2 * Copyright (c) 2002 John Rochester
3 * Copyright (c) 2013-2014 Franco Fichtner <franco@lastsummer.de>
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer,
11 * in this position and unchanged.
12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution.
15 * 3. The name of the author may not be used to endorse or promote products
16 * derived from this software without specific prior written permission
18 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
19 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
20 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
21 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
22 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
23 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
24 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
25 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
26 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
27 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
29 * $FreeBSD: src/usr.bin/makewhatis/makewhatis.c,v 1.9 2002/09/04 23:29:04 dwmalone Exp $
33 #include <sys/types.h>
34 #include <sys/param.h>
35 #include <sys/queue.h>
44 #include <stringlist.h>
48 #define DEFAULT_MANPATH "/usr/share/man"
49 #define LINE_ALLOC 4096
51 static char blank[] = "";
54 * Information collected about each man page alias.
57 RB_ENTRY(page_alias) entry;
65 * Information collected about each unique man page.
68 RB_HEAD(page_alias_tree, page_alias) head;
69 RB_ENTRY(page_info) entry;
73 static RB_HEAD(page_info_tree, page_info) page_head = RB_INITIALIZER(&page_head);
76 * Sorts page info by inode number.
79 infosort(const struct page_info *a, const struct page_info *b)
81 return (memcmp(&a->inode, &b->inode, sizeof(a->inode)));
84 RB_PROTOTYPE(page_info_tree, page_info, entry, infosort);
85 RB_GENERATE(page_info_tree, page_info, entry, infosort);
88 * Sorts page alias first by suffix, then name.
91 aliassort(const struct page_alias *a, const struct page_alias *b)
93 int ret = strcmp(a->suffix, b->suffix);
98 return (strcmp(a->name, b->name));
101 RB_PROTOTYPE(page_alias_tree, page_alias, entry, aliassort);
102 RB_GENERATE(page_alias_tree, page_alias, entry, aliassort);
105 * An entry kept for each visited directory.
110 SLIST_ENTRY(visited_dir) next;
114 * an expanding string
117 char * content; /* the start of the buffer */
118 char * end; /* just past the end of the content */
119 char * last; /* the last allocated character */
123 * Removes the last amount characters from the sbuf.
125 #define sbuf_retract(sbuf, amount) \
126 ((sbuf)->end -= (amount))
128 * Returns the length of the sbuf content.
130 #define sbuf_length(sbuf) \
131 ((sbuf)->end - (sbuf)->content)
133 typedef char *edited_copy(char *from, char *to, int length);
135 static int append; /* -a flag: append to existing whatis */
136 static int verbose; /* -v flag: be verbose with warnings */
137 static int indent = 24; /* -i option: description indentation */
138 static const char *whatis_name="whatis";/* -n option: the name */
139 static char *common_output; /* -o option: the single output file */
140 static char *locale; /* user's locale if -L is used */
141 static char *lang_locale; /* short form of locale */
142 static const char *machine;
144 static int exit_code; /* exit code to use when finished */
145 static SLIST_HEAD(, visited_dir) visited_dirs =
146 SLIST_HEAD_INITIALIZER(visited_dirs);
149 * While the whatis line is being formed, it is stored in whatis_proto.
150 * When finished, it is reformatted into whatis_final and then appended
153 static struct sbuf *whatis_proto;
154 static struct sbuf *whatis_final;
155 static StringList *whatis_lines; /* collected output lines */
157 static char tmp_file[MAXPATHLEN]; /* path of temporary file, if any */
159 /* A set of possible names for the NAME man page section */
160 static const char *name_section_titles[] = {
161 "NAME", "Name", "NAMN", "BEZEICHNUNG", "\xcc\xbe\xbe\xce",
162 "\xee\xe1\xfa\xf7\xe1\xee\xe9\xe5", NULL
165 /* A subset of the mdoc(7) commands to ignore */
166 static char mdoc_commands[] = "ArDvErEvFlLiNmPa";
169 * Frees a struct page_info and its content.
172 free_page_info(struct page_info *info)
174 struct page_alias *alias;
176 while ((alias = RB_ROOT(&info->head))) {
177 RB_REMOVE(page_alias_tree, &info->head, alias);
178 free(alias->filename);
188 * Allocates and fills in a new struct page_alias given the
189 * full file name of the man page and its dirent.
190 * If the file is not a man page, nothing is added.
193 new_page_alias(struct page_info *info, char *filename, struct dirent *dirent)
196 struct page_alias *alias;
200 basename_length = strlen(dirent->d_name);
201 suffix = &dirent->d_name[basename_length];
202 gzipped = basename_length >= 4 &&
203 strcmp(&dirent->d_name[basename_length - 3], ".gz") == 0;
209 if (--suffix == dirent->d_name || !isalnum(*suffix)) {
213 warnx("%s: invalid man page name",
221 alias = malloc(sizeof(*alias));
226 alias->name = strdup(dirent->d_name);
227 alias->filename = strdup(filename);
228 alias->suffix = strdup(suffix);
229 alias->gzipped = gzipped;
231 if (!alias->name || !alias->filename || !alias->suffix) {
235 RB_INSERT(page_alias_tree, &info->head, alias);
239 * Reset an sbuf's length to 0.
242 sbuf_clear(struct sbuf *sbuf)
244 sbuf->end = sbuf->content;
248 * Allocate a new sbuf.
253 struct sbuf *sbuf = (struct sbuf *) malloc(sizeof(struct sbuf));
254 sbuf->content = malloc(LINE_ALLOC);
255 sbuf->last = sbuf->content + LINE_ALLOC - 1;
261 * Ensure that there is enough room in the sbuf for nchars more characters.
264 sbuf_need(struct sbuf *sbuf, int nchars)
267 size_t size, cntsize;
269 /* double the size of the allocation until the buffer is big enough */
270 while (sbuf->end + nchars > sbuf->last) {
271 size = sbuf->last + 1 - sbuf->content;
273 cntsize = sbuf->end - sbuf->content;
275 new_content = malloc(size);
276 memcpy(new_content, sbuf->content, cntsize);
278 sbuf->content = new_content;
279 sbuf->end = new_content + cntsize;
280 sbuf->last = new_content + size - 1;
285 * Appends a string of a given length to the sbuf.
288 sbuf_append(struct sbuf *sbuf, const char *text, int length)
291 sbuf_need(sbuf, length);
292 memcpy(sbuf->end, text, length);
298 * Appends a null-terminated string to the sbuf.
301 sbuf_append_str(struct sbuf *sbuf, char *text)
303 sbuf_append(sbuf, text, strlen(text));
307 * Appends an edited null-terminated string to the sbuf.
310 sbuf_append_edited(struct sbuf *sbuf, char *text, edited_copy copy)
312 int length = strlen(text);
314 sbuf_need(sbuf, length);
315 sbuf->end = copy(text, sbuf->end, length);
320 * Strips any of a set of chars from the end of the sbuf.
323 sbuf_strip(struct sbuf *sbuf, const char *set)
325 while (sbuf->end > sbuf->content && strchr(set, sbuf->end[-1]) != NULL)
330 * Returns the null-terminated string built by the sbuf.
333 sbuf_content(struct sbuf *sbuf)
336 return(sbuf->content);
340 trap_signal(int sig __unused)
342 if (tmp_file[0] != '\0')
348 * Attempts to open an output file. Returns NULL if unsuccessful.
351 open_output(char *name)
355 whatis_lines = sl_init();
357 char line[LINE_ALLOC];
359 output = fopen(name, "r");
360 if (output == NULL) {
365 while (fgets(line, sizeof line, output) != NULL) {
366 line[strlen(line) - 1] = '\0';
367 sl_add(whatis_lines, strdup(line));
370 if (common_output == NULL) {
371 snprintf(tmp_file, sizeof tmp_file, "%s.tmp", name);
374 output = fopen(name, "w");
375 if (output == NULL) {
384 linesort(const void *a, const void *b)
386 return(strcmp((*(const char * const *)a), (*(const char * const *)b)));
390 * Writes the unique sorted lines to the output file.
393 finish_output(FILE *output, char *name)
398 qsort(whatis_lines->sl_str, whatis_lines->sl_cur, sizeof(char *),
400 for (i = 0; i < whatis_lines->sl_cur; i++) {
401 char *line = whatis_lines->sl_str[i];
402 if (i > 0 && strcmp(line, prev) == 0)
409 sl_free(whatis_lines, 1);
410 if (common_output == NULL) {
411 rename(tmp_file, name);
417 open_whatis(char *mandir)
419 char filename[MAXPATHLEN];
421 snprintf(filename, sizeof filename, "%s/%s", mandir, whatis_name);
422 return(open_output(filename));
426 finish_whatis(FILE *output, char *mandir)
428 char filename[MAXPATHLEN];
430 snprintf(filename, sizeof filename, "%s/%s", mandir, whatis_name);
431 finish_output(output, filename);
435 * Tests to see if the given directory has already been visited.
438 already_visited(char *dir)
441 struct visited_dir *visit;
443 if (stat(dir, &st) < 0) {
448 SLIST_FOREACH(visit, &visited_dirs, next) {
449 if (visit->inode == st.st_ino &&
450 visit->device == st.st_dev) {
451 warnx("already visited %s", dir);
455 visit = (struct visited_dir *) malloc(sizeof(struct visited_dir));
456 visit->device = st.st_dev;
457 visit->inode = st.st_ino;
458 SLIST_INSERT_HEAD(&visited_dirs, visit, next);
463 * Removes trailing spaces from a string, returning a pointer to just
464 * beyond the new last character.
469 char *rhs = &str[strlen(str)];
470 while (--rhs > str && isspace(*rhs))
477 * Returns a pointer to the next non-space character in the string.
482 while (*s != '\0' && isspace(*s))
488 * Returns whether the string contains only digits.
491 only_digits(char *line)
493 if (!isdigit(*line++))
495 while (isdigit(*line))
497 return(*line == '\0');
501 * Returns whether the line is of one of the forms:
505 * assuming that section_start is ".Sh".
508 name_section_line(char *line, const char *section_start)
513 if (strncmp(line, section_start, 3) != 0)
515 line = skip_spaces(line + 3);
516 rhs = trim_rhs(line);
522 for (title = name_section_titles; *title != NULL; title++)
523 if (strcmp(*title, line) == 0)
529 * Copies characters while removing the most common nroff/troff
531 * \(em, \(mi, \s[+-N], \&
532 * \fF, \f(fo, \f[font]
533 * \*s, \*(st, \*[stringvar]
536 de_nroff_copy(char *from, char *to, int fromlen)
538 char *from_end = &from[fromlen];
539 while (from < from_end) {
544 if (strncmp(&from[1], "em", 2) == 0 ||
545 strncmp(&from[1], "mi", 2) == 0) {
553 while (isdigit(*from))
560 else if (*from == '[') {
561 while (*++from != ']' && from < from_end)
579 * Appends a string with the nroff formatting removed.
582 add_nroff(char *text)
584 sbuf_append_edited(whatis_proto, text, de_nroff_copy);
588 * Appends "name(suffix), " to whatis_final.
591 add_whatis_name(char *name, char *suffix)
594 sbuf_append_str(whatis_final, name);
595 sbuf_append(whatis_final, "(", 1);
596 sbuf_append_str(whatis_final, suffix);
597 sbuf_append(whatis_final, "), ", 3);
602 * Processes an old-style man(7) line. This ignores commands with only
603 * a single number argument.
606 process_man_line(char *line)
609 while (isalpha(*++line))
611 line = skip_spaces(line);
612 if (only_digits(line))
615 line = skip_spaces(line);
618 sbuf_append(whatis_proto, " ", 1);
628 process_mdoc_macro(char *line)
630 static const struct mdoc_text list[] = {
631 { ".At", "AT&T UNIX" },
632 { ".Bsx", "BSD/OS" },
634 { ".Dx", "DragonFly" },
635 { ".Fx", "FreeBSD" },
637 { ".Ox", "OpenBSD" },
642 for (i = 0; i < sizeof(list) / sizeof(list[0]); ++i) {
643 if (!strcmp(line, list[i].mdoc)) {
644 sbuf_append(whatis_proto, list[i].text,
645 strlen(list[i].text));
646 sbuf_append(whatis_proto, " ", 1);
655 * Processes a new-style mdoc(7) line.
658 process_mdoc_line(char *line)
662 char *line_end = &line[strlen(line)];
663 int orig_length = sbuf_length(whatis_proto);
668 if (line[0] != '.' || !isupper(line[1]) || !islower(line[2])) {
669 add_nroff(skip_spaces(line));
670 sbuf_append(whatis_proto, " ", 1);
673 if (process_mdoc_macro(line)) {
676 xref = strncmp(line, ".Xr", 3) == 0;
678 while ((line = skip_spaces(line)) < line_end) {
682 next = strchr(next, '"');
685 memmove(next, next + 1, strlen(next));
692 next = strpbrk(line, " \t");
697 if (isupper(*line) && islower(line[1]) && line[2] == '\0') {
698 if (strcmp(line, "Ns") == 0) {
703 if (strstr(mdoc_commands, line) != NULL) {
708 if (arg > 0 && strchr(",.:;?!)]", *line) == 0) {
710 sbuf_append(whatis_proto, "(", 1);
712 sbuf_append(whatis_proto, ")", 1);
716 sbuf_append(whatis_proto, " ", 1);
722 if (sbuf_length(whatis_proto) > orig_length)
723 sbuf_append(whatis_proto, " ", 1);
726 enum { STATE_UNKNOWN, STATE_MANSTYLE, STATE_MDOCNAME, STATE_MDOCDESC };
729 * Processes a man page source into a single whatis line and adds it
733 process_page(struct page_info *info)
739 int state = STATE_UNKNOWN;
740 struct page_alias *alias;
743 * Only read the page once for each inode. It's
744 * safe to assume that page->list is set.
746 alias = RB_MIN(page_alias_tree, &info->head);
749 fprintf(stderr, "\treading %s\n", alias->filename);
752 sbuf_clear(whatis_proto);
753 if ((in = gzopen(alias->filename, "r")) == NULL) {
754 warn("%s", alias->filename);
758 while (gzgets(in, buffer, sizeof buffer) != NULL) {
760 if (strncmp(line, ".\\\"", 3) == 0) /* ignore comments */
764 * haven't reached the NAME section yet.
767 if (name_section_line(line, ".SH"))
768 state = STATE_MANSTYLE;
769 else if (name_section_line(line, ".Sh"))
770 state = STATE_MDOCNAME;
773 * Inside an old-style .SH NAME section.
776 if (strncmp(line, ".SH", 3) == 0)
778 if (strncmp(line, ".SS", 3) == 0)
781 if (strcmp(line, ".") == 0)
783 if (strncmp(line, ".IX", 3) == 0) {
785 line = skip_spaces(line);
787 process_man_line(line);
790 * Inside a new-style .Sh NAME section (the .Nm part).
794 if (strncmp(line, ".Nm", 3) == 0) {
795 process_mdoc_line(line);
798 if (strcmp(line, ".") == 0)
800 sbuf_append(whatis_proto, "- ", 2);
801 state = STATE_MDOCDESC;
805 * Inside a new-style .Sh NAME section (after the .Nm-s).
808 if (strncmp(line, ".Sh", 3) == 0)
811 if (strcmp(line, ".") == 0)
813 process_mdoc_line(line);
819 sbuf_strip(whatis_proto, " \t.-");
820 line = sbuf_content(whatis_proto);
822 * line now contains the appropriate data, but without
823 * the proper indentation or the section appended to each name.
825 descr = strstr(line, " - ");
827 descr = strchr(line, ' ');
831 "\tignoring junk description \"%s\"\n",
840 sbuf_clear(whatis_final);
841 RB_FOREACH(alias, page_alias_tree, &info->head) {
843 * This won't append names stored in `line'.
844 * The reason for that is that we cannot be sure
845 * which section they belong to unless we have
846 * a real alias (via MLINKS) in this list.
848 add_whatis_name(alias->name, alias->suffix);
851 char *arg, *text = line;
857 * See if there are names in the manual that
858 * are not in the alias list provided by the
859 * MLINKS. We may want to add those as well.
861 RB_FOREACH(alias, page_alias_tree, &info->head)
862 sl_add(names, alias->name);
866 text = strchr(text, ',');
869 if (!sl_find(names, arg)) {
870 fprintf(stderr, "\tpage alias \"%s\" "
871 "may be missing\n", arg);
881 sbuf_retract(whatis_final, 2); /* remove last ", " */
882 while (sbuf_length(whatis_final) < indent)
883 sbuf_append(whatis_final, " ", 1);
884 sbuf_append(whatis_final, " - ", 3);
885 sbuf_append_str(whatis_final, skip_spaces(descr));
886 sl_add(whatis_lines, strdup(sbuf_content(whatis_final)));
890 * Processes a single man section.
893 process_section(char *section_dir)
895 struct dirent **entries;
896 struct page_info *info;
901 fprintf(stderr, " %s\n", section_dir);
904 * scan the man section directory for pages
906 nentries = scandir(section_dir, &entries, NULL, alphasort);
908 warn("%s", section_dir);
914 * collect information about man pages
916 for (i = 0; i < nentries; i++) {
917 struct page_info ref;
921 if (asprintf(&filename, "%s/%s", section_dir,
922 entries[i]->d_name) < 0) {
926 if (stat(filename, &st) < 0) {
927 warn("%s", filename);
928 goto process_section_next;
931 if (!S_ISREG(st.st_mode)) {
932 if (verbose && !S_ISDIR(st.st_mode))
933 warnx("%s: not a regular file", filename);
934 goto process_section_next;
937 ref.inode = st.st_ino;
939 info = RB_FIND(page_info_tree, &page_head, &ref);
941 info = malloc(sizeof(*info));
946 bzero(info, sizeof(*info));
947 info->inode = st.st_ino;
948 RB_INIT(&info->head);
950 RB_INSERT(page_info_tree, &page_head, info);
953 new_page_alias(info, filename, entries[i]);
955 process_section_next:
964 * Returns whether the directory entry is a man page section.
967 select_sections(const struct dirent *entry)
969 const char *p = &entry->d_name[3];
971 if (strncmp(entry->d_name, "man", 3) != 0)
981 * Processes a single top-level man directory by finding all the
982 * sub-directories named man* and processing each one in turn.
985 process_mandir(char *dir_name)
987 struct dirent **entries;
988 struct page_info *info;
994 if (already_visited(dir_name))
997 fprintf(stderr, "man directory %s\n", dir_name);
998 nsections = scandir(dir_name, &entries, select_sections, alphasort);
1000 warn("%s", dir_name);
1004 if (common_output == NULL && (fp = open_whatis(dir_name)) == NULL)
1006 for (i = 0; i < nsections; i++) {
1007 char section_dir[MAXPATHLEN];
1008 snprintf(section_dir, sizeof section_dir, "%s/%s", dir_name,
1009 entries[i]->d_name);
1010 process_section(section_dir);
1011 snprintf(section_dir, sizeof section_dir, "%s/%s/%s", dir_name,
1012 entries[i]->d_name, machine);
1013 if (stat(section_dir, &st) == 0 && S_ISDIR(st.st_mode))
1014 process_section(section_dir);
1020 * process and free all pages
1022 while ((info = RB_ROOT(&page_head))) {
1023 RB_REMOVE(page_info_tree, &page_head, info);
1025 free_page_info(info);
1028 if (common_output == NULL)
1029 finish_whatis(fp, dir_name);
1033 * Processes one argument, which may be a colon-separated list of
1037 process_argument(const char *arg)
1045 err(1, "out of memory");
1046 while ((dir = strsep(&parg, ":")) != NULL) {
1047 if (locale != NULL) {
1048 asprintf(&mandir, "%s/%s", dir, locale);
1049 process_mandir(mandir);
1051 if (lang_locale != NULL) {
1052 asprintf(&mandir, "%s/%s", dir, lang_locale);
1053 process_mandir(mandir);
1057 process_mandir(dir);
1065 main(int argc, char **argv)
1070 while ((opt = getopt(argc, argv, "ai:n:o:vL")) != -1) {
1076 indent = atoi(optarg);
1079 whatis_name = optarg;
1082 common_output = optarg;
1088 locale = getenv("LC_ALL");
1090 locale = getenv("LC_CTYPE");
1092 locale = getenv("LANG");
1093 if (locale != NULL) {
1094 char *sep = strchr(locale, '_');
1095 if (sep != NULL && isupper(sep[1]) &&
1097 asprintf(&lang_locale, "%.*s%s",
1098 (int)(sep - locale),
1104 fprintf(stderr, "usage: %s [-a] [-i indent] [-n name] [-o output_file] [-v] [-L] [directories...]\n", argv[0]);
1109 signal(SIGINT, trap_signal);
1110 signal(SIGHUP, trap_signal);
1111 signal(SIGQUIT, trap_signal);
1112 signal(SIGTERM, trap_signal);
1113 SLIST_INIT(&visited_dirs);
1114 whatis_proto = new_sbuf();
1115 whatis_final = new_sbuf();
1117 if ((machine = getenv("MACHINE")) == NULL)
1120 if (common_output != NULL && (fp = open_output(common_output)) == NULL)
1121 err(1, "%s", common_output);
1122 if (optind == argc) {
1123 const char *manpath = getenv("MANPATH");
1124 if (manpath == NULL)
1125 manpath = DEFAULT_MANPATH;
1126 process_argument(manpath);
1128 while (optind < argc)
1129 process_argument(argv[optind++]);
1131 if (common_output != NULL)
1132 finish_output(fp, common_output);