-/* $Id: mansearch.c,v 1.26 2014/04/09 21:50:08 schwarze Exp $ */
+/* $Id: mansearch.c,v 1.48 2014/09/03 18:09:14 schwarze Exp $ */
/*
* Copyright (c) 2012 Kristaps Dzonsons <kristaps@bsd.lv>
* Copyright (c) 2013, 2014 Ingo Schwarze <schwarze@openbsd.org>
* 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 <sys/mman.h>
+#include <sys/types.h>
#include <assert.h>
#include <fcntl.h>
#include <string.h>
#include <unistd.h>
-#ifdef HAVE_OHASH
+#if HAVE_OHASH
#include <ohash.h>
#else
#include "compat_ohash.h"
#endif
#include <sqlite3.h>
+#ifndef SQLITE_DETERMINISTIC
+#define SQLITE_DETERMINISTIC 0
+#endif
#include "mandoc.h"
#include "mandoc_aux.h"
} 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 id; /* identifier in database */
+ uint64_t pageid; /* identifier in database */
char *desc; /* manual page description */
- int form; /* 0 == catpage */
+ int form; /* bit field: formatted, zipped? */
};
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,
int argc, sqlite3_value **argv);
static char *sql_statement(const struct expr *);
+
+int
+mansearch_setup(int start)
+{
+ static void *pagecache;
+ int c;
+
+#define PC_PAGESIZE 1280
+#define PC_NUMPAGES 256
+
+ if (start) {
+ if (NULL != pagecache) {
+ fprintf(stderr, "pagecache already enabled\n");
+ return((int)MANDOCLEVEL_BADARG);
+ }
+
+ pagecache = mmap(NULL, PC_PAGESIZE * PC_NUMPAGES,
+ PROT_READ | PROT_WRITE,
+ MAP_SHARED | MAP_ANON, -1, 0);
+
+ if (MAP_FAILED == pagecache) {
+ perror("mmap");
+ pagecache = NULL;
+ return((int)MANDOCLEVEL_SYSERR);
+ }
+
+ c = sqlite3_config(SQLITE_CONFIG_PAGECACHE,
+ pagecache, PC_PAGESIZE, PC_NUMPAGES);
+
+ if (SQLITE_OK == c)
+ return((int)MANDOCLEVEL_OK);
+
+ fprintf(stderr, "pagecache: %s\n", sqlite3_errstr(c));
+
+ } else if (NULL == pagecache) {
+ fprintf(stderr, "pagecache missing\n");
+ return((int)MANDOCLEVEL_BADARG);
+ }
+
+ if (-1 == munmap(pagecache, PC_PAGESIZE * PC_NUMPAGES)) {
+ perror("munmap");
+ pagecache = NULL;
+ return((int)MANDOCLEVEL_SYSERR);
+ }
+
+ pagecache = NULL;
+ return((int)MANDOCLEVEL_OK);
+}
+
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;
- int64_t id;
+ int64_t pageid;
uint64_t outbit, iterbit;
char buf[PATH_MAX];
char *sql;
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.key_offset = offsetof(struct match, id);
+ info.free = hash_free;
+ info.key_offset = offsetof(struct match, pageid);
*sz = cur = maxres = 0;
sql = NULL;
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;
*/
if (NULL == getcwd(buf, PATH_MAX)) {
- perror(NULL);
+ perror("getcwd");
goto out;
} else if (-1 == (fd = open(buf, O_RDONLY, 0))) {
perror(buf);
} 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);
*/
c = sqlite3_create_function(db, "match", 2,
- SQLITE_ANY, NULL, sql_match, NULL, NULL);
+ SQLITE_UTF8 | SQLITE_DETERMINISTIC,
+ NULL, sql_match, NULL, NULL);
assert(SQLITE_OK == c);
c = sqlite3_create_function(db, "regexp", 2,
- SQLITE_ANY, NULL, sql_regexp, NULL, NULL);
+ SQLITE_UTF8 | SQLITE_DETERMINISTIC,
+ NULL, sql_regexp, NULL, NULL);
assert(SQLITE_OK == c);
j = 1;
SQL_BIND_BLOB(db, s, j, ep->regexp);
} else
SQL_BIND_TEXT(db, s, j, ep->substr);
- if (0 == (TYPE_Nd & ep->bits))
+ if (0 == ((TYPE_Nd | TYPE_Nm) & ep->bits))
SQL_BIND_INT64(db, s, j, ep->bits);
}
* distribution of buckets in the table.
*/
while (SQLITE_ROW == (c = sqlite3_step(s))) {
- id = sqlite3_column_int64(s, 2);
- idx = ohash_lookup_memory
- (&htab, (char *)&id,
- sizeof(uint64_t), (uint32_t)id);
+ pageid = sqlite3_column_int64(s, 2);
+ idx = ohash_lookup_memory(&htab,
+ (char *)&pageid, sizeof(uint64_t),
+ (uint32_t)pageid);
if (NULL != ohash_find(&htab, idx))
continue;
mp = mandoc_calloc(1, sizeof(struct match));
- mp->id = id;
+ mp->pageid = pageid;
mp->form = sqlite3_column_int(s, 1);
if (TYPE_Nd == outbit)
- mp->desc = mandoc_strdup(
+ mp->desc = mandoc_strdup((const char *)
sqlite3_column_text(s, 0));
ohash_insert(&htab, idx, mp);
}
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));
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->sec = 10;
mpage->form = mp->form;
- buildnames(mpage, db, s, mp->id,
+ buildnames(mpage, db, s, mp->pageid,
paths->paths[i], mp->form);
mpage->output = TYPE_Nd & outbit ?
mp->desc : outbit ?
- buildoutput(db, s2, mp->id, outbit) : NULL;
+ buildoutput(db, s2, mp->pageid, outbit) : NULL;
free(mp);
cur++;
sqlite3_close(db);
ohash_delete(&htab);
}
+ qsort(*res, cur, sizeof(struct manpage), manpage_compare);
rc = 1;
out:
- exprfree(e);
- if (-1 != fd)
+ if (-1 != fd) {
+ if (-1 == fchdir(fd))
+ perror(buf);
close(fd);
+ }
+ exprfree(e);
free(sql);
*sz = cur;
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;
+ diff = mp1->sec - mp2->sec;
+ return(diff ? diff : strcasecmp(mp1->names, mp2->names));
+}
+
static void
buildnames(struct manpage *mpage, sqlite3 *db, sqlite3_stmt *s,
- uint64_t id, const char *path, int form)
+ uint64_t pageid, const char *path, int form)
{
char *newnames, *prevsec, *prevarch;
const char *oldnames, *sep1, *name, *sec, *sep2, *arch, *fsec;
+ const char *gzip;
size_t i;
int c;
mpage->names = NULL;
prevsec = prevarch = NULL;
i = 1;
- SQL_BIND_INT64(db, s, i, id);
+ SQL_BIND_INT64(db, s, i, pageid);
while (SQLITE_ROW == (c = sqlite3_step(s))) {
/* Decide whether we already have some names. */
/* 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 != mpage->file)
continue;
- if (form) {
+ if (form & FORM_SRC) {
sep1 = "man";
fsec = sec;
} else {
sep1 = "cat";
fsec = "0";
}
+ if (form & FORM_GZ)
+ gzip = ".gz";
+ else
+ gzip = "";
sep2 = '\0' == *arch ? "" : "/";
- mandoc_asprintf(&mpage->file, "%s/%s%s%s%s/%s.%s",
- path, sep1, sec, sep2, arch, name, fsec);
+ mandoc_asprintf(&mpage->file, "%s/%s%s%s%s/%s.%s%s",
+ path, sep1, sec, sep2, arch, name, fsec, gzip);
}
if (SQLITE_DONE != c)
fprintf(stderr, "%s\n", sqlite3_errmsg(db));
}
static char *
-buildoutput(sqlite3 *db, sqlite3_stmt *s, uint64_t id, uint64_t outbit)
+buildoutput(sqlite3 *db, sqlite3_stmt *s, uint64_t pageid, uint64_t outbit)
{
char *output, *newoutput;
const char *oldoutput, *sep1, *data;
output = NULL;
i = 1;
- SQL_BIND_INT64(db, s, i, id);
+ SQL_BIND_INT64(db, s, i, pageid);
SQL_BIND_INT64(db, s, i, outbit);
while (SQLITE_ROW == (c = sqlite3_step(s))) {
if (NULL == output) {
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);
size_t sz;
int needop;
- sql = mandoc_strdup("SELECT * FROM mpages WHERE ");
+ sql = mandoc_strdup(
+ "SELECT desc, form, pageid FROM mpages WHERE ");
sz = strlen(sql);
for (needop = 0; NULL != e; e = e->next) {
? (NULL == e->substr
? "desc REGEXP ?"
: "desc MATCH ?")
+ : TYPE_Nm == e->bits
+ ? (NULL == e->substr
+ ? "pageid IN (SELECT pageid FROM names "
+ "WHERE name REGEXP ?)"
+ : e->equal
+ ? "pageid IN (SELECT pageid FROM names "
+ "WHERE name = ?)"
+ : "pageid IN (SELECT pageid FROM names "
+ "WHERE name MATCH ?)")
: (NULL == e->substr
- ? "id IN (SELECT pageid FROM keys "
+ ? "pageid IN (SELECT pageid FROM keys "
"WHERE key REGEXP ? AND bits & ?)"
- : "id IN (SELECT pageid FROM keys "
+ : "pageid IN (SELECT pageid FROM keys "
"WHERE key MATCH ? AND bits & ?)"), 1);
if (e->close)
sql_append(&sql, &sz, ")", e->close);
static struct expr *
exprcomp(const struct mansearch *search, int argc, char *argv[])
{
+ uint64_t mask;
int i, toopen, logic, igncase, toclose;
- struct expr *first, *next, *cur;
+ struct expr *first, *prev, *cur, *next;
first = cur = NULL;
logic = igncase = toclose = 0;
- toopen = 1;
+ toopen = NULL != search->sec || NULL != search->arch;
for (i = 0; i < argc; i++) {
if (0 == strcmp("(", argv[i])) {
first = next;
else
cur->next = next;
+ prev = cur = next;
/*
* Searching for descriptions must be split out
* not in the keys table.
*/
- if (TYPE_Nd & next->bits && ~TYPE_Nd & next->bits) {
- cur = mandoc_calloc(1, sizeof(struct expr));
- memcpy(cur, next, sizeof(struct expr));
- next->open = 1;
- next->bits = TYPE_Nd;
- next->next = cur;
- cur->bits &= ~TYPE_Nd;
+ for (mask = TYPE_Nm; mask <= TYPE_Nd; mask <<= 1) {
+ if (mask & cur->bits && ~mask & cur->bits) {
+ next = mandoc_calloc(1,
+ sizeof(struct expr));
+ memcpy(next, cur, sizeof(struct expr));
+ prev->open = 1;
+ cur->bits = mask;
+ cur->next = next;
+ cur = next;
+ cur->bits &= ~mask;
+ }
+ }
+ prev->and = (1 == logic);
+ prev->open += toopen;
+ if (cur != prev)
cur->close = 1;
- } else
- cur = next;
- next->and = (1 == logic);
- next->open += toopen;
+
toopen = logic = igncase = 0;
}
if (toopen || logic || igncase || toclose)
goto fail;
- cur->close++;
- cur = exprspec(cur, TYPE_arch, search->arch, "^(%s|any)$");
- exprspec(cur, TYPE_sec, search->sec, "^%s$");
+ if (NULL != search->sec || NULL != search->arch)
+ cur->close++;
+ if (NULL != search->arch)
+ cur = exprspec(cur, TYPE_arch, search->arch, "^(%s|any)$");
+ if (NULL != search->sec)
+ exprspec(cur, TYPE_sec, search->sec, "^%s$");
return(first);
char *cp;
int irc;
- if (NULL == value)
- return(cur);
-
mandoc_asprintf(&cp, format, value);
cur->next = mandoc_calloc(1, sizeof(struct expr));
cur = cur->next;
{
char errbuf[BUFSIZ];
struct expr *e;
- char *key, *v;
+ char *key, *val;
uint64_t iterbit;
int i, irc;
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.
}
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 *
}
static void
-hash_free(void *p, size_t sz, void *arg)
+hash_free(void *p, void *arg)
{
free(p);
git.ckatri.com / mandoc.git / blobdiff