From 721a2add2f11faecac901ca5bf039c65daab1d40 Mon Sep 17 00:00:00 2001
From: Alexandre Julliard <julliard@winehq.org>
Date: Wed, 12 Feb 2020 09:04:59 +0100
Subject: [PATCH] wrc: Use external NLS files for codepage conversions.

Signed-off-by: Alexandre Julliard <julliard@winehq.org>
---
 tools/wrc/utils.c | 122 +++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 104 insertions(+), 18 deletions(-)

diff --git a/tools/wrc/utils.c b/tools/wrc/utils.c
index 8c8a36faf58..78cc09c54b1 100644
--- a/tools/wrc/utils.c
+++ b/tools/wrc/utils.c
@@ -303,23 +303,118 @@ int is_valid_codepage(int id)
     return IsValidCodePage( id );
 }
 
-static int wrc_mbstowcs( int codepage, int flags, const char *src, int srclen, WCHAR *dst, int dstlen )
+static WCHAR *codepage_to_unicode( int codepage, const char *src, int srclen, int *dstlen )
 {
-    return MultiByteToWideChar( codepage, flags, src, srclen, dst, dstlen );
+    WCHAR *dst = xmalloc( (srclen + 1) * sizeof(WCHAR) );
+    DWORD ret = MultiByteToWideChar( codepage, MB_ERR_INVALID_CHARS, src, srclen, dst, srclen );
+    if (!ret) return NULL;
+    dst[ret] = 0;
+    *dstlen = ret;
+    return dst;
 }
 
 #else  /* _WIN32 */
 
-#include "wine/unicode.h"
+struct nls_info
+{
+    unsigned short  codepage;
+    unsigned short  unidef;
+    unsigned short  trans_unidef;
+    unsigned short *cp2uni;
+    unsigned short *dbcs_offsets;
+};
+
+static struct nls_info nlsinfo[128];
+
+static void init_nls_info( struct nls_info *info, unsigned short *ptr )
+{
+    unsigned short hdr_size = ptr[0];
+
+    info->codepage      = ptr[1];
+    info->unidef        = ptr[4];
+    info->trans_unidef  = ptr[6];
+    ptr += hdr_size;
+    info->cp2uni = ++ptr;
+    ptr += 256;
+    if (*ptr++) ptr += 256;  /* glyph table */
+    info->dbcs_offsets  = *ptr ? ptr + 1 : NULL;
+}
+
+static const struct nls_info *get_nls_info( unsigned int codepage )
+{
+    struct stat st;
+    unsigned short *data;
+    char *path;
+    unsigned int i;
+    int fd;
+
+    for (i = 0; i < ARRAY_SIZE(nlsinfo) && nlsinfo[i].codepage; i++)
+        if (nlsinfo[i].codepage == codepage) return &nlsinfo[i];
+
+    assert( i < ARRAY_SIZE(nlsinfo) );
+
+    for (i = 0; nlsdirs[i]; i++)
+    {
+        path = strmake( "%s/c_%03u.nls", nlsdirs[i], codepage );
+        if ((fd = open( path, O_RDONLY )) != -1) break;
+        free( path );
+    }
+    if (!nlsdirs[i]) return NULL;
+
+    fstat( fd, &st );
+    data = xmalloc( st.st_size );
+    if (read( fd, data, st.st_size ) != st.st_size) error( "failed to load %s\n", path );
+    close( fd );
+    free( path );
+    init_nls_info( &nlsinfo[i], data );
+    return &nlsinfo[i];
+}
 
 int is_valid_codepage(int cp)
 {
-    return cp == CP_UTF8 || wine_cp_get_table(cp);
+    return cp == CP_UTF8 || get_nls_info( cp );
 }
 
-static int wrc_mbstowcs( int codepage, int flags, const char *src, int srclen, WCHAR *dst, int dstlen )
+static WCHAR *codepage_to_unicode( int codepage, const char *src, int srclen, int *dstlen )
 {
-    return wine_cp_mbstowcs( wine_cp_get_table( codepage ), flags, src, srclen, dst, dstlen );
+    const struct nls_info *info = get_nls_info( codepage );
+    unsigned int i;
+    WCHAR dbch, *dst = xmalloc( (srclen + 1) * sizeof(WCHAR) );
+
+    if (!info) error( "codepage %u not supported\n", codepage );
+
+    if (info->dbcs_offsets)
+    {
+        for (i = 0; srclen; i++, srclen--, src++)
+        {
+            unsigned short off = info->dbcs_offsets[(unsigned char)*src];
+            if (off)
+            {
+                if (srclen == 1) return NULL;
+                dbch = (src[0] << 8) | (unsigned char)src[1];
+                src++;
+                srclen--;
+                dst[i] = info->dbcs_offsets[off + (unsigned char)*src];
+                if (dst[i] == info->unidef && dbch != info->trans_unidef) return NULL;
+            }
+            else
+            {
+                dst[i] = info->cp2uni[(unsigned char)*src];
+                if (dst[i] == info->unidef && *src != info->trans_unidef) return NULL;
+            }
+        }
+    }
+    else
+    {
+        for (i = 0; i < srclen; i++)
+        {
+            dst[i] = info->cp2uni[(unsigned char)src[i]];
+            if (dst[i] == info->unidef && src[i] != info->trans_unidef) return NULL;
+        }
+    }
+    dst[i] = 0;
+    *dstlen = i;
+    return dst;
 }
 
 #endif  /* _WIN32 */
@@ -448,7 +543,6 @@ static char *unicode_to_utf8( const WCHAR *src, int srclen, int *dstlen )
 string_t *convert_string_unicode( const string_t *str, int codepage )
 {
     string_t *ret = xmalloc(sizeof(*ret));
-    int res;
 
     ret->type = str_unicode;
     ret->loc = str->loc;
@@ -460,16 +554,9 @@ string_t *convert_string_unicode( const string_t *str, int codepage )
         if (codepage == CP_UTF8)
             ret->str.wstr = utf8_to_unicode( str->str.cstr, str->size, &ret->size );
         else
-        {
-            ret->str.wstr = xmalloc( (str->size + 1) * sizeof(WCHAR) );
-            res = wrc_mbstowcs( codepage, MB_ERR_INVALID_CHARS, str->str.cstr, str->size,
-                                ret->str.wstr, str->size );
-            if (res == -2)
-                parser_error( "Invalid character in string '%.*s' for codepage %u",
-                              str->size, str->str.cstr, codepage );
-            ret->size = res;
-            ret->str.wstr[ret->size] = 0;
-        }
+            ret->str.wstr = codepage_to_unicode( codepage, str->str.cstr, str->size, &ret->size );
+        if (!ret->str.wstr) parser_error( "Invalid character in string '%.*s' for codepage %u",
+                                          str->size, str->str.cstr, codepage );
     }
     else
     {
@@ -698,6 +785,5 @@ int get_language_codepage( unsigned short lang, unsigned short sublang )
     }
 
     if (cp == -1) cp = defcp;
-    assert( cp <= 0 || is_valid_codepage(cp) );
     return cp;
 }