view src/hash.c @ 934:c925bacdda60

[xemacs-hg @ 2002-07-29 09:21:12 by michaels] 2002-07-17 Marcus Crestani <crestani@informatik.uni-tuebingen.de> Markus Kaltenbach <makalten@informatik.uni-tuebingen.de> Mike Sperber <mike@xemacs.org> configure flag to turn these changes on: --use-kkcc First we added a dumpable flag to lrecord_implementation. It shows, if the object is dumpable and should be processed by the dumper. * lrecord.h (struct lrecord_implementation): added dumpable flag (MAKE_LRECORD_IMPLEMENTATION): fitted the different makro definitions to the new lrecord_implementation and their calls. Then we changed mark_object, that it no longer needs a mark method for those types that have pdump descritions. * alloc.c: (mark_object): If the object has a description, the new mark algorithm is called, and the object is marked according to its description. Otherwise it uses the mark method like before. These procedures mark objects according to their descriptions. They are modeled on the corresponding pdumper procedures. (mark_with_description): (get_indirect_count): (structure_size): (mark_struct_contents): These procedures still call mark_object, this is needed while there are Lisp_Objects without descriptions left. We added pdump descriptions for many Lisp_Objects: * extents.c: extent_auxiliary_description * database.c: database_description * gui.c: gui_item_description * scrollbar.c: scrollbar_instance_description * toolbar.c: toolbar_button_description * event-stream.c: command_builder_description * mule-charset.c: charset_description * device-msw.c: devmode_description * dialog-msw.c: mswindows_dialog_id_description * eldap.c: ldap_description * postgresql.c: pgconn_description pgresult_description * tooltalk.c: tooltalk_message_description tooltalk_pattern_description * ui-gtk.c: emacs_ffi_description emacs_gtk_object_description * events.c: * events.h: * event-stream.c: * event-Xt.c: * event-gtk.c: * event-tty.c: To write a pdump description for Lisp_Event, we converted every struct in the union event to a Lisp_Object. So we created nine new Lisp_Objects: Lisp_Key_Data, Lisp_Button_Data, Lisp_Motion_Data, Lisp_Process_Data, Lisp_Timeout_Data, Lisp_Eval_Data, Lisp_Misc_User_Data, Lisp_Magic_Data, Lisp_Magic_Eval_Data. We also wrote makro selectors and mutators for the fields of the new designed Lisp_Event and added everywhere these new abstractions. We implemented XD_UNION support in (mark_with_description), so we can describe exspecially console/device specific data with XD_UNION. To describe with XD_UNION, we added a field to these objects, which holds the variant type of the object. This field is initialized in the appendant constructor. The variant is an integer, it has also to be described in an description, if XD_UNION is used. XD_UNION is used in following descriptions: * console.c: console_description (get_console_variant): returns the variant (create_console): added variant initialization * console.h (console_variant): the different console types * console-impl.h (struct console): added enum console_variant contype * device.c: device_description (Fmake_device): added variant initialization * device-impl.h (struct device): added enum console_variant devtype * objects.c: image_instance_description font_instance_description (Fmake_color_instance): added variant initialization (Fmake_font_instance): added variant initialization * objects-impl.h (struct Lisp_Color_Instance): added color_instance_type * objects-impl.h (struct Lisp_Font_Instance): added font_instance_type * process.c: process_description (make_process_internal): added variant initialization * process.h (process_variant): the different process types
author michaels
date Mon, 29 Jul 2002 09:21:25 +0000
parents fdefd0186b75
children e22b0213b713
line wrap: on
line source

/* Hash tables.
   Copyright (C) 1992, 1993, 1994 Free Software Foundation, Inc.

This file is part of XEmacs.

XEmacs is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2, or (at your option) any
later version.

XEmacs 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 General Public License
for more details.

You should have received a copy of the GNU General Public License
along with XEmacs; see the file COPYING.  If not, write to
the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.  */

/* Synched up with: Not in FSF. */

#include <config.h>
#include "lisp.h"
#include "hash.h"

#define NULL_ENTRY ((void *) 0xdeadbeef)

#define COMFORTABLE_SIZE(size) (21 * (size) / 16)

#define KEYS_DIFFER_P(old, new, testfun) \
  (((old) != (new)) && (!(testfun) || !(testfun) ((old),(new))))

static void rehash (hentry *harray, struct hash_table *ht, Elemcount size);

Hashcode
memory_hash (const void *xv, Bytecount size)
{
  Hashcode h = 0;
  unsigned const char *x = (unsigned const char *) xv;

  if (!x) return 0;

  while (size--)
    {
      Hashcode g;
      h = (h << 4) + *x++;
      if ((g = h & 0xf0000000) != 0)
	h = (h ^ (g >> 24)) ^ g;
    }

  return h;
}

Hashcode
string_hash (const char *xv)
{
  Hashcode h = 0;
  unsigned const char *x = (unsigned const char *) xv;

  if (!x) return 0;

  while (*x)
    {
      Hashcode g;
      h = (h << 4) + *x++;
      if ((g = h & 0xf0000000) != 0)
	h = (h ^ (g >> 24)) ^ g;
    }

  return h;
}

/* Return a suitable size for a hash table, with at least SIZE slots. */
static Elemcount
hash_table_size (Elemcount requested_size)
{
  /* Return some prime near, but greater than or equal to, SIZE.
     Decades from the time of writing, someone will have a system large
     enough that the list below will be too short... */
  static const Elemcount primes [] =
  {
    19, 29, 41, 59, 79, 107, 149, 197, 263, 347, 457, 599, 787, 1031,
    1361, 1777, 2333, 3037, 3967, 5167, 6719, 8737, 11369, 14783,
    19219, 24989, 32491, 42257, 54941, 71429, 92861, 120721, 156941,
    204047, 265271, 344857, 448321, 582821, 757693, 985003, 1280519,
    1664681, 2164111, 2813353, 3657361, 4754591, 6180989, 8035301,
    10445899, 13579681, 17653589, 22949669, 29834603, 38784989,
    50420551, 65546729, 85210757, 110774011, 144006217, 187208107,
    243370577, 316381771, 411296309, 534685237, 695090819, 903618083,
    1174703521, 1527114613, 1985248999 /* , 2580823717UL, 3355070839UL */
  };
  /* We've heard of binary search. */
  int low, high;
  for (low = 0, high = countof (primes) - 1; high - low > 1;)
    {
      /* Loop Invariant: size < primes [high] */
      int mid = (low + high) / 2;
      if (primes [mid] < requested_size)
	low = mid;
      else
	high = mid;
    }
  return primes [high];
}

const void *
gethash (const void *key, struct hash_table *hash_table, const void **ret_value)
{
  if (!key)
    {
      *ret_value = hash_table->zero_entry;
      return (void *) hash_table->zero_set;
    }
  else
    {
      hentry *harray = hash_table->harray;
      hash_table_test_function test_function = hash_table->test_function;
      Elemcount size = hash_table->size;
      Hashcode hcode_initial =
	hash_table->hash_function ?
	hash_table->hash_function (key) :
	(Hashcode) key;
      Elemcount hcode = (Elemcount) (hcode_initial % size);
      hentry *e = &harray [hcode];
      const void *e_key = e->key;

      if (e_key ?
	  KEYS_DIFFER_P (e_key, key, test_function) :
	  e->contents == NULL_ENTRY)
	{
	  Elemcount h2 = size - 2;
	  Elemcount incr = (Elemcount) (1 + (hcode_initial % h2));
	  do
	    {
	      hcode += incr; if (hcode >= size) hcode -= size;
	      e = &harray [hcode];
	      e_key = e->key;
	    }
	  while (e_key ?
		 KEYS_DIFFER_P (e_key, key, test_function) :
		 e->contents == NULL_ENTRY);
	}

      *ret_value = e->contents;
      return e->key;
    }
}

void
clrhash (struct hash_table *hash_table)
{
  memset (hash_table->harray, 0, sizeof (hentry) * hash_table->size);
  hash_table->zero_entry = 0;
  hash_table->zero_set   = 0;
  hash_table->fullness   = 0;
}

void
free_hash_table (struct hash_table *hash_table)
{
  xfree (hash_table->harray);
  xfree (hash_table);
}

struct hash_table*
make_hash_table (Elemcount size)
{
  struct hash_table *hash_table = xnew_and_zero (struct hash_table);
  hash_table->size = hash_table_size (COMFORTABLE_SIZE (size));
  hash_table->harray = xnew_array (hentry, hash_table->size);
  clrhash (hash_table);
  return hash_table;
}

struct hash_table *
make_general_hash_table (Elemcount size,
			hash_table_hash_function hash_function,
			hash_table_test_function test_function)
{
  struct hash_table* hash_table = make_hash_table (size);
  hash_table->hash_function = hash_function;
  hash_table->test_function = test_function;
  return hash_table;
}

static void
grow_hash_table (struct hash_table *hash_table, Elemcount new_size)
{
  Elemcount old_size   = hash_table->size;
  hentry     *old_harray = hash_table->harray;

  hash_table->size   = hash_table_size (new_size);
  hash_table->harray = xnew_array (hentry, hash_table->size);

  /* do the rehash on the "grown" table */
  {
    long old_zero_set    = hash_table->zero_set;
    void *old_zero_entry = hash_table->zero_entry;
    clrhash (hash_table);
    hash_table->zero_set   = old_zero_set;
    hash_table->zero_entry = old_zero_entry;
    rehash (old_harray, hash_table, old_size);
  }

  xfree (old_harray);
}

void
puthash (const void *key, void *contents, struct hash_table *hash_table)
{
  if (!key)
    {
      hash_table->zero_entry = contents;
      hash_table->zero_set = 1;
    }
  else
    {
      hash_table_test_function test_function = hash_table->test_function;
      Elemcount size = hash_table->size;
      hentry *harray   = hash_table->harray;
      Hashcode hcode_initial =
	hash_table->hash_function ?
	hash_table->hash_function (key) :
	(Hashcode) key;
      Elemcount hcode = (Elemcount) (hcode_initial % size);
      Elemcount h2 = size - 2;
      Elemcount incr = (Elemcount) (1 + (hcode_initial % h2));
      const void *e_key = harray [hcode].key;
      const void *oldcontents;

      if (e_key && KEYS_DIFFER_P (e_key, key, test_function))
	{
	  do
	    {
	      hcode += incr; if (hcode >= size) hcode -= size;
	      e_key = harray [hcode].key;
	    }
	  while (e_key && KEYS_DIFFER_P (e_key, key, test_function));
	}
      oldcontents = harray [hcode].contents;
      harray [hcode].key = key;
      harray [hcode].contents = contents;
      /* If the entry that we used was a deleted entry,
	 check for a non deleted entry of the same key,
	 then delete it. */
      if (!e_key && oldcontents == NULL_ENTRY)
	{
	  hentry *e;

	  do
	    {
	      hcode += incr; if (hcode >= size) hcode -= size;
	      e = &harray [hcode];
	      e_key = e->key;
	    }
	  while (e_key ?
		 KEYS_DIFFER_P (e_key, key, test_function):
		 e->contents == NULL_ENTRY);

	  if (e_key)
	    {
	      e->key = 0;
	      e->contents = NULL_ENTRY;
	    }
	}

      /* only increment the fullness when we used up a new hentry */
      if (!e_key || KEYS_DIFFER_P (e_key, key, test_function))
	{
	  Elemcount comfortable_size = COMFORTABLE_SIZE (++(hash_table->fullness));
	  if (hash_table->size < comfortable_size)
	    grow_hash_table (hash_table, comfortable_size + 1);
	}
    }
}

static void
rehash (hentry *harray, struct hash_table *hash_table, Elemcount size)
{
  hentry *limit = harray + size;
  hentry *e;
  for (e = harray; e < limit; e++)
    {
      if (e->key)
	puthash (e->key, e->contents, hash_table);
    }
}

void
remhash (const void *key, struct hash_table *hash_table)
{
  if (!key)
    {
      hash_table->zero_entry = 0;
      hash_table->zero_set = 0;
    }
  else
    {
      hentry *harray = hash_table->harray;
      hash_table_test_function test_function = hash_table->test_function;
      Elemcount size = hash_table->size;
      Hashcode hcode_initial =
	(hash_table->hash_function) ?
	(hash_table->hash_function (key)) :
	((Hashcode) key);
      Elemcount hcode = (Elemcount) (hcode_initial % size);
      hentry *e = &harray [hcode];
      const void *e_key = e->key;

      if (e_key ?
	  KEYS_DIFFER_P (e_key, key, test_function) :
	  e->contents == NULL_ENTRY)
	{
	  Elemcount h2 = size - 2;
	  Elemcount incr = (Elemcount) (1 + (hcode_initial % h2));
	  do
	    {
	      hcode += incr; if (hcode >= size) hcode -= size;
	      e = &harray [hcode];
	      e_key = e->key;
	    }
	  while (e_key?
		 KEYS_DIFFER_P (e_key, key, test_function):
		 e->contents == NULL_ENTRY);
	}
      if (e_key)
	{
	  e->key = 0;
	  e->contents = NULL_ENTRY;
	  /* Note: you can't do fullness-- here, it breaks the world. */
	}
    }
}

void
maphash (maphash_function mf, struct hash_table *hash_table, void *arg)
{
  hentry *e;
  hentry *limit;

  if (hash_table->zero_set)
    {
      if (mf (0, hash_table->zero_entry, arg))
	return;
    }

  for (e = hash_table->harray, limit = e + hash_table->size; e < limit; e++)
    {
      if (e->key && mf (e->key, e->contents, arg))
	return;
    }
}

void
map_remhash (remhash_predicate predicate, struct hash_table *hash_table, void *arg)
{
  hentry *e;
  hentry *limit;

  if (hash_table->zero_set && predicate (0, hash_table->zero_entry, arg))
    {
      hash_table->zero_set = 0;
      hash_table->zero_entry = 0;
    }

  for (e = hash_table->harray, limit = e + hash_table->size; e < limit; e++)
    if (predicate (e->key, e->contents, arg))
      {
        e->key = 0;
        e->contents = NULL_ENTRY;
      }
}