1 /* Copyright (C) 1995-1999, 2000-2004 Free Software Foundation, Inc.
2 Contributed by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1995.
4 This program is free software; you can redistribute it and/or modify it
5 under the terms of the GNU Library General Public License as published
6 by the Free Software Foundation; either version 2, or (at your option)
7 any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 Library General Public License for more details.
14 You should have received a copy of the GNU Library General Public
15 License along with this program; if not, write to the Free Software
16 Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307,
17 USA. */
19 /* Tell glibc's <string.h> to provide a prototype for stpcpy().
20 This must come before <config.h> because <config.h> may include
21 <features.h>, and once <features.h> has been included, it's too late. */
22 #ifndef _GNU_SOURCE
23 # define _GNU_SOURCE 1
24 #endif
26 #ifdef HAVE_CONFIG_H
27 # include <config.h>
28 #endif
30 #include <string.h>
32 #if defined _LIBC || defined HAVE_ARGZ_H
33 # include <argz.h>
34 #endif
35 #include <ctype.h>
36 #include <sys/types.h>
37 #include <stdlib.h>
39 #include "loadinfo.h"
41 /* On some strange systems still no definition of NULL is found. Sigh! */
42 #ifndef NULL
43 # if defined __STDC__ && __STDC__
44 # define NULL ((void *) 0)
45 # else
46 # define NULL 0
47 # endif
48 #endif
50 /* @@ end of prolog @@ */
52 #ifdef _LIBC
53 /* Rename the non ANSI C functions. This is required by the standard
54 because some ANSI C functions will require linking with this object
55 file and the name space must not be polluted. */
56 # ifndef stpcpy
57 # define stpcpy(dest, src) __stpcpy(dest, src)
58 # endif
59 #else
60 # ifndef HAVE_STPCPY
61 static char *stpcpy (char *dest, const char *src);
62 # endif
63 #endif
65 /* Pathname support.
66 ISSLASH(C) tests whether C is a directory separator character.
67 IS_ABSOLUTE_PATH(P) tests whether P is an absolute path. If it is not,
68 it may be concatenated to a directory pathname.
69 */
70 #if defined _WIN32 || defined __WIN32__ || defined __CYGWIN__ || defined __EMX__ || defined __DJGPP__
71 /* Win32, Cygwin, OS/2, DOS */
72 # define ISSLASH(C) ((C) == '/' || (C) == '\\')
73 # define HAS_DEVICE(P) \
74 ((((P)[0] >= 'A' && (P)[0] <= 'Z') || ((P)[0] >= 'a' && (P)[0] <= 'z')) \
75 && (P)[1] == ':')
76 # define IS_ABSOLUTE_PATH(P) (ISSLASH ((P)[0]) || HAS_DEVICE (P))
77 #else
78 /* Unix */
79 # define ISSLASH(C) ((C) == '/')
80 # define IS_ABSOLUTE_PATH(P) ISSLASH ((P)[0])
81 #endif
83 /* Define function which are usually not available. */
85 #if !defined _LIBC && !defined HAVE___ARGZ_COUNT
86 /* Returns the number of strings in ARGZ. */
87 static size_t
88 argz_count__ (const char *argz, size_t len)
89 {
90 size_t count = 0;
91 while (len > 0)
92 {
93 size_t part_len = strlen (argz);
94 argz += part_len + 1;
95 len -= part_len + 1;
96 count++;
97 }
98 return count;
99 }
100 # undef __argz_count
101 # define __argz_count(argz, len) argz_count__ (argz, len)
102 #else
103 # ifdef _LIBC
104 # define __argz_count(argz, len) INTUSE(__argz_count) (argz, len)
105 # endif
106 #endif /* !_LIBC && !HAVE___ARGZ_COUNT */
108 #if !defined _LIBC && !defined HAVE___ARGZ_STRINGIFY
109 /* Make '\0' separated arg vector ARGZ printable by converting all the '\0's
110 except the last into the character SEP. */
111 static void
112 argz_stringify__ (char *argz, size_t len, int sep)
113 {
114 while (len > 0)
115 {
116 size_t part_len = strlen (argz);
117 argz += part_len;
118 len -= part_len + 1;
119 if (len > 0)
120 *argz++ = sep;
121 }
122 }
123 # undef __argz_stringify
124 # define __argz_stringify(argz, len, sep) argz_stringify__ (argz, len, sep)
125 #else
126 # ifdef _LIBC
127 # define __argz_stringify(argz, len, sep) \
128 INTUSE(__argz_stringify) (argz, len, sep)
129 # endif
130 #endif /* !_LIBC && !HAVE___ARGZ_STRINGIFY */
132 #if !defined _LIBC && !defined HAVE___ARGZ_NEXT
133 static char *
134 argz_next__ (char *argz, size_t argz_len, const char *entry)
135 {
136 if (entry)
137 {
138 if (entry < argz + argz_len)
139 entry = strchr (entry, '\0') + 1;
141 return entry >= argz + argz_len ? NULL : (char *) entry;
142 }
143 else
144 if (argz_len > 0)
145 return argz;
146 else
147 return 0;
148 }
149 # undef __argz_next
150 # define __argz_next(argz, len, entry) argz_next__ (argz, len, entry)
151 #endif /* !_LIBC && !HAVE___ARGZ_NEXT */
154 /* Return number of bits set in X. */
155 static inline int
156 pop (int x)
157 {
158 /* We assume that no more than 16 bits are used. */
159 x = ((x & ~0x5555) >> 1) + (x & 0x5555);
160 x = ((x & ~0x3333) >> 2) + (x & 0x3333);
161 x = ((x >> 4) + x) & 0x0f0f;
162 x = ((x >> 8) + x) & 0xff;
164 return x;
165 }
167 \f
168 struct loaded_l10nfile *
169 _nl_make_l10nflist (struct loaded_l10nfile **l10nfile_list,
170 const char *dirlist, size_t dirlist_len,
171 int mask, const char *language, const char *territory,
172 const char *codeset, const char *normalized_codeset,
173 const char *modifier, const char *special,
174 const char *sponsor, const char *revision,
175 const char *filename, int do_allocate)
176 {
177 char *abs_filename;
178 struct loaded_l10nfile **lastp;
179 struct loaded_l10nfile *retval;
180 char *cp;
181 size_t dirlist_count;
182 size_t entries;
183 int cnt;
185 /* If LANGUAGE contains an absolute directory specification, we ignore
186 DIRLIST. */
187 if (IS_ABSOLUTE_PATH (language))
188 dirlist_len = 0;
190 /* Allocate room for the full file name. */
191 abs_filename = (char *) malloc (dirlist_len
192 + strlen (language)
193 + ((mask & TERRITORY) != 0
194 ? strlen (territory) + 1 : 0)
195 + ((mask & XPG_CODESET) != 0
196 ? strlen (codeset) + 1 : 0)
197 + ((mask & XPG_NORM_CODESET) != 0
198 ? strlen (normalized_codeset) + 1 : 0)
199 + (((mask & XPG_MODIFIER) != 0
200 || (mask & CEN_AUDIENCE) != 0)
201 ? strlen (modifier) + 1 : 0)
202 + ((mask & CEN_SPECIAL) != 0
203 ? strlen (special) + 1 : 0)
204 + (((mask & CEN_SPONSOR) != 0
205 || (mask & CEN_REVISION) != 0)
206 ? (1 + ((mask & CEN_SPONSOR) != 0
207 ? strlen (sponsor) : 0)
208 + ((mask & CEN_REVISION) != 0
209 ? strlen (revision) + 1 : 0)) : 0)
210 + 1 + strlen (filename) + 1);
212 if (abs_filename == NULL)
213 return NULL;
215 /* Construct file name. */
216 cp = abs_filename;
217 if (dirlist_len > 0)
218 {
219 memcpy (cp, dirlist, dirlist_len);
220 __argz_stringify (cp, dirlist_len, PATH_SEPARATOR);
221 cp += dirlist_len;
222 cp[-1] = '/';
223 }
225 cp = stpcpy (cp, language);
227 if ((mask & TERRITORY) != 0)
228 {
229 *cp++ = '_';
230 cp = stpcpy (cp, territory);
231 }
232 if ((mask & XPG_CODESET) != 0)
233 {
234 *cp++ = '.';
235 cp = stpcpy (cp, codeset);
236 }
237 if ((mask & XPG_NORM_CODESET) != 0)
238 {
239 *cp++ = '.';
240 cp = stpcpy (cp, normalized_codeset);
241 }
242 if ((mask & (XPG_MODIFIER | CEN_AUDIENCE)) != 0)
243 {
244 /* This component can be part of both syntaces but has different
245 leading characters. For CEN we use `+', else `@'. */
246 *cp++ = (mask & CEN_AUDIENCE) != 0 ? '+' : '@';
247 cp = stpcpy (cp, modifier);
248 }
249 if ((mask & CEN_SPECIAL) != 0)
250 {
251 *cp++ = '+';
252 cp = stpcpy (cp, special);
253 }
254 if ((mask & (CEN_SPONSOR | CEN_REVISION)) != 0)
255 {
256 *cp++ = ',';
257 if ((mask & CEN_SPONSOR) != 0)
258 cp = stpcpy (cp, sponsor);
259 if ((mask & CEN_REVISION) != 0)
260 {
261 *cp++ = '_';
262 cp = stpcpy (cp, revision);
263 }
264 }
266 *cp++ = '/';
267 stpcpy (cp, filename);
269 /* Look in list of already loaded domains whether it is already
270 available. */
271 lastp = l10nfile_list;
272 for (retval = *l10nfile_list; retval != NULL; retval = retval->next)
273 if (retval->filename != NULL)
274 {
275 int compare = strcmp (retval->filename, abs_filename);
276 if (compare == 0)
277 /* We found it! */
278 break;
279 if (compare < 0)
280 {
281 /* It's not in the list. */
282 retval = NULL;
283 break;
284 }
286 lastp = &retval->next;
287 }
289 if (retval != NULL || do_allocate == 0)
290 {
291 free (abs_filename);
292 return retval;
293 }
295 dirlist_count = (dirlist_len > 0 ? __argz_count (dirlist, dirlist_len) : 1);
297 /* Allocate a new loaded_l10nfile. */
298 retval =
299 (struct loaded_l10nfile *)
300 malloc (sizeof (*retval)
301 + (((dirlist_count << pop (mask)) + (dirlist_count > 1 ? 1 : 0))
302 * sizeof (struct loaded_l10nfile *)));
303 if (retval == NULL)
304 return NULL;
306 retval->filename = abs_filename;
308 /* We set retval->data to NULL here; it is filled in later.
309 Setting retval->decided to 1 here means that retval does not
310 correspond to a real file (dirlist_count > 1) or is not worth
311 looking up (if an unnormalized codeset was specified). */
312 retval->decided = (dirlist_count > 1
313 || ((mask & XPG_CODESET) != 0
314 && (mask & XPG_NORM_CODESET) != 0));
315 retval->data = NULL;
317 retval->next = *lastp;
318 *lastp = retval;
320 entries = 0;
321 /* Recurse to fill the inheritance list of RETVAL.
322 If the DIRLIST is a real list (i.e. DIRLIST_COUNT > 1), the RETVAL
323 entry does not correspond to a real file; retval->filename contains
324 colons. In this case we loop across all elements of DIRLIST and
325 across all bit patterns dominated by MASK.
326 If the DIRLIST is a single directory or entirely redundant (i.e.
327 DIRLIST_COUNT == 1), we loop across all bit patterns dominated by
328 MASK, excluding MASK itself.
329 In either case, we loop down from MASK to 0. This has the effect
330 that the extra bits in the locale name are dropped in this order:
331 first the modifier, then the territory, then the codeset, then the
332 normalized_codeset. */
333 for (cnt = dirlist_count > 1 ? mask : mask - 1; cnt >= 0; --cnt)
334 if ((cnt & ~mask) == 0
335 && ((cnt & CEN_SPECIFIC) == 0 || (cnt & XPG_SPECIFIC) == 0)
336 && ((cnt & XPG_CODESET) == 0 || (cnt & XPG_NORM_CODESET) == 0))
337 {
338 if (dirlist_count > 1)
339 {
340 /* Iterate over all elements of the DIRLIST. */
341 char *dir = NULL;
343 while ((dir = __argz_next ((char *) dirlist, dirlist_len, dir))
344 != NULL)
345 retval->successor[entries++]
346 = _nl_make_l10nflist (l10nfile_list, dir, strlen (dir) + 1,
347 cnt, language, territory, codeset,
348 normalized_codeset, modifier, special,
349 sponsor, revision, filename, 1);
350 }
351 else
352 retval->successor[entries++]
353 = _nl_make_l10nflist (l10nfile_list, dirlist, dirlist_len,
354 cnt, language, territory, codeset,
355 normalized_codeset, modifier, special,
356 sponsor, revision, filename, 1);
357 }
358 retval->successor[entries] = NULL;
360 return retval;
361 }
362 \f
363 /* Normalize codeset name. There is no standard for the codeset
364 names. Normalization allows the user to use any of the common
365 names. The return value is dynamically allocated and has to be
366 freed by the caller. */
367 const char *
368 _nl_normalize_codeset (const char *codeset, size_t name_len)
369 {
370 int len = 0;
371 int only_digit = 1;
372 char *retval;
373 char *wp;
374 size_t cnt;
376 for (cnt = 0; cnt < name_len; ++cnt)
377 if (isalnum ((unsigned char) codeset[cnt]))
378 {
379 ++len;
381 if (isalpha ((unsigned char) codeset[cnt]))
382 only_digit = 0;
383 }
385 retval = (char *) malloc ((only_digit ? 3 : 0) + len + 1);
387 if (retval != NULL)
388 {
389 if (only_digit)
390 wp = stpcpy (retval, "iso");
391 else
392 wp = retval;
394 for (cnt = 0; cnt < name_len; ++cnt)
395 if (isalpha ((unsigned char) codeset[cnt]))
396 *wp++ = tolower ((unsigned char) codeset[cnt]);
397 else if (isdigit ((unsigned char) codeset[cnt]))
398 *wp++ = codeset[cnt];
400 *wp = '\0';
401 }
403 return (const char *) retval;
404 }
407 /* @@ begin of epilog @@ */
409 /* We don't want libintl.a to depend on any other library. So we
410 avoid the non-standard function stpcpy. In GNU C Library this
411 function is available, though. Also allow the symbol HAVE_STPCPY
412 to be defined. */
413 #if !_LIBC && !HAVE_STPCPY
414 static char *
415 stpcpy (char *dest, const char *src)
416 {
417 while ((*dest++ = *src++) != '\0')
418 /* Do nothing. */ ;
419 return dest - 1;
420 }
421 #endif