X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flibpspp%2Fi18n.c;h=59665bee63b6a44ba6910eea288d3ca8b1ae7b12;hb=c0d0b8039a3cd836398a76a0a884d6166ed761e0;hp=1463a4eecf667e43dd8fcbbedfa35f0d527ad4d4;hpb=e23e02f4ab1b3c3160dc084e241a647636e7a6b9;p=pspp-builds.git

diff --git a/src/libpspp/i18n.c b/src/libpspp/i18n.c
index 1463a4ee..59665bee 100644
--- a/src/libpspp/i18n.c
+++ b/src/libpspp/i18n.c
@@ -1,5 +1,5 @@
 /* PSPP - a program for statistical analysis.
-   Copyright (C) 2006, 2009 Free Software Foundation, Inc.
+   Copyright (C) 2006, 2009, 2010 Free Software Foundation, Inc.
 
    This program is free software: you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
@@ -15,26 +15,37 @@
    along with this program.  If not, see <http://www.gnu.org/licenses/>. */
 
 #include <config.h>
-#include <xalloc.h>
+
+#include "libpspp/i18n.h"
+
 #include <assert.h>
+#include <errno.h>
+#include <iconv.h>
+#include <langinfo.h>
+#include <libintl.h>
 #include <locale.h>
-#include <stdlib.h>
 #include <stdio.h>
+#include <stdlib.h>
 #include <string.h>
-#include <iconv.h>
-#include <errno.h>
-#include "assertion.h"
-#include "hmapx.h"
-#include "hash-functions.h"
-
-#include "i18n.h"
 
-#include <localcharset.h>
-#include "xstrndup.h"
-
-#if HAVE_NL_LANGINFO
-#include <langinfo.h>
-#endif
+#include "libpspp/assertion.h"
+#include "libpspp/hmapx.h"
+#include "libpspp/hash-functions.h"
+#include "libpspp/pool.h"
+#include "libpspp/str.h"
+#include "libpspp/version.h"
+
+#include "gl/localcharset.h"
+#include "gl/xalloc.h"
+#include "gl/relocatable.h"
+#include "gl/xstrndup.h"
+
+struct converter
+ {
+    char *tocode;
+    char *fromcode;
+    iconv_t conv;
+  };
 
 static char *default_encoding;
 static struct hmapx map;
@@ -43,69 +54,148 @@ static struct hmapx map;
 static iconv_t
 create_iconv (const char* tocode, const char* fromcode)
 {
-  iconv_t conv;
+  size_t hash;
   struct hmapx_node *node;
-  size_t hash ;
-  char *key = alloca (strlen (tocode) + strlen (fromcode) + 2);
-
-  strcpy (key, tocode);
-  strcat (key, "\n"); /* hopefully no encoding names contain '\n' */
-  strcat (key, fromcode);
-
-  hash = hsh_hash_string (key);
-
-  node = hmapx_first_with_hash (&map, hash);
-
-  if (!node)
+  struct converter *converter;
+  assert (fromcode);
+
+  hash = hash_string (tocode, hash_string (fromcode, 0));
+  HMAPX_FOR_EACH_WITH_HASH (converter, node, hash, &map)
+    if (!strcmp (tocode, converter->tocode)
+        && !strcmp (fromcode, converter->fromcode))
+      return converter->conv;
+
+  converter = xmalloc (sizeof *converter);
+  converter->tocode = xstrdup (tocode);
+  converter->fromcode = xstrdup (fromcode);
+  converter->conv = iconv_open (tocode, fromcode);
+  hmapx_insert (&map, converter, hash);
+
+  /* I don't think it's safe to translate this string or to use messaging
+     as the converters have not yet been set up */
+  if ( (iconv_t) -1 == converter->conv && 0 != strcmp (tocode, fromcode))
     {
-      conv = iconv_open (tocode, fromcode);
-
-      /* I don't think it's safe to translate this string or to use messaging
-	 as the convertors have not yet been set up */
-      if ( (iconv_t) -1 == conv && 0 != strcmp (tocode, fromcode))
-	{
-	  const int err = errno;
-	  fprintf (stderr,
-		   "Warning: "
-		   "cannot create a convertor for \"%s\" to \"%s\": %s\n",
-		   fromcode, tocode, strerror (err));
-	}
-
-      hmapx_insert (&map, conv, hash);
-    }
-  else
-    {
-      conv = hmapx_node_data (node);
+      const int err = errno;
+      fprintf (stderr,
+               "Warning: "
+               "cannot create a converter for `%s' to `%s': %s\n",
+               fromcode, tocode, strerror (err));
     }
 
-  return conv;
+  return converter->conv;
 }
 
-/* Return a string based on TEXT converted according to HOW.
-   If length is not -1, then it must be the number of bytes in TEXT.
-   The returned string must be freed when no longer required.
-*/
+
+/* Similar to recode_string_pool, but allocates the returned value on the heap
+   instead of in a pool.  It is the caller's responsibility to free the
+   returned value. */
 char *
 recode_string (const char *to, const char *from,
 	       const char *text, int length)
 {
-  char *outbuf = 0;
-  size_t outbufferlength;
-  size_t result;
-  char *op ;
-  size_t inbytes = 0;
-  size_t outbytes ;
-  iconv_t conv ;
+  return recode_string_pool (to, from, text, length, NULL);
+}
 
+
+/* Uses CONV to convert the INBYTES starting at IP into the OUTBYTES starting
+   at OP, and appends a null terminator to the output.
+
+   Returns the output length if successful, -1 if the output buffer is too
+   small. */
+static ssize_t
+try_recode (iconv_t conv,
+            const char *ip, size_t inbytes,
+            char *op_, size_t outbytes)
+{
   /* FIXME: Need to ensure that this char is valid in the target encoding */
   const char fallbackchar = '?';
+  char *op = op_;
+
+  /* Put the converter into the initial shift state, in case there was any
+     state information left over from its last usage. */
+  iconv (conv, NULL, 0, NULL, 0);
+
+  while (iconv (conv, (ICONV_CONST char **) &ip, &inbytes,
+                &op, &outbytes) == -1)
+    switch (errno)
+      {
+      case EINVAL:
+        if (outbytes < 2)
+          return -1;
+        *op++ = fallbackchar;
+        *op = '\0';
+        return op - op_;
+
+      case EILSEQ:
+        if (outbytes == 0)
+          return -1;
+        *op++ = fallbackchar;
+        outbytes--;
+        ip++;
+        inbytes--;
+        break;
+
+      case E2BIG:
+        return -1;
+
+      default:
+        /* should never happen */
+        fprintf (stderr, "Character conversion error: %s\n", strerror (errno));
+        NOT_REACHED ();
+        break;
+      }
+
+  if (outbytes == 0)
+    return -1;
+
+  *op = '\0';
+  return op - op_;
+}
+
+/* Converts the string TEXT, which should be encoded in FROM-encoding, to a
+   dynamically allocated string in TO-encoding.  Any characters which cannot be
+   converted will be represented by '?'.
+
+   LENGTH should be the length of the string or -1, if null terminated.
+
+   The returned string will be allocated on POOL.
+
+   This function's behaviour differs from that of g_convert_with_fallback
+   provided by GLib.  The GLib function will fail (returns NULL) if any part of
+   the input string is not valid in the declared input encoding.  This function
+   however perseveres even in the presence of badly encoded input. */
+char *
+recode_string_pool (const char *to, const char *from,
+                    const char *text, int length, struct pool *pool)
+{
+  struct substring out;
 
   if ( text == NULL )
     return NULL;
 
   if ( length == -1 )
-     length = strlen(text);
+     length = strlen (text);
+
+  out = recode_substring_pool (to, from, ss_buffer (text, length), pool);
+  return out.string;
+}
 
+/* Converts the string TEXT, which should be encoded in FROM-encoding, to a
+   dynamically allocated string in TO-encoding.  Any characters which cannot be
+   converted will be represented by '?'.
+
+   The returned string will be null-terminated and allocated on POOL.
+
+   This function's behaviour differs from that of g_convert_with_fallback
+   provided by GLib.  The GLib function will fail (returns NULL) if any part of
+   the input string is not valid in the declared input encoding.  This function
+   however perseveres even in the presence of badly encoded input. */
+struct substring
+recode_substring_pool (const char *to, const char *from,
+                       struct substring text, struct pool *pool)
+{
+  size_t outbufferlength;
+  iconv_t conv ;
 
   if (to == NULL)
     to = default_encoding;
@@ -113,89 +203,116 @@ recode_string (const char *to, const char *from,
   if (from == NULL)
     from = default_encoding;
 
-  for ( outbufferlength = 1 ; outbufferlength != 0; outbufferlength <<= 1 )
-    if ( outbufferlength > length)
-      break;
-
-  outbuf = xmalloc(outbufferlength);
-  op = outbuf;
-
-  outbytes = outbufferlength;
-  inbytes = length;
-
-
   conv = create_iconv (to, from);
 
-  do {
-    const char *ip = text;
-    result = iconv (conv, (ICONV_CONST char **) &text, &inbytes,
-		   &op, &outbytes);
+  if ( (iconv_t) -1 == conv )
+    {
+      struct substring out;
+      ss_alloc_substring_pool (&out, text, pool);
+      return out;
+    }
 
-    if ( -1 == result )
+  for ( outbufferlength = 1 ; outbufferlength != 0; outbufferlength <<= 1 )
+    if ( outbufferlength > text.length)
       {
-	int the_error = errno;
-
-	switch (the_error)
-	  {
-	  case EILSEQ:
-	  case EINVAL:
-	    if ( outbytes > 0 )
-	      {
-		*op++ = fallbackchar;
-		outbytes--;
-		text++;
-		inbytes--;
-		break;
-	      }
-	    /* Fall through */
-	  case E2BIG:
-	    free (outbuf);
-	    outbufferlength <<= 1;
-	    outbuf = xmalloc (outbufferlength);
-	    op = outbuf;
-	    outbytes = outbufferlength;
-	    inbytes = length;
-	    text = ip;
-	    break;
-	  default:
-	    /* should never happen */
-	    NOT_REACHED ();
-	    break;
-	  }
+        char *output = pool_malloc (pool, outbufferlength);
+        ssize_t output_len = try_recode (conv, text.string, text.length,
+                                         output, outbufferlength);
+        if (output_len >= 0)
+          return ss_buffer (output, output_len);
+        pool_free (pool, output);
       }
-  } while ( -1 == result );
 
-  if (outbytes == 0 )
-    {
-      char *const oldaddr = outbuf;
-      outbuf = xrealloc (outbuf, outbufferlength + 1);
+  NOT_REACHED ();
+}
 
-      op += (outbuf - oldaddr) ;
-    }
+void
+i18n_init (void)
+{
+  setlocale (LC_CTYPE, "");
+  setlocale (LC_MESSAGES, "");
+#if HAVE_LC_PAPER
+  setlocale (LC_PAPER, "");
+#endif
+  bindtextdomain (PACKAGE, relocate(locale_dir));
+  textdomain (PACKAGE);
 
-  *op = '\0';
+  assert (default_encoding == NULL);
+  default_encoding = xstrdup (locale_charset ());
 
-  return outbuf;
+  hmapx_init (&map);
 }
 
 
+const char *
+get_default_encoding (void)
+{
+  return default_encoding;
+}
+
 void
-i18n_init (void)
+set_default_encoding (const char *enc)
 {
   free (default_encoding);
-  default_encoding = strdup (locale_charset ());
-
-  hmapx_init (&map);
+  default_encoding = xstrdup (enc);
 }
 
 
+/* Attempts to set the encoding from a locale name
+   returns true if successfull.
+   This function does not (should not!) alter the current locale.
+*/
+bool
+set_encoding_from_locale (const char *loc)
+{
+  bool ok = true;
+  char *c_encoding;
+  char *loc_encoding;
+  char *tmp = xstrdup (setlocale (LC_CTYPE, NULL));
+
+  setlocale (LC_CTYPE, "C");
+  c_encoding = xstrdup (locale_charset ());
+
+  setlocale (LC_CTYPE, loc);
+  loc_encoding = xstrdup (locale_charset ());
+
+
+  if ( 0 == strcmp (loc_encoding, c_encoding))
+    {
+      ok = false;
+    }
+
+
+  setlocale (LC_CTYPE, tmp);
+
+  free (tmp);
+
+  if (ok)
+    {
+      free (default_encoding);
+      default_encoding = loc_encoding;
+    }
+  else
+    free (loc_encoding);
+
+  free (c_encoding);
+
+  return ok;
+}
+
 void
 i18n_done (void)
 {
   struct hmapx_node *node;
-  iconv_t conv;
-  HMAPX_FOR_EACH (conv, node, &map)
-    iconv_close (conv);
+  struct converter *cvtr;
+
+  HMAPX_FOR_EACH (cvtr, node, &map)
+    {
+      free (cvtr->tocode);
+      free (cvtr->fromcode);
+      iconv_close (cvtr->conv);
+      free (cvtr);
+    }
 
   hmapx_destroy (&map);
 
@@ -205,6 +322,19 @@ i18n_done (void)
 
 
 
+bool
+valid_encoding (const char *enc)
+{
+  iconv_t conv = iconv_open (UTF8, enc);
+
+  if ( conv == (iconv_t) -1)
+    return false;
+
+  iconv_close (conv);
+
+  return true;
+}
+
 
 /* Return the system local's idea of the
    decimal seperator character */
@@ -213,7 +343,7 @@ get_system_decimal (void)
 {
   char radix_char;
 
-  char *ol = strdup (setlocale (LC_NUMERIC, NULL));
+  char *ol = xstrdup (setlocale (LC_NUMERIC, NULL));
   setlocale (LC_NUMERIC, "");
 
 #if HAVE_NL_LANGINFO