Fix usage of char2wchar/wchar2char. Changes:

- pg_wchar and wchar_t could have different size, so char2wchar
  doesn't call pg_mb2wchar_with_len to prevent out-of-bound
  memory bug
- make char2wchar/wchar2char symmetric, now they should not be
  called with C-locale because mbstowcs/wcstombs oftenly doesn't
  work correct with C-locale.
- Text parser uses pg_mb2wchar_with_len directly in case of
  C-locale and multibyte encoding

Per bug report by Hiroshi Inoue <inoue@tpf.co.jp> and
following discussion.

Backpatch up to 8.2 when multybyte support was implemented in tsearch.

diff --git a/src/backend/tsearch/wparser_def.c b/src/backend/tsearch/wparser_def.c
index a4143549756b5d25239c3ee83bce77136eb5bda4..8d7cc1b8d504d391314e8b8348933a2a6fc57fcf 100644 (file)
--- a/src/backend/tsearch/wparser_def.c
+++ b/src/backend/tsearch/wparser_def.c
@@ -7,7 +7,7 @@
  *
  *
  * IDENTIFICATION
- *       $PostgreSQL: pgsql/src/backend/tsearch/wparser_def.c,v 1.20 2009/01/15 16:33:59 teodor Exp $
+ *       $PostgreSQL: pgsql/src/backend/tsearch/wparser_def.c,v 1.21 2009/03/02 15:10:09 teodor Exp $
  *
  *-------------------------------------------------------------------------
  */
@@ -240,12 +240,12 @@ typedef struct TParser
        int                     lenstr;                 /* length of mbstring */
 #ifdef USE_WIDE_UPPER_LOWER
        wchar_t    *wstr;                       /* wide character string */
-       int                     lenwstr;                /* length of wsting */
+       pg_wchar   *pgwstr;                     /* wide character string for C-locale */
+       bool            usewide;
 #endif
 
        /* State of parse */
        int                     charmaxlen;
-       bool            usewide;
        TParserPosition *state;
        bool            ignore;
        bool            wanthost;
@@ -299,13 +299,24 @@ TParserInit(char *str, int len)
        if (prs->charmaxlen > 1)
        {
                prs->usewide = true;
-               prs->wstr = (wchar_t *) palloc(sizeof(wchar_t) * (prs->lenstr + 1));
-               prs->lenwstr = char2wchar(prs->wstr, prs->lenstr + 1,
-                                                                 prs->str, prs->lenstr);
+               if ( lc_ctype_is_c() )
+               {
+                       /*
+                        * char2wchar doesn't work for C-locale and
+                        * sizeof(pg_wchar) could be not equal to sizeof(wchar_t)
+                        */
+                       prs->pgwstr = (pg_wchar*) palloc(sizeof(pg_wchar) * (prs->lenstr + 1));
+                       pg_mb2wchar_with_len(prs->str, prs->pgwstr, prs->lenstr);
+               }
+               else
+               {
+                       prs->wstr = (wchar_t *) palloc(sizeof(wchar_t) * (prs->lenstr + 1));
+                       char2wchar(prs->wstr, prs->lenstr + 1, prs->str, prs->lenstr);
+               }
        }
        else
-#endif
                prs->usewide = false;
+#endif
 
        prs->state = newTParserPosition(NULL);
        prs->state->state = TPS_Base;
@@ -331,6 +342,8 @@ TParserClose(TParser *prs)
 #ifdef USE_WIDE_UPPER_LOWER
        if (prs->wstr)
                pfree(prs->wstr);
+       if (prs->pgwstr)
+               pfree(prs->pgwstr);
 #endif
 
        pfree(prs);
@@ -338,10 +351,12 @@ TParserClose(TParser *prs)
 
 /*
  * Character-type support functions, equivalent to is* macros, but
- * working with any possible encodings and locales. Note,
- * that with multibyte encoding and C-locale isw* function may fail
- * or give wrong result. Note 2: multibyte encoding and C-locale
- * often are used for Asian languages
+ * working with any possible encodings and locales. Notes:
+ *  - with multibyte encoding and C-locale isw* function may fail
+ *    or give wrong result. 
+ *  - multibyte encoding and C-locale often are used for 
+ *    Asian languages.
+ *  - if locale is C the we use pgwstr instead of wstr
  */
 
 #ifdef USE_WIDE_UPPER_LOWER
@@ -352,14 +367,14 @@ p_is##type(TParser *prs) {                                                                                                        \
        Assert( prs->state );                                                                                                   \
        if ( prs->usewide )                                                                                                             \
        {                                                                                                                                               \
-               if ( lc_ctype_is_c() )                                                                                          \
-                       return is##type( 0xff & *( prs->wstr + prs->state->poschar) );  \
+               if ( prs->pgwstr )                                                                                                      \
+                       return is##type( 0xff & *( prs->pgwstr + prs->state->poschar) );\
                                                                                                                                                        \
                return isw##type( *(wint_t*)( prs->wstr + prs->state->poschar ) );      \
        }                                                                                                                                               \
                                                                                                                                                        \
        return is##type( *(unsigned char*)( prs->str + prs->state->posbyte ) ); \
-}      \
+}                                                                                                                                                      \
                                                                                                                                                        \
 static int                                                                                                                                     \
 p_isnot##type(TParser *prs) {                                                                                          \
@@ -373,9 +388,9 @@ p_isalnum(TParser *prs)
 
        if (prs->usewide)
        {
-               if (lc_ctype_is_c())
+               if (prs->pgwstr)
                {
-                       unsigned int c = *(prs->wstr + prs->state->poschar);
+                       unsigned int c = *(prs->pgwstr + prs->state->poschar);
 
                        /*
                         * any non-ascii symbol with multibyte encoding with C-locale is
@@ -405,9 +420,9 @@ p_isalpha(TParser *prs)
 
        if (prs->usewide)
        {
-               if (lc_ctype_is_c())
+               if (prs->pgwstr)
                {
-                       unsigned int c = *(prs->wstr + prs->state->poschar);
+                       unsigned int c = *(prs->pgwstr + prs->state->poschar);
 
                        /*
                         * any non-ascii symbol with multibyte encoding with C-locale is

diff --git a/src/backend/utils/mb/mbutils.c b/src/backend/utils/mb/mbutils.c
index c3cf7f5db69f31abed4d7988545f7d6e1034340a..f5ba80d101d5e6f40f4cc9ff772a0261c04f4d39 100644 (file)
--- a/src/backend/utils/mb/mbutils.c
+++ b/src/backend/utils/mb/mbutils.c
@@ -4,7 +4,7 @@
  * (currently mule internal code (mic) is used)
  * Tatsuo Ishii
  *
- * $PostgreSQL: pgsql/src/backend/utils/mb/mbutils.c,v 1.78 2009/01/22 10:09:48 mha Exp $
+ * $PostgreSQL: pgsql/src/backend/utils/mb/mbutils.c,v 1.79 2009/03/02 15:10:09 teodor Exp $
  */
 #include "postgres.h"
 
@@ -601,7 +601,10 @@ wchar2char(char *to, const wchar_t *from, size_t tolen)
        }
        else
 #endif   /* WIN32 */
+       {
+               Assert( !lc_ctype_is_c() );
                result = wcstombs(to, from, tolen);
+       }
        return result;
 }
 
@@ -647,22 +650,12 @@ char2wchar(wchar_t *to, size_t tolen, const char *from, size_t fromlen)
        else
 #endif   /* WIN32 */
        {
-               if (lc_ctype_is_c())
-               {
-                       /*
-                        * pg_mb2wchar_with_len always adds trailing '\0', so 'to' should be
-                        * allocated with sufficient space
-                        */
-                       result = pg_mb2wchar_with_len(from, (pg_wchar *) to, fromlen);
-               }
-               else
-               {
-                       /* mbstowcs requires ending '\0' */
-                       char       *str = pnstrdup(from, fromlen);
+               /* mbstowcs requires ending '\0' */
+               char       *str = pnstrdup(from, fromlen);
 
-                       result = mbstowcs(to, str, tolen);
-                       pfree(str);
-               }
+               Assert( !lc_ctype_is_c() );
+               result = mbstowcs(to, str, tolen);
+               pfree(str);
        }
 
        if (result == -1)