=================================================================== RCS file: /cvs/mandoc/mansearch.c,v retrieving revision 1.33 retrieving revision 1.51 diff -u -p -r1.33 -r1.51 --- mandoc/mansearch.c 2014/04/17 19:20:01 1.33 +++ mandoc/mansearch.c 2014/11/27 01:58:21 1.51 @@ -1,4 +1,4 @@ -/* $Id: mansearch.c,v 1.33 2014/04/17 19:20:01 schwarze Exp $ */ +/* $Id: mansearch.c,v 1.51 2014/11/27 01:58:21 schwarze Exp $ */ /* * Copyright (c) 2012 Kristaps Dzonsons * Copyright (c) 2013, 2014 Ingo Schwarze @@ -15,11 +15,11 @@ * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. */ -#ifdef HAVE_CONFIG_H #include "config.h" -#endif #include +#include + #include #include #include @@ -32,12 +32,15 @@ #include #include -#ifdef HAVE_OHASH +#if HAVE_OHASH #include #else #include "compat_ohash.h" #endif #include +#ifndef SQLITE_DETERMINISTIC +#define SQLITE_DETERMINISTIC 0 +#endif #include "mandoc.h" #include "mandoc_aux.h" @@ -64,19 +67,21 @@ extern const char *const mansearch_keynames[]; } while (0) struct expr { - uint64_t bits; /* type-mask */ - const char *substr; /* to search for, if applicable */ regex_t regexp; /* compiled regexp, if applicable */ + const char *substr; /* to search for, if applicable */ + struct expr *next; /* next in sequence */ + uint64_t bits; /* type-mask */ + int equal; /* equality, not subsring match */ int open; /* opening parentheses before */ int and; /* logical AND before */ int close; /* closing parentheses after */ - struct expr *next; /* next in sequence */ }; struct match { uint64_t pageid; /* identifier in database */ + uint64_t bits; /* name type mask */ char *desc; /* manual page description */ - int form; /* 0 == catpage */ + int form; /* bit field: formatted, zipped? */ }; static void buildnames(struct manpage *, sqlite3 *, @@ -85,14 +90,15 @@ static void buildnames(struct manpage *, sqlite3 *, static char *buildoutput(sqlite3 *, sqlite3_stmt *, uint64_t, uint64_t); static void *hash_alloc(size_t, void *); -static void hash_free(void *, size_t, void *); -static void *hash_halloc(size_t, void *); -static struct expr *exprcomp(const struct mansearch *, +static void hash_free(void *, void *); +static void *hash_calloc(size_t, size_t, void *); +static struct expr *exprcomp(const struct mansearch *, int, char *[]); static void exprfree(struct expr *); static struct expr *exprspec(struct expr *, uint64_t, const char *, const char *); static struct expr *exprterm(const struct mansearch *, char *, int); +static int manpage_compare(const void *, const void *); static void sql_append(char **sql, size_t *sz, const char *newstr, int count); static void sql_match(sqlite3_context *context, @@ -101,6 +107,7 @@ static void sql_regexp(sqlite3_context *context, int argc, sqlite3_value **argv); static char *sql_statement(const struct expr *); + int mansearch_setup(int start) { @@ -117,7 +124,8 @@ mansearch_setup(int start) } pagecache = mmap(NULL, PC_PAGESIZE * PC_NUMPAGES, - PROT_READ | PROT_WRITE, MAP_ANON, -1, 0); + PROT_READ | PROT_WRITE, + MAP_SHARED | MAP_ANON, -1, 0); if (MAP_FAILED == pagecache) { perror("mmap"); @@ -152,7 +160,6 @@ int mansearch(const struct mansearch *search, const struct manpaths *paths, int argc, char *argv[], - const char *outkey, struct manpage **res, size_t *sz) { int fd, rc, c, indexbit; @@ -170,11 +177,9 @@ mansearch(const struct mansearch *search, unsigned int idx; size_t i, j, cur, maxres; - memset(&info, 0, sizeof(struct ohash_info)); - - info.halloc = hash_halloc; + info.calloc = hash_calloc; info.alloc = hash_alloc; - info.hfree = hash_free; + info.free = hash_free; info.key_offset = offsetof(struct match, pageid); *sz = cur = maxres = 0; @@ -190,11 +195,11 @@ mansearch(const struct mansearch *search, goto out; outbit = 0; - if (NULL != outkey) { + if (NULL != search->outkey) { for (indexbit = 0, iterbit = 1; indexbit < mansearch_keymax; indexbit++, iterbit <<= 1) { - if (0 == strcasecmp(outkey, + if (0 == strcasecmp(search->outkey, mansearch_keynames[indexbit])) { outbit = iterbit; break; @@ -235,11 +240,10 @@ mansearch(const struct mansearch *search, } else if (-1 == chdir(paths->paths[i])) { perror(paths->paths[i]); continue; - } + } - c = sqlite3_open_v2 - (MANDOC_DB, &db, - SQLITE_OPEN_READONLY, NULL); + c = sqlite3_open_v2(MANDOC_DB, &db, + SQLITE_OPEN_READONLY, NULL); if (SQLITE_OK != c) { perror(MANDOC_DB); @@ -288,9 +292,9 @@ mansearch(const struct mansearch *search, */ while (SQLITE_ROW == (c = sqlite3_step(s))) { pageid = sqlite3_column_int64(s, 2); - idx = ohash_lookup_memory - (&htab, (char *)&pageid, - sizeof(uint64_t), (uint32_t)pageid); + idx = ohash_lookup_memory(&htab, + (char *)&pageid, sizeof(uint64_t), + (uint32_t)pageid); if (NULL != ohash_find(&htab, idx)) continue; @@ -298,8 +302,9 @@ mansearch(const struct mansearch *search, mp = mandoc_calloc(1, sizeof(struct match)); mp->pageid = pageid; mp->form = sqlite3_column_int(s, 1); + mp->bits = sqlite3_column_int64(s, 3); if (TYPE_Nd == outbit) - mp->desc = mandoc_strdup( + mp->desc = mandoc_strdup((const char *) sqlite3_column_text(s, 0)); ohash_insert(&htab, idx, mp); } @@ -309,15 +314,16 @@ mansearch(const struct mansearch *search, sqlite3_finalize(s); - c = sqlite3_prepare_v2(db, - "SELECT * FROM mlinks WHERE pageid=?" - " ORDER BY sec, arch, name", + c = sqlite3_prepare_v2(db, + "SELECT sec, arch, name, pageid FROM mlinks " + "WHERE pageid=? ORDER BY sec, arch, name", -1, &s, NULL); if (SQLITE_OK != c) fprintf(stderr, "%s\n", sqlite3_errmsg(db)); c = sqlite3_prepare_v2(db, - "SELECT * FROM keys WHERE pageid=? AND bits & ?", + "SELECT bits, key, pageid FROM keys " + "WHERE pageid=? AND bits & ?", -1, &s2, NULL); if (SQLITE_OK != c) fprintf(stderr, "%s\n", sqlite3_errmsg(db)); @@ -327,10 +333,13 @@ mansearch(const struct mansearch *search, mp = ohash_next(&htab, &idx)) { if (cur + 1 > maxres) { maxres += 1024; - *res = mandoc_realloc - (*res, maxres * sizeof(struct manpage)); + *res = mandoc_reallocarray(*res, + maxres, sizeof(struct manpage)); } mpage = *res + cur; + mpage->ipath = i; + mpage->bits = mp->bits; + mpage->sec = 10; mpage->form = mp->form; buildnames(mpage, db, s, mp->pageid, paths->paths[i], mp->form); @@ -346,7 +355,16 @@ mansearch(const struct mansearch *search, sqlite3_finalize(s2); sqlite3_close(db); ohash_delete(&htab); + + /* + * In man(1) mode, prefer matches in earlier trees + * over matches in later trees. + */ + + if (cur && search->firstmatch) + break; } + qsort(*res, cur, sizeof(struct manpage), manpage_compare); rc = 1; out: if (-1 != fd) { @@ -360,6 +378,32 @@ out: return(rc); } +void +mansearch_free(struct manpage *res, size_t sz) +{ + size_t i; + + for (i = 0; i < sz; i++) { + free(res[i].file); + free(res[i].names); + free(res[i].output); + } + free(res); +} + +static int +manpage_compare(const void *vp1, const void *vp2) +{ + const struct manpage *mp1, *mp2; + int diff; + + mp1 = vp1; + mp2 = vp2; + return( (diff = mp2->bits - mp1->bits) ? diff : + (diff = mp1->sec - mp2->sec) ? diff : + strcasecmp(mp1->names, mp2->names)); +} + static void buildnames(struct manpage *mpage, sqlite3 *db, sqlite3_stmt *s, uint64_t pageid, const char *path, int form) @@ -388,10 +432,15 @@ buildnames(struct manpage *mpage, sqlite3 *db, sqlite3 /* Fetch the next name. */ - sec = sqlite3_column_text(s, 0); - arch = sqlite3_column_text(s, 1); - name = sqlite3_column_text(s, 2); + sec = (const char *)sqlite3_column_text(s, 0); + arch = (const char *)sqlite3_column_text(s, 1); + name = (const char *)sqlite3_column_text(s, 2); + /* Remember the first section found. */ + + if (9 < mpage->sec && '1' <= *sec && '9' >= *sec) + mpage->sec = (*sec - '1') + 1; + /* If the section changed, append the old one. */ if (NULL != prevsec && @@ -423,28 +472,28 @@ buildnames(struct manpage *mpage, sqlite3 *db, sqlite3 /* Also save the first file name encountered. */ - if (NULL != mpage->file) + if (mpage->file != NULL) continue; - if (form) { + if (form & FORM_SRC) { sep1 = "man"; fsec = sec; } else { sep1 = "cat"; fsec = "0"; } - sep2 = '\0' == *arch ? "" : "/"; + sep2 = *arch == '\0' ? "" : "/"; mandoc_asprintf(&mpage->file, "%s/%s%s%s%s/%s.%s", path, sep1, sec, sep2, arch, name, fsec); } - if (SQLITE_DONE != c) + if (c != SQLITE_DONE) fprintf(stderr, "%s\n", sqlite3_errmsg(db)); sqlite3_reset(s); /* Append one final section to the names. */ - if (NULL != prevsec) { - sep2 = '\0' == *prevarch ? "" : "/"; + if (prevsec != NULL) { + sep2 = *prevarch == '\0' ? "" : "/"; mandoc_asprintf(&newnames, "%s(%s%s%s)", mpage->names, prevsec, sep2, prevarch); free(mpage->names); @@ -474,7 +523,7 @@ buildoutput(sqlite3 *db, sqlite3_stmt *s, uint64_t pag oldoutput = output; sep1 = " # "; } - data = sqlite3_column_text(s, 1); + data = (const char *)sqlite3_column_text(s, 1); mandoc_asprintf(&newoutput, "%s%s%s", oldoutput, sep1, data); free(output); @@ -542,7 +591,10 @@ sql_statement(const struct expr *e) size_t sz; int needop; - sql = mandoc_strdup("SELECT * FROM mpages WHERE "); + sql = mandoc_strdup(e->equal ? + "SELECT desc, form, pageid, bits " + "FROM mpages NATURAL JOIN names WHERE " : + "SELECT desc, form, pageid, 0 FROM mpages WHERE "); sz = strlen(sql); for (needop = 0; NULL != e; e = e->next) { @@ -561,6 +613,8 @@ sql_statement(const struct expr *e) ? (NULL == e->substr ? "pageid IN (SELECT pageid FROM names " "WHERE name REGEXP ?)" + : e->equal + ? "name = ? " : "pageid IN (SELECT pageid FROM names " "WHERE name MATCH ?)") : (NULL == e->substr @@ -702,7 +756,7 @@ exprterm(const struct mansearch *search, char *buf, in { char errbuf[BUFSIZ]; struct expr *e; - char *key, *v; + char *key, *val; uint64_t iterbit; int i, irc; @@ -711,41 +765,55 @@ exprterm(const struct mansearch *search, char *buf, in e = mandoc_calloc(1, sizeof(struct expr)); - /*"whatis" mode uses an opaque string and default fields. */ - - if (MANSEARCH_WHATIS & search->flags) { + if (search->argmode == ARG_NAME) { + e->bits = TYPE_Nm; e->substr = buf; - e->bits = search->deftype; + e->equal = 1; return(e); } /* - * If no =~ is specified, search with equality over names and - * descriptions. - * If =~ begins the phrase, use name and description fields. + * Separate macro keys from search string. + * If needed, request regular expression handling + * by setting e->substr to NULL. */ - if (NULL == (v = strpbrk(buf, "=~"))) { + if (search->argmode == ARG_WORD) { + e->bits = TYPE_Nm; + e->substr = NULL; + mandoc_asprintf(&val, "[[:<:]]%s[[:>:]]", buf); + cs = 0; + } else if ((val = strpbrk(buf, "=~")) == NULL) { + e->bits = TYPE_Nm | TYPE_Nd; e->substr = buf; - e->bits = search->deftype; - return(e); - } else if (v == buf) - e->bits = search->deftype; - - if ('~' == *v++) { + } else { + if (val == buf) + e->bits = TYPE_Nm | TYPE_Nd; + if ('=' == *val) + e->substr = val + 1; + *val++ = '\0'; if (NULL != strstr(buf, "arch")) cs = 0; - if (0 != (irc = regcomp(&e->regexp, v, - REG_EXTENDED | REG_NOSUB | (cs ? 0 : REG_ICASE)))) { + } + + /* Compile regular expressions. */ + + if (NULL == e->substr) { + irc = regcomp(&e->regexp, val, + REG_EXTENDED | REG_NOSUB | (cs ? 0 : REG_ICASE)); + if (search->argmode == ARG_WORD) + free(val); + if (irc) { regerror(irc, &e->regexp, errbuf, sizeof(errbuf)); fprintf(stderr, "regcomp: %s\n", errbuf); free(e); return(NULL); } - } else - e->substr = v; - v[-1] = '\0'; + } + if (e->bits) + return(e); + /* * Parse out all possible fields. * If the field doesn't resolve, bail. @@ -788,10 +856,10 @@ exprfree(struct expr *p) } static void * -hash_halloc(size_t sz, void *arg) +hash_calloc(size_t nmemb, size_t sz, void *arg) { - return(mandoc_calloc(sz, 1)); + return(mandoc_calloc(nmemb, sz)); } static void * @@ -802,7 +870,7 @@ hash_alloc(size_t sz, void *arg) } static void -hash_free(void *p, size_t sz, void *arg) +hash_free(void *p, void *arg) { free(p);