aboutsummaryrefslogtreecommitdiffstats
path: root/gnu/lib/libg++/g++-include/VOHSet.ccP
diff options
context:
space:
mode:
Diffstat (limited to 'gnu/lib/libg++/g++-include/VOHSet.ccP')
-rw-r--r--gnu/lib/libg++/g++-include/VOHSet.ccP308
1 files changed, 308 insertions, 0 deletions
diff --git a/gnu/lib/libg++/g++-include/VOHSet.ccP b/gnu/lib/libg++/g++-include/VOHSet.ccP
new file mode 100644
index 000000000000..c5d4557a4cb4
--- /dev/null
+++ b/gnu/lib/libg++/g++-include/VOHSet.ccP
@@ -0,0 +1,308 @@
+// This may look like C code, but it is really -*- C++ -*-
+/*
+Copyright (C) 1988 Free Software Foundation
+ written by Doug Lea (dl@rocky.oswego.edu)
+ based on code by Doug Schmidt
+
+This file is part of the GNU C++ Library. This library is free
+software; you can redistribute it and/or modify it under the terms of
+the GNU Library General Public License as published by the Free
+Software Foundation; either version 2 of the License, or (at your
+option) any later version. This library is distributed in the hope
+that it will be useful, but WITHOUT ANY WARRANTY; without even the
+implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR
+PURPOSE. See the GNU Library General Public License for more details.
+You should have received a copy of the GNU Library General Public
+License along with this library; if not, write to the Free Software
+Foundation, 675 Mass Ave, Cambridge, MA 02139, USA.
+*/
+
+#ifdef __GNUG__
+#pragma implementation
+#endif
+#include <stream.h>
+#include "<T>.VOHSet.h"
+
+
+/* codes for status fields */
+#define EMPTYCELL 0
+#define VALIDCELL 1
+#define DELETEDCELL 2
+
+
+<T>VOHSet::<T>VOHSet(int sz)
+{
+// The size of the hash table is always the smallest power of 2 >= the size
+// indicated by the user. This allows several optimizations, including
+// the use of actual double hashing and elimination of the mod instruction.
+
+ size = 1;
+ while (size < sz) size <<= 1;
+ tab = new <T>[size];
+ status = new char[size];
+ for (int i = 0; i < size; ++i) status[i] = EMPTYCELL;
+ count = cnt = 0;
+}
+
+<T>VOHSet::<T>VOHSet(<T>VOHSet& a)
+{
+ tab = new <T>[size = a.size];
+ status = new char[size];
+ for (int i = 0; i < size; ++i) status[i] = EMPTYCELL;
+ count = cnt = 0;
+ for (Pix p = a.first(); p; a.next(p)) add(a(p));
+}
+
+Pix <T>VOHSet::seek(<T&> key)
+{
+// Uses ordered double hashing to perform a search of the table.
+// This greatly speeds up the average-case time for an unsuccessful search.
+
+ unsigned hashval = <T>HASH(key);
+
+ // We can avoid the mod operation since size is a power of 2.
+ unsigned h = hashval & (size - 1);
+
+ // The increment must be odd, since all odd numbers are relatively
+ // prime to a power of 2!!
+ unsigned inc = ((((hashval / size) << 1) + 1) & (size - 1));
+
+ // There is always at least 1 empty cell, so this loop is guaranteed to halt!
+ while (status[h] != EMPTYCELL)
+ {
+ int cmp = <T>CMP (key, tab[h]);
+ if (cmp == 0)
+ {
+ if (status[h] == VALIDCELL)
+ return Pix(&tab[h]);
+ else
+ return 0;
+ }
+ else if (cmp > 0)
+ return 0;
+ else
+ h = ((h + inc) & (size - 1));
+ }
+ return 0;
+}
+
+// This adds an item if it doesn't already exist. By performing the initial
+// comparison we assure that the table always contains at least 1 empty
+// spot. This speeds up later searching by a constant factor.
+// The insertion algorithm uses ordered double hashing. See Standish's
+// 1980 ``Data Structure's Techniques'' book for details.
+
+Pix <T>VOHSet::add(<T&> x)
+{
+ if (size <= cnt+1)
+ resize();
+
+ unsigned hashval = <T>HASH(x);
+ unsigned h = hashval & (size - 1);
+
+ if (status[h] != VALIDCELL) // save some work if possible
+ {
+ if (status[h] == EMPTYCELL)
+ cnt++;
+ count++;
+ tab[h] = x;
+ status[h] = VALIDCELL;
+ return Pix(&tab[h]);
+ }
+ int cmp = <T>CMP(x, tab[h]);
+ if (cmp == 0)
+ return Pix(&tab[h]);
+
+ <T> item = x;
+ Pix mypix = 0;
+ unsigned inc = ((((hashval / size) << 1) + 1) & (size - 1));
+
+ for (;;)
+ {
+ if (cmp < 0)
+ {
+ <T> temp = tab[h];
+ tab[h] = item;
+ item = temp;
+ if (mypix == 0) mypix = Pix(&tab[h]);
+ inc = ((((<T>HASH(item) / size) << 1) + 1) & (size - 1));
+ h = ((h + inc) & (size - 1));
+ if (status[h] != EMPTYCELL) cmp = <T>CMP(item, tab[h]);
+ }
+ else
+ h = ((h + inc) & (size - 1));
+ if (status[h] != VALIDCELL)
+ {
+ if (status[h] == EMPTYCELL)
+ cnt++;
+ count++;
+ tab[h] = item;
+ status[h] = VALIDCELL;
+ return (mypix == 0)? Pix(&tab[h]) : mypix;
+ }
+ }
+}
+
+
+void <T>VOHSet::del(<T&> key)
+{
+// This performs a deletion by marking the item's status field.
+// Note that we only decrease the count, *not* the cnt, since this
+// would cause trouble for subsequent steps in the algorithm. See
+// Reingold and Hanson's ``Data Structure's'' book for a justification
+// of this approach.
+
+ unsigned hashval = <T>HASH(key);
+ unsigned h = hashval & (size - 1);
+ unsigned inc = ((((hashval / size) << 1) + 1) & (size - 1));
+
+ while (status[h] != EMPTYCELL)
+ {
+ int cmp = <T>CMP(key, tab[h]);
+ if (cmp > 0)
+ h = ((h + inc) & (size - 1));
+ else if (status[h] == VALIDCELL && cmp == 0)
+ {
+ status[h] = DELETEDCELL;
+ count--;
+ return;
+ }
+ else
+ return;
+ }
+}
+
+void <T>VOHSet::clear()
+{
+ for (int i = 0; i < size; ++i) status[i] = EMPTYCELL;
+ count = cnt = 0;
+}
+
+void <T>VOHSet::resize(int newsize)
+{
+ if (newsize <= count)
+ newsize = count;
+ int s = 1;
+ while (s <= newsize) s <<= 1;
+ newsize = s;
+
+ <T>* oldtab = tab;
+ char* oldstatus = status;
+ int oldsize = size;
+ tab = new <T>[size = newsize];
+ status = new char[size];
+ for (int i = 0; i < size; ++i) status[i] = EMPTYCELL;
+ count = cnt = 0;
+
+ for (i = 0; i < oldsize; ++i) if (oldstatus[i] == VALIDCELL) add(oldtab[i]);
+ delete [] oldtab;
+ delete oldstatus;
+}
+
+Pix <T>VOHSet::first()
+{
+ for (int pos = 0; pos < size; ++pos)
+ if (status[pos] == VALIDCELL) return Pix(&tab[pos]);
+ return 0;
+}
+
+void <T>VOHSet::next(Pix& i)
+{
+ if (i == 0) return;
+ int pos = ((unsigned)i - (unsigned)tab) / sizeof(<T>) + 1;
+ for (; pos < size; ++pos)
+ if (status[pos] == VALIDCELL)
+ {
+ i = Pix(&tab[pos]);
+ return;
+ }
+ i = 0;
+}
+
+
+int <T>VOHSet:: operator == (<T>VOHSet& b)
+{
+ if (count != b.count)
+ return 0;
+ else
+ {
+ for (int i = 0; i < size; ++i)
+ if (status[i] == VALIDCELL && b.seek(tab[i]) == 0)
+ return 0;
+ for (i = 0; i < b.size; ++i)
+ if (b.status[i] == VALIDCELL && seek(b.tab[i]) == 0)
+ return 0;
+ return 1;
+ }
+}
+
+int <T>VOHSet:: operator != (<T>VOHSet& b)
+{
+ return !(*this == b);
+}
+
+int <T>VOHSet::operator <= (<T>VOHSet& b)
+{
+ if (count > b.count)
+ return 0;
+ else
+ {
+ for (int i = 0; i < size; ++i)
+ if (status[i] == VALIDCELL && b.seek(tab[i]) == 0)
+ return 0;
+ return 1;
+ }
+}
+
+void <T>VOHSet::operator |= (<T>VOHSet& b)
+{
+ if (&b == this || b.count == 0)
+ return;
+ for (int i = 0; i < b.size; ++i)
+ if (b.status[i] == VALIDCELL) add(b.tab[i]);
+}
+
+void <T>VOHSet::operator &= (<T>VOHSet& b)
+{
+ if (&b == this || count == 0)
+ return;
+ for (int i = 0; i < size; ++i)
+ {
+ if (status[i] == VALIDCELL && b.seek(tab[i]) == 0)
+ {
+ status[i] = DELETEDCELL;
+ --count;
+ }
+ }
+}
+
+void <T>VOHSet::operator -= (<T>VOHSet& b)
+{
+ for (int i = 0; i < size; ++i)
+ {
+ if (status[i] == VALIDCELL && b.seek(tab[i]) != 0)
+ {
+ status[i] = DELETEDCELL;
+ --count;
+ }
+ }
+}
+
+int <T>VOHSet::OK()
+{
+ int v = tab != 0;
+ v &= status != 0;
+ int n = 0;
+ for (int i = 0; i < size; ++i)
+ {
+ if (status[i] == VALIDCELL) ++n;
+ else if (status[i] != DELETEDCELL && status[i] != EMPTYCELL)
+ v = 0;
+ }
+ v &= n == count;
+ if (!v) error("invariant failure");
+ return v;
+}
+
+
+