2 * Copyright (c) 2002 John Rochester
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer,
10 * in this position and unchanged.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 * notice, this list of conditions and the following disclaimer in the
13 * documentation and/or other materials provided with the distribution.
14 * 3. The name of the author may not be used to endorse or promote products
15 * derived from this software without specific prior written permission
17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
18 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
19 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
20 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
21 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
22 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
23 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
24 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
25 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
26 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28 * $FreeBSD: src/usr.bin/catman/catman.c,v 1.9 2003/06/10 02:18:00 ache Exp $
29 * $DragonFly: src/usr.bin/catman/catman.c,v 1.2 2004/05/17 17:44:26 cpressey Exp $
32 #include <sys/types.h>
34 #include <sys/param.h>
47 #define DEFAULT_MANPATH "/usr/share/man"
49 #define TOP_LEVEL_DIR 0 /* signifies a top-level man directory */
50 #define MAN_SECTION_DIR 1 /* signifies a man section directory */
51 #define UNKNOWN 2 /* signifies an unclassifiable directory */
53 #define TEST_EXISTS 0x01
55 #define TEST_FILE 0x04
56 #define TEST_READABLE 0x08
57 #define TEST_WRITABLE 0x10
58 #define TEST_EXECUTABLE 0x20
60 static int verbose
; /* -v flag: be verbose with warnings */
61 static int pretend
; /* -n, -p flags: print out what would be done
62 instead of actually doing it */
63 static int force
; /* -f flag: force overwriting all cat pages */
64 static int rm_junk
; /* -r flag: remove garbage pages */
65 static char *locale
; /* user's locale if -L is used */
66 static char *lang_locale
; /* short form of locale */
67 static int exit_code
; /* exit code to use when finished */
70 * -T argument for nroff
72 static const char *nroff_device
= "ascii";
75 * Mapping from locale to nroff device
77 static const char *locale_device
[] = {
79 "ISO8859-1", "latin1",
80 "ISO8859-15", "latin1",
84 #define BZ2_CMD "bzip2"
85 #define BZ2_EXT ".bz2"
86 #define BZ2CAT_CMD "bz"
90 enum Ziptype
{NONE
, BZIP
, GZIP
};
93 static gid_t gids
[NGROUPS_MAX
];
95 static int starting_dir
;
96 static char tmp_file
[MAXPATHLEN
];
100 * A hashtable is an array of chains composed of this entry structure.
106 struct hash_entry
*next
;
109 #define HASHTABLE_ALLOC 16384 /* allocation for hashtable (power of 2) */
110 #define HASH_MASK (HASHTABLE_ALLOC - 1)
112 static struct hash_entry
*visited
[HASHTABLE_ALLOC
];
113 static struct hash_entry
*links
[HASHTABLE_ALLOC
];
116 * Inserts a string into a hashtable keyed by inode & device number.
119 insert_hashtable(struct hash_entry
**table
, ino_t inode_number
,
120 dev_t device_number
, const char *data
)
122 struct hash_entry
*new_entry
;
123 struct hash_entry
**chain
;
125 new_entry
= malloc(sizeof(struct hash_entry
));
126 if (new_entry
== NULL
)
127 err(1, "can't insert into hashtable");
128 chain
= &table
[inode_number
& HASH_MASK
];
129 new_entry
->inode_number
= inode_number
;
130 new_entry
->device_number
= device_number
;
131 new_entry
->data
= data
;
132 new_entry
->next
= *chain
;
137 * Finds a string in a hashtable keyed by inode & device number.
140 find_hashtable(struct hash_entry
**table
, ino_t inode_number
,
143 struct hash_entry
*chain
;
145 chain
= table
[inode_number
& HASH_MASK
];
146 while (chain
!= NULL
) {
147 if (chain
->inode_number
== inode_number
&&
148 chain
->device_number
== device_number
)
156 trap_signal(int sig __unused
)
158 if (tmp_file
[0] != '\0')
164 * Deals with junk files in the man or cat section directories.
167 junk(const char *mandir
, const char *name
, const char *reason
)
170 fprintf(stderr
, "%s/%s: %s\n", mandir
, name
, reason
);
172 fprintf(stderr
, "rm %s/%s\n", mandir
, name
);
173 if (!pretend
&& unlink(name
) < 0)
174 warn("%s/%s", mandir
, name
);
179 * Returns TOP_LEVEL_DIR for .../man, MAN_SECTION_DIR for .../manXXX,
180 * and UNKNOWN for everything else.
183 directory_type(char *dir
)
188 p
= strrchr(dir
, '/');
189 if (p
== NULL
|| p
[1] != '\0')
197 if (strncmp(p
, "man", 3) == 0) {
200 return TOP_LEVEL_DIR
;
201 while (isalnum((unsigned char)*p
) || *p
== '_') {
203 return(MAN_SECTION_DIR
);
210 * Tests whether the given file name (without a preceding path)
211 * is a proper man page name (like "mk-amd-map.8.gz").
212 * Only alphanumerics and '_' are allowed after the last '.' and
213 * the last '.' can't be the first or last characters.
216 is_manpage_name(char *name
)
218 char *lastdot
= NULL
;
221 for (n
= name
; *n
!= '\0'; n
++) {
240 return(lastdot
> name
&& lastdot
+ 1 < n
);
244 is_bzipped(char *name
)
246 int len
= strlen(name
);
247 return(len
>= 5 && strcmp(&name
[len
- 4], BZ2_EXT
) == 0);
251 is_gzipped(char *name
)
253 int len
= strlen(name
);
254 return(len
>= 4 && strcmp(&name
[len
- 3], GZ_EXT
) == 0);
258 * Converts manXXX to catXXX.
261 get_cat_section(char *section
)
265 cat_section
= strdup(section
);
266 strncpy(cat_section
, "cat", 3);
271 * Converts .../man/manXXX to .../man.
274 get_mandir(char *section
)
279 slash
= strrchr(section
, '/');
280 mandir
= malloc(slash
- section
+ 1);
281 strncpy(mandir
, section
, slash
- section
);
282 mandir
[slash
- section
] = '\0';
287 * Tests to see if the given directory has already been visited.
290 already_visited(char *mandir
, char *dir
, int count_visit
)
294 if (stat(dir
, &st
) < 0) {
296 warn("%s/%s", mandir
, dir
);
302 if (find_hashtable(visited
, st
.st_ino
, st
.st_dev
) != NULL
) {
304 warnx("already visited %s/%s", mandir
, dir
);
306 warnx("already visited %s", dir
);
310 insert_hashtable(visited
, st
.st_ino
, st
.st_dev
, "");
315 * Returns a set of TEST_* bits describing a file's type and permissions.
316 * If mod_time isn't NULL, it will contain the file's modification time.
319 test_path(char *name
, time_t *mod_time
)
323 if (stat(name
, &test_st
) < 0)
325 result
= TEST_EXISTS
;
326 if (mod_time
!= NULL
)
327 *mod_time
= test_st
.st_mtime
;
328 if (S_ISDIR(test_st
.st_mode
))
330 else if (S_ISREG(test_st
.st_mode
))
332 if (test_st
.st_uid
== uid
) {
333 test_st
.st_mode
>>= 6;
336 for (i
= 0; i
< ngids
; i
++) {
337 if (test_st
.st_gid
== gids
[i
]) {
338 test_st
.st_mode
>>= 3;
343 if (test_st
.st_mode
& S_IROTH
)
344 result
|= TEST_READABLE
;
345 if (test_st
.st_mode
& S_IWOTH
)
346 result
|= TEST_WRITABLE
;
347 if (test_st
.st_mode
& S_IXOTH
)
348 result
|= TEST_EXECUTABLE
;
353 * Checks whether a file is a symbolic link.
356 is_symlink(char *path
)
360 return(lstat(path
, &st
) >= 0 && S_ISLNK(st
.st_mode
));
364 * Tests to see if the given directory can be written to.
367 check_writable(char *mandir
)
369 if (verbose
&& !(test_path(mandir
, NULL
) & TEST_WRITABLE
))
370 fprintf(stderr
, "%s: not writable - will only be able to write "
371 "to existing cat directories\n", mandir
);
375 * If the directory exists, attempt to make it writable, otherwise
376 * attempt to create it.
379 make_writable_dir(char *mandir
, char *dir
)
383 if ((test
= test_path(dir
, NULL
)) != 0) {
384 if (!(test
& TEST_WRITABLE
) && chmod(dir
, 0755) < 0) {
385 warn("%s/%s: chmod", mandir
, dir
);
390 if (verbose
|| pretend
)
391 fprintf(stderr
, "mkdir %s\n", dir
);
394 if (mkdir(dir
, 0755) < 0) {
395 warn("%s/%s: mkdir", mandir
, dir
);
405 * Processes a single man page source by using nroff to create
406 * the preformatted cat page.
409 process_page(char *mandir
, char *src
, char *cat
, enum Ziptype zipped
)
411 int src_test
, cat_test
;
412 time_t src_mtime
, cat_mtime
;
413 char cmd
[MAXPATHLEN
];
416 const char *link_name
;
418 src_test
= test_path(src
, &src_mtime
);
419 if (!(src_test
& (TEST_FILE
|TEST_READABLE
))) {
420 if (!(src_test
& TEST_DIR
)) {
421 warnx("%s/%s: unreadable", mandir
, src
);
423 if (rm_junk
&& is_symlink(src
))
424 junk(mandir
, src
, "bogus symlink");
428 src_dev
= test_st
.st_dev
;
429 src_ino
= test_st
.st_ino
;
430 cat_test
= test_path(cat
, &cat_mtime
);
431 if (cat_test
& (TEST_FILE
|TEST_READABLE
)) {
432 if (!force
&& cat_mtime
>= src_mtime
) {
434 fprintf(stderr
, "\t%s/%s: up to date\n",
440 * Is the man page a link to one we've already processed?
442 if ((link_name
= find_hashtable(links
, src_ino
, src_dev
)) != NULL
) {
443 if (verbose
|| pretend
)
444 fprintf(stderr
, "%slink %s -> %s\n",
445 verbose
? "\t" : "", cat
, link_name
);
447 link(link_name
, cat
);
450 insert_hashtable(links
, src_ino
, src_dev
, strdup(cat
));
451 if (verbose
|| pretend
) {
452 fprintf(stderr
, "%sformat %s -> %s\n",
453 verbose
? "\t" : "", src
, cat
);
457 snprintf(tmp_file
, sizeof tmp_file
, "%s.tmp", cat
);
458 snprintf(cmd
, sizeof cmd
,
459 "%scat %s | tbl | nroff -T%s -man | col | %s > %s.tmp",
460 zipped
== BZIP
? BZ2CAT_CMD
: zipped
== GZIP
? GZCAT_CMD
: "",
462 zipped
== BZIP
? BZ2_CMD
: zipped
== GZIP
? GZ_CMD
: "cat",
464 if (system(cmd
) != 0)
465 err(1, "formatting pipeline");
466 if (rename(tmp_file
, cat
) < 0)
472 * Scan the man section directory for pages and process each one,
473 * then check for junk in the corresponding cat section.
476 scan_section(char *mandir
, char *section
, char *cat_section
)
478 struct dirent
**entries
;
479 char **expected
= NULL
;
485 char page_path
[MAXPATHLEN
];
486 char cat_path
[MAXPATHLEN
];
487 char zip_path
[MAXPATHLEN
];
490 * scan the man section directory for pages
492 npages
= scandir(section
, &entries
, NULL
, alphasort
);
494 warn("%s/%s", mandir
, section
);
498 if (verbose
|| rm_junk
) {
500 * Maintain a list of all cat pages that should exist,
501 * corresponding to existing man pages.
503 expected
= (char **) calloc(npages
, sizeof(char *));
505 for (i
= 0; i
< npages
; free(entries
[i
++])) {
506 page_name
= entries
[i
]->d_name
;
507 snprintf(page_path
, sizeof page_path
, "%s/%s", section
,
509 if (!is_manpage_name(page_name
)) {
510 if (!(test_path(page_path
, NULL
) & TEST_DIR
)) {
511 junk(mandir
, page_path
,
512 "invalid man page name");
516 zipped
= is_bzipped(page_name
) ? BZIP
:
517 is_gzipped(page_name
) ? GZIP
: NONE
;
518 if (zipped
!= NONE
) {
519 snprintf(cat_path
, sizeof cat_path
, "%s/%s",
520 cat_section
, page_name
);
521 if (expected
!= NULL
)
522 expected
[nexpected
++] = strdup(page_name
);
523 process_page(mandir
, page_path
, cat_path
, zipped
);
526 * We've got an uncompressed man page,
527 * check to see if there's a (preferred)
530 snprintf(zip_path
, sizeof zip_path
, "%s%s",
532 if (test_path(zip_path
, NULL
) != 0) {
533 junk(mandir
, page_path
,
534 "man page unused due to existing " GZ_EXT
);
538 "warning, %s is uncompressed\n",
541 snprintf(cat_path
, sizeof cat_path
, "%s/%s",
542 cat_section
, page_name
);
543 if (expected
!= NULL
) {
544 asprintf(&expected
[nexpected
++],
547 process_page(mandir
, page_path
, cat_path
, NONE
);
552 if (expected
== NULL
)
555 * scan cat sections for junk
557 npages
= scandir(cat_section
, &entries
, NULL
, alphasort
);
559 for (i
= 0; i
< npages
; free(entries
[i
++])) {
560 const char *junk_reason
;
563 page_name
= entries
[i
]->d_name
;
564 if (strcmp(page_name
, ".") == 0 || strcmp(page_name
, "..") == 0)
567 * Keep the index into the expected cat page list
568 * ahead of the name we've found.
570 while (e
< nexpected
&&
571 (cmp
= strcmp(page_name
, expected
[e
])) > 0)
575 /* we have an unexpected page */
576 if (!is_manpage_name(page_name
)) {
577 junk_reason
= "invalid cat page name";
578 } else if (!is_gzipped(page_name
) && e
+ 1 < nexpected
&&
579 strncmp(page_name
, expected
[e
+ 1], strlen(page_name
)) == 0 &&
580 strlen(expected
[e
+ 1]) == strlen(page_name
) + 3) {
581 junk_reason
= "cat page unused due to existing " GZ_EXT
;
583 junk_reason
= "cat page without man page";
584 snprintf(cat_path
, sizeof cat_path
, "%s/%s", cat_section
,
586 junk(mandir
, cat_path
, junk_reason
);
589 while (e
< nexpected
)
596 * Processes a single man section.
599 process_section(char *mandir
, char *section
)
603 if (already_visited(mandir
, section
, 1))
606 fprintf(stderr
, " section %s\n", section
);
607 cat_section
= get_cat_section(section
);
608 if (make_writable_dir(mandir
, cat_section
))
609 scan_section(mandir
, section
, cat_section
);
613 select_sections(struct dirent
*entry
)
615 return(directory_type(entry
->d_name
) == MAN_SECTION_DIR
);
619 * Processes a single top-level man directory. If section isn't NULL,
620 * it will only process that section sub-directory, otherwise it will
621 * process all of them.
624 process_mandir(char *dir_name
, char *section
)
626 fchdir(starting_dir
);
627 if (already_visited(NULL
, dir_name
, section
== NULL
))
629 check_writable(dir_name
);
631 fprintf(stderr
, "man directory %s\n", dir_name
);
633 fprintf(stderr
, "cd %s\n", dir_name
);
634 if (chdir(dir_name
) < 0) {
635 warn("%s: chdir", dir_name
);
639 if (section
!= NULL
) {
640 process_section(dir_name
, section
);
642 struct dirent
**entries
;
646 nsections
= scandir(".", &entries
, select_sections
, alphasort
);
648 warn("%s", dir_name
);
652 for (i
= 0; i
< nsections
; i
++) {
653 process_section(dir_name
, entries
[i
]->d_name
);
661 * Processes one argument, which may be a colon-separated list of
665 process_argument(const char *arg
)
673 err(1, "out of memory");
674 while ((dir
= strsep(&parg
, ":")) != NULL
) {
675 switch (directory_type(dir
)) {
677 if (locale
!= NULL
) {
678 asprintf(&mandir
, "%s/%s", dir
, locale
);
679 process_mandir(mandir
, NULL
);
681 if (lang_locale
!= NULL
) {
682 asprintf(&mandir
, "%s/%s", dir
,
684 process_mandir(mandir
, NULL
);
688 process_mandir(dir
, NULL
);
691 case MAN_SECTION_DIR
: {
692 mandir
= get_mandir(dir
);
693 process_mandir(mandir
, dir
);
697 warnx("%s: directory name not in proper man form", dir
);
705 determine_locale(void)
709 if ((locale
= setlocale(LC_CTYPE
, "")) == NULL
) {
710 warnx("-L option used, but no locale found\n");
713 sep
= strchr(locale
, '_');
714 if (sep
!= NULL
&& isupper(sep
[1]) && isupper(sep
[2]))
715 asprintf(&lang_locale
, "%.*s%s", sep
- locale
, locale
, &sep
[3]);
716 sep
= nl_langinfo(CODESET
);
717 if (sep
!= NULL
&& *sep
!= '\0' && strcmp(sep
, "US-ASCII") != 0) {
720 for (i
= 0; locale_device
[i
] != NULL
; i
+= 2) {
721 if (strcmp(sep
, locale_device
[i
]) == 0) {
722 nroff_device
= locale_device
[i
+ 1];
728 if (lang_locale
!= NULL
)
729 fprintf(stderr
, "short locale is %s\n", lang_locale
);
730 fprintf(stderr
, "nroff device is %s\n", nroff_device
);
737 fprintf(stderr
, "usage: %s [-fLnrv] [directories...]\n", getprogname());
742 main(int argc
, char **argv
)
746 if ((uid
= getuid()) == 0) {
747 fprintf(stderr
, "don't run %s as root, use:\n echo", argv
[0]);
748 for (optind
= 0; optind
< argc
; optind
++)
749 fprintf(stderr
, " %s", argv
[optind
]);
750 fprintf(stderr
, " | nice -5 su -m man\n");
753 while ((opt
= getopt(argc
, argv
, "vnfLrh")) != -1) {
775 ngids
= getgroups(NGROUPS_MAX
, gids
);
776 if ((starting_dir
= open(".", 0)) < 0)
779 signal(SIGINT
, trap_signal
);
780 signal(SIGHUP
, trap_signal
);
781 signal(SIGQUIT
, trap_signal
);
782 signal(SIGTERM
, trap_signal
);
783 if (optind
== argc
) {
784 const char *manpath
= getenv("MANPATH");
786 manpath
= DEFAULT_MANPATH
;
787 process_argument(manpath
);
789 while (optind
< argc
)
790 process_argument(argv
[optind
++]);