add hash functions for locale, latin1, and utf8 strings

* libguile/hash.c (scm_i_locale_string_hash)
  (scm_i_latin1_string_hash, scm_i_utf8_string_hash): New functions.

diff --git a/libguile/hash.c b/libguile/hash.c
index e1591d2..0dcd1c2 100644 (file)
--- a/libguile/hash.c
+++ b/libguile/hash.c
@@ -22,6 +22,12 @@
 # include <config.h>
 #endif
 
+#ifdef HAVE_WCHAR_H
+#include <wchar.h>
+#endif
+
+#include <unistr.h>
+
 #include "libguile/_scm.h"
 #include "libguile/chars.h"
 #include "libguile/ports.h"
@@ -64,6 +70,79 @@ scm_i_string_hash (SCM str)
   return h;
 }
 
+unsigned long 
+scm_i_locale_string_hash (const char *str, size_t len)
+{
+#ifdef HAVE_WCHAR_H
+  mbstate_t state;
+  wchar_t c;
+  size_t byte_idx = 0, nbytes;
+  unsigned long h = 0;
+
+  if (len == (size_t) -1)
+    len = strlen (str);
+
+  while ((nbytes = mbrtowc (&c, str + byte_idx, len - byte_idx, &state)) > 0)
+    {
+      if (nbytes >= (size_t) -2)
+        /* Invalid input string; punt.  */
+        return scm_i_string_hash (scm_from_locale_stringn (str, len));
+
+      h = (unsigned long) c + h * 37;
+      byte_idx += nbytes;
+    }
+
+  return h;
+#else
+  return scm_i_string_hash (scm_from_locale_stringn (str, len));
+#endif
+}
+
+unsigned long 
+scm_i_latin1_string_hash (const char *str, size_t len)
+{
+  const scm_t_uint8 *ustr = (const scm_t_uint8 *) str;
+  size_t i = 0;
+  unsigned long h = 0;
+  
+  if (len == (size_t) -1)
+    len = strlen (str);
+
+  for (; i < len; i++)
+    h = (unsigned long) ustr[i] + h * 37;
+
+  return h;
+}
+
+unsigned long 
+scm_i_utf8_string_hash (const char *str, size_t len)
+{
+  const scm_t_uint8 *ustr = (const scm_t_uint8 *) str;
+  size_t byte_idx = 0;
+  unsigned long h = 0;
+  
+  if (len == (size_t) -1)
+    len = strlen (str);
+
+  while (byte_idx < len)
+    {
+      ucs4_t c;
+      int nbytes;
+
+      nbytes = u8_mbtouc (&c, ustr + byte_idx, len - byte_idx);
+      if (nbytes == 0)
+        break;
+      else if (nbytes < 0)
+        /* Bad UTF-8; punt.  */
+        return scm_i_string_hash (scm_from_utf8_stringn (str, len));
+
+      h = (unsigned long) c + h * 37;
+      byte_idx += nbytes;
+    }
+
+  return h;
+}
+
 
 /* Dirk:FIXME:: why downcase for characters? (2x: scm_hasher, scm_ihashv) */
 /* Dirk:FIXME:: scm_hasher could be made static. */

diff --git a/libguile/hash.h b/libguile/hash.h
index 2ebc053..3077486 100644 (file)
--- a/libguile/hash.h
+++ b/libguile/hash.h
@@ -3,7 +3,7 @@
 #ifndef SCM_HASH_H
 #define SCM_HASH_H
 
-/* Copyright (C) 1995,1996,2000, 2006, 2008 Free Software Foundation, Inc.
+/* Copyright (C) 1995,1996,2000, 2006, 2008, 2011 Free Software Foundation, Inc.
  *
  * This library is free software; you can redistribute it and/or
  * modify it under the terms of the GNU Lesser General Public License
@@ -28,6 +28,13 @@
 \f
 
 SCM_API unsigned long scm_string_hash (const unsigned char *str, size_t len);
+SCM_INTERNAL unsigned long scm_i_locale_string_hash (const char *str,
+                                                     size_t len);
+SCM_INTERNAL unsigned long scm_i_latin1_string_hash (const  char *str,
+                                                     size_t len);
+SCM_INTERNAL unsigned long scm_i_utf8_string_hash (const char *str,
+                                                   size_t len);
+
 SCM_INTERNAL unsigned long scm_i_string_hash (SCM str);
 SCM_API unsigned long scm_hasher (SCM obj, unsigned long n, size_t d);
 SCM_API unsigned long scm_ihashq (SCM obj, unsigned long n);