trunk/Src/wptNLS.c

/* wptNLS.cpp - W32 Native Language Support
 *      Copyright (C) 2001, 2002, 2003 Timo Schulz
 *      Copyright (C) 1995-1999 Free Software Foundation, Inc.
 *
 * This code based on the simple-gettext.c code from the GnuPG 
 * by Ulrich Drepper.
 *
 * WinPT is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License 
 * as published by the Free Software Foundation; either version 2 
 * of the License, or (at your option) any later version.
 *  
 * WinPT is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License 
 * along with WinPT; if not, write to the Free Software Foundation,
 * Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA 
 */

#ifdef HAVE_CONFIG_H
#include <config.h>
#endif

#include <stdio.h>
#include <stdio.h>
#include <string.h>
#include <ctype.h>
#include <errno.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <windows.h>

#include "wptTypes.h"
#include "wptNLS.h"

/* Missing W32 functions. */
static char *
w32_stpcpy (char *a,const char *b)
{
    while (*b)
        *a++ = *b++;
    *a = 0;
    return (char*)a;
}


/* The magic number of the GNU message catalog format.  */
#define MAGIC         0x950412de
#define MAGIC_SWAPPED 0xde120495

/* Revision number of the currently used .mo (binary) file format.  */
#define MO_REVISION_NUMBER 0


/* Header for binary .mo file format.  */
struct mo_file_header {
  u32 magic; /* The magic number.       */  
  u32 revision; /* The revision number of the file format.  */
  u32 nstrings; /* The number of strings pairs.  */  
  u32 orig_tab_offset; /* Offset of table with start offsets of original 
                          strings.  */
  u32 trans_tab_offset; /* Offset of table with start offsets of translation
                           strings.  */  
  u32 hash_tab_size; /* Size of hashing table.  */  
  u32 hash_tab_offset; /* Offset of first hashing entry.  */
};

struct string_desc {
  u32 length; /* Length of addressed string.  */  
  u32 offset; /* Offset of string in file.      */
};

struct loaded_domain {
  char *data;
  int must_swap;
  u32 nstrings;
  char *mapped;
  struct string_desc *orig_tab;
  struct string_desc *trans_tab;
  u32 hash_size;
  u32 *hash_tab;
};

static struct loaded_domain *the_domain;

static u32
do_swap_u32( u32 i )
{
  return (i << 24) | ((i & 0xff00) << 8) | ((i >> 8) & 0xff00) | (i >> 24);
}

#define SWAPIT(flag, data) ((flag) ? do_swap_u32(data) : (data) )


/* We assume to have `unsigned long int' value with at least 32 bits.  */
#define HASHWORDBITS 32

/* The so called `hashpjw' function by P.J. Weinberger
   [see Aho/Sethi/Ullman, COMPILERS: Principles, Techniques and Tools,
   1986, 1987 Bell Telephone Laboratories, Inc.]  */

static u32
hash_string( const char *str_param )
{
    unsigned long int hval, g;
    const char *str = str_param;

    hval = 0;
    while (*str != '\0') {
        hval <<= 4;
        hval += (unsigned long int) *str++;     
        g = hval & ((unsigned long int) 0xf << (HASHWORDBITS - 4));
        if (g != 0) {
            hval ^= g >> (HASHWORDBITS - 8);
            hval ^= g;  
        }
    }
    return hval;
} /* hash_string */


static struct loaded_domain *
load_domain( const char *filename )
{
    FILE *fp;
    size_t size;
    struct stat st;
    struct mo_file_header *data = NULL;
    struct loaded_domain *domain = NULL;
    size_t to_read;
    char *read_ptr;

    fp = fopen( filename, "rb" );
    if( !fp )
       return NULL; /* can't open the file */
    /* we must know about the size of the file */
    if( fstat( fileno(fp ), &st )
        || (size = (size_t)st.st_size) != st.st_size
        || size < sizeof (struct mo_file_header) ) {
        fclose( fp );   
        return NULL;
    }

    data = (struct mo_file_header *) malloc( size );
    if( !data ) {
        fclose( fp );
        return NULL; /* out of memory */
    }

    to_read = size;
    read_ptr = (char *) data;
    do {
        long int nb = fread( read_ptr, 1, to_read, fp );
        if( nb < to_read ) {
            fclose (fp);
            free(data); 
            return NULL; /* read error */
                
        }
        read_ptr += nb; 
        to_read -= nb;
    } while( to_read > 0 );
    fclose (fp);

    /* Using the magic number we can test whether it really is a message
     * catalog file.  */
    if( data->magic != MAGIC && data->magic != MAGIC_SWAPPED ) {
        /* The magic number is wrong: not a message catalog file.  */
        free( data );
        return NULL;
    }

    domain = (struct loaded_domain *)calloc( 1, sizeof *domain );
    if( !domain )  {
        free( data );   
        return NULL;
    }
    domain->data = (char *) data;
    domain->must_swap = data->magic != MAGIC;

    /* Fill in the information about the available tables.  */
    switch( SWAPIT(domain->must_swap, data->revision) ) {
      case 0:
          domain->nstrings = SWAPIT(domain->must_swap, data->nstrings);
          domain->orig_tab = (struct string_desc *)
              ((char *) data + SWAPIT(domain->must_swap, data->orig_tab_offset));
          domain->trans_tab = (struct string_desc *)
              ((char *) data + SWAPIT(domain->must_swap, data->trans_tab_offset));
          domain->hash_size = SWAPIT(domain->must_swap, data->hash_tab_size);   
          domain->hash_tab = (u32 *)    
              ((char *) data + SWAPIT(domain->must_swap, data->hash_tab_offset));       
          break;

      default: /* This is an invalid revision.  */
          free( data );
          free( domain );
          return NULL;
    }

    /* allocate an array to keep track of code page mappings */
    domain->mapped = (char *)calloc( 1, domain->nstrings );
    if( !domain->mapped ) {
        free( data );
        free( domain ); 
        return NULL;
    }

    return domain;
} /* load_domain */


/****************
 * Set the file used for translations.  Pass a NULL to disable
 * translation.  A new filename may be set at anytime.
 * WARNING: After changing the filename you shoudl not access any data
 *          retrieved by gettext().
 */
int
set_gettext_file( const char *filename, const char *nls_dir )
{
    struct loaded_domain *domain = NULL;

    if( filename && *filename ) {
        if( filename[0] == '/'     
            || ( isalpha(filename[0])            
            && filename[1] == ':'       
            && (filename[2] == '/' || filename[2] == '\\') )
           ) {
            /* absolute path - use it as is */      
            domain = load_domain( filename );
        }
        else { /* relative path - append ".mo" and get dir from the environment */
            char *buf = NULL;
            char *dir;
                
            dir = strdup( nls_dir );
            if( dir && (buf= (char *)malloc(strlen(dir)+strlen(filename)+1+3+1)) ) {
                strcpy(w32_stpcpy(w32_stpcpy(w32_stpcpy( buf, dir),"/"), filename),".mo");
                domain = load_domain( buf );                    
                free(buf);
            }
            free(dir);
        }
        if( !domain )
            return -1;
    }

    if( the_domain ) {
        free( the_domain->data );
        free( the_domain->mapped );
        free( the_domain );     
        the_domain = NULL;
    }
    the_domain = domain;
    return 0;
} /* set_gettext_file */


static const char*
get_string( struct loaded_domain *domain, u32 idx )
{
    char *p = domain->data + SWAPIT(domain->must_swap,
                                    domain->trans_tab[idx].offset);
    if( !domain->mapped[idx] ) {        
        domain->mapped[idx] = 1;                
    }
    return (const char*)p;
} /* get_string */


const char *
gettext( const char *msgid )
{
    struct loaded_domain *domain;
    size_t act = 0;
    size_t top, bottom;

    if( !(domain = the_domain) )        
        goto not_found;

    /* Locate the MSGID and its translation.  */
    if( domain->hash_size > 2 && domain->hash_tab ) {           
        /* Use the hashing table.  */           
        u32 len = strlen (msgid);       
        u32 hash_val = hash_string (msgid);     
        u32 idx = hash_val % domain->hash_size; 
        u32 incr = 1 + (hash_val % (domain->hash_size - 2));    
        u32 nstr = SWAPIT (domain->must_swap, domain->hash_tab[idx]);

        if ( !nstr ) /* Hash table entry is empty.  */  
            goto not_found;

                
        if( SWAPIT(domain->must_swap,   
            domain->orig_tab[nstr - 1].length) == len   
            && !strcmp( msgid,  
            domain->data + SWAPIT(domain->must_swap,
                                    domain->orig_tab[nstr - 1].offset)) )
                                    return get_string( domain, nstr - 1 );
        for(;;) {
            if (idx >= domain->hash_size - incr)
                idx -= domain->hash_size - incr;
            else
                idx += incr;
            nstr = SWAPIT(domain->must_swap, domain->hash_tab[idx]);
            if( !nstr ) 
                goto not_found; /* Hash table entry is empty.  */

            if ( SWAPIT(domain->must_swap,
                                domain->orig_tab[nstr - 1].length) == len
                                && !strcmp (msgid,
                                domain->data + SWAPIT(domain->must_swap,
                                           domain->orig_tab[nstr - 1].offset)))
                        
                                           return get_string( domain, nstr-1 );
        } /* NOTREACHED */
    }

    /* Now we try the default method:  binary search in the sorted
       array of messages.  */
    bottom = 0;
    top = domain->nstrings;
    while( bottom < top ) {             
        int cmp_val;
        
        act = (bottom + top) / 2;       
        cmp_val = strcmp(msgid, domain->data + SWAPIT(domain->must_swap,
                        domain->orig_tab[act].offset));
        if (cmp_val < 0)
            top = act;
        else if (cmp_val > 0)
            bottom = act + 1;
        else    
            return get_string( domain, act );
    }

not_found:
    return msgid;
} /* gettext */

/* Map the user specific language ID to a
   gettext conform language string.
   Example: LANG_GERMAN -> "de" */
const char*
get_gettext_langid (void)
{
    struct {
        const char *id;
        int langid;
    } lang_table[] = {
        {"de", LANG_GERMAN},
        {NULL, 0},
    };
    LANGID lang;
    int i;
    
    lang = GetUserDefaultLangID ();
    
    for (i=0; lang_table[i].id; i++) {
        if (PRIMARYLANGID (lang) == lang_table[i].langid)
            return lang_table[i].id;
    }
    return NULL;
}


/* XXX: this has to be placed in a .c file but not here */
#include <shlobj.h>

void _SHFree (void *p) 
{
    IMalloc *pm; 
    SHGetMalloc(&pm);
    if (pm) {
        pm->lpVtbl->Free(pm,p);
        pm->lpVtbl->Release(pm);
    } 
}
1	/* wptNLS.cpp - W32 Native Language Support
2	* Copyright (C) 2001, 2002, 2003 Timo Schulz
3	* Copyright (C) 1995-1999 Free Software Foundation, Inc.
4	*
5	* This code based on the simple-gettext.c code from the GnuPG
6	* by Ulrich Drepper.
7	*
8	* WinPT is free software; you can redistribute it and/or
9	* modify it under the terms of the GNU General Public License
10	* as published by the Free Software Foundation; either version 2
11	* of the License, or (at your option) any later version.
12	*
13	* WinPT is distributed in the hope that it will be useful,
14	* but WITHOUT ANY WARRANTY; without even the implied warranty of
15	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16	* General Public License for more details.
17	*
18	* You should have received a copy of the GNU General Public License
19	* along with WinPT; if not, write to the Free Software Foundation,
20	* Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
21	*/
22
23	#ifdef HAVE_CONFIG_H
24	#include <config.h>
25	#endif
26
27	#include <stdio.h>
28	#include <stdio.h>
29	#include <string.h>
30	#include <ctype.h>
31	#include <errno.h>
32	#include <sys/types.h>
33	#include <sys/stat.h>
34	#include <windows.h>
35
36	#include "wptTypes.h"
37	#include "wptNLS.h"
38
39	/* Missing W32 functions. */
40	static char *
41	w32_stpcpy (char a,const char b)
42	{
43	while (*b)
44	a++ = b++;
45	*a = 0;
46	return (char*)a;
47	}
48
49
50	/* The magic number of the GNU message catalog format. */
51	#define MAGIC 0x950412de
52	#define MAGIC_SWAPPED 0xde120495
53
54	/* Revision number of the currently used .mo (binary) file format. */
55	#define MO_REVISION_NUMBER 0
56
57
58	/* Header for binary .mo file format. */
59	struct mo_file_header {
60	u32 magic; /* The magic number. */
61	u32 revision; /* The revision number of the file format. */
62	u32 nstrings; /* The number of strings pairs. */
63	u32 orig_tab_offset; /* Offset of table with start offsets of original
64	strings. */
65	u32 trans_tab_offset; /* Offset of table with start offsets of translation
66	strings. */
67	u32 hash_tab_size; /* Size of hashing table. */
68	u32 hash_tab_offset; /* Offset of first hashing entry. */
69	};
70
71	struct string_desc {
72	u32 length; /* Length of addressed string. */
73	u32 offset; /* Offset of string in file. */
74	};
75
76	struct loaded_domain {
77	char *data;
78	int must_swap;
79	u32 nstrings;
80	char *mapped;
81	struct string_desc *orig_tab;
82	struct string_desc *trans_tab;
83	u32 hash_size;
84	u32 *hash_tab;
85	};
86
87	static struct loaded_domain *the_domain;
88
89	static u32
90	do_swap_u32( u32 i )
91	{
92	return (i << 24) \| ((i & 0xff00) << 8) \| ((i >> 8) & 0xff00) \| (i >> 24);
93	}
94
95	#define SWAPIT(flag, data) ((flag) ? do_swap_u32(data) : (data) )
96
97
98	/* We assume to have `unsigned long int' value with at least 32 bits. */
99	#define HASHWORDBITS 32
100
101	/* The so called `hashpjw' function by P.J. Weinberger
102	[see Aho/Sethi/Ullman, COMPILERS: Principles, Techniques and Tools,
103	1986, 1987 Bell Telephone Laboratories, Inc.] */
104
105	static u32
106	hash_string( const char *str_param )
107	{
108	unsigned long int hval, g;
109	const char *str = str_param;
110
111	hval = 0;
112	while (*str != '\0') {
113	hval <<= 4;
114	hval += (unsigned long int) *str++;
115	g = hval & ((unsigned long int) 0xf << (HASHWORDBITS - 4));
116	if (g != 0) {
117	hval ^= g >> (HASHWORDBITS - 8);
118	hval ^= g;
119	}
120	}
121	return hval;
122	} /* hash_string */
123
124
125	static struct loaded_domain *
126	load_domain( const char *filename )
127	{
128	FILE *fp;
129	size_t size;
130	struct stat st;
131	struct mo_file_header *data = NULL;
132	struct loaded_domain *domain = NULL;
133	size_t to_read;
134	char *read_ptr;
135
136	fp = fopen( filename, "rb" );
137	if( !fp )
138	return NULL; /* can't open the file */
139	/* we must know about the size of the file */
140	if( fstat( fileno(fp ), &st )
141	\|\| (size = (size_t)st.st_size) != st.st_size
142	\|\| size < sizeof (struct mo_file_header) ) {
143	fclose( fp );
144	return NULL;
145	}
146
147	data = (struct mo_file_header *) malloc( size );
148	if( !data ) {
149	fclose( fp );
150	return NULL; /* out of memory */
151	}
152
153	to_read = size;
154	read_ptr = (char *) data;
155	do {
156	long int nb = fread( read_ptr, 1, to_read, fp );
157	if( nb < to_read ) {
158	fclose (fp);
159	free(data);
160	return NULL; /* read error */
161
162	}
163	read_ptr += nb;
164	to_read -= nb;
165	} while( to_read > 0 );
166	fclose (fp);
167
168	/* Using the magic number we can test whether it really is a message
169	* catalog file. */
170	if( data->magic != MAGIC && data->magic != MAGIC_SWAPPED ) {
171	/* The magic number is wrong: not a message catalog file. */
172	free( data );
173	return NULL;
174	}
175
176	domain = (struct loaded_domain )calloc( 1, sizeof domain );
177	if( !domain ) {
178	free( data );
179	return NULL;
180	}
181	domain->data = (char *) data;
182	domain->must_swap = data->magic != MAGIC;
183
184	/* Fill in the information about the available tables. */
185	switch( SWAPIT(domain->must_swap, data->revision) ) {
186	case 0:
187	domain->nstrings = SWAPIT(domain->must_swap, data->nstrings);
188	domain->orig_tab = (struct string_desc *)
189	((char *) data + SWAPIT(domain->must_swap, data->orig_tab_offset));
190	domain->trans_tab = (struct string_desc *)
191	((char *) data + SWAPIT(domain->must_swap, data->trans_tab_offset));
192	domain->hash_size = SWAPIT(domain->must_swap, data->hash_tab_size);
193	domain->hash_tab = (u32 *)
194	((char *) data + SWAPIT(domain->must_swap, data->hash_tab_offset));
195	break;
196
197	default: /* This is an invalid revision. */
198	free( data );
199	free( domain );
200	return NULL;
201	}
202
203	/* allocate an array to keep track of code page mappings */
204	domain->mapped = (char *)calloc( 1, domain->nstrings );
205	if( !domain->mapped ) {
206	free( data );
207	free( domain );
208	return NULL;
209	}
210
211	return domain;
212	} /* load_domain */
213
214
215	/****************
216	* Set the file used for translations. Pass a NULL to disable
217	* translation. A new filename may be set at anytime.
218	* WARNING: After changing the filename you shoudl not access any data
219	* retrieved by gettext().
220	*/
221	int
222	set_gettext_file( const char filename, const char nls_dir )
223	{
224	struct loaded_domain *domain = NULL;
225
226	if( filename && *filename ) {
227	if( filename[0] == '/'
228	\|\| ( isalpha(filename[0])
229	&& filename[1] == ':'
230	&& (filename[2] == '/' \|\| filename[2] == '\\') )
231	) {
232	/* absolute path - use it as is */
233	domain = load_domain( filename );
234	}
235	else { /* relative path - append ".mo" and get dir from the environment */
236	char *buf = NULL;
237	char *dir;
238
239	dir = strdup( nls_dir );
240	if( dir && (buf= (char *)malloc(strlen(dir)+strlen(filename)+1+3+1)) ) {
241	strcpy(w32_stpcpy(w32_stpcpy(w32_stpcpy( buf, dir),"/"), filename),".mo");
242	domain = load_domain( buf );
243	free(buf);
244	}
245	free(dir);
246	}
247	if( !domain )
248	return -1;
249	}
250
251	if( the_domain ) {
252	free( the_domain->data );
253	free( the_domain->mapped );
254	free( the_domain );
255	the_domain = NULL;
256	}
257	the_domain = domain;
258	return 0;
259	} /* set_gettext_file */
260
261
262	static const char*
263	get_string( struct loaded_domain *domain, u32 idx )
264	{
265	char *p = domain->data + SWAPIT(domain->must_swap,
266	domain->trans_tab[idx].offset);
267	if( !domain->mapped[idx] ) {
268	domain->mapped[idx] = 1;
269	}
270	return (const char*)p;
271	} /* get_string */
272
273
274	const char *
275	gettext( const char *msgid )
276	{
277	struct loaded_domain *domain;
278	size_t act = 0;
279	size_t top, bottom;
280
281	if( !(domain = the_domain) )
282	goto not_found;
283
284	/* Locate the MSGID and its translation. */
285	if( domain->hash_size > 2 && domain->hash_tab ) {
286	/* Use the hashing table. */
287	u32 len = strlen (msgid);
288	u32 hash_val = hash_string (msgid);
289	u32 idx = hash_val % domain->hash_size;
290	u32 incr = 1 + (hash_val % (domain->hash_size - 2));
291	u32 nstr = SWAPIT (domain->must_swap, domain->hash_tab[idx]);
292
293	if ( !nstr ) /* Hash table entry is empty. */
294	goto not_found;
295
296
297	if( SWAPIT(domain->must_swap,
298	domain->orig_tab[nstr - 1].length) == len
299	&& !strcmp( msgid,
300	domain->data + SWAPIT(domain->must_swap,
301	domain->orig_tab[nstr - 1].offset)) )
302	return get_string( domain, nstr - 1 );
303	for(;;) {
304	if (idx >= domain->hash_size - incr)
305	idx -= domain->hash_size - incr;
306	else
307	idx += incr;
308	nstr = SWAPIT(domain->must_swap, domain->hash_tab[idx]);
309	if( !nstr )
310	goto not_found; /* Hash table entry is empty. */
311
312	if ( SWAPIT(domain->must_swap,
313	domain->orig_tab[nstr - 1].length) == len
314	&& !strcmp (msgid,
315	domain->data + SWAPIT(domain->must_swap,
316	domain->orig_tab[nstr - 1].offset)))
317
318	return get_string( domain, nstr-1 );
319	} /* NOTREACHED */
320	}
321
322	/* Now we try the default method: binary search in the sorted
323	array of messages. */
324	bottom = 0;
325	top = domain->nstrings;
326	while( bottom < top ) {
327	int cmp_val;
328
329	act = (bottom + top) / 2;
330	cmp_val = strcmp(msgid, domain->data + SWAPIT(domain->must_swap,
331	domain->orig_tab[act].offset));
332	if (cmp_val < 0)
333	top = act;
334	else if (cmp_val > 0)
335	bottom = act + 1;
336	else
337	return get_string( domain, act );
338	}
339
340	not_found:
341	return msgid;
342	} /* gettext */
343
344	/* Map the user specific language ID to a
345	gettext conform language string.
346	Example: LANG_GERMAN -> "de" */
347	const char*
348	get_gettext_langid (void)
349	{
350	struct {
351	const char *id;
352	int langid;
353	} lang_table[] = {
354	{"de", LANG_GERMAN},
355	{NULL, 0},
356	};
357	LANGID lang;
358	int i;
359
360	lang = GetUserDefaultLangID ();
361
362	for (i=0; lang_table[i].id; i++) {
363	if (PRIMARYLANGID (lang) == lang_table[i].langid)
364	return lang_table[i].id;
365	}
366	return NULL;
367	}
368
369
370	/* XXX: this has to be placed in a .c file but not here */
371	#include <shlobj.h>
372
373	void _SHFree (void *p)
374	{
375	IMalloc *pm;
376	SHGetMalloc(&pm);
377	if (pm) {
378	pm->lpVtbl->Free(pm,p);
379	pm->lpVtbl->Release(pm);
380	}
381	}