1 /* $OpenBSD: mansearch.c,v 1.50 2016/07/09 15:23:36 schwarze Exp $ */
3 * Copyright (c) 2012 Kristaps Dzonsons <kristaps@bsd.lv>
4 * Copyright (c) 2013-2017 Ingo Schwarze <schwarze@openbsd.org>
6 * Permission to use, copy, modify, and distribute this software for any
7 * purpose with or without fee is hereby granted, provided that the above
8 * copyright notice and this permission notice appear in all copies.
10 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES
11 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
12 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR
13 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
14 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
15 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
16 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
21 #include <sys/types.h>
40 #include "mandoc_aux.h"
41 #include "mandoc_ohash.h"
43 #include "mansearch.h"
48 struct dbm_match match; /* Match type and expression. */
49 uint64_t bits; /* Type mask. */
50 /* Used for OR and AND groups: */
51 struct expr *next; /* Next child in the parent group. */
52 struct expr *child; /* First child in this group. */
53 enum { EXPR_TERM, EXPR_OR, EXPR_AND } type;
56 const char *const mansearch_keynames[KEY_MAX] = {
57 "arch", "sec", "Xr", "Ar", "Fa", "Fl", "Dv", "Fn",
58 "Ic", "Pa", "Cm", "Li", "Em", "Cd", "Va", "Ft",
59 "Tn", "Er", "Ev", "Sy", "Sh", "In", "Ss", "Ox",
60 "An", "Mt", "St", "Bx", "At", "Nx", "Fx", "Lk",
61 "Ms", "Bsx", "Dx", "Rs", "Vt", "Lb", "Nm", "Nd"
65 static struct ohash *manmerge(struct expr *, struct ohash *);
66 static struct ohash *manmerge_term(struct expr *, struct ohash *);
67 static struct ohash *manmerge_or(struct expr *, struct ohash *);
68 static struct ohash *manmerge_and(struct expr *, struct ohash *);
69 static char *buildnames(const struct dbm_page *);
70 static char *buildoutput(size_t, struct dbm_page *);
71 static size_t lstlen(const char *, size_t);
72 static void lstcat(char *, size_t *, const char *, const char *);
73 static int lstmatch(const char *, const char *);
74 static struct expr *exprcomp(const struct mansearch *,
75 int, char *[], int *);
76 static struct expr *expr_and(const struct mansearch *,
77 int, char *[], int *);
78 static struct expr *exprterm(const struct mansearch *,
79 int, char *[], int *);
80 static void exprfree(struct expr *);
81 static int manpage_compare(const void *, const void *);
85 mansearch(const struct mansearch *search,
86 const struct manpaths *paths,
87 int argc, char *argv[],
88 struct manpage **res, size_t *sz)
93 struct dbm_page *page;
94 struct manpage *mpage;
96 size_t cur, i, maxres, outkey;
98 int argi, chdir_status, getcwd_status, im;
101 if ((e = exprcomp(search, argc, argv, &argi)) == NULL) {
110 if (search->outkey != NULL)
111 for (im = 0; im < KEY_MAX; im++)
112 if (0 == strcasecmp(search->outkey,
113 mansearch_keynames[im])) {
119 * Remember the original working directory, if possible.
120 * This will be needed if the second or a later directory
121 * is given as a relative path.
122 * Do not error out if the current directory is not
123 * searchable: Maybe it won't be needed after all.
126 if (getcwd(buf, PATH_MAX) == NULL) {
128 (void)strlcpy(buf, strerror(errno), sizeof(buf));
133 * Loop over the directories (containing databases) for us to
135 * Don't let missing/bad databases/directories phase us.
136 * In each, try to open the resident database and, if it opens,
137 * scan it for our match expression.
141 for (i = 0; i < paths->sz; i++) {
142 if (chdir_status && paths->paths[i][0] != '/') {
143 if ( ! getcwd_status) {
144 warnx("%s: getcwd: %s", paths->paths[i], buf);
146 } else if (chdir(buf) == -1) {
151 if (chdir(paths->paths[i]) == -1) {
152 warn("%s", paths->paths[i]);
157 if (dbm_open(MANDOC_DB) == -1) {
159 warn("%s/%s", paths->paths[i], MANDOC_DB);
163 if ((htab = manmerge(e, NULL)) == NULL) {
168 for (rp = ohash_first(htab, &slot); rp != NULL;
169 rp = ohash_next(htab, &slot)) {
170 page = dbm_page_get(rp->page);
172 if (lstmatch(search->sec, page->sect) == 0 ||
173 lstmatch(search->arch, page->arch) == 0)
176 if (cur + 1 > maxres) {
178 *res = mandoc_reallocarray(*res,
179 maxres, sizeof(**res));
182 mandoc_asprintf(&mpage->file, "%s/%s",
183 paths->paths[i], page->file + 1);
184 mpage->names = buildnames(page);
185 mpage->output = buildoutput(outkey, page);
187 mpage->bits = rp->bits;
188 mpage->sec = *page->sect - '0';
189 if (mpage->sec < 0 || mpage->sec > 9)
191 mpage->form = *page->file;
200 * In man(1) mode, prefer matches in earlier trees
201 * over matches in later trees.
204 if (cur && search->firstmatch)
207 qsort(*res, cur, sizeof(struct manpage), manpage_compare);
208 if (chdir_status && getcwd_status && chdir(buf) == -1)
216 * Merge the results for the expression tree rooted at e
217 * into the the result list htab.
219 static struct ohash *
220 manmerge(struct expr *e, struct ohash *htab)
224 return manmerge_term(e, htab);
226 return manmerge_or(e->child, htab);
228 return manmerge_and(e->child, htab);
234 static struct ohash *
235 manmerge_term(struct expr *e, struct ohash *htab)
237 struct dbm_res res, *rp;
243 htab = mandoc_malloc(sizeof(*htab));
244 mandoc_ohash_init(htab, 4, offsetof(struct dbm_res, page));
247 for (im = 0, ib = 1; im < KEY_MAX; im++, ib <<= 1) {
248 if ((e->bits & ib) == 0)
253 dbm_page_byarch(&e->match);
256 dbm_page_bysect(&e->match);
259 dbm_page_byname(&e->match);
262 dbm_page_bydesc(&e->match);
265 dbm_page_bymacro(im - 2, &e->match);
270 * When hashing for deduplication, use the unique
271 * page ID itself instead of a hash function;
272 * that is quite efficient.
276 res = dbm_page_next();
279 slot = ohash_lookup_memory(htab,
280 (char *)&res, sizeof(res.page), res.page);
281 if ((rp = ohash_find(htab, slot)) != NULL) {
282 rp->bits |= res.bits;
285 rp = mandoc_malloc(sizeof(*rp));
287 ohash_insert(htab, slot, rp);
293 static struct ohash *
294 manmerge_or(struct expr *e, struct ohash *htab)
297 htab = manmerge(e, htab);
303 static struct ohash *
304 manmerge_and(struct expr *e, struct ohash *htab)
306 struct ohash *hand, *h1, *h2;
308 unsigned int slot1, slot2;
310 /* Evaluate the first term of the AND clause. */
312 hand = manmerge(e, NULL);
314 while ((e = e->next) != NULL) {
316 /* Evaluate the next term and prepare for ANDing. */
318 h2 = manmerge(e, NULL);
319 if (ohash_entries(h2) < ohash_entries(hand)) {
324 hand = mandoc_malloc(sizeof(*hand));
325 mandoc_ohash_init(hand, 4, offsetof(struct dbm_res, page));
327 /* Keep all pages that are in both result sets. */
329 for (res = ohash_first(h1, &slot1); res != NULL;
330 res = ohash_next(h1, &slot1)) {
331 if (ohash_find(h2, ohash_lookup_memory(h2,
332 (char *)res, sizeof(res->page),
336 ohash_insert(hand, ohash_lookup_memory(hand,
337 (char *)res, sizeof(res->page),
341 /* Discard the merged results. */
343 for (res = ohash_first(h2, &slot2); res != NULL;
344 res = ohash_next(h2, &slot2))
352 /* Merge the result of the AND into htab. */
357 for (res = ohash_first(hand, &slot1); res != NULL;
358 res = ohash_next(hand, &slot1)) {
359 slot2 = ohash_lookup_memory(htab,
360 (char *)res, sizeof(res->page), res->page);
361 if (ohash_find(htab, slot2) == NULL)
362 ohash_insert(htab, slot2, res);
367 /* Discard the merged result. */
375 mansearch_free(struct manpage *res, size_t sz)
379 for (i = 0; i < sz; i++) {
388 manpage_compare(const void *vp1, const void *vp2)
390 const struct manpage *mp1, *mp2;
395 return (diff = mp2->bits - mp1->bits) ? diff :
396 (diff = mp1->sec - mp2->sec) ? diff :
397 strcasecmp(mp1->names, mp2->names);
401 buildnames(const struct dbm_page *page)
406 sz = lstlen(page->name, 2) + 1 + lstlen(page->sect, 2) +
407 (page->arch == NULL ? 0 : 1 + lstlen(page->arch, 2)) + 2;
408 buf = mandoc_malloc(sz);
410 lstcat(buf, &i, page->name, ", ");
412 lstcat(buf, &i, page->sect, ", ");
413 if (page->arch != NULL) {
415 lstcat(buf, &i, page->arch, ", ");
424 * Count the buffer space needed to print the NUL-terminated
425 * list of NUL-terminated strings, when printing sep separator
426 * characters between strings.
429 lstlen(const char *cp, size_t sep)
433 for (sz = 0;; sz++) {
438 } else if (cp[0] < ' ')
446 * Print the NUL-terminated list of NUL-terminated strings
447 * into the buffer, seperating strings with sep.
450 lstcat(char *buf, size_t *i, const char *cp, const char *sep)
461 } else if (cp[0] >= ' ')
468 * Return 1 if the string *want occurs in any of the strings
469 * in the NUL-terminated string list *have, or 0 otherwise.
470 * If either argument is NULL or empty, assume no filtering
471 * is desired and return 1.
474 lstmatch(const char *want, const char *have)
476 if (want == NULL || have == NULL || *have == '\0')
478 while (*have != '\0') {
479 if (strcasestr(have, want) != NULL)
481 have = strchr(have, '\0') + 1;
487 * Build a list of values taken by the macro im in the manual page.
490 buildoutput(size_t im, struct dbm_page *page)
492 const char *oldoutput, *sep, *input;
493 char *output, *newoutput, *value;
498 return mandoc_strdup(page->desc);
516 sz = lstlen(input, 3) + 1;
517 output = mandoc_malloc(sz);
519 lstcat(output, &i, input, " # ");
526 dbm_macro_bypage(im - 2, page->addr);
527 while ((value = dbm_macro_next()) != NULL) {
528 if (output == NULL) {
535 mandoc_asprintf(&newoutput, "%s%s%s", oldoutput, sep, value);
543 * Compile a set of string tokens into an expression.
544 * Tokens in "argv" are assumed to be individual expression atoms (e.g.,
545 * "(", "foo=bar", etc.).
548 exprcomp(const struct mansearch *search, int argc, char *argv[], int *argi)
550 struct expr *parent, *child;
551 int needterm, nested;
553 if ((nested = *argi) == argc)
556 parent = child = NULL;
557 while (*argi < argc) {
558 if (strcmp(")", argv[*argi]) == 0) {
560 warnx("missing term "
561 "before closing parenthesis");
565 warnx("ignoring unmatched right parenthesis");
569 if (strcmp("-o", argv[*argi]) == 0) {
572 warnx("ignoring -o after %s",
575 warnx("ignoring initial -o");
583 child = expr_and(search, argc, argv, argi);
586 if (parent == NULL) {
587 parent = mandoc_calloc(1, sizeof(*parent));
588 parent->type = EXPR_OR;
590 parent->child = child;
592 child->next = expr_and(search, argc, argv, argi);
595 if (needterm && *argi)
596 warnx("ignoring trailing %s", argv[*argi - 1]);
597 return parent == NULL ? child : parent;
601 expr_and(const struct mansearch *search, int argc, char *argv[], int *argi)
603 struct expr *parent, *child;
607 parent = child = NULL;
608 while (*argi < argc) {
609 if (strcmp(")", argv[*argi]) == 0) {
611 warnx("missing term "
612 "before closing parenthesis");
616 if (strcmp("-o", argv[*argi]) == 0)
618 if (strcmp("-a", argv[*argi]) == 0) {
621 warnx("ignoring -a after %s",
624 warnx("ignoring initial -a");
633 child = exprterm(search, argc, argv, argi);
639 if (parent == NULL) {
640 parent = mandoc_calloc(1, sizeof(*parent));
641 parent->type = EXPR_AND;
643 parent->child = child;
645 child->next = exprterm(search, argc, argv, argi);
646 if (child->next != NULL) {
651 if (needterm && *argi)
652 warnx("ignoring trailing %s", argv[*argi - 1]);
653 return parent == NULL ? child : parent;
657 exprterm(const struct mansearch *search, int argc, char *argv[], int *argi)
665 if (strcmp("(", argv[*argi]) == 0) {
667 e = exprcomp(search, argc, argv, argi);
669 assert(strcmp(")", argv[*argi]) == 0);
672 warnx("unclosed parenthesis");
676 if (strcmp("-i", argv[*argi]) == 0 && *argi + 1 < argc) {
682 e = mandoc_calloc(1, sizeof(*e));
688 if (search->argmode == ARG_NAME) {
690 e->match.type = DBM_EXACT;
691 e->match.str = argv[(*argi)++];
696 * Separate macro keys from search string.
697 * If needed, request regular expression handling.
700 if (search->argmode == ARG_WORD) {
702 e->match.type = DBM_REGEX;
704 mandoc_asprintf(&val, "[[:<:]]%s[[:>:]]", argv[*argi]);
706 mandoc_asprintf(&val, "\\<%s\\>", argv[*argi]);
708 mandoc_asprintf(&val,
709 "(^|[^a-zA-Z01-9_])%s([^a-zA-Z01-9_]|$)", argv[*argi]);
712 } else if ((val = strpbrk(argv[*argi], "=~")) == NULL) {
713 e->bits = TYPE_Nm | TYPE_Nd;
714 e->match.type = DBM_SUB;
715 e->match.str = argv[*argi];
717 if (val == argv[*argi])
718 e->bits = TYPE_Nm | TYPE_Nd;
720 e->match.type = DBM_SUB;
721 e->match.str = val + 1;
723 e->match.type = DBM_REGEX;
725 if (strstr(argv[*argi], "arch") != NULL)
729 /* Compile regular expressions. */
731 if (e->match.type == DBM_REGEX) {
732 e->match.re = mandoc_malloc(sizeof(*e->match.re));
733 irc = regcomp(e->match.re, val,
734 REG_EXTENDED | REG_NOSUB | (cs ? 0 : REG_ICASE));
736 regerror(irc, e->match.re, errbuf, sizeof(errbuf));
737 warnx("regcomp /%s/: %s", val, errbuf);
739 if (search->argmode == ARG_WORD)
755 * Parse out all possible fields.
756 * If the field doesn't resolve, bail.
759 while (NULL != (key = strsep(&argv[*argi], ","))) {
762 for (i = 0, iterbit = 1; i < KEY_MAX; i++, iterbit <<= 1) {
763 if (0 == strcasecmp(key, mansearch_keynames[i])) {
769 if (strcasecmp(key, "any"))
770 warnx("treating unknown key "
771 "\"%s\" as \"any\"", key);
781 exprfree(struct expr *e)
785 if (e->child != NULL)