X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flibpspp%2Fi18n.c;h=6cb721a993df60ecd9fc3d4707a9d7b2d19dbeb4;hb=c555358da9c0b1747b9b236ba28ec55efdc89852;hp=e08ba2804bbf4d6b0dbe7c6cf81b820c7a239462;hpb=1221be814f2f1d67c49849512e0938b059a58009;p=pspp

diff --git a/src/libpspp/i18n.c b/src/libpspp/i18n.c
index e08ba2804b..6cb721a993 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
@@ -15,36 +15,34 @@
    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 <libintl.h>
-#include <iconv.h>
-#include <errno.h>
-#include <relocatable.h>
-#include "assertion.h"
-#include "hmapx.h"
-#include "hash-functions.h"
-#include "pool.h"
-
-#include "i18n.h"
 
-#include "version.h"
+#include "libpspp/assertion.h"
+#include "libpspp/hmapx.h"
+#include "libpspp/hash-functions.h"
+#include "libpspp/pool.h"
+#include "libpspp/version.h"
 
-#include <localcharset.h>
-#include "xstrndup.h"
-
-#if HAVE_NL_LANGINFO
-#include <langinfo.h>
-#endif
+#include "gl/localcharset.h"
+#include "gl/xalloc.h"
+#include "gl/relocatable.h"
+#include "gl/xstrndup.h"
 
 struct converter
-  {
-    const char *tocode;
-    const char *fromcode;
+ {
+    char *tocode;
+    char *fromcode;
     iconv_t conv;
   };
 
@@ -73,19 +71,23 @@ create_iconv (const char* tocode, const char* fromcode)
   hmapx_insert (&map, converter, hash);
 
   /* I don't think it's safe to translate this string or to use messaging
-     as the convertors have not yet been set up */
+     as the converters have not yet been set up */
   if ( (iconv_t) -1 == converter->conv && 0 != strcmp (tocode, fromcode))
     {
       const int err = errno;
       fprintf (stderr,
                "Warning: "
-               "cannot create a convertor for \"%s\" to \"%s\": %s\n",
+               "cannot create a converter for `%s' to `%s': %s\n",
                fromcode, tocode, strerror (err));
     }
 
   return converter->conv;
 }
 
+
+/* 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)
@@ -94,26 +96,78 @@ recode_string (const char *to, const char *from,
 }
 
 
-/* Return a string based on TEXT which must be encoded using FROM.
-   The returned string will be encoded in TO.
-   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.
-*/
+/* Uses CONV to convert the INBYTES starting at IP into the OUTBYTES starting
+   at OP, and appends a null terminator to the output.
+
+   Returns true if successful, false if the output buffer is too small. */
+static bool
+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 = '?';
+
+  /* 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 false;
+        *op++ = fallbackchar;
+        *op++ = '\0';
+        return true;
+
+      case EILSEQ:
+        if (outbytes == 0)
+          return false;
+        *op++ = fallbackchar;
+        outbytes--;
+        ip++;
+        inbytes--;
+        break;
+
+      case E2BIG:
+        return false;
+
+      default:
+        /* should never happen */
+        fprintf (stderr, "Character conversion error: %s\n", strerror (errno));
+        NOT_REACHED ();
+        break;
+      }
+
+  if (outbytes == 0)
+    return false;
+
+  *op = '\0';
+  return true;
+}
+
+/* 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)
+                    const char *text, int length, struct pool *pool)
 {
-  char *outbuf = 0;
   size_t outbufferlength;
-  size_t result;
-  char *op ;
-  size_t inbytes = 0;
-  size_t outbytes ;
   iconv_t conv ;
 
-  /* FIXME: Need to ensure that this char is valid in the target encoding */
-  const char fallbackchar = '?';
-
   if ( text == NULL )
     return NULL;
 
@@ -126,76 +180,23 @@ recode_string_pool (const char *to, const char *from,
   if (from == NULL)
     from = default_encoding;
 
-  for ( outbufferlength = 1 ; outbufferlength != 0; outbufferlength <<= 1 )
-    if ( outbufferlength > length)
-      break;
-
-  outbuf = pool_malloc (pool, outbufferlength);
-  op = outbuf;
-
-  outbytes = outbufferlength;
-  inbytes = length;
-
-
   conv = create_iconv (to, from);
 
   if ( (iconv_t) -1 == conv )
-	return xstrdup (text);
-
-  do {
-    const char *ip = text;
-    result = iconv (conv, (ICONV_CONST char **) &text, &inbytes,
-		   &op, &outbytes);
+    return xstrdup (text);
 
-    if ( -1 == result )
+  for ( outbufferlength = 1 ; outbufferlength != 0; outbufferlength <<= 1 )
+    if ( outbufferlength > 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 = pool_malloc (pool, outbufferlength);
-	    op = outbuf;
-	    outbytes = outbufferlength;
-	    inbytes = length;
-	    text = ip;
-	    break;
-	  default:
-	    /* should never happen */
-            fprintf (stderr, "Character conversion error: %s\n", strerror (the_error));
-	    NOT_REACHED ();
-	    break;
-	  }
+        char *output = pool_malloc (pool, outbufferlength);
+        if (try_recode (conv, text, length, output, outbufferlength))
+          return output;
+        pool_free (pool, output);
       }
-  } while ( -1 == result );
-
-  if (outbytes == 0 )
-    {
-      char *const oldaddr = outbuf;
-      outbuf = pool_realloc (pool, outbuf, outbufferlength + 1);
-
-      op += (outbuf - oldaddr) ;
-    }
 
-  *op = '\0';
-
-  return outbuf;
+  NOT_REACHED ();
 }
 
-
 void
 i18n_init (void)
 {
@@ -279,8 +280,11 @@ i18n_done (void)
 {
   struct hmapx_node *node;
   struct converter *cvtr;
+
   HMAPX_FOR_EACH (cvtr, node, &map)
     {
+      free (cvtr->tocode);
+      free (cvtr->fromcode);
       iconv_close (cvtr->conv);
       free (cvtr);
     }