1 /* locate -- search databases for filenames that match patterns
2 Copyright (C) 1994, 96, 98, 99, 2000, 2003 Free Software Foundation, Inc.
4 This program is free software; you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation; either version 2, or (at your option)
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
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program; if not, write to the Free Software
16 Foundation, Inc., 9 Temple Place - Suite 330, Boston, MA 02111-1307,
20 /* Usage: locate [options] pattern...
22 Scan a pathname list for the full pathname of a file, given only
23 a piece of the name (possibly containing shell globbing metacharacters).
24 The list has been processed with front-compression, which reduces
25 the list size by a factor of 4-5.
26 Recognizes two database formats, old and new. The old format is
27 bigram coded, which reduces space by a further 20-25% and uses the
28 following encoding of the database bytes:
30 0-28 likeliest differential counts + offset (14) to make nonnegative
31 30 escape code for out-of-range count to follow in next halfword
32 128-255 bigram codes (the 128 most common, as determined by `updatedb')
33 32-127 single character (printable) ASCII remainder
35 Uses a novel two-tiered string search technique:
37 First, match a metacharacter-free subpattern and a partial pathname
38 BACKWARDS to avoid full expansion of the pathname list.
39 The time savings is 40-50% over forward matching, which cannot efficiently
40 handle overlapped search patterns and compressed path remainders.
42 Then, match the actual shell glob-style regular expression (if in this form)
43 against the candidate pathnames using the slower shell filename
46 Described more fully in Usenix ;login:, Vol 8, No 1,
47 February/March, 1983, p. 8.
49 Written by James A. Woods <jwoods@adobe.com>.
50 Modified by David MacKenzie <djm@gnu.ai.mit.edu>. */
55 #include <sys/types.h>
64 #if defined(HAVE_STRING_H) || defined(STDC_HEADERS)
87 # define _(Text) gettext (Text)
90 #define textdomain(Domain)
91 #define bindtextdomain(Package, Directory)
94 # define N_(String) gettext_noop (String)
96 /* We used to use (String) instead of just String, but HP-UX 11.23 for
97 * ia64 has what seems to be a compiler bug, because it refuses input
98 * like: static const char buf[] = ("string");i
100 # define N_(String) String
103 #include "locatedb.h"
106 /* Note that this evaluates C many times. */
108 # define TOUPPER(Ch) toupper (Ch)
109 # define TOLOWER(Ch) tolower (Ch)
111 # define TOUPPER(Ch) (islower (Ch) ? toupper (Ch) : (Ch))
112 # define TOLOWER(Ch) (isupper (Ch) ? tolower (Ch) : (Ch))
115 typedef enum {false, true} boolean
;
117 /* Warn if a database is older than this. 8 days allows for a weekly
118 update that takes up to a day to perform. */
119 #define WARN_NUMBER_UNITS (8)
120 /* Printable name of units used in WARN_SECONDS */
121 static const char warn_name_units
[] = N_("days");
122 #define SECONDS_PER_UNIT (60 * 60 * 24)
124 #define WARN_SECONDS ((SECONDS_PER_UNIT) * (WARN_NUMBER_UNITS))
126 /* Check for existence of files before printing them out? */
127 int check_existence
= 0;
129 char *next_element ();
134 /* Read in a 16-bit int, high byte first (network byte order). */
144 x
|= (fgetc (fp
) & 0xff);
148 /* Return a pointer to the last character in a static copy of the last
149 glob-free subpattern in NAME,
150 with '\0' prepended for a fast backwards pre-match. */
153 last_literal_end (name
)
156 static char *globfree
= NULL
; /* A copy of the subpattern in NAME. */
157 static size_t gfalloc
= 0; /* Bytes allocated for `globfree'. */
158 register char *subp
; /* Return value. */
159 register char *p
; /* Search location in NAME. */
161 /* Find the end of the subpattern.
162 Skip trailing metacharacters and [] ranges. */
163 for (p
= name
+ strlen (name
) - 1; p
>= name
&& strchr ("*?]", *p
) != NULL
;
167 while (p
>= name
&& *p
!= '[')
173 if (p
- name
+ 3 > gfalloc
)
175 gfalloc
= p
- name
+ 3 + 64; /* Room to grow. */
176 globfree
= xrealloc (globfree
, gfalloc
);
181 /* If the pattern has only metacharacters, make every path match the
182 subpattern, so it gets checked the slow way. */
183 if (p
== name
&& strchr ("?*[]", *p
) != NULL
)
188 /* Find the start of the metacharacter-free subpattern. */
189 for (endmark
= p
; p
>= name
&& strchr ("]*?", *p
) == NULL
; p
--)
191 /* Copy the subpattern into globfree. */
192 for (++p
; p
<= endmark
; )
195 *subp
-- = '\0'; /* Null terminate, though it's not needed. */
202 * Read bytes from FP into the buffer at offset OFFSET in (*BUF),
203 * until we reach DELIMITER or end-of-file. We reallocate the buffer
204 * as necessary, altering (*BUF) and (*SIZ) as appropriate. No assumption
205 * is made regarding the content of the data (i.e. the implementation is
206 * 8-bit clean, the only delimiter is DELIMITER).
208 * Written Fri May 23 18:41:16 2003 by James Youngman, because getstr()
209 * has been removed from gnulib.
211 * We call the function locate_read_str() to avoid a name clash with the curses
214 static int locate_read_str(char **buf
, size_t *siz
, FILE *fp
, int delimiter
, int offs
)
220 nread
= getdelim(&p
, &sz
, delimiter
, fp
);
225 needed
= offs
+ nread
;
228 char *pnew
= realloc(*buf
, needed
);
231 return -1; /* FAIL */
239 memcpy((*buf
)+offs
, p
, nread
);
246 /* Print the entries in DBFILE that match shell globbing pattern PATHPART.
247 Return the number of entries printed. */
250 locate (pathpart
, dbfile
, ignore_case
)
251 char *pathpart
, *dbfile
;
254 /* The pathname database. */
258 /* Number of bytes read from an entry. */
261 /* true if PATHPART contains globbing metacharacters. */
263 /* The end of the last glob-free subpattern in PATHPART. */
266 /* The current input database entry. */
268 /* Amount allocated for it. */
271 /* The length of the prefix shared with the previous database entry. */
273 /* Where in `path' to stop the backward search for the last character
274 in the subpattern. Set according to `count'. */
277 /* true if we found a fast match (of patend) on the previous path. */
278 boolean prev_fast_match
= false;
279 /* The return value. */
282 /* true if reading a bigram-encoded database. */
283 boolean old_format
= false;
284 /* For the old database format,
285 the first and second characters of the most common bigrams. */
286 char bigram1
[128], bigram2
[128];
288 /* To check the age of the database. */
292 if (stat (dbfile
, &st
) || (fp
= fopen (dbfile
, "r")) == NULL
)
294 error (0, errno
, "%s", dbfile
);
298 if (now
- st
.st_mtime
> WARN_SECONDS
)
301 warning: database `fred' is more than 8 days old */
302 error (0, 0, _("warning: database `%s' is more than %d %s old"),
303 dbfile
, WARN_NUMBER_UNITS
, _(warn_name_units
));
306 pathsize
= 1026; /* Increased as necessary by locate_read_str. */
307 path
= xmalloc (pathsize
);
309 nread
= fread (path
, 1, sizeof (LOCATEDB_MAGIC
), fp
);
310 if (nread
!= sizeof (LOCATEDB_MAGIC
)
311 || memcmp (path
, LOCATEDB_MAGIC
, sizeof (LOCATEDB_MAGIC
)))
314 /* Read the list of the most common bigrams in the database. */
316 for (i
= 0; i
< 128; i
++)
318 bigram1
[i
] = getc (fp
);
319 bigram2
[i
] = getc (fp
);
324 /* If we ignore case,
325 convert it to lower first so we don't have to do it every time */
327 for (patend
=pathpart
;*patend
;++patend
){
328 *patend
=TOLOWER(*patend
);
333 globflag
= strchr (pathpart
, '*') || strchr (pathpart
, '?')
334 || strchr (pathpart
, '[');
336 patend
= last_literal_end (pathpart
);
341 register char *s
; /* Scan the path we read in. */
345 /* Get the offset in the path where this path info starts. */
346 if (c
== LOCATEDB_OLD_ESCAPE
)
347 count
+= getw (fp
) - LOCATEDB_OLD_OFFSET
;
349 count
+= c
- LOCATEDB_OLD_OFFSET
;
351 /* Overlay the old path with the remainder of the new. */
352 for (s
= path
+ count
; (c
= getc (fp
)) > LOCATEDB_OLD_ESCAPE
;)
354 *s
++ = c
; /* An ordinary character. */
357 /* Bigram markers have the high bit set. */
366 if (c
== LOCATEDB_ESCAPE
)
367 count
+= (short)get_short (fp
);
373 if (count
> strlen(path
))
375 /* This should not happen generally , but since we're
376 * reading in data which is outside our control, we
379 error(1, 0, _("locate database `%s' is corrupt or invalid"), dbfile
);
382 /* Overlay the old path with the remainder of the new. */
383 nread
= locate_read_str (&path
, &pathsize
, fp
, 0, count
);
387 s
= path
+ count
+ nread
- 2; /* Move to the last char in path. */
388 assert (s
[0] != '\0');
389 assert (s
[1] == '\0'); /* Our terminator. */
390 assert (s
[2] == '\0'); /* Added by locate_read_str. */
393 /* If the previous path matched, scan the whole path for the last
394 char in the subpattern. If not, the shared prefix doesn't match
395 the pattern, so don't scan it for the last char. */
396 cutoff
= prev_fast_match
? path
: path
+ count
;
398 /* Search backward starting at the end of the path we just read in,
399 for the character at the end of the last glob-free subpattern
401 for(prev_fast_match
=false; s
>=cutoff
; s
--)
403 char *s2
; /* Scan the path we read in. */
404 register char *p2
; /* Scan `patend'. */
406 /* Fast first char check. */
409 if(TOLOWER(*s
)!=*patend
)
416 for(s2
=s
-1, p2
=patend
-1; *p2
!='\0' && TOLOWER(*s2
)==*p2
; s2
--, p2
--);
418 for(s2
=s
-1, p2
=patend
-1; *p2
!='\0' && *s2
==*p2
; s2
--, p2
--);
422 /* Success on the fast match. Compare the whole pattern
423 if it contains globbing characters. */
424 prev_fast_match
= true;
429 if(fnmatch(pathpart
,basename(path
),FNM_CASEFOLD
)!=0)
433 if(fnmatch(pathpart
,basename(path
),0)!=0)
436 if(check_existence
&& stat(path
,&st
)!=0)
446 error (0, errno
, "%s", dbfile
);
449 if (fclose (fp
) == EOF
)
451 error (0, errno
, "%s", dbfile
);
458 extern char *version_string
;
460 /* The name this program was run with. */
467 fprintf (stream
, _("\
468 Usage: %s [-d path | --database=path] [-e | --existing]\n\
469 [-i | --ignore-case] [--version] [--help] pattern...\n"),
471 fputs (_("\nReport bugs to <bug-findutils@gnu.org>.\n"), stream
);
474 static struct option
const longopts
[] =
476 {"database", required_argument
, NULL
, 'd'},
477 {"existing", no_argument
, NULL
, 'e'},
478 {"ignore-case", no_argument
, NULL
, 'i'},
479 {"help", no_argument
, NULL
, 'h'},
480 {"version", no_argument
, NULL
, 'v'},
481 {NULL
, no_argument
, NULL
, 0}
490 int fnmatch_flags
= 0;
494 program_name
= argv
[0];
496 #ifdef HAVE_SETLOCALE
497 setlocale (LC_ALL
, "");
499 bindtextdomain (PACKAGE
, LOCALEDIR
);
500 textdomain (PACKAGE
);
502 dbpath
= getenv ("LOCATE_PATH");
508 while ((optc
= getopt_long (argc
, argv
, "d:ei", longopts
, (int *) 0)) != -1)
521 fnmatch_flags
|= FNM_CASEFOLD
;
529 printf (_("GNU locate version %s\n"), version_string
);
544 for (; optind
< argc
; optind
++)
547 next_element (dbpath
); /* Initialize. */
548 while ((e
= next_element ((char *) NULL
)) != NULL
)
549 found
|= locate (argv
[optind
], e
, ignore_case
);