length of filename as a decimal number, then prepends the
+ * the filename to the file contents.
+ */
+ if (strncmp(filename, "#1/", 3) == 0) {
+ /* Parse the time, owner, mode, size fields. */
+ /* This must occur before _read_ahead is called again. */
+ ar_parse_common_header(ar, entry, h);
+
+ /* Parse the size of the name, adjust the file size. */
+ number = ar_atol10(h + AR_name_offset + 3, AR_name_size - 3);
+ bsd_name_length = (size_t)number;
+ /* Guard against the filename + trailing NUL
+ * overflowing a size_t and against the filename size
+ * being larger than the entire entry. */
+ if (number > (uint64_t)(bsd_name_length + 1)
+ || (off_t)bsd_name_length > ar->entry_bytes_remaining) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
+ "Bad input file size");
+ return (ARCHIVE_FATAL);
+ }
+ ar->entry_bytes_remaining -= bsd_name_length;
+ /* Adjust file size reported to client. */
+ archive_entry_set_size(entry, ar->entry_bytes_remaining);
+
+ /* Read the long name into memory. */
+ if ((b = __archive_read_ahead(a, bsd_name_length, NULL)) == NULL) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
+ "Truncated input file");
+ return (ARCHIVE_FATAL);
+ }
+ __archive_read_consume(a, bsd_name_length);
+
+ /* Store it in the entry. */
+ p = (char *)malloc(bsd_name_length + 1);
+ if (p == NULL) {
+ archive_set_error(&a->archive, ENOMEM,
+ "Can't allocate fname buffer");
+ return (ARCHIVE_FATAL);
+ }
+ strncpy(p, b, bsd_name_length);
+ p[bsd_name_length] = '\0';
+ archive_entry_copy_pathname(entry, p);
+ free(p);
+ return (ARCHIVE_OK);
+ }
+
+ /*
+ * "/" is the SVR4/GNU archive symbol table.
+ */
+ if (strcmp(filename, "/") == 0) {
+ archive_entry_copy_pathname(entry, "/");
+ /* Parse the time, owner, mode, size fields. */
+ r = ar_parse_common_header(ar, entry, h);
+ /* Force the file type to a regular file. */
+ archive_entry_set_filetype(entry, AE_IFREG);
+ return (r);
+ }
+
+ /*
+ * "__.SYMDEF" is a BSD archive symbol table.
+ */
+ if (strcmp(filename, "__.SYMDEF") == 0) {
+ archive_entry_copy_pathname(entry, filename);
+ /* Parse the time, owner, mode, size fields. */
+ return (ar_parse_common_header(ar, entry, h));
+ }
+
+ /*
+ * Otherwise, this is a standard entry. The filename
+ * has already been trimmed as much as possible, based
+ * on our current knowledge of the format.
+ */
+ archive_entry_copy_pathname(entry, filename);
+ return (ar_parse_common_header(ar, entry, h));
+}
+
+static int
+ar_parse_common_header(struct ar *ar, struct archive_entry *entry,
+ const char *h)
+{
+ uint64_t n;
+
+ /* Copy remaining header */
+ archive_entry_set_mtime(entry,
+ (time_t)ar_atol10(h + AR_date_offset, AR_date_size), 0L);
+ archive_entry_set_uid(entry,
+ (uid_t)ar_atol10(h + AR_uid_offset, AR_uid_size));
+ archive_entry_set_gid(entry,
+ (gid_t)ar_atol10(h + AR_gid_offset, AR_gid_size));
+ archive_entry_set_mode(entry,
+ (mode_t)ar_atol8(h + AR_mode_offset, AR_mode_size));
+ n = ar_atol10(h + AR_size_offset, AR_size_size);
+
+ ar->entry_offset = 0;
+ ar->entry_padding = n % 2;
+ archive_entry_set_size(entry, n);
+ ar->entry_bytes_remaining = n;
+ return (ARCHIVE_OK);
+}
+
+static int
+archive_read_format_ar_read_data(struct archive_read *a,
+ const void **buff, size_t *size, off_t *offset)
+{
+ ssize_t bytes_read;
+ struct ar *ar;
+
+ ar = (struct ar *)(a->format->data);
+
+ if (ar->entry_bytes_remaining > 0) {
+ *buff = __archive_read_ahead(a, 1, &bytes_read);
+ if (bytes_read == 0) {
+ archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
+ "Truncated ar archive");
+ return (ARCHIVE_FATAL);
+ }
+ if (bytes_read < 0)
+ return (ARCHIVE_FATAL);
+ if (bytes_read > ar->entry_bytes_remaining)
+ bytes_read = (ssize_t)ar->entry_bytes_remaining;
+ *size = bytes_read;
+ *offset = ar->entry_offset;
+ ar->entry_offset += bytes_read;
+ ar->entry_bytes_remaining -= bytes_read;
+ __archive_read_consume(a, (size_t)bytes_read);
+ return (ARCHIVE_OK);
+ } else {
+ while (ar->entry_padding > 0) {
+ *buff = __archive_read_ahead(a, 1, &bytes_read);
+ if (bytes_read <= 0)
+ return (ARCHIVE_FATAL);
+ if (bytes_read > ar->entry_padding)
+ bytes_read = (ssize_t)ar->entry_padding;
+ __archive_read_consume(a, (size_t)bytes_read);
+ ar->entry_padding -= bytes_read;
+ }
+ *buff = NULL;
+ *size = 0;
+ *offset = ar->entry_offset;
+ return (ARCHIVE_EOF);
+ }
+}
+
+static int
+archive_read_format_ar_skip(struct archive_read *a)
+{
+ off_t bytes_skipped;
+ struct ar* ar;
+
+ ar = (struct ar *)(a->format->data);
+
+ bytes_skipped = __archive_read_skip(a,
+ ar->entry_bytes_remaining + ar->entry_padding);
+ if (bytes_skipped < 0)
+ return (ARCHIVE_FATAL);
+
+ ar->entry_bytes_remaining = 0;
+ ar->entry_padding = 0;
+
+ return (ARCHIVE_OK);
+}
+
+static int
+ar_parse_gnu_filename_table(struct archive_read *a)
+{
+ struct ar *ar;
+ char *p;
+ size_t size;
+
+ ar = (struct ar*)(a->format->data);
+ size = ar->strtab_size;
+
+ for (p = ar->strtab; p < ar->strtab + size - 1; ++p) {
+ if (*p == '/') {
+ *p++ = '\0';
+ if (*p != '\n')
+ goto bad_string_table;
+ *p = '\0';
+ }
+ }
+ /*
+ * GNU ar always pads the table to an even size.
+ * The pad character is either '\n' or '`'.
+ */
+ if (p != ar->strtab + size && *p != '\n' && *p != '`')
+ goto bad_string_table;
+
+ /* Enforce zero termination. */
+ ar->strtab[size - 1] = '\0';
+
+ return (ARCHIVE_OK);
+
+bad_string_table:
+ archive_set_error(&a->archive, EINVAL,
+ "Invalid string table");
+ free(ar->strtab);
+ ar->strtab = NULL;
+ return (ARCHIVE_WARN);
+}
+
+static uint64_t
+ar_atol8(const char *p, unsigned char_cnt)
+{
+ uint64_t l, limit, last_digit_limit;
+ unsigned int digit, base;
+
+ base = 8;
+ limit = UINT64_MAX / base;
+ last_digit_limit = UINT64_MAX % base;
+
+ while ((*p == ' ' || *p == '\t') && char_cnt-- > 0)
+ p++;
+
+ l = 0;
+ digit = *p - '0';
+ while (*p >= '0' && digit < base && char_cnt-- > 0) {
+ if (l>limit || (l == limit && digit > last_digit_limit)) {
+ l = UINT64_MAX; /* Truncate on overflow. */
+ break;
+ }
+ l = (l * base) + digit;
+ digit = *++p - '0';
+ }
+ return (l);
+}
+
+static uint64_t
+ar_atol10(const char *p, unsigned char_cnt)
+{
+ uint64_t l, limit, last_digit_limit;
+ unsigned int base, digit;
+
+ base = 10;
+ limit = UINT64_MAX / base;
+ last_digit_limit = UINT64_MAX % base;
+
+ while ((*p == ' ' || *p == '\t') && char_cnt-- > 0)
+ p++;
+ l = 0;
+ digit = *p - '0';
+ while (*p >= '0' && digit < base && char_cnt-- > 0) {
+ if (l > limit || (l == limit && digit > last_digit_limit)) {
+ l = UINT64_MAX; /* Truncate on overflow. */
+ break;
+ }
+ l = (l * base) + digit;
+ digit = *++p - '0';
+ }
+ return (l);
+}
diff --git a/libarchive/archive_read_support_format_cpio.c b/libarchive/archive_read_support_format_cpio.c
new file mode 100644
index 000000000000..ed579b90cb37
--- /dev/null
+++ b/libarchive/archive_read_support_format_cpio.c
@@ -0,0 +1,777 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_format_cpio.c 201163 2009-12-29 05:50:34Z kientzle $");
+
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+/* #include <stdint.h> */ /* See archive_platform.h */
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+
+#include "archive.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+struct cpio_bin_header {
+ unsigned char c_magic[2];
+ unsigned char c_dev[2];
+ unsigned char c_ino[2];
+ unsigned char c_mode[2];
+ unsigned char c_uid[2];
+ unsigned char c_gid[2];
+ unsigned char c_nlink[2];
+ unsigned char c_rdev[2];
+ unsigned char c_mtime[4];
+ unsigned char c_namesize[2];
+ unsigned char c_filesize[4];
+};
+
+struct cpio_odc_header {
+ char c_magic[6];
+ char c_dev[6];
+ char c_ino[6];
+ char c_mode[6];
+ char c_uid[6];
+ char c_gid[6];
+ char c_nlink[6];
+ char c_rdev[6];
+ char c_mtime[11];
+ char c_namesize[6];
+ char c_filesize[11];
+};
+
+struct cpio_newc_header {
+ char c_magic[6];
+ char c_ino[8];
+ char c_mode[8];
+ char c_uid[8];
+ char c_gid[8];
+ char c_nlink[8];
+ char c_mtime[8];
+ char c_filesize[8];
+ char c_devmajor[8];
+ char c_devminor[8];
+ char c_rdevmajor[8];
+ char c_rdevminor[8];
+ char c_namesize[8];
+ char c_crc[8];
+};
+
+struct links_entry {
+ struct links_entry *next;
+ struct links_entry *previous;
+ int links;
+ dev_t dev;
+ int64_t ino;
+ char *name;
+};
+
+#define CPIO_MAGIC 0x13141516
+struct cpio {
+ int magic;
+ int (*read_header)(struct archive_read *, struct cpio *,
+ struct archive_entry *, size_t *, size_t *);
+ struct links_entry *links_head;
+ struct archive_string entry_name;
+ struct archive_string entry_linkname;
+ off_t entry_bytes_remaining;
+ off_t entry_offset;
+ off_t entry_padding;
+};
+
+static int64_t atol16(const char *, unsigned);
+static int64_t atol8(const char *, unsigned);
+static int archive_read_format_cpio_bid(struct archive_read *);
+static int archive_read_format_cpio_cleanup(struct archive_read *);
+static int archive_read_format_cpio_read_data(struct archive_read *,
+ const void **, size_t *, off_t *);
+static int archive_read_format_cpio_read_header(struct archive_read *,
+ struct archive_entry *);
+static int be4(const unsigned char *);
+static int find_odc_header(struct archive_read *);
+static int find_newc_header(struct archive_read *);
+static int header_bin_be(struct archive_read *, struct cpio *,
+ struct archive_entry *, size_t *, size_t *);
+static int header_bin_le(struct archive_read *, struct cpio *,
+ struct archive_entry *, size_t *, size_t *);
+static int header_newc(struct archive_read *, struct cpio *,
+ struct archive_entry *, size_t *, size_t *);
+static int header_odc(struct archive_read *, struct cpio *,
+ struct archive_entry *, size_t *, size_t *);
+static int is_octal(const char *, size_t);
+static int is_hex(const char *, size_t);
+static int le4(const unsigned char *);
+static void record_hardlink(struct cpio *cpio, struct archive_entry *entry);
+
+int
+archive_read_support_format_cpio(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct cpio *cpio;
+ int r;
+
+ cpio = (struct cpio *)malloc(sizeof(*cpio));
+ if (cpio == NULL) {
+ archive_set_error(&a->archive, ENOMEM, "Can't allocate cpio data");
+ return (ARCHIVE_FATAL);
+ }
+ memset(cpio, 0, sizeof(*cpio));
+ cpio->magic = CPIO_MAGIC;
+
+ r = __archive_read_register_format(a,
+ cpio,
+ "cpio",
+ archive_read_format_cpio_bid,
+ NULL,
+ archive_read_format_cpio_read_header,
+ archive_read_format_cpio_read_data,
+ NULL,
+ archive_read_format_cpio_cleanup);
+
+ if (r != ARCHIVE_OK)
+ free(cpio);
+ return (ARCHIVE_OK);
+}
+
+
+static int
+archive_read_format_cpio_bid(struct archive_read *a)
+{
+ const void *h;
+ const unsigned char *p;
+ struct cpio *cpio;
+ int bid;
+
+ cpio = (struct cpio *)(a->format->data);
+
+ if ((h = __archive_read_ahead(a, 6, NULL)) == NULL)
+ return (-1);
+
+ p = (const unsigned char *)h;
+ bid = 0;
+ if (memcmp(p, "070707", 6) == 0) {
+ /* ASCII cpio archive (odc, POSIX.1) */
+ cpio->read_header = header_odc;
+ bid += 48;
+ /*
+ * XXX TODO: More verification; Could check that only octal
+ * digits appear in appropriate header locations. XXX
+ */
+ } else if (memcmp(p, "070701", 6) == 0) {
+ /* ASCII cpio archive (SVR4 without CRC) */
+ cpio->read_header = header_newc;
+ bid += 48;
+ /*
+ * XXX TODO: More verification; Could check that only hex
+ * digits appear in appropriate header locations. XXX
+ */
+ } else if (memcmp(p, "070702", 6) == 0) {
+ /* ASCII cpio archive (SVR4 with CRC) */
+ /* XXX TODO: Flag that we should check the CRC. XXX */
+ cpio->read_header = header_newc;
+ bid += 48;
+ /*
+ * XXX TODO: More verification; Could check that only hex
+ * digits appear in appropriate header locations. XXX
+ */
+ } else if (p[0] * 256 + p[1] == 070707) {
+ /* big-endian binary cpio archives */
+ cpio->read_header = header_bin_be;
+ bid += 16;
+ /* Is more verification possible here? */
+ } else if (p[0] + p[1] * 256 == 070707) {
+ /* little-endian binary cpio archives */
+ cpio->read_header = header_bin_le;
+ bid += 16;
+ /* Is more verification possible here? */
+ } else
+ return (ARCHIVE_WARN);
+
+ return (bid);
+}
+
+static int
+archive_read_format_cpio_read_header(struct archive_read *a,
+ struct archive_entry *entry)
+{
+ struct cpio *cpio;
+ const void *h;
+ size_t namelength;
+ size_t name_pad;
+ int r;
+
+ cpio = (struct cpio *)(a->format->data);
+ r = (cpio->read_header(a, cpio, entry, &namelength, &name_pad));
+
+ if (r < ARCHIVE_WARN)
+ return (r);
+
+ /* Read name from buffer. */
+ h = __archive_read_ahead(a, namelength + name_pad, NULL);
+ if (h == NULL)
+ return (ARCHIVE_FATAL);
+ __archive_read_consume(a, namelength + name_pad);
+ archive_strncpy(&cpio->entry_name, (const char *)h, namelength);
+ archive_entry_set_pathname(entry, cpio->entry_name.s);
+ cpio->entry_offset = 0;
+
+ /* If this is a symlink, read the link contents. */
+ if (archive_entry_filetype(entry) == AE_IFLNK) {
+ h = __archive_read_ahead(a, cpio->entry_bytes_remaining, NULL);
+ if (h == NULL)
+ return (ARCHIVE_FATAL);
+ __archive_read_consume(a, cpio->entry_bytes_remaining);
+ archive_strncpy(&cpio->entry_linkname, (const char *)h,
+ cpio->entry_bytes_remaining);
+ archive_entry_set_symlink(entry, cpio->entry_linkname.s);
+ cpio->entry_bytes_remaining = 0;
+ }
+
+ /* XXX TODO: If the full mode is 0160200, then this is a Solaris
+ * ACL description for the following entry. Read this body
+ * and parse it as a Solaris-style ACL, then read the next
+ * header. XXX */
+
+ /* Compare name to "TRAILER!!!" to test for end-of-archive. */
+ if (namelength == 11 && strcmp((const char *)h, "TRAILER!!!") == 0) {
+ /* TODO: Store file location of start of block. */
+ archive_clear_error(&a->archive);
+ return (ARCHIVE_EOF);
+ }
+
+ /* Detect and record hardlinks to previously-extracted entries. */
+ record_hardlink(cpio, entry);
+
+ return (r);
+}
+
+static int
+archive_read_format_cpio_read_data(struct archive_read *a,
+ const void **buff, size_t *size, off_t *offset)
+{
+ ssize_t bytes_read;
+ struct cpio *cpio;
+
+ cpio = (struct cpio *)(a->format->data);
+ if (cpio->entry_bytes_remaining > 0) {
+ *buff = __archive_read_ahead(a, 1, &bytes_read);
+ if (bytes_read <= 0)
+ return (ARCHIVE_FATAL);
+ if (bytes_read > cpio->entry_bytes_remaining)
+ bytes_read = cpio->entry_bytes_remaining;
+ *size = bytes_read;
+ *offset = cpio->entry_offset;
+ cpio->entry_offset += bytes_read;
+ cpio->entry_bytes_remaining -= bytes_read;
+ __archive_read_consume(a, bytes_read);
+ return (ARCHIVE_OK);
+ } else {
+ while (cpio->entry_padding > 0) {
+ *buff = __archive_read_ahead(a, 1, &bytes_read);
+ if (bytes_read <= 0)
+ return (ARCHIVE_FATAL);
+ if (bytes_read > cpio->entry_padding)
+ bytes_read = cpio->entry_padding;
+ __archive_read_consume(a, bytes_read);
+ cpio->entry_padding -= bytes_read;
+ }
+ *buff = NULL;
+ *size = 0;
+ *offset = cpio->entry_offset;
+ return (ARCHIVE_EOF);
+ }
+}
+
+/*
+ * Skip forward to the next cpio newc header by searching for the
+ * 07070[12] string. This should be generalized and merged with
+ * find_odc_header below.
+ */
+static int
+is_hex(const char *p, size_t len)
+{
+ while (len-- > 0) {
+ if ((*p >= '0' && *p <= '9')
+ || (*p >= 'a' && *p <= 'f')
+ || (*p >= 'A' && *p <= 'F'))
+ ++p;
+ else
+ return (0);
+ }
+ return (1);
+}
+
+static int
+find_newc_header(struct archive_read *a)
+{
+ const void *h;
+ const char *p, *q;
+ size_t skip, skipped = 0;
+ ssize_t bytes;
+
+ for (;;) {
+ h = __archive_read_ahead(a, sizeof(struct cpio_newc_header), &bytes);
+ if (h == NULL)
+ return (ARCHIVE_FATAL);
+ p = h;
+ q = p + bytes;
+
+ /* Try the typical case first, then go into the slow search.*/
+ if (memcmp("07070", p, 5) == 0
+ && (p[5] == '1' || p[5] == '2')
+ && is_hex(p, sizeof(struct cpio_newc_header)))
+ return (ARCHIVE_OK);
+
+ /*
+ * Scan ahead until we find something that looks
+ * like an odc header.
+ */
+ while (p + sizeof(struct cpio_newc_header) <= q) {
+ switch (p[5]) {
+ case '1':
+ case '2':
+ if (memcmp("07070", p, 5) == 0
+ && is_hex(p, sizeof(struct cpio_newc_header))) {
+ skip = p - (const char *)h;
+ __archive_read_consume(a, skip);
+ skipped += skip;
+ if (skipped > 0) {
+ archive_set_error(&a->archive,
+ 0,
+ "Skipped %d bytes before "
+ "finding valid header",
+ (int)skipped);
+ return (ARCHIVE_WARN);
+ }
+ return (ARCHIVE_OK);
+ }
+ p += 2;
+ break;
+ case '0':
+ p++;
+ break;
+ default:
+ p += 6;
+ break;
+ }
+ }
+ skip = p - (const char *)h;
+ __archive_read_consume(a, skip);
+ skipped += skip;
+ }
+}
+
+static int
+header_newc(struct archive_read *a, struct cpio *cpio,
+ struct archive_entry *entry, size_t *namelength, size_t *name_pad)
+{
+ const void *h;
+ const struct cpio_newc_header *header;
+ int r;
+
+ r = find_newc_header(a);
+ if (r < ARCHIVE_WARN)
+ return (r);
+
+ /* Read fixed-size portion of header. */
+ h = __archive_read_ahead(a, sizeof(struct cpio_newc_header), NULL);
+ if (h == NULL)
+ return (ARCHIVE_FATAL);
+ __archive_read_consume(a, sizeof(struct cpio_newc_header));
+
+ /* Parse out hex fields. */
+ header = (const struct cpio_newc_header *)h;
+
+ if (memcmp(header->c_magic, "070701", 6) == 0) {
+ a->archive.archive_format = ARCHIVE_FORMAT_CPIO_SVR4_NOCRC;
+ a->archive.archive_format_name = "ASCII cpio (SVR4 with no CRC)";
+ } else if (memcmp(header->c_magic, "070702", 6) == 0) {
+ a->archive.archive_format = ARCHIVE_FORMAT_CPIO_SVR4_CRC;
+ a->archive.archive_format_name = "ASCII cpio (SVR4 with CRC)";
+ } else {
+ /* TODO: Abort here? */
+ }
+
+ archive_entry_set_devmajor(entry, atol16(header->c_devmajor, sizeof(header->c_devmajor)));
+ archive_entry_set_devminor(entry, atol16(header->c_devminor, sizeof(header->c_devminor)));
+ archive_entry_set_ino(entry, atol16(header->c_ino, sizeof(header->c_ino)));
+ archive_entry_set_mode(entry, atol16(header->c_mode, sizeof(header->c_mode)));
+ archive_entry_set_uid(entry, atol16(header->c_uid, sizeof(header->c_uid)));
+ archive_entry_set_gid(entry, atol16(header->c_gid, sizeof(header->c_gid)));
+ archive_entry_set_nlink(entry, atol16(header->c_nlink, sizeof(header->c_nlink)));
+ archive_entry_set_rdevmajor(entry, atol16(header->c_rdevmajor, sizeof(header->c_rdevmajor)));
+ archive_entry_set_rdevminor(entry, atol16(header->c_rdevminor, sizeof(header->c_rdevminor)));
+ archive_entry_set_mtime(entry, atol16(header->c_mtime, sizeof(header->c_mtime)), 0);
+ *namelength = atol16(header->c_namesize, sizeof(header->c_namesize));
+ /* Pad name to 2 more than a multiple of 4. */
+ *name_pad = (2 - *namelength) & 3;
+
+ /*
+ * Note: entry_bytes_remaining is at least 64 bits and
+ * therefore guaranteed to be big enough for a 33-bit file
+ * size.
+ */
+ cpio->entry_bytes_remaining =
+ atol16(header->c_filesize, sizeof(header->c_filesize));
+ archive_entry_set_size(entry, cpio->entry_bytes_remaining);
+ /* Pad file contents to a multiple of 4. */
+ cpio->entry_padding = 3 & -cpio->entry_bytes_remaining;
+ return (r);
+}
+
+/*
+ * Skip forward to the next cpio odc header by searching for the
+ * 070707 string. This is a hand-optimized search that could
+ * probably be easily generalized to handle all character-based
+ * cpio variants.
+ */
+static int
+is_octal(const char *p, size_t len)
+{
+ while (len-- > 0) {
+ if (*p < '0' || *p > '7')
+ return (0);
+ ++p;
+ }
+ return (1);
+}
+
+static int
+find_odc_header(struct archive_read *a)
+{
+ const void *h;
+ const char *p, *q;
+ size_t skip, skipped = 0;
+ ssize_t bytes;
+
+ for (;;) {
+ h = __archive_read_ahead(a, sizeof(struct cpio_odc_header), &bytes);
+ if (h == NULL)
+ return (ARCHIVE_FATAL);
+ p = h;
+ q = p + bytes;
+
+ /* Try the typical case first, then go into the slow search.*/
+ if (memcmp("070707", p, 6) == 0
+ && is_octal(p, sizeof(struct cpio_odc_header)))
+ return (ARCHIVE_OK);
+
+ /*
+ * Scan ahead until we find something that looks
+ * like an odc header.
+ */
+ while (p + sizeof(struct cpio_odc_header) <= q) {
+ switch (p[5]) {
+ case '7':
+ if (memcmp("070707", p, 6) == 0
+ && is_octal(p, sizeof(struct cpio_odc_header))) {
+ skip = p - (const char *)h;
+ __archive_read_consume(a, skip);
+ skipped += skip;
+ if (skipped > 0) {
+ archive_set_error(&a->archive,
+ 0,
+ "Skipped %d bytes before "
+ "finding valid header",
+ (int)skipped);
+ return (ARCHIVE_WARN);
+ }
+ return (ARCHIVE_OK);
+ }
+ p += 2;
+ break;
+ case '0':
+ p++;
+ break;
+ default:
+ p += 6;
+ break;
+ }
+ }
+ skip = p - (const char *)h;
+ __archive_read_consume(a, skip);
+ skipped += skip;
+ }
+}
+
+static int
+header_odc(struct archive_read *a, struct cpio *cpio,
+ struct archive_entry *entry, size_t *namelength, size_t *name_pad)
+{
+ const void *h;
+ int r;
+ const struct cpio_odc_header *header;
+
+ a->archive.archive_format = ARCHIVE_FORMAT_CPIO_POSIX;
+ a->archive.archive_format_name = "POSIX octet-oriented cpio";
+
+ /* Find the start of the next header. */
+ r = find_odc_header(a);
+ if (r < ARCHIVE_WARN)
+ return (r);
+
+ /* Read fixed-size portion of header. */
+ h = __archive_read_ahead(a, sizeof(struct cpio_odc_header), NULL);
+ if (h == NULL)
+ return (ARCHIVE_FATAL);
+ __archive_read_consume(a, sizeof(struct cpio_odc_header));
+
+ /* Parse out octal fields. */
+ header = (const struct cpio_odc_header *)h;
+
+ archive_entry_set_dev(entry, atol8(header->c_dev, sizeof(header->c_dev)));
+ archive_entry_set_ino(entry, atol8(header->c_ino, sizeof(header->c_ino)));
+ archive_entry_set_mode(entry, atol8(header->c_mode, sizeof(header->c_mode)));
+ archive_entry_set_uid(entry, atol8(header->c_uid, sizeof(header->c_uid)));
+ archive_entry_set_gid(entry, atol8(header->c_gid, sizeof(header->c_gid)));
+ archive_entry_set_nlink(entry, atol8(header->c_nlink, sizeof(header->c_nlink)));
+ archive_entry_set_rdev(entry, atol8(header->c_rdev, sizeof(header->c_rdev)));
+ archive_entry_set_mtime(entry, atol8(header->c_mtime, sizeof(header->c_mtime)), 0);
+ *namelength = atol8(header->c_namesize, sizeof(header->c_namesize));
+ *name_pad = 0; /* No padding of filename. */
+
+ /*
+ * Note: entry_bytes_remaining is at least 64 bits and
+ * therefore guaranteed to be big enough for a 33-bit file
+ * size.
+ */
+ cpio->entry_bytes_remaining =
+ atol8(header->c_filesize, sizeof(header->c_filesize));
+ archive_entry_set_size(entry, cpio->entry_bytes_remaining);
+ cpio->entry_padding = 0;
+ return (r);
+}
+
+static int
+header_bin_le(struct archive_read *a, struct cpio *cpio,
+ struct archive_entry *entry, size_t *namelength, size_t *name_pad)
+{
+ const void *h;
+ const struct cpio_bin_header *header;
+
+ a->archive.archive_format = ARCHIVE_FORMAT_CPIO_BIN_LE;
+ a->archive.archive_format_name = "cpio (little-endian binary)";
+
+ /* Read fixed-size portion of header. */
+ h = __archive_read_ahead(a, sizeof(struct cpio_bin_header), NULL);
+ if (h == NULL)
+ return (ARCHIVE_FATAL);
+ __archive_read_consume(a, sizeof(struct cpio_bin_header));
+
+ /* Parse out binary fields. */
+ header = (const struct cpio_bin_header *)h;
+
+ archive_entry_set_dev(entry, header->c_dev[0] + header->c_dev[1] * 256);
+ archive_entry_set_ino(entry, header->c_ino[0] + header->c_ino[1] * 256);
+ archive_entry_set_mode(entry, header->c_mode[0] + header->c_mode[1] * 256);
+ archive_entry_set_uid(entry, header->c_uid[0] + header->c_uid[1] * 256);
+ archive_entry_set_gid(entry, header->c_gid[0] + header->c_gid[1] * 256);
+ archive_entry_set_nlink(entry, header->c_nlink[0] + header->c_nlink[1] * 256);
+ archive_entry_set_rdev(entry, header->c_rdev[0] + header->c_rdev[1] * 256);
+ archive_entry_set_mtime(entry, le4(header->c_mtime), 0);
+ *namelength = header->c_namesize[0] + header->c_namesize[1] * 256;
+ *name_pad = *namelength & 1; /* Pad to even. */
+
+ cpio->entry_bytes_remaining = le4(header->c_filesize);
+ archive_entry_set_size(entry, cpio->entry_bytes_remaining);
+ cpio->entry_padding = cpio->entry_bytes_remaining & 1; /* Pad to even. */
+ return (ARCHIVE_OK);
+}
+
+static int
+header_bin_be(struct archive_read *a, struct cpio *cpio,
+ struct archive_entry *entry, size_t *namelength, size_t *name_pad)
+{
+ const void *h;
+ const struct cpio_bin_header *header;
+
+ a->archive.archive_format = ARCHIVE_FORMAT_CPIO_BIN_BE;
+ a->archive.archive_format_name = "cpio (big-endian binary)";
+
+ /* Read fixed-size portion of header. */
+ h = __archive_read_ahead(a, sizeof(struct cpio_bin_header), NULL);
+ if (h == NULL)
+ return (ARCHIVE_FATAL);
+ __archive_read_consume(a, sizeof(struct cpio_bin_header));
+
+ /* Parse out binary fields. */
+ header = (const struct cpio_bin_header *)h;
+ archive_entry_set_dev(entry, header->c_dev[0] * 256 + header->c_dev[1]);
+ archive_entry_set_ino(entry, header->c_ino[0] * 256 + header->c_ino[1]);
+ archive_entry_set_mode(entry, header->c_mode[0] * 256 + header->c_mode[1]);
+ archive_entry_set_uid(entry, header->c_uid[0] * 256 + header->c_uid[1]);
+ archive_entry_set_gid(entry, header->c_gid[0] * 256 + header->c_gid[1]);
+ archive_entry_set_nlink(entry, header->c_nlink[0] * 256 + header->c_nlink[1]);
+ archive_entry_set_rdev(entry, header->c_rdev[0] * 256 + header->c_rdev[1]);
+ archive_entry_set_mtime(entry, be4(header->c_mtime), 0);
+ *namelength = header->c_namesize[0] * 256 + header->c_namesize[1];
+ *name_pad = *namelength & 1; /* Pad to even. */
+
+ cpio->entry_bytes_remaining = be4(header->c_filesize);
+ archive_entry_set_size(entry, cpio->entry_bytes_remaining);
+ cpio->entry_padding = cpio->entry_bytes_remaining & 1; /* Pad to even. */
+ return (ARCHIVE_OK);
+}
+
+static int
+archive_read_format_cpio_cleanup(struct archive_read *a)
+{
+ struct cpio *cpio;
+
+ cpio = (struct cpio *)(a->format->data);
+ /* Free inode->name map */
+ while (cpio->links_head != NULL) {
+ struct links_entry *lp = cpio->links_head->next;
+
+ if (cpio->links_head->name)
+ free(cpio->links_head->name);
+ free(cpio->links_head);
+ cpio->links_head = lp;
+ }
+ archive_string_free(&cpio->entry_name);
+ free(cpio);
+ (a->format->data) = NULL;
+ return (ARCHIVE_OK);
+}
+
+static int
+le4(const unsigned char *p)
+{
+ return ((p[0]<<16) + (p[1]<<24) + (p[2]<<0) + (p[3]<<8));
+}
+
+
+static int
+be4(const unsigned char *p)
+{
+ return ((p[0]<<24) + (p[1]<<16) + (p[2]<<8) + (p[3]));
+}
+
+/*
+ * Note that this implementation does not (and should not!) obey
+ * locale settings; you cannot simply substitute strtol here, since
+ * it does obey locale.
+ */
+static int64_t
+atol8(const char *p, unsigned char_cnt)
+{
+ int64_t l;
+ int digit;
+
+ l = 0;
+ while (char_cnt-- > 0) {
+ if (*p >= '0' && *p <= '7')
+ digit = *p - '0';
+ else
+ return (l);
+ p++;
+ l <<= 3;
+ l |= digit;
+ }
+ return (l);
+}
+
+static int64_t
+atol16(const char *p, unsigned char_cnt)
+{
+ int64_t l;
+ int digit;
+
+ l = 0;
+ while (char_cnt-- > 0) {
+ if (*p >= 'a' && *p <= 'f')
+ digit = *p - 'a' + 10;
+ else if (*p >= 'A' && *p <= 'F')
+ digit = *p - 'A' + 10;
+ else if (*p >= '0' && *p <= '9')
+ digit = *p - '0';
+ else
+ return (l);
+ p++;
+ l <<= 4;
+ l |= digit;
+ }
+ return (l);
+}
+
+static void
+record_hardlink(struct cpio *cpio, struct archive_entry *entry)
+{
+ struct links_entry *le;
+ dev_t dev;
+ int64_t ino;
+
+ if (archive_entry_nlink(entry) <= 1)
+ return;
+
+ dev = archive_entry_dev(entry);
+ ino = archive_entry_ino64(entry);
+
+ /*
+ * First look in the list of multiply-linked files. If we've
+ * already dumped it, convert this entry to a hard link entry.
+ */
+ for (le = cpio->links_head; le; le = le->next) {
+ if (le->dev == dev && le->ino == ino) {
+ archive_entry_copy_hardlink(entry, le->name);
+
+ if (--le->links <= 0) {
+ if (le->previous != NULL)
+ le->previous->next = le->next;
+ if (le->next != NULL)
+ le->next->previous = le->previous;
+ if (cpio->links_head == le)
+ cpio->links_head = le->next;
+ free(le->name);
+ free(le);
+ }
+
+ return;
+ }
+ }
+
+ le = (struct links_entry *)malloc(sizeof(struct links_entry));
+ if (le == NULL)
+ __archive_errx(1, "Out of memory adding file to list");
+ if (cpio->links_head != NULL)
+ cpio->links_head->previous = le;
+ le->next = cpio->links_head;
+ le->previous = NULL;
+ cpio->links_head = le;
+ le->dev = dev;
+ le->ino = ino;
+ le->links = archive_entry_nlink(entry) - 1;
+ le->name = strdup(archive_entry_pathname(entry));
+ if (le->name == NULL)
+ __archive_errx(1, "Out of memory adding file to list");
+}
diff --git a/libarchive/archive_read_support_format_empty.c b/libarchive/archive_read_support_format_empty.c
new file mode 100644
index 000000000000..518fdcb4913f
--- /dev/null
+++ b/libarchive/archive_read_support_format_empty.c
@@ -0,0 +1,93 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_format_empty.c 191524 2009-04-26 18:24:14Z kientzle $");
+
+#include "archive.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+
+static int archive_read_format_empty_bid(struct archive_read *);
+static int archive_read_format_empty_read_data(struct archive_read *,
+ const void **, size_t *, off_t *);
+static int archive_read_format_empty_read_header(struct archive_read *,
+ struct archive_entry *);
+int
+archive_read_support_format_empty(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ int r;
+
+ r = __archive_read_register_format(a,
+ NULL,
+ NULL,
+ archive_read_format_empty_bid,
+ NULL,
+ archive_read_format_empty_read_header,
+ archive_read_format_empty_read_data,
+ NULL,
+ NULL);
+
+ return (r);
+}
+
+
+static int
+archive_read_format_empty_bid(struct archive_read *a)
+{
+ ssize_t avail;
+
+ (void)__archive_read_ahead(a, 1, &avail);
+ if (avail != 0)
+ return (-1);
+ return (1);
+}
+
+static int
+archive_read_format_empty_read_header(struct archive_read *a,
+ struct archive_entry *entry)
+{
+ (void)a; /* UNUSED */
+ (void)entry; /* UNUSED */
+
+ a->archive.archive_format = ARCHIVE_FORMAT_EMPTY;
+ a->archive.archive_format_name = "Empty file";
+
+ return (ARCHIVE_EOF);
+}
+
+static int
+archive_read_format_empty_read_data(struct archive_read *a,
+ const void **buff, size_t *size, off_t *offset)
+{
+ (void)a; /* UNUSED */
+ (void)buff; /* UNUSED */
+ (void)size; /* UNUSED */
+ (void)offset; /* UNUSED */
+
+ return (ARCHIVE_EOF);
+}
diff --git a/libarchive/archive_read_support_format_iso9660.c b/libarchive/archive_read_support_format_iso9660.c
new file mode 100644
index 000000000000..f35f0ea80bf3
--- /dev/null
+++ b/libarchive/archive_read_support_format_iso9660.c
@@ -0,0 +1,2961 @@
+/*-
+ * Copyright (c) 2003-2007 Tim Kientzle
+ * Copyright (c) 2009 Andreas Henriksson <andreas@fatal.se>
+ * Copyright (c) 2009-2011 Michihiro NAKAJIMA
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "archive_platform.h"
+__FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_format_iso9660.c 201246 2009-12-30 05:30:35Z kientzle $");
+
+#ifdef HAVE_ERRNO_H
+#include <errno.h>
+#endif
+/* #include <stdint.h> */ /* See archive_platform.h */
+#include <stdio.h>
+#ifdef HAVE_STDLIB_H
+#include <stdlib.h>
+#endif
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+#include <time.h>
+#ifdef HAVE_ZLIB_H
+#include <zlib.h>
+#endif
+
+#include "archive.h"
+#include "archive_endian.h"
+#include "archive_entry.h"
+#include "archive_private.h"
+#include "archive_read_private.h"
+#include "archive_string.h"
+
+/*
+ * An overview of ISO 9660 format:
+ *
+ * Each disk is laid out as follows:
+ * * 32k reserved for private use
+ * * Volume descriptor table. Each volume descriptor
+ * is 2k and specifies basic format information.
+ * The "Primary Volume Descriptor" (PVD) is defined by the
+ * standard and should always be present; other volume
+ * descriptors include various vendor-specific extensions.
+ * * Files and directories. Each file/dir is specified by
+ * an "extent" (starting sector and length in bytes).
+ * Dirs are just files with directory records packed one
+ * after another. The PVD contains a single dir entry
+ * specifying the location of the root directory. Everything
+ * else follows from there.
+ *
+ * This module works by first reading the volume descriptors, then
+ * building a list of directory entries, sorted by starting
+ * sector. At each step, I look for the earliest dir entry that
+ * hasn't yet been read, seek forward to that location and read
+ * that entry. If it's a dir, I slurp in the new dir entries and
+ * add them to the heap; if it's a regular file, I return the
+ * corresponding archive_entry and wait for the client to request
+ * the file body. This strategy allows us to read most compliant
+ * CDs with a single pass through the data, as required by libarchive.
+ */
+#define LOGICAL_BLOCK_SIZE 2048
+#define SYSTEM_AREA_BLOCK 16
+
+/* Structure of on-disk primary volume descriptor. */
+#define PVD_type_offset 0
+#define PVD_type_size 1
+#define PVD_id_offset (PVD_type_offset + PVD_type_size)
+#define PVD_id_size 5
+#define PVD_version_offset (PVD_id_offset + PVD_id_size)
+#define PVD_version_size 1
+#define PVD_reserved1_offset (PVD_version_offset + PVD_version_size)
+#define PVD_reserved1_size 1
+#define PVD_system_id_offset (PVD_reserved1_offset + PVD_reserved1_size)
+#define PVD_system_id_size 32
+#define PVD_volume_id_offset (PVD_system_id_offset + PVD_system_id_size)
+#define PVD_volume_id_size 32
+#define PVD_reserved2_offset (PVD_volume_id_offset + PVD_volume_id_size)
+#define PVD_reserved2_size 8
+#define PVD_volume_space_size_offset (PVD_reserved2_offset + PVD_reserved2_size)
+#define PVD_volume_space_size_size 8
+#define PVD_reserved3_offset (PVD_volume_space_size_offset + PVD_volume_space_size_size)
+#define PVD_reserved3_size 32
+#define PVD_volume_set_size_offset (PVD_reserved3_offset + PVD_reserved3_size)
+#define PVD_volume_set_size_size 4
+#define PVD_volume_sequence_number_offset (PVD_volume_set_size_offset + PVD_volume_set_size_size)
+#define PVD_volume_sequence_number_size 4
+#define PVD_logical_block_size_offset (PVD_volume_sequence_number_offset + PVD_volume_sequence_number_size)
+#define PVD_logical_block_size_size 4
+#define PVD_path_table_size_offset (PVD_logical_block_size_offset + PVD_logical_block_size_size)
+#define PVD_path_table_size_size 8
+#define PVD_type_1_path_table_offset (PVD_path_table_size_offset + PVD_path_table_size_size)
+#define PVD_type_1_path_table_size 4
+#define PVD_opt_type_1_path_table_offset (PVD_type_1_path_table_offset + PVD_type_1_path_table_size)
+#define PVD_opt_type_1_path_table_size 4
+#define PVD_type_m_path_table_offset (PVD_opt_type_1_path_table_offset + PVD_opt_type_1_path_table_size)
+#define PVD_type_m_path_table_size 4
+#define PVD_opt_type_m_path_table_offset (PVD_type_m_path_table_offset + PVD_type_m_path_table_size)
+#define PVD_opt_type_m_path_table_size 4
+#define PVD_root_directory_record_offset (PVD_opt_type_m_path_table_offset + PVD_opt_type_m_path_table_size)
+#define PVD_root_directory_record_size 34
+#define PVD_volume_set_id_offset (PVD_root_directory_record_offset + PVD_root_directory_record_size)
+#define PVD_volume_set_id_size 128
+#define PVD_publisher_id_offset (PVD_volume_set_id_offset + PVD_volume_set_id_size)
+#define PVD_publisher_id_size 128
+#define PVD_preparer_id_offset (PVD_publisher_id_offset + PVD_publisher_id_size)
+#define PVD_preparer_id_size 128
+#define PVD_application_id_offset (PVD_preparer_id_offset + PVD_preparer_id_size)
+#define PVD_application_id_size 128
+#define PVD_copyright_file_id_offset (PVD_application_id_offset + PVD_application_id_size)
+#define PVD_copyright_file_id_size 37
+#define PVD_abstract_file_id_offset (PVD_copyright_file_id_offset + PVD_copyright_file_id_size)
+#define PVD_abstract_file_id_size 37
+#define PVD_bibliographic_file_id_offset (PVD_abstract_file_id_offset + PVD_abstract_file_id_size)
+#define PVD_bibliographic_file_id_size 37
+#define PVD_creation_date_offset (PVD_bibliographic_file_id_offset + PVD_bibliographic_file_id_size)
+#define PVD_creation_date_size 17
+#define PVD_modification_date_offset (PVD_creation_date_offset + PVD_creation_date_size)
+#define PVD_modification_date_size 17
+#define PVD_expiration_date_offset (PVD_modification_date_offset + PVD_modification_date_size)
+#define PVD_expiration_date_size 17
+#define PVD_effective_date_offset (PVD_expiration_date_offset + PVD_expiration_date_size)
+#define PVD_effective_date_size 17
+#define PVD_file_structure_version_offset (PVD_effective_date_offset + PVD_effective_date_size)
+#define PVD_file_structure_version_size 1
+#define PVD_reserved4_offset (PVD_file_structure_version_offset + PVD_file_structure_version_size)
+#define PVD_reserved4_size 1
+#define PVD_application_data_offset (PVD_reserved4_offset + PVD_reserved4_size)
+#define PVD_application_data_size 512
+#define PVD_reserved5_offset (PVD_application_data_offset + PVD_application_data_size)
+#define PVD_reserved5_size (2048 - PVD_reserved5_offset)
+
+/* TODO: It would make future maintenance easier to just hardcode the
+ * above values. In particular, ECMA119 states the offsets as part of
+ * the standard. That would eliminate the need for the following check.*/
+#if PVD_reserved5_offset != 1395
+#error PVD offset and size definitions are wrong.
+#endif
+
+
+/* Structure of optional on-disk supplementary volume descriptor. */
+#define SVD_type_offset 0
+#define SVD_type_size 1
+#define SVD_id_offset (SVD_type_offset + SVD_type_size)
+#define SVD_id_size 5
+#define SVD_version_offset (SVD_id_offset + SVD_id_size)
+#define SVD_version_size 1
+/* ... */
+#define SVD_reserved1_offset 72
+#define SVD_reserved1_size 8
+#define SVD_volume_space_size_offset 80
+#define SVD_volume_space_size_size 8
+#define SVD_escape_sequences_offset (SVD_volume_space_size_offset + SVD_volume_space_size_size)
+#define SVD_escape_sequences_size 32
+/* ... */
+#define SVD_logical_block_size_offset 128
+#define SVD_logical_block_size_size 4
+#define SVD_type_L_path_table_offset 140
+#define SVD_type_M_path_table_offset 148
+/* ... */
+#define SVD_root_directory_record_offset 156
+#define SVD_root_directory_record_size 34
+#define SVD_file_structure_version_offset 881
+#define SVD_reserved2_offset 882
+#define SVD_reserved2_size 1
+#define SVD_reserved3_offset 1395
+#define SVD_reserved3_size 653
+/* ... */
+/* FIXME: validate correctness of last SVD entry offset. */
+
+/* Structure of an on-disk directory record. */
+/* Note: ISO9660 stores each multi-byte integer twice, once in
+ * each byte order. The sizes here are the size of just one
+ * of the two integers. (This is why the offset of a field isn't
+ * the same as the offset+size of the previous field.) */
+#define DR_length_offset 0
+#define DR_length_size 1
+#define DR_ext_attr_length_offset 1
+#define DR_ext_attr_length_size 1
+#define DR_extent_offset 2
+#define DR_extent_size 4
+#define DR_size_offset 10
+#define DR_size_size 4
+#define DR_date_offset 18
+#define DR_date_size 7
+#define DR_flags_offset 25
+#define DR_flags_size 1
+#define DR_file_unit_size_offset 26
+#define DR_file_unit_size_size 1
+#define DR_interleave_offset 27
+#define DR_interleave_size 1
+#define DR_volume_sequence_number_offset 28
+#define DR_volume_sequence_number_size 2
+#define DR_name_len_offset 32
+#define DR_name_len_size 1
+#define DR_name_offset 33
+
+#ifdef HAVE_ZLIB_H
+static const unsigned char zisofs_magic[8] = {
+ 0x37, 0xE4, 0x53, 0x96, 0xC9, 0xDB, 0xD6, 0x07
+};
+
+struct zisofs {
+ /* Set 1 if this file compressed by paged zlib */
+ int pz;
+ int pz_log2_bs; /* Log2 of block size */
+ uint64_t pz_uncompressed_size;
+
+ int initialized;
+ unsigned char *uncompressed_buffer;
+ size_t uncompressed_buffer_size;
+
+ uint32_t pz_offset;
+ unsigned char header[16];
+ size_t header_avail;
+ int header_passed;
+ unsigned char *block_pointers;
+ size_t block_pointers_alloc;
+ size_t block_pointers_size;
+ size_t block_pointers_avail;
+ size_t block_off;
+ uint32_t block_avail;
+
+ z_stream stream;
+ int stream_valid;
+};
+#else
+struct zisofs {
+ /* Set 1 if this file compressed by paged zlib */
+ int pz;
+};
+#endif
+
+struct content {
+ uint64_t offset;/* Offset on disk. */
+ uint64_t size; /* File size in bytes. */
+ struct content *next;
+};
+
+/* In-memory storage for a directory record. */
+struct file_info {
+ struct file_info *use_next;
+ struct file_info *parent;
+ struct file_info *next;
+ struct file_info *re_next;
+ int subdirs;
+ uint64_t key; /* Heap Key. */
+ uint64_t offset; /* Offset on disk. */
+ uint64_t size; /* File size in bytes. */
+ uint32_t ce_offset; /* Offset of CE. */
+ uint32_t ce_size; /* Size of CE. */
+ char rr_moved; /* Flag to rr_moved. */
+ char rr_moved_has_re_only;
+ char re; /* Having RRIP "RE" extension. */
+ char re_descendant;
+ uint64_t cl_offset; /* Having RRIP "CL" extension. */
+ int birthtime_is_set;
+ time_t birthtime; /* File created time. */
+ time_t mtime; /* File last modified time. */
+ time_t atime; /* File last accessed time. */
+ time_t ctime; /* File attribute change time. */
+ uint64_t rdev; /* Device number. */
+ mode_t mode;
+ uid_t uid;
+ gid_t gid;
+ int64_t number;
+ int nlinks;
+ struct archive_string name; /* Pathname */
+ char name_continues; /* Non-zero if name continues */
+ struct archive_string symlink;
+ char symlink_continues; /* Non-zero if link continues */
+ /* Set 1 if this file compressed by paged zlib(zisofs) */
+ int pz;
+ int pz_log2_bs; /* Log2 of block size */
+ uint64_t pz_uncompressed_size;
+ /* Set 1 if this file is multi extent. */
+ int multi_extent;
+ struct {
+ struct content *first;
+ struct content **last;
+ } contents;
+ struct {
+ struct file_info *first;
+ struct file_info **last;
+ } rede_files;
+ /* To check a ininity loop. */
+ struct file_info *loop_by;
+};
+
+struct heap_queue {
+ struct file_info **files;
+ int allocated;
+ int used;
+};
+
+struct iso9660 {
+ int magic;
+#define ISO9660_MAGIC 0x96609660
+
+ int opt_support_joliet;
+ int opt_support_rockridge;
+
+ struct archive_string pathname;
+ char seenRockridge; /* Set true if RR extensions are used. */
+ char seenSUSP; /* Set true if SUSP is beging used. */
+ char seenJoliet;
+
+ unsigned char suspOffset;
+ struct file_info *rr_moved;
+ struct read_ce_queue {
+ struct read_ce_req {
+ uint64_t offset;/* Offset of CE on disk. */
+ struct file_info *file;
+ } *reqs;
+ int cnt;
+ int allocated;
+ } read_ce_req;
+
+ int64_t previous_number;
+ struct archive_string previous_pathname;
+
+ struct file_info *use_files;
+ struct heap_queue pending_files;
+ struct {
+ struct file_info *first;
+ struct file_info **last;
+ } cache_files;
+ struct {
+ struct file_info *first;
+ struct file_info **last;
+ } re_files;
+
+ uint64_t current_position;
+ ssize_t logical_block_size;
+ uint64_t volume_size; /* Total size of volume in bytes. */
+ int32_t volume_block;/* Total size of volume in logical blocks. */
+
+ struct vd {
+ int location; /* Location of Extent. */
+ uint32_t size;
+ } primary, joliet;
+
+ off_t entry_sparse_offset;
+ int64_t entry_bytes_remaining;
+ struct zisofs entry_zisofs;
+ struct content *entry_content;
+};
+
+static int archive_read_format_iso9660_bid(struct archive_read *);
+static int archive_read_format_iso9660_options(struct archive_read *,
+ const char *, const char *);
+static int archive_read_format_iso9660_cleanup(struct archive_read *);
+static int archive_read_format_iso9660_read_data(struct archive_read *,
+ const void **, size_t *, off_t *);
+static int archive_read_format_iso9660_read_data_skip(struct archive_read *);
+static int archive_read_format_iso9660_read_header(struct archive_read *,
+ struct archive_entry *);
+static const char *build_pathname(struct archive_string *, struct file_info *);
+#if DEBUG
+static void dump_isodirrec(FILE *, const unsigned char *isodirrec);
+#endif
+static time_t time_from_tm(struct tm *);
+static time_t isodate17(const unsigned char *);
+static time_t isodate7(const unsigned char *);
+static int isBootRecord(struct iso9660 *, const unsigned char *);
+static int isVolumePartition(struct iso9660 *, const unsigned char *);
+static int isVDSetTerminator(struct iso9660 *, const unsigned char *);
+static int isJolietSVD(struct iso9660 *, const unsigned char *);
+static int isSVD(struct iso9660 *, const unsigned char *);
+static int isEVD(struct iso9660 *, const unsigned char *);
+static int isPVD(struct iso9660 *, const unsigned char *);
+static int next_cache_entry(struct archive_read *, struct iso9660 *,
+ struct file_info **);
+static int next_entry_seek(struct archive_read *a, struct iso9660 *iso9660,
+ struct file_info **pfile);
+static struct file_info *
+ parse_file_info(struct archive_read *a,
+ struct file_info *parent, const unsigned char *isodirrec);
+static int parse_rockridge(struct archive_read *a,
+ struct file_info *file, const unsigned char *start,
+ const unsigned char *end);
+static int register_CE(struct archive_read *a, int32_t location,
+ struct file_info *file);
+static int read_CE(struct archive_read *a, struct iso9660 *iso9660);
+static void parse_rockridge_NM1(struct file_info *,
+ const unsigned char *, int);
+static void parse_rockridge_SL1(struct file_info *,
+ const unsigned char *, int);
+static void parse_rockridge_TF1(struct file_info *,
+ const unsigned char *, int);
+static void parse_rockridge_ZF1(struct file_info *,
+ const unsigned char *, int);
+static void register_file(struct iso9660 *, struct file_info *);
+static void release_files(struct iso9660 *);
+static unsigned toi(const void *p, int n);
+static inline void re_add_entry(struct iso9660 *, struct file_info *);
+static inline struct file_info * re_get_entry(struct iso9660 *);
+static inline int rede_add_entry(struct file_info *);
+static inline struct file_info * rede_get_entry(struct file_info *);
+static inline void cache_add_entry(struct iso9660 *iso9660,
+ struct file_info *file);
+static inline struct file_info *cache_get_entry(struct iso9660 *iso9660);
+static void heap_add_entry(struct heap_queue *heap,
+ struct file_info *file, uint64_t key);
+static struct file_info *heap_get_entry(struct heap_queue *heap);
+
+#define add_entry(iso9660, file) \
+ heap_add_entry(&((iso9660)->pending_files), file, file->offset)
+#define next_entry(iso9660) \
+ heap_get_entry(&((iso9660)->pending_files))
+
+int
+archive_read_support_format_iso9660(struct archive *_a)
+{
+ struct archive_read *a = (struct archive_read *)_a;
+ struct iso9660 *iso9660;
+ int r;
+
+ iso9660 = (struct iso9660 *)malloc(sizeof(*iso9660));
+ if (iso9660 == NULL) {
+ archive_set_error(&a->archive, ENOMEM, "Can't allocate iso9660 data");
+ return (ARCHIVE_FATAL);
+ }
+ memset(iso9660, 0, sizeof(*iso9660));
+ iso9660->magic = ISO9660_MAGIC;
+ iso9660->cache_files.first = NULL;
+ iso9660->cache_files.last = &(iso9660->cache_files.first);
+ iso9660->re_files.first = NULL;
+ iso9660->re_files.last = &(iso9660->re_files.first);
+ /* Enable to support Joliet extensions by default. */
+ iso9660->opt_support_joliet = 1;
+ /* Enable to support Rock Ridge extensions by default. */
+ iso9660->opt_support_rockridge = 1;
+
+ r = __archive_read_register_format(a,
+ iso9660,
+ "iso9660",
+ archive_read_format_iso9660_bid,
+ archive_read_format_iso9660_options,
+ archive_read_format_iso9660_read_header,
+ archive_read_format_iso9660_read_data,
+ archive_read_format_iso9660_read_data_skip,
+ archive_read_format_iso9660_cleanup);
+
+ if (r != ARCHIVE_OK) {
+ free(iso9660);
+ return (r);
+ }
+ return (ARCHIVE_OK);
+}
+
+
+static int
+archive_read_format_iso9660_bid(struct archive_read *a)
+{
+ struct iso9660 *iso9660;
+ ssize_t bytes_read;
+ const void *h;
+ const unsigned char *p;
+ int seenTerminator;
+
+ iso9660 = (struct iso9660 *)(a->format->data);
+
+ /*
+ * Skip the first 32k (reserved area) and get the first
+ * 8 sectors of the volume descriptor table. Of course,
+ * if the I/O layer gives us more, we'll take it.
+ */
+#define RESERVED_AREA (SYSTEM_AREA_BLOCK * LOGICAL_BLOCK_SIZE)
+ h = __archive_read_ahead(a,
+ RESERVED_AREA + 8 * LOGICAL_BLOCK_SIZE,
+ &bytes_read);
+ if (h == NULL)
+ return (-1);
+ p = (const unsigned char *)h;
+
+ /* Skip the reserved area. */
+ bytes_read -= RESERVED_AREA;
+ p += RESERVED_AREA;
+
+ /* Check each volume descriptor. */
+ seenTerminator = 0;
+ for (; bytes_read > LOGICAL_BLOCK_SIZE;
+ bytes_read -= LOGICAL_BLOCK_SIZE, p += LOGICAL_BLOCK_SIZE) {
+ /* Do not handle undefined Volume Descriptor Type. */
+ if (p[0] >= 4 && p[0] <= 254)
+ return (0);
+ /* Standard Identifier must be "CD001" */
+ if (memcmp(p + 1, "CD001", 5) != 0)
+ return (0);
+ if (!iso9660->primary.location) {
+ if (isPVD(iso9660, p))
+ continue;
+ }
+ if (!iso9660->joliet.location) {
+ if (isJolietSVD(iso9660, p))
+ continue;
+ }
+ if (isBootRecord(iso9660, p))
+ continue;
+ if (isEVD(iso9660, p))
+ continue;
+ if (isSVD(iso9660, p))
+ continue;
+ if (isVolumePartition(iso9660, p))
+ continue;
+ if (isVDSetTerminator(iso9660, p)) {
+ seenTerminator = 1;
+ break;
+ }
+ return (0);
+ }
+ /*
+ * ISO 9660 format must have Primary Volume Descriptor and
+ * Volume Descriptor Set Terminator.
+ */
+ if (seenTerminator && iso9660->primary.location > 16)
+ return (48);
+
+ /* We didn't find a valid PVD; return a bid of zero. */
+ return (0);
+}
+
+static int
+archive_read_format_iso9660_options(struct archive_read *a,
+ const char *key, const char *val)
+{
+ struct iso9660 *iso9660;
+
+ iso9660 = (struct iso9660 *)(a->format->data);
+
+ if (strcmp(key, "joliet") == 0) {
+ if (val == NULL || strcmp(val, "off") == 0 ||
+ strcmp(val, "ignore") == 0 ||
+ strcmp(val, "disable") == 0 ||
+ strcmp(val, "0") == 0)
+ iso9660->opt_support_joliet = 0;
+ else
+ iso9660->opt_support_joliet = 1;
+ return (ARCHIVE_OK);
+ }
+ if (strcmp(key, "rockridge") == 0 ||
+ strcmp(key, "Rockridge") == 0) {
+ iso9660->opt_support_rockridge = val != NULL;
+ return (ARCHIVE_OK);
+ }
+
+ /* Note: The "warn" return is just to inform the options
+ * supervisor that we didn't handle it. It will generate
+ * a suitable error if noone used this option. */
+ return (ARCHIVE_WARN);
+}
+
+static int
+isBootRecord(struct iso9660 *iso9660, const unsigned char *h)
+{
+ (void)iso9660; /* UNUSED */
+
+ /* Type of the Volume Descriptor Boot Record must be 0. */
+ if (h[0] != 0)
+ return (0);
+
+ /* Volume Descriptor Version must be 1. */
+ if (h[6] != 1)
+ return (0);
+
+ return (1);
+}
+
+static int
+isVolumePartition(struct iso9660 *iso9660, const unsigned char *h)
+{
+ int32_t location;
+
+ /* Type of the Volume Partition Descriptor must be 3. */
+ if (h[0] != 3)
+ return (0);
+
+ /* Volume Descriptor Version must be 1. */
+ if (h[6] != 1)
+ return (0);
+ /* Unused Field */
+ if (h[7] != 0)
+ return (0);
+
+ location = archive_le32dec(h + 72);
+ if (location <= SYSTEM_AREA_BLOCK ||
+ location >= iso9660->volume_block)
+ return (0);
+ if ((uint32_t)location != archive_be32dec(h + 76))
+ return (0);
+
+ return (1);
+}
+
+static int
+isVDSetTerminator(struct iso9660 *iso9660, const unsigned char *h)
+{
+ int i;
+
+ (void)iso9660; /* UNUSED */
+
+ /* Type of the Volume Descriptor Set Terminator must be 255. */
+ if (h[0] != 255)
+ return (0);
+
+ /* Volume Descriptor Version must be 1. */
+ if (h[6] != 1)
+ return (0);
+
+ /* Reserved field must be 0. */
+ for (i = 7; i < 2048; ++i)
+ if (h[i] != 0)
+ return (0);
+
+ return (1);
+}
+
+static int
+isJolietSVD(struct iso9660 *iso9660, const unsigned char *h)
+{
+ const unsigned char *p;
+ ssize_t logical_block_size;
+ int32_t volume_block;
+
+ /* Check if current sector is a kind of Supplementary Volume
+ * Descriptor. */
+ if (!isSVD(iso9660, h))
+ return (0);
+
+ /* FIXME: do more validations according to joliet spec. */
+
+ /* check if this SVD contains joliet extension! */
+ p = h + SVD_escape_sequences_offset;
+ /* N.B. Joliet spec says p[1] == '\\', but.... */
+ if (p[0] == '%' && p[1] == '/') {
+ int level = 0;
+
+ if (p[2] == '@')
+ level = 1;
+ else if (p[2] == 'C')
+ level = 2;
+ else if (p[2] == 'E')
+ level = 3;
+ else /* not joliet */
+ return (0);
+
+ iso9660->seenJoliet = level;
+
+ } else /* not joliet */
+ return (0);
+
+ logical_block_size =
+ archive_le16dec(h + SVD_logical_block_size_offset);
+ volume_block = archive_le32dec(h + SVD_volume_space_size_offset);
+
+ iso9660->logical_block_size = logical_block_size;
+ iso9660->volume_block = volume_block;
+ iso9660->volume_size = logical_block_size * (uint64_t)volume_block;
+ /* Read Root Directory Record in Volume Descriptor. */
+ p = h + SVD_root_directory_record_offset;
+ iso9660->joliet.location = archive_le32dec(p + DR_extent_offset);
+ iso9660->joliet.size = archive_le32dec(p + DR_size_offset);
+
+ return (48);
+}
+
+static int
+isSVD(struct iso9660 *iso9660, const unsigned char *h)
+{
+ const unsigned char *p;
+ ssize_t logical_block_size;
+ int32_t volume_block;
+ int32_t location;
+ int i;
+
+ (void)iso9660; /* UNUSED */
+
+ /* Type 2 means it's a SVD. */
+ if (h[SVD_type_offset] != 2)
+ return (0);
+
+ /* Reserved field must be 0. */
+ for (i = 0; i < SVD_reserved1_size; ++i)
+ if (h[SVD_reserved1_offset + i] != 0)
+ return (0);
+ for (i = 0; i < SVD_reserved2_size; ++i)
+ if (h[SVD_reserved2_offset + i] != 0)
+ return (0);
+ for (i = 0; i < SVD_reserved3_size; ++i)
+ if (h[SVD_reserved3_offset + i] != 0)
+ return (0);
+
+ /* File structure version must be 1 for ISO9660/ECMA119. */
+ if (h[SVD_file_structure_version_offset] != 1)
+ return (0);
+
+ logical_block_size =
+ archive_le16dec(h + SVD_logical_block_size_offset);
+ if (logical_block_size <= 0)
+ return (0);
+
+ volume_block = archive_le32dec(h + SVD_volume_space_size_offset);
+ if (volume_block <= SYSTEM_AREA_BLOCK+4)
+ return (0);
+
+ /* Location of Occurrence of Type L Path Table must be
+ * available location,
+ * >= SYSTEM_AREA_BLOCK(16) + 2 and < Volume Space Size. */