1 /*
2 * straycats.c: find and process stray cat files
3 *
4 * Copyright (C) 1994, 1995 Graeme W. Wilford. (Wilf.)
5 * Copyright (C) 2001, 2002, 2003, 2004, 2006, 2007, 2008, 2010, 2011
6 * Colin Watson.
7 *
8 * This file is part of man-db.
9 *
10 * man-db is free software; you can redistribute it and/or modify it
11 * under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 of the License, or
13 * (at your option) any later version.
14 *
15 * man-db is distributed in the hope that it will be useful, but
16 * WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
19 *
20 * You should have received a copy of the GNU General Public License
21 * along with man-db; if not, write to the Free Software Foundation,
22 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
23 *
24 * Tue May 3 21:24:51 BST 1994 Wilf. (G.Wilford@ee.surrey.ac.uk)
25 */
26
27 #ifdef HAVE_CONFIG_H
28 # include "config.h"
29 #endif /* HAVE_CONFIG_H */
30
31 #include <assert.h>
32 #include <stdbool.h>
33 #include <string.h>
34 #include <stdlib.h>
35 #include <stdio.h>
36 #include <signal.h>
37 #include <errno.h>
38 #include <sys/types.h>
39 #include <sys/stat.h>
40 #include <unistd.h>
41 #include <dirent.h>
42
43 #include "canonicalize.h"
44 #include "dirname.h"
45 #include "error.h"
46 #include "gl_array_list.h"
47 #include "gl_xlist.h"
48 #include "xalloc.h"
49
50 #include "gettext.h"
51 #define _(String) gettext (String)
52
53 #include "manconfig.h"
54
55 #include "appendstr.h"
56 #include "compression.h"
57 #include "debug.h"
58 #include "filenames.h"
59 #include "glcontainers.h"
60 #include "pipeline.h"
61 #include "decompress.h"
62 #include "encodings.h"
63 #include "orderfiles.h"
64 #include "sandbox.h"
65 #include "security.h"
66 #include "util.h"
67
68 #include "mydbm.h"
69 #include "db_storage.h"
70
71 #include "descriptions.h"
72 #include "lexgrog.h"
73 #include "manp.h"
74 #include "manconv_client.h"
75 #include "straycats.h"
76 #include "ult_src.h"
77
78 extern man_sandbox *sandbox;
79
80 static char *catdir, *mandir;
81
82 static int check_for_stray (MYDBM_FILE dbf)
83 {
84 DIR *cdir;
85 struct dirent *catlist;
86 gl_list_t names;
87 const char *name;
88 size_t lenman, lencat;
89 int strays = 0;
90
91 cdir = opendir (catdir);
92 if (!cdir) {
93 error (0, errno, _("can't search directory %s"), catdir);
94 return 0;
95 }
96
97 names = new_string_list (GL_ARRAY_LIST, false);
98
99 while ((catlist = readdir (cdir)) != NULL) {
100 if (*catlist->d_name == '.' &&
101 strlen (catlist->d_name) < (size_t) 3)
102 continue;
103 gl_list_add_last (names, xstrdup (catlist->d_name));
104 }
105 closedir (cdir);
106
107 order_files (catdir, &names);
108
109 mandir = appendstr (mandir, "/", (void *) 0);
110 catdir = appendstr (catdir, "/", (void *) 0);
111 lenman = strlen (mandir);
112 lencat = strlen (catdir);
113
114 GL_LIST_FOREACH (names, name) {
115 struct mandata *info;
116 char *ext, *section = NULL;
117 short found;
118 struct stat buf;
119 struct compression *comp;
120
121 info = XZALLOC (struct mandata);
122
123 *(mandir + lenman) = *(catdir + lencat) = '\0';
124 mandir = appendstr (mandir, name, (void *) 0);
125 catdir = appendstr (catdir, name, (void *) 0);
126
127 ext = strrchr (mandir, '.');
128 if (!ext) {
129 if (quiet < 2)
130 error (0, 0,
131 _("warning: %s: "
132 "ignoring bogus filename"),
133 catdir);
134 goto next;
135 } else if (comp_info (ext, false)) {
136 *ext = '\0';
137 info->comp = xstrdup (ext + 1);
138 }
139
140 ext = strrchr (mandir, '.');
141 *(mandir + lenman - 1) = '\0';
142 section = xstrdup (strrchr (mandir, '/') + 4);
143 *(mandir + lenman - 1) = '/';
144
145 /* check for bogosity */
146
147 if (!ext || strncmp (ext + 1, section, strlen (section)) != 0) {
148 if (quiet < 2)
149 error (0, 0,
150 _("warning: %s: "
151 "ignoring bogus filename"),
152 catdir);
153 goto next;
154 }
155
156 /*
157 * now that we've stripped off the cat compression
158 * extension (if it has one), we can try some of ours.
159 */
160
161 debug ("Testing for existence: %s\n", mandir);
162
163 if (stat (mandir, &buf) == 0)
164 found = 1;
165 else if ((comp = comp_file (mandir))) {
166 found = 1;
167 free (comp->stem);
168 } else
169 found = 0;
170
171 if (!found) {
172 decompress *decomp;
173 struct mandata *exists;
174 lexgrog lg;
175 char *lang, *page_encoding;
176 char *mandir_base;
177 pipecmd *col_cmd;
178 char *col_locale;
179 char *fullpath;
180
181 /* we have a straycat. Need to filter it and get
182 its whatis (if necessary) */
183
184 lg.whatis = 0;
185 *(ext++) = '\0';
186 info->ext = xstrdup (ext);
187
188 /* see if we already have it, before going any
189 further */
190 mandir_base = base_name (mandir);
191 exists = dblookup_exact (dbf, mandir_base, info->ext,
192 true);
193 if (exists &&
194 compare_ids (STRAY_CAT, exists->id, false) >= 0)
195 goto next_exists;
196 debug ("%s(%s) is not in the db.\n",
197 mandir_base, info->ext);
198
199 /* fill in the missing parts of the structure */
200 info->sec = xstrdup (section);
201 info->id = STRAY_CAT;
202 info->filter = xstrdup ("-");
203 info->mtime.tv_sec = 0;
204 info->mtime.tv_nsec = 0;
205
206 drop_effective_privs ();
207 decomp = decompress_open (catdir, 0);
208 regain_effective_privs ();
209 if (!decomp) {
210 error (0, errno, _("can't open %s"), catdir);
211 goto next_exists;
212 }
213
214 lang = lang_dir (mandir);
215 page_encoding = get_page_encoding (lang);
216 if (page_encoding)
217 add_manconv (decompress_get_pipeline (decomp),
218 page_encoding, "UTF-8");
219 free (page_encoding);
220 free (lang);
221
222 col_cmd = pipecmd_new_argstr
223 (get_def_user ("col", PROG_COL));
224 pipecmd_arg (col_cmd, "-bx");
225 col_locale = find_charset_locale ("UTF-8");
226 if (col_locale) {
227 pipecmd_setenv (col_cmd, "LC_CTYPE",
228 col_locale);
229 free (col_locale);
230 }
231 pipecmd_pre_exec (col_cmd, sandbox_load, sandbox_free,
232 sandbox);
233 pipeline_command (decompress_get_pipeline (decomp),
234 col_cmd);
235
236 fullpath = canonicalize_file_name (catdir);
237 if (!fullpath)
238 gripe_canonicalize_failed (catdir);
239 else {
240 char *catdir_base;
241
242 free (fullpath);
243 drop_effective_privs ();
244 decompress_start (decomp);
245 regain_effective_privs ();
246
247 strays++;
248
249 lg.type = CATPAGE;
250 catdir_base = base_name (catdir);
251 if (find_name_decompressed (decomp,
252 catdir_base,
253 &lg)) {
254 gl_list_t descs, trace;
255 strays++;
256 descs = parse_descriptions
257 (mandir_base, lg.whatis);
258 trace = new_string_list (GL_ARRAY_LIST,
259 true);
260 gl_list_add_last (trace,
261 xstrdup (catdir));
262 store_descriptions (dbf, descs, info,
263 NULL, mandir_base,
264 trace);
265 gl_list_free (trace);
266 gl_list_free (descs);
267 } else if (quiet < 2)
268 error (0, 0, _("warning: %s: whatis parse for %s(%s) failed"),
269 catdir, mandir_base, info->sec);
270 free (catdir_base);
271 }
272
273 free (lg.whatis);
274 decompress_free (decomp);
275 next_exists:
276 free_mandata_struct (exists);
277 free (mandir_base);
278 }
279 next:
280 free (section);
281 free_mandata_struct (info);
282 }
283 gl_list_free (names);
284 return strays;
285 }
286
287 static int open_catdir (MYDBM_FILE dbf)
288 {
289 DIR *cdir;
290 struct dirent *catlist;
291 size_t catlen, manlen;
292 int strays = 0;
293
294 cdir = opendir (catdir);
295 if (!cdir) {
296 error (0, errno, _("can't search directory %s"), catdir);
297 return 0;
298 }
299
300 if (!quiet)
301 printf (_("Checking for stray cats under %s...\n"), catdir);
302
303 catdir = appendstr (catdir, "/", (void *) 0);
304 mandir = appendstr (mandir, "/", (void *) 0);
305 catlen = strlen (catdir);
306 manlen = strlen (mandir);
307
308 /* should make this case insensitive */
309 while ((catlist = readdir (cdir))) {
310 char *t1;
311
312 if (strncmp (catlist->d_name, "cat", 3) != 0)
313 continue;
314
315 catdir = appendstr (catdir, catlist->d_name, (void *) 0);
316 mandir = appendstr (mandir, catlist->d_name, (void *) 0);
317
318 *(t1 = mandir + manlen) = 'm';
319 *(t1 + 2) = 'n';
320
321 strays += check_for_stray (dbf);
322
323 *(catdir + catlen) = *(mandir + manlen) = '\0';
324 }
325 closedir (cdir);
326 return strays;
327 }
328
329 int straycats (MYDBM_FILE dbf, const char *manpath)
330 {
331 char *catpath;
332 int strays;
333
334 assert (dbf->file);
335
336 catpath = get_catpath (manpath, SYSTEM_CAT | USER_CAT);
337
338 /* look in the usual catpath location */
339 mandir = xstrdup (manpath);
340 catdir = xstrdup (manpath);
341 strays = open_catdir (dbf);
342
343 /* look in the alternate catpath location if we have one
344 and it's different from the usual catpath */
345
346 if (catpath)
347 debug ("catpath: %s, manpath: %s\n", catpath, manpath);
348
349 if (catpath && strcmp (catpath, manpath) != 0) {
350 *mandir = *catdir = '\0';
351 mandir = appendstr (mandir, manpath, (void *) 0);
352 catdir = appendstr (catdir, catpath, (void *) 0);
353 strays += open_catdir (dbf);
354 }
355
356 free (mandir);
357 free (catdir);
358
359 free (catpath);
360
361 return strays;
362 }