1 |
/* |
2 |
** This program and library is free software; you can redistribute it and/or |
3 |
** modify it under the terms of the GNU (Library) General Public License |
4 |
** as published by the Free Software Foundation; either version 2 |
5 |
** of the License, or any later version. |
6 |
** |
7 |
** This program is distributed in the hope that it will be useful, |
8 |
** but WITHOUT ANY WARRANTY; without even the implied warranty of |
9 |
** MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
10 |
** GNU (Library) General Public License for more details. |
11 |
** |
12 |
** You should have received a copy of the GNU (Library) General Public License |
13 |
** long with this program; if not, write to the Free Software |
14 |
** Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. |
15 |
**-------------------------------------------------------------------- |
16 |
** |
17 |
** change sprintf to snprintf to avoid corruption, |
18 |
** added safestrcpy() macro to avoid corruption from strcpy overflow, |
19 |
** and use MAXKEYLEN as string length vs. literal "34" |
20 |
** SRE 11/17/99 |
21 |
** |
22 |
** 2000-11 jruiz,rasc some redesign |
23 |
** 2001-04-07 rasc fixed FileRule pathname |
24 |
** |
25 |
*/ |
26 |
|
27 |
#include "swish.h" |
28 |
#include "mem.h" |
29 |
#include "string.h" |
30 |
#include "index.h" |
31 |
#include "hash.h" |
32 |
#include "file.h" |
33 |
#include "list.h" |
34 |
#include "fs.h" |
35 |
#include "check.h" |
36 |
#include "error.h" |
37 |
#include "xml.h" |
38 |
#include "txt.h" |
39 |
#include "parse_conffile.h" |
40 |
#include "swish_qsort.h" |
41 |
|
42 |
typedef struct |
43 |
{ |
44 |
int currentsize; |
45 |
int maxsize; |
46 |
char **filenames; |
47 |
} |
48 |
DOCENTRYARRAY; |
49 |
|
50 |
|
51 |
|
52 |
#define MAXKEYLEN 34 /* Hash key -- allow for 64 bit inodes */ |
53 |
|
54 |
|
55 |
static int get_rules( char *name, StringList *sl, PATH_LIST *pathlist ); |
56 |
static int check_FileTests( char *path, PATH_LIST *test ); |
57 |
static void indexadir(SWISH *, char *); |
58 |
static void indexafile(SWISH *, char *); |
59 |
static void printfile(SWISH *, char *); |
60 |
static void printfiles(SWISH *, DOCENTRYARRAY *); |
61 |
static void printdirs(SWISH *, DOCENTRYARRAY *); |
62 |
static DOCENTRYARRAY *adddocentry(DOCENTRYARRAY * e, char *filename); |
63 |
static void split_path(char *path, char **directory, char **file); |
64 |
|
65 |
/* |
66 |
-- init structures for this module |
67 |
*/ |
68 |
|
69 |
void initModule_FS(SWISH * sw) |
70 |
{ |
71 |
struct MOD_FS *fs; |
72 |
|
73 |
fs = (struct MOD_FS *) emalloc(sizeof(struct MOD_FS)); |
74 |
|
75 |
memset(fs, 0, sizeof(struct MOD_FS)); |
76 |
sw->FS = fs; |
77 |
|
78 |
} |
79 |
|
80 |
|
81 |
/* |
82 |
-- release all wired memory for this module |
83 |
*/ |
84 |
|
85 |
void freeModule_FS(SWISH * sw) |
86 |
{ |
87 |
struct MOD_FS *fs = sw->FS; |
88 |
|
89 |
/* Free fs parameters */ |
90 |
|
91 |
free_regex_list( &fs->filerules.pathname ); |
92 |
free_regex_list( &fs->filerules.dirname ); |
93 |
free_regex_list( &fs->filerules.filename ); |
94 |
free_regex_list( &fs->filerules.dircontains ); |
95 |
free_regex_list( &fs->filerules.title ); |
96 |
|
97 |
free_regex_list( &fs->filematch.pathname ); |
98 |
free_regex_list( &fs->filematch.dirname ); |
99 |
free_regex_list( &fs->filematch.filename ); |
100 |
free_regex_list( &fs->filematch.dircontains ); |
101 |
free_regex_list( &fs->filematch.title ); |
102 |
|
103 |
|
104 |
/* free module data */ |
105 |
efree(fs); |
106 |
sw->FS = NULL; |
107 |
|
108 |
return; |
109 |
} |
110 |
|
111 |
|
112 |
/* |
113 |
-- Config Directives |
114 |
-- Configuration directives for this Module |
115 |
-- return: 0/1 = none/config applied |
116 |
Aug 1, 2001 -- these probably should be pre-compiled regular expressions, |
117 |
and their memory should be freed on exit. moseley |
118 |
*/ |
119 |
|
120 |
|
121 |
|
122 |
int configModule_FS(SWISH * sw, StringList * sl) |
123 |
{ |
124 |
struct MOD_FS *fs = sw->FS; |
125 |
char *w0 = sl->word[0]; |
126 |
|
127 |
if (strcasecmp(w0, "FileRules") == 0) |
128 |
return get_rules( w0, sl, &fs->filerules ); |
129 |
|
130 |
if (strcasecmp(w0, "FileMatch") == 0) |
131 |
return get_rules( w0, sl, &fs->filematch ); |
132 |
|
133 |
|
134 |
if (strcasecmp(w0, "FollowSymLinks") == 0) |
135 |
{ |
136 |
fs->followsymlinks = getYesNoOrAbort(sl, 1, 1); |
137 |
return 1; |
138 |
} |
139 |
|
140 |
return 0; /* not one of our parameters */ |
141 |
|
142 |
} |
143 |
|
144 |
static int get_rules( char *name, StringList *sl, PATH_LIST *pathlist ) |
145 |
{ |
146 |
char *w1; |
147 |
char *both; |
148 |
int regex_pattern = 0; |
149 |
|
150 |
|
151 |
if (sl->n < 4) |
152 |
{ |
153 |
printf("err: Wrong number of parameters in %s\n", name); |
154 |
return 0; |
155 |
} |
156 |
|
157 |
|
158 |
/* For "is" make sure it matches the entire pattern */ |
159 |
/* A bit ugly */ |
160 |
|
161 |
if ( strcasecmp(sl->word[2], "is") == 0 ) |
162 |
{ |
163 |
int i; |
164 |
/* make patterns match the full string */ |
165 |
for ( i = 3; i < sl->n; i++ ) |
166 |
{ |
167 |
int len = strlen( sl->word[i] ); |
168 |
char *new; |
169 |
char *old = sl->word[i]; |
170 |
|
171 |
if ( (strcasecmp( old, "not" ) == 0) && i < sl->n-1 ) |
172 |
continue; |
173 |
|
174 |
new = emalloc( len + 3 ); |
175 |
|
176 |
if ( sl->word[i][0] != '^' ) |
177 |
{ |
178 |
strcpy( new, "^" ); |
179 |
strcat( new, sl->word[i] ); |
180 |
} |
181 |
else |
182 |
strcpy( new, sl->word[i] ); |
183 |
|
184 |
if ( sl->word[i][len-1] != '$' ) |
185 |
strcat( new, "$" ); |
186 |
|
187 |
sl->word[i] = new; |
188 |
efree( old ); |
189 |
} |
190 |
|
191 |
} |
192 |
|
193 |
else if ( strcasecmp(sl->word[2], "regex") == 0 ) |
194 |
regex_pattern++; |
195 |
|
196 |
|
197 |
else if ( !(strcasecmp(sl->word[2], "contains") == 0) ) |
198 |
{ |
199 |
printf("err: %s must be followed by [is|contains|regex]\n", name); |
200 |
return 0; |
201 |
} |
202 |
|
203 |
w1 = sl->word[1]; |
204 |
|
205 |
both = emalloc( strlen( name ) + strlen( w1 ) + 2 ); |
206 |
strcpy( both, name ); |
207 |
strcat( both, " "); |
208 |
strcat( both, w1 ); |
209 |
|
210 |
|
211 |
|
212 |
if ( strcasecmp(w1, "pathname") == 0 ) |
213 |
add_regex_patterns( both, &pathlist->pathname, &(sl->word)[3], regex_pattern ); |
214 |
|
215 |
else if ( strcasecmp(w1, "filename") == 0 ) |
216 |
add_regex_patterns( both, &pathlist->filename, &(sl->word)[3], regex_pattern ); |
217 |
|
218 |
else if ( strcasecmp(w1, "dirname") == 0 ) |
219 |
add_regex_patterns( both, &pathlist->dirname, &(sl->word)[3], regex_pattern ); |
220 |
|
221 |
else if ( strcasecmp(w1, "title") == 0 ) |
222 |
add_regex_patterns( both, &pathlist->title, &(sl->word)[3], regex_pattern ); |
223 |
|
224 |
else if ( strcasecmp(w1, "directory") == 0 ) |
225 |
add_regex_patterns( both, &pathlist->dircontains, &(sl->word)[3], regex_pattern ); |
226 |
|
227 |
else |
228 |
{ |
229 |
printf("err: '%s' - invalid parameter '%s'\n", both, w1 ); |
230 |
return 0; |
231 |
} |
232 |
|
233 |
|
234 |
efree( both ); |
235 |
return 1; |
236 |
} |
237 |
|
238 |
|
239 |
|
240 |
|
241 |
|
242 |
|
243 |
|
244 |
|
245 |
/* Have we already indexed a file or directory? |
246 |
** This function is used to avoid multiple index entries |
247 |
** or endless looping due to symbolic links. |
248 |
*/ |
249 |
|
250 |
static int fs_already_indexed(SWISH * sw, char *path) |
251 |
{ |
252 |
#ifndef NO_SYMBOLIC_FILE_LINKS |
253 |
struct dev_ino *p; |
254 |
struct stat buf; |
255 |
char key[MAXKEYLEN]; /* Hash key -- allow for 64 bit inodes */ |
256 |
unsigned hashval; |
257 |
|
258 |
if (stat(path, &buf)) |
259 |
return 0; |
260 |
|
261 |
/* Create hash key: string contains device and inode. */ |
262 |
/* Avoid snprintf -> MAXKEYLEN is big enough for two longs |
263 |
snprintf( key, MAXKEYLEN, "%lx/%lx", (unsigned long)buf.st_dev, |
264 |
(unsigned long)buf.st_ino ); |
265 |
*/ |
266 |
sprintf(key, "%lx/%lx", (unsigned long) buf.st_dev, (unsigned long) buf.st_ino); |
267 |
|
268 |
hashval = bighash(key); /* Search hash for this file. */ |
269 |
for (p = sw->Index->inode_hash[hashval]; p != NULL; p = p->next) |
270 |
if (p->dev == buf.st_dev && p->ino == buf.st_ino) |
271 |
{ /* We found it. */ |
272 |
if (sw->verbose >= 3) |
273 |
printf("Skipping %s: %s\n", path, "Already indexed."); |
274 |
return 1; |
275 |
} |
276 |
|
277 |
/* Not found, make new entry. */ |
278 |
p = (struct dev_ino *) Mem_ZoneAlloc(sw->Index->entryZone,sizeof(struct dev_ino)); |
279 |
|
280 |
p->dev = buf.st_dev; |
281 |
p->ino = buf.st_ino; |
282 |
p->next = sw->Index->inode_hash[hashval]; |
283 |
sw->Index->inode_hash[hashval] = p; /* Aug 1, 2001 -- this is not freed */ |
284 |
#endif |
285 |
|
286 |
return 0; |
287 |
} |
288 |
|
289 |
|
290 |
/* Recursively goes into a directory and calls the word-indexing |
291 |
** functions for each file that's found. |
292 |
*/ |
293 |
|
294 |
static void indexadir(SWISH * sw, char *dir) |
295 |
{ |
296 |
int allgoodfiles = 0; |
297 |
DIR *dfd; |
298 |
|
299 |
#ifdef NEXTSTEP |
300 |
struct direct *dp; |
301 |
#else |
302 |
struct dirent *dp; |
303 |
#endif |
304 |
int pathbuflen; |
305 |
char *pathname; |
306 |
DOCENTRYARRAY *sortfilelist = NULL; |
307 |
DOCENTRYARRAY *sortdirlist = NULL; |
308 |
int dirlen = strlen( dir ); |
309 |
struct MOD_FS *fs = sw->FS; |
310 |
|
311 |
if (fs_already_indexed(sw, dir)) |
312 |
return; |
313 |
|
314 |
/* and another stat if not set to follow symlinks */ |
315 |
if (!fs->followsymlinks && islink(dir)) |
316 |
return; |
317 |
|
318 |
|
319 |
/* logic is not well defined - here we only check dirname */ |
320 |
/* but that bypasses pathname checks -- but that's checked per-file */ |
321 |
/* This allows one to override File* directory checks */ |
322 |
/* but allows a pathname check to be limited to full paths */ |
323 |
/* This also means you can avoid indexing an entire directory tree with FileRules dirname, */ |
324 |
/* but using a FileRules pathname allows recursion into the directory */ |
325 |
|
326 |
/* Reject entire directory due to FileRules dirname */ |
327 |
if ( *dir && match_regex_list( dir, fs->filerules.dirname ) ) |
328 |
return; |
329 |
|
330 |
|
331 |
|
332 |
|
333 |
/* Handle "FileRules directory" directive */ |
334 |
/* - Check all files within the directory before proceeding -- means reading the directory twice */ |
335 |
/* - All files are checked. */ |
336 |
|
337 |
if (fs->filematch.dircontains || fs->filerules.dircontains ) |
338 |
{ |
339 |
if ((dfd = opendir(dir)) == NULL) |
340 |
{ |
341 |
if ( sw->verbose ) |
342 |
progwarnno("Failed to open dir '%s' :", dir ); |
343 |
return; |
344 |
} |
345 |
|
346 |
while ((dp = readdir(dfd)) != NULL) |
347 |
{ |
348 |
if ( match_regex_list( dp->d_name, fs->filerules.dircontains ) ) |
349 |
{ |
350 |
closedir( dfd ); |
351 |
return; /* doesn't recurse */ |
352 |
} |
353 |
|
354 |
if ( match_regex_list( dp->d_name, fs->filematch.dircontains ) ) |
355 |
{ |
356 |
allgoodfiles++; |
357 |
break; |
358 |
} |
359 |
|
360 |
} |
361 |
closedir(dfd); |
362 |
} |
363 |
|
364 |
|
365 |
/* Now, build list of files and directories */ |
366 |
|
367 |
pathbuflen = MAXFILELEN; |
368 |
pathname = (char *) emalloc( pathbuflen + 1 ); |
369 |
|
370 |
if ((dfd = opendir(dir)) == NULL) |
371 |
{ |
372 |
if ( sw->verbose ) |
373 |
progwarnno("Failed to open dir '%s' :", dir ); |
374 |
return; |
375 |
} |
376 |
|
377 |
|
378 |
if ( dirlen == 1 && *dir == '/' ) /* case of root dir */ |
379 |
dirlen = 0; |
380 |
|
381 |
while ((dp = readdir(dfd)) != NULL) |
382 |
{ |
383 |
int filelen = strlen( dp->d_name ); |
384 |
|
385 |
/* For security reasons, don't index dot files */ |
386 |
/* Check for hidden under Windows? */ |
387 |
|
388 |
if ((dp->d_name)[0] == '.') |
389 |
continue; |
390 |
|
391 |
|
392 |
/* Build full path to file */ |
393 |
|
394 |
/* reallocate filename buffer, if needed (dir + path + '/' ) */ |
395 |
if ( dirlen + filelen + 1 > pathbuflen ) |
396 |
{ |
397 |
pathbuflen = dirlen + filelen + 200; |
398 |
pathname = (char *) erealloc(pathname, pathbuflen + 1); |
399 |
} |
400 |
|
401 |
if ( dirlen ) |
402 |
memcpy(pathname, dir, dirlen); |
403 |
|
404 |
pathname[dirlen] = '/'; // Add path separator |
405 |
memcpy(pathname + dirlen + 1, dp->d_name, filelen); |
406 |
pathname[ dirlen + filelen + 1] = '\0'; |
407 |
|
408 |
/* Check if the path is a symlink */ |
409 |
if ( !fs->followsymlinks && islink( pathname ) ) |
410 |
continue; |
411 |
|
412 |
|
413 |
if ( isdirectory(pathname) ) |
414 |
{ |
415 |
sortdirlist = (DOCENTRYARRAY *) adddocentry(sortdirlist, pathname); |
416 |
} |
417 |
else |
418 |
{ |
419 |
if (fs_already_indexed(sw, pathname)) |
420 |
continue; |
421 |
|
422 |
if ( allgoodfiles || check_FileTests( pathname, &fs->filematch ) ) |
423 |
{ |
424 |
sortfilelist = (DOCENTRYARRAY *) adddocentry(sortfilelist, pathname); |
425 |
continue; |
426 |
} |
427 |
|
428 |
|
429 |
if (!isoksuffix(dp->d_name, sw->suffixlist)) |
430 |
continue; |
431 |
|
432 |
|
433 |
/* Check FileRules for rejects */ |
434 |
if ( check_FileTests( pathname, &fs->filerules ) ) |
435 |
continue; |
436 |
|
437 |
sortfilelist = (DOCENTRYARRAY *) adddocentry(sortfilelist, pathname); |
438 |
} |
439 |
} |
440 |
|
441 |
efree(pathname); |
442 |
|
443 |
closedir(dfd); |
444 |
|
445 |
printfiles(sw, sortfilelist); |
446 |
printdirs(sw, sortdirlist); |
447 |
} |
448 |
|
449 |
/* Calls the word-indexing function for a single file. |
450 |
*/ |
451 |
|
452 |
static void indexafile(SWISH * sw, char *path) |
453 |
{ |
454 |
struct MOD_FS *fs = sw->FS; |
455 |
|
456 |
|
457 |
if (!fs->followsymlinks && islink(path)) |
458 |
return; |
459 |
|
460 |
|
461 |
/* This only means "IndexDir test.html test.html test.html" will only index test.html once */ |
462 |
if (fs_already_indexed(sw, path)) |
463 |
return; |
464 |
|
465 |
/* Check for File|Pathmatch, and index if any match */ |
466 |
if ( check_FileTests( path, &fs->filematch ) ) |
467 |
{ |
468 |
printfile(sw, path); |
469 |
return; |
470 |
} |
471 |
|
472 |
/* This is likely faster, so do it first */ |
473 |
if (!isoksuffix(path, sw->suffixlist)) |
474 |
return; |
475 |
|
476 |
/* Check FileRules for rejects */ |
477 |
if ( check_FileTests( path, &fs->filerules ) ) |
478 |
return; |
479 |
|
480 |
|
481 |
/* Passed all tests, so index */ |
482 |
printfile(sw, path); |
483 |
} |
484 |
|
485 |
/********************************************************** |
486 |
* Process FileTests |
487 |
* |
488 |
* Returns 1 = something matched |
489 |
* |
490 |
**********************************************************/ |
491 |
static int check_FileTests( char *path, PATH_LIST *test ) |
492 |
{ |
493 |
char *dir; |
494 |
char *file; |
495 |
|
496 |
|
497 |
if ( match_regex_list( path, test->pathname ) ) |
498 |
return 1; |
499 |
|
500 |
if ( !( test->dirname || test->filename ) ) |
501 |
return 0; |
502 |
|
503 |
split_path( path, &dir, &file ); |
504 |
|
505 |
if ( *dir && match_regex_list( dir, test->dirname ) ) |
506 |
{ |
507 |
efree( dir ); |
508 |
efree( file ); |
509 |
return 1; |
510 |
} |
511 |
|
512 |
if ( *file && match_regex_list( file, test->filename ) ) |
513 |
{ |
514 |
efree( dir ); |
515 |
efree( file ); |
516 |
return 1; |
517 |
} |
518 |
|
519 |
efree( dir ); |
520 |
efree( file ); |
521 |
return 0; |
522 |
} |
523 |
|
524 |
/*************************************************** |
525 |
* Note that this is mostly a duplicate of above, |
526 |
* but was designed to work with both path and URLs |
527 |
* |
528 |
* Probably should settle on one |
529 |
* Also, this returns "" on empty dirs, where above returns " " |
530 |
* Mar 2002 -- and is only called by fs.c... |
531 |
* May 2002, moved to fs.c. Why isn't basename library used for this? |
532 |
***************************************************/ |
533 |
|
534 |
static void split_path(char *path, char **directory, char **file) |
535 |
{ |
536 |
char *p1, |
537 |
*p2, |
538 |
*p3; |
539 |
|
540 |
/* look for last DIRDELIMITER (FS) and last / (HTTP) */ |
541 |
//p1 = strrchr( path, DIRDELIMITER); |
542 |
p1 = strrchr( path, '/'); |
543 |
p2 = strrchr( path, '/'); |
544 |
|
545 |
if (p1 && p2) |
546 |
{ /* if both are found, use the longest. */ |
547 |
if (p1 >= p2) |
548 |
p3 = p1; |
549 |
else |
550 |
p3 = p2; |
551 |
} else if (p1 && !p2) |
552 |
p3 = p1; |
553 |
else if (!p1 && p2) |
554 |
p3 = p2; |
555 |
else |
556 |
p3 = NULL; |
557 |
|
558 |
/* Set directory */ |
559 |
if (!p3) |
560 |
*directory = (char *) estrdup((char *) ""); |
561 |
else |
562 |
{ |
563 |
char c = *++p3; |
564 |
|
565 |
*p3 = '\0'; |
566 |
*directory = (char *) estrdup((char *) path); |
567 |
*p3 = c; |
568 |
path = p3; |
569 |
} |
570 |
|
571 |
*file = (char *) estrdup((char *) path); |
572 |
} |
573 |
|
574 |
|
575 |
|
576 |
|
577 |
/* Indexes the words in the file |
578 |
*/ |
579 |
|
580 |
static void printfile(SWISH * sw, char *filename) |
581 |
{ |
582 |
char *s; |
583 |
FileProp *fprop; |
584 |
|
585 |
|
586 |
if (filename) |
587 |
{ |
588 |
if (sw->verbose >= 3) |
589 |
{ |
590 |
/* Only display file name */ |
591 |
if ((s = (char *) strrchr(filename, '/')) == NULL) |
592 |
printf(" %s", filename); |
593 |
else |
594 |
printf(" %s", s + 1); |
595 |
fflush(stdout); |
596 |
} |
597 |
|
598 |
|
599 |
fprop = file_properties(filename, filename, sw); |
600 |
|
601 |
do_index_file(sw, fprop); |
602 |
|
603 |
free_file_properties(fprop); |
604 |
} |
605 |
} |
606 |
|
607 |
/* 2001-08 Jose Ruiz */ |
608 |
/* function for comparing filenames to get all filenames in a dir sorted |
609 |
** Original addsortentry used strcmp - So, I use the same routine here |
610 |
** What about Win32? |
611 |
*/ |
612 |
int compfilenames(const void *s1, const void *s2) |
613 |
{ |
614 |
char *r1 = *(char * const *) s1; |
615 |
char *r2 = *(char * const *) s2; |
616 |
|
617 |
return strcmp(r1,r2); |
618 |
} |
619 |
|
620 |
/* Indexes the words in all the files in the array of files |
621 |
** The array is sorted alphabetically |
622 |
*/ |
623 |
|
624 |
static void printfiles(SWISH * sw, DOCENTRYARRAY * e) |
625 |
{ |
626 |
int i; |
627 |
|
628 |
if (e) |
629 |
{ |
630 |
/* 2001-08 sorting of filenames moved here - Do we really |
631 |
** need to sort them? - Adjust it in config.h */ |
632 |
if(e->currentsize) |
633 |
{ |
634 |
if(SORT_FILENAMES) |
635 |
{ |
636 |
swish_qsort(e->filenames, e->currentsize, sizeof(char *), compfilenames); |
637 |
} |
638 |
} |
639 |
|
640 |
for (i = 0; i < e->currentsize; i++) |
641 |
{ |
642 |
printfile(sw, e->filenames[i]); |
643 |
efree( e->filenames[i] ); |
644 |
} |
645 |
|
646 |
/* free the array and filenames */ |
647 |
efree(e->filenames); |
648 |
efree(e); |
649 |
} |
650 |
} |
651 |
|
652 |
/* Prints out the directory names as things are getting indexed. |
653 |
** Calls indexadir() so directories in the array are indexed, |
654 |
** in alphabetical order... |
655 |
*/ |
656 |
|
657 |
void printdirs(SWISH * sw, DOCENTRYARRAY * e) |
658 |
{ |
659 |
int i; |
660 |
|
661 |
if (e) |
662 |
{ |
663 |
/* 2001-08 sorting of dirs moved here - Do we really |
664 |
** need to sort them? - Adjust it in config.h */ |
665 |
if(e->currentsize) |
666 |
{ |
667 |
if(SORT_FILENAMES) |
668 |
{ |
669 |
swish_qsort(e->filenames, e->currentsize, sizeof(char *), compfilenames); |
670 |
} |
671 |
} |
672 |
|
673 |
for (i = 0; i < e->currentsize; i++) |
674 |
{ |
675 |
if (sw->verbose >= 3) |
676 |
printf("\nIn dir \"%s\":\n", e->filenames[i]); |
677 |
else if (sw->verbose >= 2) |
678 |
printf("Checking dir \"%s\"...\n", e->filenames[i]); |
679 |
|
680 |
indexadir(sw, e->filenames[i]); |
681 |
efree(e->filenames[i]); |
682 |
} |
683 |
efree(e->filenames); |
684 |
efree(e); |
685 |
} |
686 |
} |
687 |
|
688 |
/* Stores file names in alphabetical order so they can be |
689 |
** indexed alphabetically. No big whoop. |
690 |
*/ |
691 |
|
692 |
static DOCENTRYARRAY *adddocentry(DOCENTRYARRAY * e, char *filename) |
693 |
{ |
694 |
if (e == NULL) |
695 |
{ |
696 |
e = (DOCENTRYARRAY *) emalloc(sizeof(DOCENTRYARRAY)); |
697 |
e->maxsize = VERYBIGHASHSIZE; /* Put what you like */ |
698 |
e->filenames = (char **) emalloc(e->maxsize * sizeof(char *)); |
699 |
|
700 |
e->currentsize = 1; |
701 |
e->filenames[0] = (char *) estrdup(filename); |
702 |
} |
703 |
else |
704 |
{ |
705 |
if ((e->currentsize + 1) == e->maxsize) |
706 |
{ |
707 |
e->maxsize += 1000; |
708 |
e->filenames = (char **) erealloc(e->filenames, e->maxsize * sizeof(char *)); |
709 |
} |
710 |
e->filenames[e->currentsize++] = (char *) estrdup(filename); |
711 |
} |
712 |
return e; |
713 |
} |
714 |
|
715 |
|
716 |
|
717 |
|
718 |
/********************************************************/ |
719 |
/* "Public" functions */ |
720 |
/********************************************************/ |
721 |
|
722 |
void fs_indexpath(SWISH * sw, char *path) |
723 |
{ |
724 |
|
725 |
normalize_path( path ); /* flip backslashes and remove trailing slash */ |
726 |
|
727 |
|
728 |
if (isdirectory(path)) |
729 |
{ |
730 |
if (sw->verbose >= 2) |
731 |
printf("\nChecking dir \"%s\"...\n", path); |
732 |
|
733 |
indexadir(sw, path); |
734 |
} |
735 |
|
736 |
else if (isfile(path)) |
737 |
{ |
738 |
if (sw->verbose >= 2) |
739 |
printf("\nChecking file \"%s\"...\n", path); |
740 |
indexafile(sw, path); |
741 |
} |
742 |
else |
743 |
progwarnno("Invalid path '%s': ", path); |
744 |
} |
745 |
|
746 |
|
747 |
|
748 |
|
749 |
struct _indexing_data_source_def FileSystemIndexingDataSource = { |
750 |
"File-System", |
751 |
"fs", |
752 |
fs_indexpath, |
753 |
configModule_FS |
754 |
}; |