aboutsummaryrefslogtreecommitdiffstats
path: root/doc/gperf.texi
blob: e510ac97021fa0ce3b38dcb4f0ae8985ac4e4024 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
\input texinfo @c -*- texinfo -*-
@c %**start of header
@setfilename gperf.info
@settitle Perfect Hash Function Generator
@c @setchapternewpage odd
@c %**end of header

@c some day we should @include version.texi instead of defining
@c these values at hand.
@set UPDATED 26 September 2000
@set EDITION 2.7.2
@set VERSION 2.7.2
@c ---------------------

@c remove the black boxes generated in the GPL appendix.
@finalout

@c Merge functions into the concept index
@syncodeindex fn cp
@c @synindex pg cp

@dircategory Programming Tools
@direntry
* Gperf: (gperf).                Perfect Hash Function Generator.
@end direntry

@ifinfo
This file documents the features of the GNU Perfect Hash Function
Generator @value{VERSION}.

Copyright @copyright{} 1989-2000 Free Software Foundation, Inc.

Permission is granted to make and distribute verbatim copies of this
manual provided the copyright notice and this permission notice are
preserved on all copies.

@ignore
Permission is granted to process this file through TeX and print the
results, provided the printed document carries a copying permission
notice identical to this one except for the removal of this paragraph
(this paragraph not being relevant to the printed manual).

@end ignore

Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided also that the
section entitled ``GNU General Public License'' is included exactly as
in the original, and provided that the entire resulting derived work is
distributed under the terms of a permission notice identical to this
one.

Permission is granted to copy and distribute translations of this manual
into another language, under the above conditions for modified versions,
except that the section entitled ``GNU General Public License'' and this
permission notice may be included in translations approved by the Free
Software Foundation instead of in the original English.

@end ifinfo

@titlepage
@title User's Guide to @code{gperf} @value{VERSION}
@subtitle The GNU Perfect Hash Function Generator
@subtitle Edition @value{EDITION}, @value{UPDATED}
@author Douglas C. Schmidt

@page
@vskip 0pt plus 1filll
Copyright @copyright{} 1989-2000 Free Software Foundation, Inc.


Permission is granted to make and distribute verbatim copies of
this manual provided the copyright notice and this permission notice
are preserved on all copies.

Permission is granted to copy and distribute modified versions of this
manual under the conditions for verbatim copying, provided also that the
section entitled ``GNU General Public License'' is included
exactly as in the original, and provided that the entire resulting
derived work is distributed under the terms of a permission notice
identical to this one.

Permission is granted to copy and distribute translations of this manual
into another language, under the above conditions for modified versions,
except that the section entitled ``GNU General Public License'' may be
included in a translation approved by the author instead of in the
original English.
@end titlepage

@ifinfo
@node Top, Copying, (dir), (dir)
@top Introduction

This manual documents the GNU @code{gperf} perfect hash function generator
utility, focusing on its features and how to use them, and how to report
bugs.

@menu
* Copying::                     GNU @code{gperf} General Public License says
                                how you can copy and share @code{gperf}.
* Contributors::                People who have contributed to @code{gperf}.
* Motivation::                  Static search structures and GNU GPERF.
* Search Structures::           Static search structures and GNU @code{gperf}
* Description::                 High-level discussion of how GPERF functions.
* Options::                     A description of options to the program.
* Bugs::                        Known bugs and limitations with GPERF.
* Projects::                    Things still left to do.
* Implementation::              Implementation Details for GNU GPERF.
* Bibliography::                Material Referenced in this Report.

* Concept Index::               

@detailmenu --- The Detailed Node Listing ---

High-Level Description of GNU @code{gperf}

* Input Format::                Input Format to @code{gperf}
* Output Format::               Output Format for Generated C Code with @code{gperf}
* Binary Strings::              Use of NUL characters

Input Format to @code{gperf}

* Declarations::                @code{struct} Declarations and C Code Inclusion.
* Keywords::                    Format for Keyword Entries.
* Functions::                   Including Additional C Functions.

Invoking @code{gperf}

* Input Details::               Options that affect Interpretation of the Input File
* Output Language::             Specifying the Language for the Output Code
* Output Details::              Fine tuning Details in the Output Code
* Algorithmic Details::         Changing the Algorithms employed by @code{gperf}
* Verbosity::                   Informative Output

@end detailmenu
@end menu

@end ifinfo

@node Copying, Contributors, Top, Top
@unnumbered GNU GENERAL PUBLIC LICENSE
@include gpl.texinfo

@node Contributors, Motivation, Copying, Top
@unnumbered Contributors to GNU @code{gperf} Utility

@itemize @bullet
@item
@cindex Bugs 
The GNU @code{gperf} perfect hash function generator utility was
originally written in GNU C++ by Douglas C. Schmidt.  It is now also
available in a highly-portable ``old-style'' C version.  The general
idea for the perfect hash function generator was inspired by Keith
Bostic's algorithm written in C, and distributed to net.sources around
1984.  The current program is a heavily modified, enhanced, and extended
implementation of Keith's basic idea, created at the University of
California, Irvine.  Bugs, patches, and suggestions should be reported
to both @code{<bug-gnu-utils@@gnu.org>} and
@code{<gperf-bugs@@lists.sourceforge.net>}.

@item
Special thanks is extended to Michael Tiemann and Doug Lea, for
providing a useful compiler, and for giving me a forum to exhibit my
creation.

In addition, Adam de Boor and Nels Olson provided many tips and insights
that greatly helped improve the quality and functionality of @code{gperf}.

@item
A testsuite was added by Bruno Haible. He also rewrote the output
routines for better reliability.
@end itemize

@node Motivation, Search Structures, Contributors, Top
@chapter Introduction

@code{gperf} is a perfect hash function generator written in C++.  It
transforms an @var{n} element user-specified keyword set @var{W} into a
perfect hash function @var{F}.  @var{F} uniquely maps keywords in
@var{W} onto the range 0..@var{k}, where @var{k} >= @var{n}.  If @var{k}
= @var{n} then @var{F} is a @emph{minimal} perfect hash function.
@code{gperf} generates a 0..@var{k} element static lookup table and a
pair of C functions.  These functions determine whether a given
character string @var{s} occurs in @var{W}, using at most one probe into
the lookup table.

@code{gperf} currently generates the reserved keyword recognizer for
lexical analyzers in several production and research compilers and
language processing tools, including GNU C, GNU C++, GNU Pascal, GNU
Modula 3, and GNU indent.  Complete C++ source code for @code{gperf} is
available via anonymous ftp from @code{ftp://ftp.gnu.org/pub/gnu/gperf/}.
A paper describing @code{gperf}'s design and implementation in greater
detail is available in the Second USENIX C++ Conference proceedings.

@node Search Structures, Description, Motivation, Top
@chapter Static search structures and GNU @code{gperf}
@cindex Static search structure

A @dfn{static search structure} is an Abstract Data Type with certain
fundamental operations, e.g., @emph{initialize}, @emph{insert},
and @emph{retrieve}.  Conceptually, all insertions occur before any
retrievals.  In practice, @code{gperf} generates a @code{static} array
containing search set keywords and any associated attributes specified
by the user.  Thus, there is essentially no execution-time cost for the
insertions.  It is a useful data structure for representing @emph{static
search sets}.  Static search sets occur frequently in software system
applications.  Typical static search sets include compiler reserved
words, assembler instruction opcodes, and built-in shell interpreter
commands.  Search set members, called @dfn{keywords}, are inserted into
the structure only once, usually during program initialization, and are
not generally modified at run-time.

Numerous static search structure implementations exist, e.g.,
arrays, linked lists, binary search trees, digital search tries, and
hash tables.  Different approaches offer trade-offs between space
utilization and search time efficiency.  For example, an @var{n} element
sorted array is space efficient, though the average-case time
complexity for retrieval operations using binary search is
proportional to log @var{n}.  Conversely, hash table implementations
often locate a table entry in constant time, but typically impose
additional memory overhead and exhibit poor worst case performance.

@cindex Minimal perfect hash functions
@emph{Minimal perfect hash functions} provide an optimal solution for a
particular class of static search sets.  A minimal perfect hash
function is defined by two properties:

@itemize @bullet
@item 
It allows keyword recognition in a static search set using at most
@emph{one} probe into the hash table.  This represents the ``perfect''
property.
@item 
The actual memory allocated to store the keywords is precisely large
enough for the keyword set, and @emph{no larger}.  This is the
``minimal'' property.
@end itemize

For most applications it is far easier to generate @emph{perfect} hash
functions than @emph{minimal perfect} hash functions.  Moreover,
non-minimal perfect hash functions frequently execute faster than
minimal ones in practice.  This phenomena occurs since searching a
sparse keyword table increases the probability of locating a ``null''
entry, thereby reducing string comparisons.  @code{gperf}'s default
behavior generates @emph{near-minimal} perfect hash functions for
keyword sets.  However, @code{gperf} provides many options that permit
user control over the degree of minimality and perfection.

Static search sets often exhibit relative stability over time.  For
example, Ada's 63 reserved words have remained constant for nearly a
decade.  It is therefore frequently worthwhile to expend concerted
effort building an optimal search structure @emph{once}, if it
subsequently receives heavy use multiple times.  @code{gperf} removes
the drudgery associated with constructing time- and space-efficient
search structures by hand.  It has proven a useful and practical tool
for serious programming projects.  Output from @code{gperf} is currently
used in several production and research compilers, including GNU C, GNU
C++, GNU Pascal, and GNU Modula 3.  The latter two compilers are not yet
part of the official GNU distribution.  Each compiler utilizes
@code{gperf} to automatically generate static search structures that
efficiently identify their respective reserved keywords.

@node Description, Options, Search Structures, Top
@chapter High-Level Description of GNU @code{gperf}

@menu
* Input Format::                Input Format to @code{gperf}
* Output Format::               Output Format for Generated C Code with @code{gperf}
* Binary Strings::              Use of NUL characters
@end menu

The perfect hash function generator @code{gperf} reads a set of
``keywords'' from a @dfn{keyfile} (or from the standard input by
default).  It attempts to derive a perfect hashing function that
recognizes a member of the @dfn{static keyword set} with at most a
single probe into the lookup table.  If @code{gperf} succeeds in
generating such a function it produces a pair of C source code routines
that perform hashing and table lookup recognition.  All generated C code
is directed to the standard output.  Command-line options described
below allow you to modify the input and output format to @code{gperf}.

By default, @code{gperf} attempts to produce time-efficient code, with
less emphasis on efficient space utilization.  However, several options
exist that permit trading-off execution time for storage space and vice
versa.  In particular, expanding the generated table size produces a
sparse search structure, generally yielding faster searches.
Conversely, you can direct @code{gperf} to utilize a C @code{switch}
statement scheme that minimizes data space storage size.  Furthermore,
using a C @code{switch} may actually speed up the keyword retrieval time
somewhat.  Actual results depend on your C compiler, of course.

In general, @code{gperf} assigns values to the characters it is using
for hashing until some set of values gives each keyword a unique value.
A helpful heuristic is that the larger the hash value range, the easier
it is for @code{gperf} to find and generate a perfect hash function.
Experimentation is the key to getting the most from @code{gperf}.

@node Input Format, Output Format, Description, Description
@section Input Format to @code{gperf}
@cindex Format
@cindex Declaration section
@cindex Keywords section
@cindex Functions section
You can control the input keyfile format by varying certain command-line
arguments, in particular the @samp{-t} option.  The input's appearance
is similar to GNU utilities @code{flex} and @code{bison} (or UNIX
utilities @code{lex} and @code{yacc}).  Here's an outline of the general
format:

@example
@group
declarations
%%
keywords
%%
functions
@end group
@end example

@emph{Unlike} @code{flex} or @code{bison}, all sections of
@code{gperf}'s input are optional.  The following sections describe the
input format for each section.

@menu
* Declarations::                @code{struct} Declarations and C Code Inclusion.
* Keywords::                    Format for Keyword Entries.
* Functions::                   Including Additional C Functions.
@end menu

@node Declarations, Keywords, Input Format, Input Format
@subsection @code{struct} Declarations and C Code Inclusion

The keyword input file optionally contains a section for including
arbitrary C declarations and definitions, as well as provisions for
providing a user-supplied @code{struct}.  If the @samp{-t} option
@emph{is} enabled, you @emph{must} provide a C @code{struct} as the last
component in the declaration section from the keyfile file.  The first
field in this struct must be a @code{char *} or @code{const char *}
identifier called @samp{name}, although it is possible to modify this
field's name with the @samp{-K} option described below.

Here is a simple example, using months of the year and their attributes as
input:

@example
@group
struct months @{ char *name; int number; int days; int leap_days; @};
%%
january,   1, 31, 31
february,  2, 28, 29
march,     3, 31, 31
april,     4, 30, 30
may,       5, 31, 31
june,      6, 30, 30
july,      7, 31, 31
august,    8, 31, 31
september, 9, 30, 30
october,  10, 31, 31
november, 11, 30, 30
december, 12, 31, 31
@end group
@end example

@cindex @samp{%%}
Separating the @code{struct} declaration from the list of keywords and
other fields are a pair of consecutive percent signs, @samp{%%},
appearing left justified in the first column, as in the UNIX utility
@code{lex}.

@cindex @samp{%@{}
@cindex @samp{%@}}
Using a syntax similar to GNU utilities @code{flex} and @code{bison}, it
is possible to directly include C source text and comments verbatim into
the generated output file.  This is accomplished by enclosing the region
inside left-justified surrounding @samp{%@{}, @samp{%@}} pairs.  Here is
an input fragment based on the previous example that illustrates this
feature:

@example
@group
%@{
#include <assert.h>
/* This section of code is inserted directly into the output. */
int return_month_days (struct months *months, int is_leap_year);
%@}
struct months @{ char *name; int number; int days; int leap_days; @};
%%
january,   1, 31, 31
february,  2, 28, 29
march,     3, 31, 31
...
@end group
@end example

It is possible to omit the declaration section entirely.  In this case
the keyfile begins directly with the first keyword line, e.g.:

@example
@group
january,   1, 31, 31
february,  2, 28, 29
march,     3, 31, 31
april,     4, 30, 30
...
@end group
@end example

@node Keywords, Functions, Declarations, Input Format
@subsection Format for Keyword Entries

The second keyfile format section contains lines of keywords and any
associated attributes you might supply.  A line beginning with @samp{#}
in the first column is considered a comment.  Everything following the
@samp{#} is ignored, up to and including the following newline.

The first field of each non-comment line is always the key itself.  It
can be given in two ways: as a simple name, i.e., without surrounding
string quotation marks, or as a string enclosed in double-quotes, in
C syntax, possibly with backslash escapes like @code{\"} or @code{\234}
or @code{\xa8}. In either case, it must start right at the beginning
of the line, without leading whitespace.
In this context, a ``field'' is considered to extend up to, but
not include, the first blank, comma, or newline.  Here is a simple
example taken from a partial list of C reserved words:

@example
@group
# These are a few C reserved words, see the c.gperf file 
# for a complete list of ANSI C reserved words.
unsigned
sizeof
switch
signed
if
default
for
while
return
@end group
@end example

Note that unlike @code{flex} or @code{bison} the first @samp{%%} marker
may be elided if the declaration section is empty.

Additional fields may optionally follow the leading keyword.  Fields
should be separated by commas, and terminate at the end of line.  What
these fields mean is entirely up to you; they are used to initialize the
elements of the user-defined @code{struct} provided by you in the
declaration section.  If the @samp{-t} option is @emph{not} enabled
these fields are simply ignored.  All previous examples except the last
one contain keyword attributes.

@node Functions,  , Keywords, Input Format
@subsection Including Additional C Functions

The optional third section also corresponds closely with conventions
found in @code{flex} and @code{bison}.  All text in this section,
starting at the final @samp{%%} and extending to the end of the input
file, is included verbatim into the generated output file.  Naturally,
it is your responsibility to ensure that the code contained in this
section is valid C.

@node Output Format, Binary Strings, Input Format, Description
@section Output Format for Generated C Code with @code{gperf}
@cindex hash table

Several options control how the generated C code appears on the standard 
output.  Two C function are generated.  They are called @code{hash} and 
@code{in_word_set}, although you may modify their names with a command-line 
option.  Both functions require two arguments, a string, @code{char *} 
@var{str}, and a length parameter, @code{int} @var{len}.  Their default 
function prototypes are as follows:

@deftypefun {unsigned int} hash (const char * @var{str}, unsigned int @var{len})
By default, the generated @code{hash} function returns an integer value
created by adding @var{len} to several user-specified @var{str} key
positions indexed into an @dfn{associated values} table stored in a
local static array.  The associated values table is constructed
internally by @code{gperf} and later output as a static local C array
called @samp{hash_table}; its meaning and properties are described below
(@pxref{Implementation}). The relevant key positions are specified via
the @samp{-k} option when running @code{gperf}, as detailed in the
@emph{Options} section below(@pxref{Options}).
@end deftypefun

@deftypefun {} in_word_set (const char * @var{str}, unsigned int @var{len})
If @var{str} is in the keyword set, returns a pointer to that
keyword. More exactly, if the option @samp{-t} was given, it returns
a pointer to the matching keyword's structure. Otherwise it returns
@code{NULL}.
@end deftypefun

If the option @samp{-c} is not used, @var{str} must be a NUL terminated
string of exactly length @var{len}. If @samp{-c} is used, @var{str} must
simply be an array of @var{len} characters and does not need to be NUL
terminated.

The code generated for these two functions is affected by the following
options:

@table @samp
@item -t
@itemx --struct-type
Make use of the user-defined @code{struct}.

@item -S @var{total-switch-statements}
@itemx --switch=@var{total-switch-statements}
@cindex @code{switch}
Generate 1 or more C @code{switch} statement rather than use a large,
(and potentially sparse) static array.  Although the exact time and
space savings of this approach vary according to your C compiler's
degree of optimization, this method often results in smaller and faster
code.
@end table

If the @samp{-t} and @samp{-S} options are omitted, the default action
is to generate a @code{char *} array containing the keys, together with
additional null strings used for padding the array.  By experimenting
with the various input and output options, and timing the resulting C
code, you can determine the best option choices for different keyword
set characteristics.

@node Binary Strings,  , Output Format, Description
@section Use of NUL characters
@cindex NUL

By default, the code generated by @code{gperf} operates on zero
terminated strings, the usual representation of strings in C. This means
that the keywords in the input file must not contain NUL characters,
and the @var{str} argument passed to @code{hash} or @code{in_word_set}
must be NUL terminated and have exactly length @var{len}.

If option @samp{-c} is used, then the @var{str} argument does not need
to be NUL terminated. The code generated by @code{gperf} will only
access the first @var{len}, not @var{len+1}, bytes starting at @var{str}.
However, the keywords in the input file still must not contain NUL
characters.

If option @samp{-l} is used, then the hash table performs binary
comparison. The keywords in the input file may contain NUL characters,
written in string syntax as @code{\000} or @code{\x00}, and the code
generated by @code{gperf} will treat NUL like any other character.
Also, in this case the @samp{-c} option is ignored.

@node Options, Bugs, Description, Top
@chapter Invoking @code{gperf}

There are @emph{many} options to @code{gperf}.  They were added to make
the program more convenient for use with real applications.  ``On-line''
help is readily available via the @samp{-h} option.  Here is the
complete list of options.

@menu
* Input Details::               Options that affect Interpretation of the Input File
* Output Language::             Specifying the Language for the Output Code
* Output Details::              Fine tuning Details in the Output Code
* Algorithmic Details::         Changing the Algorithms employed by @code{gperf}
* Verbosity::                   Informative Output
@end menu

@node Input Details, Output Language, Options, Options
@section Options that affect Interpretation of the Input File

@table @samp
@item -e @var{keyword-delimiter-list}
@itemx --delimiters=@var{keyword-delimiter-list}
@cindex Delimiters
Allows the user to provide a string containing delimiters used to
separate keywords from their attributes.  The default is ",\n".  This
option is essential if you want to use keywords that have embedded
commas or newlines.  One useful trick is to use -e'TAB', where TAB is
the literal tab character.

@item -t
@itemx --struct-type
Allows you to include a @code{struct} type declaration for generated
code.  Any text before a pair of consecutive @samp{%%} is considered
part of the type declaration.  Keywords and additional fields may follow
this, one group of fields per line.  A set of examples for generating
perfect hash tables and functions for Ada, C, C++, Pascal, Modula 2,
Modula 3 and JavaScript reserved words are distributed with this release.
@end table

@node Output Language, Output Details, Input Details, Options
@section Options to specify the Language for the Output Code

@table @samp
@item -L @var{generated-language-name}
@itemx --language=@var{generated-language-name}
Instructs @code{gperf} to generate code in the language specified by the
option's argument.  Languages handled are currently:

@table @samp
@item KR-C
Old-style K&R C. This language is understood by old-style C compilers and
ANSI C compilers, but ANSI C compilers may flag warnings (or even errors)
because of lacking @samp{const}.

@item C
Common C. This language is understood by ANSI C compilers, and also by
old-style C compilers, provided that you @code{#define const} to empty
for compilers which don't know about this keyword.

@item ANSI-C
ANSI C. This language is understood by ANSI C compilers and C++ compilers.

@item C++
C++. This language is understood by C++ compilers.
@end table

The default is C.

@item -a
This option is supported for compatibility with previous releases of
@code{gperf}. It does not do anything.

@item -g
This option is supported for compatibility with previous releases of
@code{gperf}. It does not do anything.
@end table

@node Output Details, Algorithmic Details, Output Language, Options
@section Options for fine tuning Details in the Output Code

@table @samp
@item -K @var{key-name}
@itemx --slot-name=@var{key-name}
@cindex Slot name
This option is only useful when option @samp{-t} has been given.
By default, the program assumes the structure component identifier for
the keyword is @samp{name}.  This option allows an arbitrary choice of
identifier for this component, although it still must occur as the first
field in your supplied @code{struct}.

@item -F @var{initializers}
@itemx --initializer-suffix=@var{initializers}
@cindex Initializers
This option is only useful when option @samp{-t} has been given.
It permits to specify initializers for the structure members following
@var{key name} in empty hash table entries.  The list of initializers
should start with a comma.  By default, the emitted code will
zero-initialize structure members following @var{key name}.

@item -H @var{hash-function-name}
@itemx --hash-fn-name=@var{hash-function-name}
Allows you to specify the name for the generated hash function.  Default
name is @samp{hash}.  This option permits the use of two hash tables in
the same file.

@item -N @var{lookup-function-name}
@itemx --lookup-fn-name=@var{lookup-function-name}
Allows you to specify the name for the generated lookup function.
Default name is @samp{in_word_set}.  This option permits completely
automatic generation of perfect hash functions, especially when multiple
generated hash functions are used in the same application.

@item -Z @var{class-name}
@itemx --class-name=@var{class-name}
@cindex Class name
This option is only useful when option @samp{-L C++} has been given.  It
allows you to specify the name of generated C++ class.  Default name is
@code{Perfect_Hash}.

@item -7
@itemx --seven-bit
This option specifies that all strings that will be passed as arguments
to the generated hash function and the generated lookup function will
solely consist of 7-bit ASCII characters (characters in the range 0..127).
(Note that the ANSI C functions @code{isalnum} and @code{isgraph} do
@emph{not} guarantee that a character is in this range. Only an explicit
test like @samp{c >= 'A' && c <= 'Z'} guarantees this.) This was the
default in versions of @code{gperf} earlier than 2.7; now the default is
to assume 8-bit characters.

@item -c
@itemx --compare-strncmp
Generates C code that uses the @code{strncmp} function to perform
string comparisons.  The default action is to use @code{strcmp}.

@item -C
@itemx --readonly-tables
Makes the contents of all generated lookup tables constant, i.e.,
``readonly''.  Many compilers can generate more efficient code for this
by putting the tables in readonly memory.

@item -E
@itemx  --enum
Define constant values using an enum local to the lookup function rather
than with #defines.  This also means that different lookup functions can
reside in the same file.  Thanks to James Clark @code{<jjc@@ai.mit.edu>}.

@item -I
@itemx --includes
Include the necessary system include file, @code{<string.h>}, at the
beginning of the code.  By default, this is not done; the user must
include this header file himself to allow compilation of the code.

@item -G
@itemx --global
Generate the static table of keywords as a static global variable,
rather than hiding it inside of the lookup function (which is the
default behavior).

@item -W @var{hash-table-array-name}
@itemx --word-array-name=@var{hash-table-array-name}
@cindex Array name
Allows you to specify the name for the generated array containing the
hash table.  Default name is @samp{wordlist}.  This option permits the
use of two hash tables in the same file, even when the option @samp{-G}
is given.

@item -S @var{total-switch-statements}
@itemx --switch=@var{total-switch-statements}
@cindex @code{switch}
Causes the generated C code to use a @code{switch} statement scheme,
rather than an array lookup table.  This can lead to a reduction in both
time and space requirements for some keyfiles.  The argument to this
option determines how many @code{switch} statements are generated. A
value of 1 generates 1 @code{switch} containing all the elements, a
value of 2 generates 2 tables with 1/2 the elements in each
@code{switch}, etc.  This is useful since many C compilers cannot
correctly generate code for large @code{switch} statements. This option
was inspired in part by Keith Bostic's original C program.

@item -T
@itemx --omit-struct-type
Prevents the transfer of the type declaration to the output file.  Use
this option if the type is already defined elsewhere.

@item -p
This option is supported for compatibility with previous releases of
@code{gperf}. It does not do anything.
@end table

@node Algorithmic Details, Verbosity, Output Details, Options
@section Options for changing the Algorithms employed by @code{gperf}

@table @samp
@item -k @var{keys}
@itemx --key-positions=@var{keys}
Allows selection of the character key positions used in the keywords'
hash function. The allowable choices range between 1-126, inclusive.
The positions are separated by commas, e.g., @samp{-k 9,4,13,14};
ranges may be used, e.g., @samp{-k 2-7}; and positions may occur
in any order.  Furthermore, the meta-character '*' causes the generated
hash function to consider @strong{all} character positions in each key,
whereas '$' instructs the hash function to use the ``final character''
of a key (this is the only way to use a character position greater than
126, incidentally).

For instance, the option @samp{-k 1,2,4,6-10,'$'} generates a hash
function that considers positions 1,2,4,6,7,8,9,10, plus the last
character in each key (which may differ for each key, obviously).  Keys
with length less than the indicated key positions work properly, since
selected key positions exceeding the key length are simply not
referenced in the hash function.

@item -l
@itemx --compare-strlen
Compare key lengths before trying a string comparison.  This might cut
down on the number of string comparisons made during the lookup, since
keys with different lengths are never compared via @code{strcmp}.
However, using @samp{-l} might greatly increase the size of the
generated C code if the lookup table range is large (which implies that
the switch option @samp{-S} is not enabled), since the length table
contains as many elements as there are entries in the lookup table.
This option is mandatory for binary comparisons (@pxref{Binary Strings}).

@item -D
@itemx --duplicates
@cindex Duplicates
Handle keywords whose key position sets hash to duplicate values.
Duplicate hash values occur for two reasons:

@itemize @bullet
@item
Since @code{gperf} does not backtrack it is possible for it to process
all your input keywords without finding a unique mapping for each word.
However, frequently only a very small number of duplicates occur, and 
the majority of keys still require one probe into the table.

@item
Sometimes a set of keys may have the same names, but possess different
attributes.  With the -D option @code{gperf} treats all these keys as
part of an equivalence class and generates a perfect hash function with
multiple comparisons for duplicate keys.  It is up to you to completely
disambiguate the keywords by modifying the generated C code.  However,
@code{gperf} helps you out by organizing the output.
@end itemize

Option @samp{-D} is extremely useful for certain large or highly
redundant keyword sets, e.g., assembler instruction opcodes.
Using this option usually means that the generated hash function is no
longer perfect.  On the other hand, it permits @code{gperf} to work on
keyword sets that it otherwise could not handle.

@item -f @var{iteration-amount}
@itemx --fast=@var{iteration-amount}
Generate the perfect hash function ``fast''.  This decreases
@code{gperf}'s running time at the cost of minimizing generated
table-size.  The iteration amount represents the number of times to
iterate when resolving a collision.  `0' means iterate by the number of
keywords.  This option is probably most useful when used in conjunction
with options @samp{-D} and/or @samp{-S} for @emph{large} keyword sets.

@item -i @var{initial-value}
@itemx --initial-asso=@var{initial-value}
Provides an initial @var{value} for the associate values array.  Default
is 0.  Increasing the initial value helps inflate the final table size,
possibly leading to more time efficient keyword lookups.  Note that this
option is not particularly useful when @samp{-S} is used.  Also,
@samp{-i} is overridden when the @samp{-r} option is used.

@item -j @var{jump-value}
@itemx --jump=@var{jump-value}
@cindex Jump value
Affects the ``jump value'', i.e., how far to advance the associated
character value upon collisions.  @var{Jump-value} is rounded up to an
odd number, the default is 5.  If the @var{jump-value} is 0 @code{gperf}
jumps by random amounts.

@item -n
@itemx --no-strlen
Instructs the generator not to include the length of a keyword when
computing its hash value.  This may save a few assembly instructions in
the generated lookup table.

@item -o
@itemx --occurrence-sort
Reorders the keywords by sorting the keywords so that frequently
occuring key position set components appear first.  A second reordering
pass follows so that keys with ``already determined values'' are placed
towards the front of the keylist.  This may decrease the time required
to generate a perfect hash function for many keyword sets, and also
produce more minimal perfect hash functions.  The reason for this is
that the reordering helps prune the search time by handling inevitable
collisions early in the search process.  On the other hand, if the
number of keywords is @emph{very} large using @samp{-o} may
@emph{increase} @code{gperf}'s execution time, since collisions will
begin earlier and continue throughout the remainder of keyword
processing.  See Cichelli's paper from the January 1980 Communications
of the ACM for details.

@item -r
@itemx --random
Utilizes randomness to initialize the associated values table.  This
frequently generates solutions faster than using deterministic
initialization (which starts all associated values at 0).  Furthermore,
using the randomization option generally increases the size of the
table.  If @code{gperf} has difficultly with a certain keyword set try using
@samp{-r} or @samp{-D}.

@item -s @var{size-multiple}
@itemx --size-multiple=@var{size-multiple}
Affects the size of the generated hash table.  The numeric argument for
this option indicates ``how many times larger or smaller'' the maximum
associated value range should be, in relationship to the number of keys.
If the @var{size-multiple} is negative the maximum associated value is
calculated by @emph{dividing} it into the total number of keys.  For
example, a value of 3 means ``allow the maximum associated value to be
about 3 times larger than the number of input keys''.  

Conversely, a value of -3 means ``allow the maximum associated value to
be about 3 times smaller than the number of input keys''.  Negative
values are useful for limiting the overall size of the generated hash
table, though this usually increases the number of duplicate hash
values.

If `generate switch' option @samp{-S} is @emph{not} enabled, the maximum
associated value influences the static array table size, and a larger
table should decrease the time required for an unsuccessful search, at
the expense of extra table space.

The default value is 1, thus the default maximum associated value about
the same size as the number of keys (for efficiency, the maximum
associated value is always rounded up to a power of 2).  The actual
table size may vary somewhat, since this technique is essentially a
heuristic.  In particular, setting this value too high slows down
@code{gperf}'s runtime, since it must search through a much larger range
of values.  Judicious use of the @samp{-f} option helps alleviate this
overhead, however.
@end table

@node Verbosity,  , Algorithmic Details, Options
@section Informative Output

@table @samp
@item -h
@itemx --help
Prints a short summary on the meaning of each program option.  Aborts
further program execution.

@item -v
@itemx --version
Prints out the current version number.

@item -d
@itemx --debug
Enables the debugging option.  This produces verbose diagnostics to
``standard error'' when @code{gperf} is executing.  It is useful both for
maintaining the program and for determining whether a given set of
options is actually speeding up the search for a solution.  Some useful
information is dumped at the end of the program when the @samp{-d}
option is enabled.
@end table

@node Bugs, Projects, Options, Top
@chapter Known Bugs and Limitations with @code{gperf}

The following are some limitations with the current release of
@code{gperf}:

@itemize @bullet
@item
The @code{gperf} utility is tuned to execute quickly, and works quickly
for small to medium size data sets (around 1000 keywords).  It is
extremely useful for maintaining perfect hash functions for compiler
keyword sets.  Several recent enhancements now enable @code{gperf} to
work efficiently on much larger keyword sets (over 15,000 keywords).
When processing large keyword sets it helps greatly to have over 8 megs
of RAM.

However, since @code{gperf} does not backtrack no guaranteed solution
occurs on every run.  On the other hand, it is usually easy to obtain a
solution by varying the option parameters.  In particular, try the
@samp{-r} option, and also try changing the default arguments to the
@samp{-s} and @samp{-j} options.  To @emph{guarantee} a solution, use
the @samp{-D} and @samp{-S} options, although the final results are not
likely to be a @emph{perfect} hash function anymore!  Finally, use the
@samp{-f} option if you want @code{gperf} to generate the perfect hash
function @emph{fast}, with less emphasis on making it minimal.

@item 
The size of the generate static keyword array can get @emph{extremely}
large if the input keyword file is large or if the keywords are quite
similar.  This tends to slow down the compilation of the generated C
code, and @emph{greatly} inflates the object code size.  If this
situation occurs, consider using the @samp{-S} option to reduce data
size, potentially increasing keyword recognition time a negligible
amount.  Since many C compilers cannot correctly generated code for
large switch statements it is important to qualify the @var{-S} option
with an appropriate numerical argument that controls the number of
switch statements generated.

@item 
The maximum number of key positions selected for a given key has an
arbitrary limit of 126.  This restriction should be removed, and if
anyone considers this a problem write me and let me know so I can remove
the constraint.
@end itemize

@node Projects, Implementation, Bugs, Top
@chapter Things Still Left to Do

It should be ``relatively'' easy to replace the current perfect hash
function algorithm with a more exhaustive approach; the perfect hash
module is essential independent from other program modules.  Additional
worthwhile improvements include:

@itemize @bullet
@item 
Make the algorithm more robust.  At present, the program halts with an
error diagnostic if it can't find a direct solution and the @samp{-D}
option is not enabled.  A more comprehensive, albeit computationally
expensive, approach would employ backtracking or enable alternative
options and retry.  It's not clear how helpful this would be, in
general, since most search sets are rather small in practice.

@item 
Another useful extension involves modifying the program to generate
``minimal'' perfect hash functions (under certain circumstances, the
current version can be rather extravagant in the generated table size).
Again, this is mostly of theoretical interest, since a sparse table
often produces faster lookups, and use of the @samp{-S} @code{switch}
option can minimize the data size, at the expense of slightly longer
lookups (note that the gcc compiler generally produces good code for
@code{switch} statements, reducing the need for more complex schemes).

@item
In addition to improving the algorithm, it would also be useful to
generate a C++ class or Ada package as the code output, in addition to
the current C routines.
@end itemize

@node Implementation, Bibliography, Projects, Top
@chapter Implementation Details of GNU @code{gperf}

A paper describing the high-level description of the data structures and
algorithms used to implement @code{gperf} will soon be available.  This
paper is useful not only from a maintenance and enhancement perspective,
but also because they demonstrate several clever and useful programming
techniques, e.g., `Iteration Number' boolean arrays, double
hashing, a ``safe'' and efficient method for reading arbitrarily long
input from a file, and a provably optimal algorithm for simultaneously
determining both the minimum and maximum elements in a list.

@page

@node Bibliography, Concept Index, Implementation, Top
@chapter Bibliography

[1] Chang, C.C.: @i{A Scheme for Constructing Ordered Minimal Perfect
Hashing Functions} Information Sciences 39(1986), 187-195.
       
[2] Cichelli, Richard J. @i{Author's Response to ``On Cichelli's Minimal Perfect Hash
Functions Method''} Communications of the ACM, 23, 12(December 1980), 729.
    
[3] Cichelli, Richard J. @i{Minimal Perfect Hash Functions Made Simple}
Communications of the ACM, 23, 1(January 1980), 17-19.
           
[4] Cook, C. R. and Oldehoeft, R.R. @i{A Letter Oriented Minimal
Perfect Hashing Function} SIGPLAN Notices, 17, 9(September 1982), 18-27.

[5] Cormack, G. V. and Horspool, R. N. S. and Kaiserwerth, M.
@i{Practical Perfect Hashing} Computer Journal, 28, 1(January 1985), 54-58.
    
[6] Jaeschke, G. @i{Reciprocal Hashing: A Method for Generating Minimal
Perfect Hashing Functions} Communications of the ACM, 24, 12(December
1981), 829-833.

[7] Jaeschke, G. and Osterburg, G. @i{On Cichelli's Minimal Perfect
Hash Functions Method} Communications of the ACM, 23, 12(December 1980),
728-729.

[8] Sager, Thomas J. @i{A Polynomial Time Generator for Minimal Perfect
Hash Functions} Communications of the ACM, 28, 5(December 1985), 523-532

[9] Schmidt, Douglas C. @i{GPERF: A Perfect Hash Function Generator}
Second USENIX C++ Conference Proceedings, April 1990.

[10] Sebesta, R.W. and Taylor, M.A. @i{Minimal Perfect Hash Functions
for Reserved Word Lists}  SIGPLAN Notices, 20, 12(September 1985), 47-53.

[11] Sprugnoli, R. @i{Perfect Hashing Functions: A Single Probe
Retrieving Method for Static Sets} Communications of the ACM, 20
11(November 1977), 841-850.

[12] Stallman, Richard M. @i{Using and Porting GNU CC} Free Software Foundation,
1988.

[13] Stroustrup, Bjarne @i{The C++ Programming Language.} Addison-Wesley, 1986.

[14] Tiemann, Michael D. @i{User's Guide to GNU C++} Free Software
Foundation, 1989.

@node Concept Index,  , Bibliography, Top
@unnumbered Concept Index

@printindex cp

@contents
@bye