version 1.153, 2014/06/21 16:18:25 |
version 1.270, 2021/11/05 17:04:10 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
|
* Copyright (c) 2011-2020 Ingo Schwarze <schwarze@openbsd.org> |
* Copyright (c) 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2011, 2012, 2013, 2014 Ingo Schwarze <schwarze@openbsd.org> |
* Copyright (c) 2016 Ed Maste <emaste@freebsd.org> |
* |
* |
* Permission to use, copy, modify, and distribute this software for any |
* Permission to use, copy, modify, and distribute this software for any |
* purpose with or without fee is hereby granted, provided that the above |
* purpose with or without fee is hereby granted, provided that the above |
* copyright notice and this permission notice appear in all copies. |
* copyright notice and this permission notice appear in all copies. |
* |
* |
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES |
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHORS DISCLAIM ALL WARRANTIES |
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF |
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF |
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR |
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR |
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES |
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES |
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN |
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN |
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF |
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF |
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. |
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. |
|
* |
|
* Implementation of the makewhatis(8) program. |
*/ |
*/ |
#ifdef HAVE_CONFIG_H |
|
#include "config.h" |
#include "config.h" |
#endif |
|
|
|
|
#include <sys/types.h> |
|
#include <sys/mman.h> |
#include <sys/stat.h> |
#include <sys/stat.h> |
#include <sys/wait.h> |
|
|
|
#include <assert.h> |
#include <assert.h> |
#include <ctype.h> |
#include <ctype.h> |
|
#if HAVE_ERR |
|
#include <err.h> |
|
#endif |
#include <errno.h> |
#include <errno.h> |
#include <fcntl.h> |
#include <fcntl.h> |
|
#if HAVE_FTS |
#include <fts.h> |
#include <fts.h> |
#include <getopt.h> |
#else |
|
#include "compat_fts.h" |
|
#endif |
#include <limits.h> |
#include <limits.h> |
|
#if HAVE_SANDBOX_INIT |
|
#include <sandbox.h> |
|
#endif |
|
#include <stdarg.h> |
#include <stddef.h> |
#include <stddef.h> |
#include <stdio.h> |
#include <stdio.h> |
#include <stdint.h> |
#include <stdint.h> |
|
|
#include <string.h> |
#include <string.h> |
#include <unistd.h> |
#include <unistd.h> |
|
|
#ifdef HAVE_OHASH |
#include "mandoc_aux.h" |
#include <ohash.h> |
#include "mandoc_ohash.h" |
#else |
#include "mandoc.h" |
#include "compat_ohash.h" |
#include "roff.h" |
#endif |
|
#include <sqlite3.h> |
|
|
|
#include "mdoc.h" |
#include "mdoc.h" |
#include "man.h" |
#include "man.h" |
#include "mandoc.h" |
#include "mandoc_parse.h" |
#include "mandoc_aux.h" |
#include "manconf.h" |
#include "manpath.h" |
|
#include "mansearch.h" |
#include "mansearch.h" |
|
#include "dba_array.h" |
|
#include "dba.h" |
|
|
extern int mansearch_keymax; |
|
extern const char *const mansearch_keynames[]; |
extern const char *const mansearch_keynames[]; |
|
|
#define SQL_EXEC(_v) \ |
|
if (SQLITE_OK != sqlite3_exec(db, (_v), NULL, NULL, NULL)) \ |
|
say("", "%s: %s", (_v), sqlite3_errmsg(db)) |
|
#define SQL_BIND_TEXT(_s, _i, _v) \ |
|
if (SQLITE_OK != sqlite3_bind_text \ |
|
((_s), (_i)++, (_v), -1, SQLITE_STATIC)) \ |
|
say(mlink->file, "%s", sqlite3_errmsg(db)) |
|
#define SQL_BIND_INT(_s, _i, _v) \ |
|
if (SQLITE_OK != sqlite3_bind_int \ |
|
((_s), (_i)++, (_v))) \ |
|
say(mlink->file, "%s", sqlite3_errmsg(db)) |
|
#define SQL_BIND_INT64(_s, _i, _v) \ |
|
if (SQLITE_OK != sqlite3_bind_int64 \ |
|
((_s), (_i)++, (_v))) \ |
|
say(mlink->file, "%s", sqlite3_errmsg(db)) |
|
#define SQL_STEP(_s) \ |
|
if (SQLITE_DONE != sqlite3_step((_s))) \ |
|
say(mlink->file, "%s", sqlite3_errmsg(db)) |
|
|
|
enum op { |
enum op { |
OP_DEFAULT = 0, /* new dbs from dir list or default config */ |
OP_DEFAULT = 0, /* new dbs from dir list or default config */ |
OP_CONFFILE, /* new databases from custom config file */ |
OP_CONFFILE, /* new databases from custom config file */ |
|
|
OP_TEST /* change no databases, report potential problems */ |
OP_TEST /* change no databases, report potential problems */ |
}; |
}; |
|
|
enum form { |
|
FORM_NONE, /* format is unknown */ |
|
FORM_SRC, /* format is -man or -mdoc */ |
|
FORM_CAT /* format is cat */ |
|
}; |
|
|
|
struct str { |
struct str { |
char *rendered; /* key in UTF-8 or ASCII form */ |
|
const struct mpage *mpage; /* if set, the owning parse */ |
const struct mpage *mpage; /* if set, the owning parse */ |
uint64_t mask; /* bitmask in sequence */ |
uint64_t mask; /* bitmask in sequence */ |
char key[]; /* may contain escape sequences */ |
char key[]; /* rendered text */ |
}; |
}; |
|
|
struct inodev { |
struct inodev { |
|
|
|
|
struct mpage { |
struct mpage { |
struct inodev inodev; /* used for hashing routine */ |
struct inodev inodev; /* used for hashing routine */ |
int64_t pageid; /* pageid in mpages SQL table */ |
struct dba_array *dba; |
enum form form; /* format from file content */ |
|
char *sec; /* section from file content */ |
char *sec; /* section from file content */ |
char *arch; /* architecture from file content */ |
char *arch; /* architecture from file content */ |
char *title; /* title from file content */ |
char *title; /* title from file content */ |
char *desc; /* description from file content */ |
char *desc; /* description from file content */ |
|
struct mpage *next; /* singly linked list */ |
struct mlink *mlinks; /* singly linked list */ |
struct mlink *mlinks; /* singly linked list */ |
|
int name_head_done; |
|
enum form form; /* format from file content */ |
}; |
}; |
|
|
struct mlink { |
struct mlink { |
char file[PATH_MAX]; /* filename rel. to manpath */ |
char file[PATH_MAX]; /* filename rel. to manpath */ |
enum form dform; /* format from directory */ |
|
enum form fform; /* format from file name suffix */ |
|
char *dsec; /* section from directory */ |
char *dsec; /* section from directory */ |
char *arch; /* architecture from directory */ |
char *arch; /* architecture from directory */ |
char *name; /* name from file name (not empty) */ |
char *name; /* name from file name (not empty) */ |
|
|
struct mlink *next; /* singly linked list */ |
struct mlink *next; /* singly linked list */ |
struct mpage *mpage; /* parent */ |
struct mpage *mpage; /* parent */ |
int gzip; /* filename has a .gz suffix */ |
int gzip; /* filename has a .gz suffix */ |
|
enum form dform; /* format from directory */ |
|
enum form fform; /* format from file name suffix */ |
}; |
}; |
|
|
enum stmt { |
typedef int (*mdoc_fp)(struct mpage *, const struct roff_meta *, |
STMT_DELETE_PAGE = 0, /* delete mpage */ |
const struct roff_node *); |
STMT_INSERT_PAGE, /* insert mpage */ |
|
STMT_INSERT_LINK, /* insert mlink */ |
|
STMT_INSERT_NAME, /* insert name */ |
|
STMT_INSERT_KEY, /* insert parsed key */ |
|
STMT__MAX |
|
}; |
|
|
|
typedef int (*mdoc_fp)(struct mpage *, const struct mdoc_node *); |
|
|
|
struct mdoc_handler { |
struct mdoc_handler { |
mdoc_fp fp; /* optional handler */ |
mdoc_fp fp; /* optional handler */ |
uint64_t mask; /* set unless handler returns 0 */ |
uint64_t mask; /* set unless handler returns 0 */ |
|
int taboo; /* node flags that must not be set */ |
}; |
}; |
|
|
static void dbclose(int); |
|
static void dbadd(struct mpage *, struct mchars *); |
int mandocdb(int, char *[]); |
static void dbadd_mlink(const struct mlink *mlink); |
|
static int dbopen(int); |
static void dbadd(struct dba *, struct mpage *); |
static void dbprune(void); |
static void dbadd_mlink(const struct mlink *); |
|
static void dbprune(struct dba *); |
|
static void dbwrite(struct dba *); |
static void filescan(const char *); |
static void filescan(const char *); |
static void *hash_alloc(size_t, void *); |
#if HAVE_FTS_COMPARE_CONST |
static void hash_free(void *, void *); |
static int fts_compare(const FTSENT *const *, const FTSENT *const *); |
static void *hash_calloc(size_t, size_t, void *); |
#else |
|
static int fts_compare(const FTSENT **, const FTSENT **); |
|
#endif |
static void mlink_add(struct mlink *, const struct stat *); |
static void mlink_add(struct mlink *, const struct stat *); |
static void mlink_check(struct mpage *, struct mlink *); |
static void mlink_check(struct mpage *, struct mlink *); |
static void mlink_free(struct mlink *); |
static void mlink_free(struct mlink *); |
static void mlinks_undupe(struct mpage *); |
static void mlinks_undupe(struct mpage *); |
static void mpages_free(void); |
static void mpages_free(void); |
static void mpages_merge(struct mchars *, struct mparse *); |
static void mpages_merge(struct dba *, struct mparse *); |
static void names_check(void); |
|
static void parse_cat(struct mpage *, int); |
static void parse_cat(struct mpage *, int); |
static void parse_man(struct mpage *, const struct man_node *); |
static void parse_man(struct mpage *, const struct roff_meta *, |
static void parse_mdoc(struct mpage *, const struct mdoc_node *); |
const struct roff_node *); |
static int parse_mdoc_body(struct mpage *, const struct mdoc_node *); |
static void parse_mdoc(struct mpage *, const struct roff_meta *, |
static int parse_mdoc_head(struct mpage *, const struct mdoc_node *); |
const struct roff_node *); |
static int parse_mdoc_Fd(struct mpage *, const struct mdoc_node *); |
static int parse_mdoc_head(struct mpage *, const struct roff_meta *, |
static int parse_mdoc_Fn(struct mpage *, const struct mdoc_node *); |
const struct roff_node *); |
static int parse_mdoc_Nd(struct mpage *, const struct mdoc_node *); |
static int parse_mdoc_Fa(struct mpage *, const struct roff_meta *, |
static int parse_mdoc_Nm(struct mpage *, const struct mdoc_node *); |
const struct roff_node *); |
static int parse_mdoc_Sh(struct mpage *, const struct mdoc_node *); |
static int parse_mdoc_Fd(struct mpage *, const struct roff_meta *, |
static int parse_mdoc_Xr(struct mpage *, const struct mdoc_node *); |
const struct roff_node *); |
|
static void parse_mdoc_fname(struct mpage *, const struct roff_node *); |
|
static int parse_mdoc_Fn(struct mpage *, const struct roff_meta *, |
|
const struct roff_node *); |
|
static int parse_mdoc_Fo(struct mpage *, const struct roff_meta *, |
|
const struct roff_node *); |
|
static int parse_mdoc_Nd(struct mpage *, const struct roff_meta *, |
|
const struct roff_node *); |
|
static int parse_mdoc_Nm(struct mpage *, const struct roff_meta *, |
|
const struct roff_node *); |
|
static int parse_mdoc_Sh(struct mpage *, const struct roff_meta *, |
|
const struct roff_node *); |
|
static int parse_mdoc_Va(struct mpage *, const struct roff_meta *, |
|
const struct roff_node *); |
|
static int parse_mdoc_Xr(struct mpage *, const struct roff_meta *, |
|
const struct roff_node *); |
static void putkey(const struct mpage *, char *, uint64_t); |
static void putkey(const struct mpage *, char *, uint64_t); |
static void putkeys(const struct mpage *, |
static void putkeys(const struct mpage *, char *, size_t, uint64_t); |
const char *, size_t, uint64_t); |
|
static void putmdockey(const struct mpage *, |
static void putmdockey(const struct mpage *, |
const struct mdoc_node *, uint64_t); |
const struct roff_node *, uint64_t, int); |
static void render_key(struct mchars *, struct str *); |
#ifdef READ_ALLOWED_PATH |
static void say(const char *, const char *, ...); |
static int read_allowed(const char *); |
static int set_basedir(const char *); |
#endif |
|
static int render_string(char **, size_t *); |
|
static void say(const char *, const char *, ...) |
|
__attribute__((__format__ (__printf__, 2, 3))); |
|
static int set_basedir(const char *, int); |
static int treescan(void); |
static int treescan(void); |
static size_t utf8(unsigned int, char [7]); |
static size_t utf8(unsigned int, char [7]); |
|
|
static char tempfilename[32]; |
|
static char *progname; |
|
static int nodb; /* no database changes */ |
static int nodb; /* no database changes */ |
static int mparse_options; /* abort the parse early */ |
static int mparse_options; /* abort the parse early */ |
static int use_all; /* use all found files */ |
static int use_all; /* use all found files */ |
Line 187 static int write_utf8; /* write UTF-8 output; else A |
|
Line 184 static int write_utf8; /* write UTF-8 output; else A |
|
static int exitcode; /* to be returned by main */ |
static int exitcode; /* to be returned by main */ |
static enum op op; /* operational mode */ |
static enum op op; /* operational mode */ |
static char basedir[PATH_MAX]; /* current base directory */ |
static char basedir[PATH_MAX]; /* current base directory */ |
|
static size_t basedir_len; /* strlen(basedir) */ |
|
static struct mpage *mpage_head; /* list of distinct manual pages */ |
static struct ohash mpages; /* table of distinct manual pages */ |
static struct ohash mpages; /* table of distinct manual pages */ |
static struct ohash mlinks; /* table of directory entries */ |
static struct ohash mlinks; /* table of directory entries */ |
static struct ohash names; /* table of all names */ |
static struct ohash names; /* table of all names */ |
static struct ohash strings; /* table of all strings */ |
static struct ohash strings; /* table of all strings */ |
static sqlite3 *db = NULL; /* current database */ |
|
static sqlite3_stmt *stmts[STMT__MAX]; /* current statements */ |
|
static uint64_t name_mask; |
static uint64_t name_mask; |
|
|
static const struct mdoc_handler mdocs[MDOC_MAX] = { |
static const struct mdoc_handler mdoc_handlers[MDOC_MAX - MDOC_Dd] = { |
{ NULL, 0 }, /* Ap */ |
{ NULL, 0, NODE_NOPRT }, /* Dd */ |
{ NULL, 0 }, /* Dd */ |
{ NULL, 0, NODE_NOPRT }, /* Dt */ |
{ NULL, 0 }, /* Dt */ |
{ NULL, 0, NODE_NOPRT }, /* Os */ |
{ NULL, 0 }, /* Os */ |
{ parse_mdoc_Sh, TYPE_Sh, 0 }, /* Sh */ |
{ parse_mdoc_Sh, TYPE_Sh }, /* Sh */ |
{ parse_mdoc_head, TYPE_Ss, 0 }, /* Ss */ |
{ parse_mdoc_head, TYPE_Ss }, /* Ss */ |
{ NULL, 0, 0 }, /* Pp */ |
{ NULL, 0 }, /* Pp */ |
{ NULL, 0, 0 }, /* D1 */ |
{ NULL, 0 }, /* D1 */ |
{ NULL, 0, 0 }, /* Dl */ |
{ NULL, 0 }, /* Dl */ |
{ NULL, 0, 0 }, /* Bd */ |
{ NULL, 0 }, /* Bd */ |
{ NULL, 0, 0 }, /* Ed */ |
{ NULL, 0 }, /* Ed */ |
{ NULL, 0, 0 }, /* Bl */ |
{ NULL, 0 }, /* Bl */ |
{ NULL, 0, 0 }, /* El */ |
{ NULL, 0 }, /* El */ |
{ NULL, 0, 0 }, /* It */ |
{ NULL, 0 }, /* It */ |
{ NULL, 0, 0 }, /* Ad */ |
{ NULL, 0 }, /* Ad */ |
{ NULL, TYPE_An, 0 }, /* An */ |
{ NULL, TYPE_An }, /* An */ |
{ NULL, 0, 0 }, /* Ap */ |
{ NULL, TYPE_Ar }, /* Ar */ |
{ NULL, TYPE_Ar, 0 }, /* Ar */ |
{ NULL, TYPE_Cd }, /* Cd */ |
{ NULL, TYPE_Cd, 0 }, /* Cd */ |
{ NULL, TYPE_Cm }, /* Cm */ |
{ NULL, TYPE_Cm, 0 }, /* Cm */ |
{ NULL, TYPE_Dv }, /* Dv */ |
{ NULL, TYPE_Dv, 0 }, /* Dv */ |
{ NULL, TYPE_Er }, /* Er */ |
{ NULL, TYPE_Er, 0 }, /* Er */ |
{ NULL, TYPE_Ev }, /* Ev */ |
{ NULL, TYPE_Ev, 0 }, /* Ev */ |
{ NULL, 0 }, /* Ex */ |
{ NULL, 0, 0 }, /* Ex */ |
{ NULL, TYPE_Fa }, /* Fa */ |
{ parse_mdoc_Fa, 0, 0 }, /* Fa */ |
{ parse_mdoc_Fd, 0 }, /* Fd */ |
{ parse_mdoc_Fd, 0, 0 }, /* Fd */ |
{ NULL, TYPE_Fl }, /* Fl */ |
{ NULL, TYPE_Fl, 0 }, /* Fl */ |
{ parse_mdoc_Fn, 0 }, /* Fn */ |
{ parse_mdoc_Fn, 0, 0 }, /* Fn */ |
{ NULL, TYPE_Ft }, /* Ft */ |
{ NULL, TYPE_Ft | TYPE_Vt, 0 }, /* Ft */ |
{ NULL, TYPE_Ic }, /* Ic */ |
{ NULL, TYPE_Ic, 0 }, /* Ic */ |
{ NULL, TYPE_In }, /* In */ |
{ NULL, TYPE_In, 0 }, /* In */ |
{ NULL, TYPE_Li }, /* Li */ |
{ NULL, TYPE_Li, 0 }, /* Li */ |
{ parse_mdoc_Nd, 0 }, /* Nd */ |
{ parse_mdoc_Nd, 0, 0 }, /* Nd */ |
{ parse_mdoc_Nm, 0 }, /* Nm */ |
{ parse_mdoc_Nm, 0, 0 }, /* Nm */ |
{ NULL, 0 }, /* Op */ |
{ NULL, 0, 0 }, /* Op */ |
{ NULL, 0 }, /* Ot */ |
{ NULL, 0, 0 }, /* Ot */ |
{ NULL, TYPE_Pa }, /* Pa */ |
{ NULL, TYPE_Pa, NODE_NOSRC }, /* Pa */ |
{ NULL, 0 }, /* Rv */ |
{ NULL, 0, 0 }, /* Rv */ |
{ NULL, TYPE_St }, /* St */ |
{ NULL, TYPE_St, 0 }, /* St */ |
{ NULL, TYPE_Va }, /* Va */ |
{ parse_mdoc_Va, TYPE_Va, 0 }, /* Va */ |
{ parse_mdoc_body, TYPE_Va }, /* Vt */ |
{ parse_mdoc_Va, TYPE_Vt, 0 }, /* Vt */ |
{ parse_mdoc_Xr, 0 }, /* Xr */ |
{ parse_mdoc_Xr, 0, 0 }, /* Xr */ |
{ NULL, 0 }, /* %A */ |
{ NULL, 0, 0 }, /* %A */ |
{ NULL, 0 }, /* %B */ |
{ NULL, 0, 0 }, /* %B */ |
{ NULL, 0 }, /* %D */ |
{ NULL, 0, 0 }, /* %D */ |
{ NULL, 0 }, /* %I */ |
{ NULL, 0, 0 }, /* %I */ |
{ NULL, 0 }, /* %J */ |
{ NULL, 0, 0 }, /* %J */ |
{ NULL, 0 }, /* %N */ |
{ NULL, 0, 0 }, /* %N */ |
{ NULL, 0 }, /* %O */ |
{ NULL, 0, 0 }, /* %O */ |
{ NULL, 0 }, /* %P */ |
{ NULL, 0, 0 }, /* %P */ |
{ NULL, 0 }, /* %R */ |
{ NULL, 0, 0 }, /* %R */ |
{ NULL, 0 }, /* %T */ |
{ NULL, 0, 0 }, /* %T */ |
{ NULL, 0 }, /* %V */ |
{ NULL, 0, 0 }, /* %V */ |
{ NULL, 0 }, /* Ac */ |
{ NULL, 0, 0 }, /* Ac */ |
{ NULL, 0 }, /* Ao */ |
{ NULL, 0, 0 }, /* Ao */ |
{ NULL, 0 }, /* Aq */ |
{ NULL, 0, 0 }, /* Aq */ |
{ NULL, TYPE_At }, /* At */ |
{ NULL, TYPE_At, 0 }, /* At */ |
{ NULL, 0 }, /* Bc */ |
{ NULL, 0, 0 }, /* Bc */ |
{ NULL, 0 }, /* Bf */ |
{ NULL, 0, 0 }, /* Bf */ |
{ NULL, 0 }, /* Bo */ |
{ NULL, 0, 0 }, /* Bo */ |
{ NULL, 0 }, /* Bq */ |
{ NULL, 0, 0 }, /* Bq */ |
{ NULL, TYPE_Bsx }, /* Bsx */ |
{ NULL, TYPE_Bsx, NODE_NOSRC }, /* Bsx */ |
{ NULL, TYPE_Bx }, /* Bx */ |
{ NULL, TYPE_Bx, NODE_NOSRC }, /* Bx */ |
{ NULL, 0 }, /* Db */ |
{ NULL, 0, 0 }, /* Db */ |
{ NULL, 0 }, /* Dc */ |
{ NULL, 0, 0 }, /* Dc */ |
{ NULL, 0 }, /* Do */ |
{ NULL, 0, 0 }, /* Do */ |
{ NULL, 0 }, /* Dq */ |
{ NULL, 0, 0 }, /* Dq */ |
{ NULL, 0 }, /* Ec */ |
{ NULL, 0, 0 }, /* Ec */ |
{ NULL, 0 }, /* Ef */ |
{ NULL, 0, 0 }, /* Ef */ |
{ NULL, TYPE_Em }, /* Em */ |
{ NULL, TYPE_Em, 0 }, /* Em */ |
{ NULL, 0 }, /* Eo */ |
{ NULL, 0, 0 }, /* Eo */ |
{ NULL, TYPE_Fx }, /* Fx */ |
{ NULL, TYPE_Fx, NODE_NOSRC }, /* Fx */ |
{ NULL, TYPE_Ms }, /* Ms */ |
{ NULL, TYPE_Ms, 0 }, /* Ms */ |
{ NULL, 0 }, /* No */ |
{ NULL, 0, 0 }, /* No */ |
{ NULL, 0 }, /* Ns */ |
{ NULL, 0, 0 }, /* Ns */ |
{ NULL, TYPE_Nx }, /* Nx */ |
{ NULL, TYPE_Nx, NODE_NOSRC }, /* Nx */ |
{ NULL, TYPE_Ox }, /* Ox */ |
{ NULL, TYPE_Ox, NODE_NOSRC }, /* Ox */ |
{ NULL, 0 }, /* Pc */ |
{ NULL, 0, 0 }, /* Pc */ |
{ NULL, 0 }, /* Pf */ |
{ NULL, 0, 0 }, /* Pf */ |
{ NULL, 0 }, /* Po */ |
{ NULL, 0, 0 }, /* Po */ |
{ NULL, 0 }, /* Pq */ |
{ NULL, 0, 0 }, /* Pq */ |
{ NULL, 0 }, /* Qc */ |
{ NULL, 0, 0 }, /* Qc */ |
{ NULL, 0 }, /* Ql */ |
{ NULL, 0, 0 }, /* Ql */ |
{ NULL, 0 }, /* Qo */ |
{ NULL, 0, 0 }, /* Qo */ |
{ NULL, 0 }, /* Qq */ |
{ NULL, 0, 0 }, /* Qq */ |
{ NULL, 0 }, /* Re */ |
{ NULL, 0, 0 }, /* Re */ |
{ NULL, 0 }, /* Rs */ |
{ NULL, 0, 0 }, /* Rs */ |
{ NULL, 0 }, /* Sc */ |
{ NULL, 0, 0 }, /* Sc */ |
{ NULL, 0 }, /* So */ |
{ NULL, 0, 0 }, /* So */ |
{ NULL, 0 }, /* Sq */ |
{ NULL, 0, 0 }, /* Sq */ |
{ NULL, 0 }, /* Sm */ |
{ NULL, 0, 0 }, /* Sm */ |
{ NULL, 0 }, /* Sx */ |
{ NULL, 0, 0 }, /* Sx */ |
{ NULL, TYPE_Sy }, /* Sy */ |
{ NULL, TYPE_Sy, 0 }, /* Sy */ |
{ NULL, TYPE_Tn }, /* Tn */ |
{ NULL, TYPE_Tn, 0 }, /* Tn */ |
{ NULL, 0 }, /* Ux */ |
{ NULL, 0, NODE_NOSRC }, /* Ux */ |
{ NULL, 0 }, /* Xc */ |
{ NULL, 0, 0 }, /* Xc */ |
{ NULL, 0 }, /* Xo */ |
{ NULL, 0, 0 }, /* Xo */ |
{ parse_mdoc_head, 0 }, /* Fo */ |
{ parse_mdoc_Fo, 0, 0 }, /* Fo */ |
{ NULL, 0 }, /* Fc */ |
{ NULL, 0, 0 }, /* Fc */ |
{ NULL, 0 }, /* Oo */ |
{ NULL, 0, 0 }, /* Oo */ |
{ NULL, 0 }, /* Oc */ |
{ NULL, 0, 0 }, /* Oc */ |
{ NULL, 0 }, /* Bk */ |
{ NULL, 0, 0 }, /* Bk */ |
{ NULL, 0 }, /* Ek */ |
{ NULL, 0, 0 }, /* Ek */ |
{ NULL, 0 }, /* Bt */ |
{ NULL, 0, 0 }, /* Bt */ |
{ NULL, 0 }, /* Hf */ |
{ NULL, 0, 0 }, /* Hf */ |
{ NULL, 0 }, /* Fr */ |
{ NULL, 0, 0 }, /* Fr */ |
{ NULL, 0 }, /* Ud */ |
{ NULL, 0, 0 }, /* Ud */ |
{ NULL, TYPE_Lb }, /* Lb */ |
{ NULL, TYPE_Lb, NODE_NOSRC }, /* Lb */ |
{ NULL, 0 }, /* Lp */ |
{ NULL, 0, 0 }, /* Lp */ |
{ NULL, TYPE_Lk }, /* Lk */ |
{ NULL, TYPE_Lk, 0 }, /* Lk */ |
{ NULL, TYPE_Mt }, /* Mt */ |
{ NULL, TYPE_Mt, NODE_NOSRC }, /* Mt */ |
{ NULL, 0 }, /* Brq */ |
{ NULL, 0, 0 }, /* Brq */ |
{ NULL, 0 }, /* Bro */ |
{ NULL, 0, 0 }, /* Bro */ |
{ NULL, 0 }, /* Brc */ |
{ NULL, 0, 0 }, /* Brc */ |
{ NULL, 0 }, /* %C */ |
{ NULL, 0, 0 }, /* %C */ |
{ NULL, 0 }, /* Es */ |
{ NULL, 0, 0 }, /* Es */ |
{ NULL, 0 }, /* En */ |
{ NULL, 0, 0 }, /* En */ |
{ NULL, TYPE_Dx }, /* Dx */ |
{ NULL, TYPE_Dx, NODE_NOSRC }, /* Dx */ |
{ NULL, 0 }, /* %Q */ |
{ NULL, 0, 0 }, /* %Q */ |
{ NULL, 0 }, /* br */ |
{ NULL, 0, 0 }, /* %U */ |
{ NULL, 0 }, /* sp */ |
{ NULL, 0, 0 }, /* Ta */ |
{ NULL, 0 }, /* %U */ |
|
{ NULL, 0 }, /* Ta */ |
|
}; |
}; |
|
|
|
|
int |
int |
main(int argc, char *argv[]) |
mandocdb(int argc, char *argv[]) |
{ |
{ |
int ch, i; |
struct manconf conf; |
size_t j, sz; |
|
const char *path_arg; |
|
struct mchars *mc; |
|
struct manpaths dirs; |
|
struct mparse *mp; |
struct mparse *mp; |
struct ohash_info mpages_info, mlinks_info; |
struct dba *dba; |
|
const char *path_arg, *progname; |
|
size_t j, sz; |
|
int ch, i; |
|
|
memset(stmts, 0, STMT__MAX * sizeof(sqlite3_stmt *)); |
#if HAVE_PLEDGE |
memset(&dirs, 0, sizeof(struct manpaths)); |
if (pledge("stdio rpath wpath cpath", NULL) == -1) { |
|
warn("pledge"); |
|
return (int)MANDOCLEVEL_SYSERR; |
|
} |
|
#endif |
|
|
mpages_info.alloc = mlinks_info.alloc = hash_alloc; |
#if HAVE_SANDBOX_INIT |
mpages_info.calloc = mlinks_info.calloc = hash_calloc; |
if (sandbox_init(kSBXProfileNoInternet, SANDBOX_NAMED, NULL) == -1) { |
mpages_info.free = mlinks_info.free = hash_free; |
warnx("sandbox_init"); |
|
return (int)MANDOCLEVEL_SYSERR; |
|
} |
|
#endif |
|
|
mpages_info.key_offset = offsetof(struct mpage, inodev); |
memset(&conf, 0, sizeof(conf)); |
mlinks_info.key_offset = offsetof(struct mlink, file); |
|
|
|
progname = strrchr(argv[0], '/'); |
|
if (progname == NULL) |
|
progname = argv[0]; |
|
else |
|
++progname; |
|
|
|
/* |
/* |
* We accept a few different invocations. |
* We accept a few different invocations. |
* The CHECKOP macro makes sure that invocation styles don't |
* The CHECKOP macro makes sure that invocation styles don't |
* clobber each other. |
* clobber each other. |
*/ |
*/ |
#define CHECKOP(_op, _ch) do \ |
#define CHECKOP(_op, _ch) do \ |
if (OP_DEFAULT != (_op)) { \ |
if ((_op) != OP_DEFAULT) { \ |
fprintf(stderr, "%s: -%c: Conflicting option\n", \ |
warnx("-%c: Conflicting option", (_ch)); \ |
progname, (_ch)); \ |
|
goto usage; \ |
goto usage; \ |
} while (/*CONSTCOND*/0) |
} while (/*CONSTCOND*/0) |
|
|
|
mparse_options = MPARSE_VALIDATE; |
path_arg = NULL; |
path_arg = NULL; |
op = OP_DEFAULT; |
op = OP_DEFAULT; |
|
|
while (-1 != (ch = getopt(argc, argv, "aC:Dd:npQT:tu:v"))) |
while ((ch = getopt(argc, argv, "aC:Dd:npQT:tu:v")) != -1) |
switch (ch) { |
switch (ch) { |
case 'a': |
case 'a': |
use_all = 1; |
use_all = 1; |
Line 391 main(int argc, char *argv[]) |
|
Line 385 main(int argc, char *argv[]) |
|
mparse_options |= MPARSE_QUICK; |
mparse_options |= MPARSE_QUICK; |
break; |
break; |
case 'T': |
case 'T': |
if (strcmp(optarg, "utf8")) { |
if (strcmp(optarg, "utf8") != 0) { |
fprintf(stderr, "%s: -T%s: " |
warnx("-T%s: Unsupported output format", |
"Unsupported output format\n", |
optarg); |
progname, optarg); |
|
goto usage; |
goto usage; |
} |
} |
write_utf8 = 1; |
write_utf8 = 1; |
Line 420 main(int argc, char *argv[]) |
|
Line 413 main(int argc, char *argv[]) |
|
argc -= optind; |
argc -= optind; |
argv += optind; |
argv += optind; |
|
|
if (OP_CONFFILE == op && argc > 0) { |
#if HAVE_PLEDGE |
fprintf(stderr, "%s: -C: Too many arguments\n", |
if (nodb) { |
progname); |
if (pledge("stdio rpath", NULL) == -1) { |
|
warn("pledge"); |
|
return (int)MANDOCLEVEL_SYSERR; |
|
} |
|
} |
|
#endif |
|
|
|
if (op == OP_CONFFILE && argc > 0) { |
|
warnx("-C: Too many arguments"); |
goto usage; |
goto usage; |
} |
} |
|
|
exitcode = (int)MANDOCLEVEL_OK; |
exitcode = (int)MANDOCLEVEL_OK; |
mp = mparse_alloc(mparse_options, MANDOCLEVEL_FATAL, NULL, NULL); |
mchars_alloc(); |
mc = mchars_alloc(); |
mp = mparse_alloc(mparse_options, MANDOC_OS_OTHER, NULL); |
|
mandoc_ohash_init(&mpages, 6, offsetof(struct mpage, inodev)); |
|
mandoc_ohash_init(&mlinks, 6, offsetof(struct mlink, file)); |
|
|
ohash_init(&mpages, 6, &mpages_info); |
if (op == OP_UPDATE || op == OP_DELETE || op == OP_TEST) { |
ohash_init(&mlinks, 6, &mlinks_info); |
|
|
|
if (OP_UPDATE == op || OP_DELETE == op || OP_TEST == op) { |
|
|
|
/* |
/* |
* Most of these deal with a specific directory. |
* Most of these deal with a specific directory. |
* Jump into that directory first. |
* Jump into that directory first. |
*/ |
*/ |
if (OP_TEST != op && 0 == set_basedir(path_arg)) |
if (op != OP_TEST && set_basedir(path_arg, 1) == 0) |
goto out; |
goto out; |
|
|
if (dbopen(1)) { |
dba = nodb ? dba_new(128) : dba_read(MANDOC_DB); |
|
if (dba != NULL) { |
/* |
/* |
* The existing database is usable. Process |
* The existing database is usable. Process |
* all files specified on the command-line. |
* all files specified on the command-line. |
Line 450 main(int argc, char *argv[]) |
|
Line 451 main(int argc, char *argv[]) |
|
use_all = 1; |
use_all = 1; |
for (i = 0; i < argc; i++) |
for (i = 0; i < argc; i++) |
filescan(argv[i]); |
filescan(argv[i]); |
if (OP_TEST != op) |
if (nodb == 0) |
dbprune(); |
dbprune(dba); |
} else { |
} else { |
/* |
/* Database missing or corrupt. */ |
* Database missing or corrupt. |
if (op != OP_UPDATE || errno != ENOENT) |
* Recreate from scratch. |
say(MANDOC_DB, "%s: Automatically recreating" |
*/ |
" from scratch", strerror(errno)); |
exitcode = (int)MANDOCLEVEL_OK; |
exitcode = (int)MANDOCLEVEL_OK; |
op = OP_DEFAULT; |
op = OP_DEFAULT; |
if (0 == treescan()) |
if (treescan() == 0) |
goto out; |
goto out; |
if (0 == dbopen(0)) |
dba = dba_new(128); |
goto out; |
|
} |
} |
if (OP_DELETE != op) |
if (op != OP_DELETE) |
mpages_merge(mc, mp); |
mpages_merge(dba, mp); |
dbclose(OP_DEFAULT == op ? 0 : 1); |
if (nodb == 0) |
|
dbwrite(dba); |
|
dba_free(dba); |
} else { |
} else { |
/* |
/* |
* If we have arguments, use them as our manpaths. |
* If we have arguments, use them as our manpaths. |
* If we don't, grok from manpath(1) or however else |
* If we don't, use man.conf(5). |
* manpath_parse() wants to do it. |
|
*/ |
*/ |
if (argc > 0) { |
if (argc > 0) { |
dirs.paths = mandoc_reallocarray(NULL, |
conf.manpath.paths = mandoc_reallocarray(NULL, |
argc, sizeof(char *)); |
argc, sizeof(char *)); |
dirs.sz = (size_t)argc; |
conf.manpath.sz = (size_t)argc; |
for (i = 0; i < argc; i++) |
for (i = 0; i < argc; i++) |
dirs.paths[i] = mandoc_strdup(argv[i]); |
conf.manpath.paths[i] = mandoc_strdup(argv[i]); |
} else |
} else |
manpath_parse(&dirs, path_arg, NULL, NULL); |
manconf_parse(&conf, path_arg, NULL, NULL); |
|
|
if (0 == dirs.sz) { |
if (conf.manpath.sz == 0) { |
exitcode = (int)MANDOCLEVEL_BADARG; |
exitcode = (int)MANDOCLEVEL_BADARG; |
say("", "Empty manpath"); |
say("", "Empty manpath"); |
} |
} |
Line 493 main(int argc, char *argv[]) |
|
Line 494 main(int argc, char *argv[]) |
|
* Ignore zero-length directories and strip trailing |
* Ignore zero-length directories and strip trailing |
* slashes. |
* slashes. |
*/ |
*/ |
for (j = 0; j < dirs.sz; j++) { |
for (j = 0; j < conf.manpath.sz; j++) { |
sz = strlen(dirs.paths[j]); |
sz = strlen(conf.manpath.paths[j]); |
if (sz && '/' == dirs.paths[j][sz - 1]) |
if (sz && conf.manpath.paths[j][sz - 1] == '/') |
dirs.paths[j][--sz] = '\0'; |
conf.manpath.paths[j][--sz] = '\0'; |
if (0 == sz) |
if (sz == 0) |
continue; |
continue; |
|
|
if (j) { |
if (j) { |
ohash_init(&mpages, 6, &mpages_info); |
mandoc_ohash_init(&mpages, 6, |
ohash_init(&mlinks, 6, &mlinks_info); |
offsetof(struct mpage, inodev)); |
|
mandoc_ohash_init(&mlinks, 6, |
|
offsetof(struct mlink, file)); |
} |
} |
|
|
if (0 == set_basedir(dirs.paths[j])) |
if (set_basedir(conf.manpath.paths[j], argc > 0) == 0) |
goto out; |
continue; |
if (0 == treescan()) |
if (treescan() == 0) |
goto out; |
continue; |
if (0 == dbopen(0)) |
dba = dba_new(128); |
goto out; |
mpages_merge(dba, mp); |
|
if (nodb == 0) |
|
dbwrite(dba); |
|
dba_free(dba); |
|
|
mpages_merge(mc, mp); |
if (j + 1 < conf.manpath.sz) { |
if (warnings && !nodb && |
|
! (MPARSE_QUICK & mparse_options)) |
|
names_check(); |
|
dbclose(0); |
|
|
|
if (j + 1 < dirs.sz) { |
|
mpages_free(); |
mpages_free(); |
ohash_delete(&mpages); |
ohash_delete(&mpages); |
ohash_delete(&mlinks); |
ohash_delete(&mlinks); |
Line 526 main(int argc, char *argv[]) |
|
Line 526 main(int argc, char *argv[]) |
|
} |
} |
} |
} |
out: |
out: |
manpath_free(&dirs); |
manconf_free(&conf); |
mchars_free(mc); |
|
mparse_free(mp); |
mparse_free(mp); |
|
mchars_free(); |
mpages_free(); |
mpages_free(); |
ohash_delete(&mpages); |
ohash_delete(&mpages); |
ohash_delete(&mlinks); |
ohash_delete(&mlinks); |
return(exitcode); |
#if DEBUG_MEMORY |
|
mandoc_d_finish(); |
|
#endif |
|
return exitcode; |
usage: |
usage: |
|
progname = getprogname(); |
fprintf(stderr, "usage: %s [-aDnpQ] [-C file] [-Tutf8]\n" |
fprintf(stderr, "usage: %s [-aDnpQ] [-C file] [-Tutf8]\n" |
" %s [-aDnpQ] [-Tutf8] dir ...\n" |
" %s [-aDnpQ] [-Tutf8] dir ...\n" |
" %s [-DnpQ] [-Tutf8] -d dir [file ...]\n" |
" %s [-DnpQ] [-Tutf8] -d dir [file ...]\n" |
" %s [-Dnp] -u dir [file ...]\n" |
" %s [-Dnp] -u dir [file ...]\n" |
" %s [-Q] -t file ...\n", |
" %s [-Q] -t file ...\n", |
progname, progname, progname, |
progname, progname, progname, progname, progname); |
progname, progname); |
|
|
|
return((int)MANDOCLEVEL_BADARG); |
return (int)MANDOCLEVEL_BADARG; |
} |
} |
|
|
/* |
/* |
|
* To get a singly linked list in alpha order while inserting entries |
|
* at the beginning, process directory entries in reverse alpha order. |
|
*/ |
|
static int |
|
#if HAVE_FTS_COMPARE_CONST |
|
fts_compare(const FTSENT *const *a, const FTSENT *const *b) |
|
#else |
|
fts_compare(const FTSENT **a, const FTSENT **b) |
|
#endif |
|
{ |
|
return -strcmp((*a)->fts_name, (*b)->fts_name); |
|
} |
|
|
|
/* |
* Scan a directory tree rooted at "basedir" for manpages. |
* Scan a directory tree rooted at "basedir" for manpages. |
* We use fts(), scanning directory parts along the way for clues to our |
* We use fts(), scanning directory parts along the way for clues to our |
* section and architecture. |
* section and architecture. |
|
|
* or |
* or |
* [./]cat<section>[/<arch>]/<name>.0 |
* [./]cat<section>[/<arch>]/<name>.0 |
* |
* |
* TODO: accomodate for multi-language directories. |
* TODO: accommodate for multi-language directories. |
*/ |
*/ |
static int |
static int |
treescan(void) |
treescan(void) |
|
|
FTS *f; |
FTS *f; |
FTSENT *ff; |
FTSENT *ff; |
struct mlink *mlink; |
struct mlink *mlink; |
int dform, gzip; |
int gzip; |
|
enum form dform; |
char *dsec, *arch, *fsec, *cp; |
char *dsec, *arch, *fsec, *cp; |
const char *path; |
const char *path; |
const char *argv[2]; |
const char *argv[2]; |
|
|
argv[0] = "."; |
argv[0] = "."; |
argv[1] = (char *)NULL; |
argv[1] = NULL; |
|
|
f = fts_open((char * const *)argv, |
f = fts_open((char * const *)argv, FTS_PHYSICAL | FTS_NOCHDIR, |
FTS_PHYSICAL | FTS_NOCHDIR, NULL); |
fts_compare); |
if (NULL == f) { |
if (f == NULL) { |
exitcode = (int)MANDOCLEVEL_SYSERR; |
exitcode = (int)MANDOCLEVEL_SYSERR; |
say("", "&fts_open"); |
say("", "&fts_open"); |
return(0); |
return 0; |
} |
} |
|
|
dsec = arch = NULL; |
dsec = arch = NULL; |
dform = FORM_NONE; |
dform = FORM_NONE; |
|
|
while (NULL != (ff = fts_read(f))) { |
while ((ff = fts_read(f)) != NULL) { |
path = ff->fts_path + 2; |
path = ff->fts_path + 2; |
switch (ff->fts_info) { |
switch (ff->fts_info) { |
|
|
|
|
* then get handled just like regular files. |
* then get handled just like regular files. |
*/ |
*/ |
case FTS_SL: |
case FTS_SL: |
if (NULL == realpath(path, buf)) { |
if (realpath(path, buf) == NULL) { |
if (warnings) |
if (warnings) |
say(path, "&realpath"); |
say(path, "&realpath"); |
continue; |
continue; |
} |
} |
if (strstr(buf, basedir) != buf) { |
if (strncmp(buf, basedir, basedir_len) != 0 |
|
#ifdef READ_ALLOWED_PATH |
|
&& !read_allowed(buf) |
|
#endif |
|
) { |
if (warnings) say("", |
if (warnings) say("", |
"%s: outside base directory", buf); |
"%s: outside base directory", buf); |
continue; |
continue; |
} |
} |
/* Use logical inode to avoid mpages dupe. */ |
/* Use logical inode to avoid mpages dupe. */ |
if (-1 == stat(path, ff->fts_statp)) { |
if (stat(path, ff->fts_statp) == -1) { |
if (warnings) |
if (warnings) |
say(path, "&stat"); |
say(path, "&stat"); |
continue; |
continue; |
} |
} |
|
if ((ff->fts_statp->st_mode & S_IFMT) != S_IFREG) |
|
continue; |
/* FALLTHROUGH */ |
/* FALLTHROUGH */ |
|
|
/* |
/* |
|
|
* stored directory data and handling the filename. |
* stored directory data and handling the filename. |
*/ |
*/ |
case FTS_F: |
case FTS_F: |
if (0 == strcmp(path, MANDOC_DB)) |
if ( ! strcmp(path, MANDOC_DB)) |
continue; |
continue; |
if ( ! use_all && ff->fts_level < 2) { |
if ( ! use_all && ff->fts_level < 2) { |
if (warnings) |
if (warnings) |
|
|
} |
} |
gzip = 0; |
gzip = 0; |
fsec = NULL; |
fsec = NULL; |
while (NULL == fsec) { |
while (fsec == NULL) { |
fsec = strrchr(ff->fts_name, '.'); |
fsec = strrchr(ff->fts_name, '.'); |
if (NULL == fsec || strcmp(fsec+1, "gz")) |
if (fsec == NULL || strcmp(fsec+1, "gz")) |
break; |
break; |
gzip = 1; |
gzip = 1; |
*fsec = '\0'; |
*fsec = '\0'; |
fsec = NULL; |
fsec = NULL; |
} |
} |
if (NULL == fsec) { |
if (fsec == NULL) { |
if ( ! use_all) { |
if ( ! use_all) { |
if (warnings) |
if (warnings) |
say(path, |
say(path, |
"No filename suffix"); |
"No filename suffix"); |
continue; |
continue; |
} |
} |
} else if (0 == strcmp(++fsec, "html")) { |
} else if ( ! strcmp(++fsec, "html")) { |
if (warnings) |
if (warnings) |
say(path, "Skip html"); |
say(path, "Skip html"); |
continue; |
continue; |
} else if (0 == strcmp(fsec, "ps")) { |
} else if ( ! strcmp(fsec, "ps")) { |
if (warnings) |
if (warnings) |
say(path, "Skip ps"); |
say(path, "Skip ps"); |
continue; |
continue; |
} else if (0 == strcmp(fsec, "pdf")) { |
} else if ( ! strcmp(fsec, "pdf")) { |
if (warnings) |
if (warnings) |
say(path, "Skip pdf"); |
say(path, "Skip pdf"); |
continue; |
continue; |
} else if ( ! use_all && |
} else if ( ! use_all && |
((FORM_SRC == dform && strcmp(fsec, dsec)) || |
((dform == FORM_SRC && |
(FORM_CAT == dform && strcmp(fsec, "0")))) { |
strncmp(fsec, dsec, strlen(dsec))) || |
|
(dform == FORM_CAT && strcmp(fsec, "0")))) { |
if (warnings) |
if (warnings) |
say(path, "Wrong filename suffix"); |
say(path, "Wrong filename suffix"); |
continue; |
continue; |
|
|
continue; |
continue; |
|
|
case FTS_D: |
case FTS_D: |
/* FALLTHROUGH */ |
|
case FTS_DP: |
case FTS_DP: |
break; |
break; |
|
|
|
|
* If we're not in use_all, enforce it. |
* If we're not in use_all, enforce it. |
*/ |
*/ |
cp = ff->fts_name; |
cp = ff->fts_name; |
if (FTS_DP == ff->fts_info) |
if (ff->fts_info == FTS_DP) { |
|
dform = FORM_NONE; |
|
dsec = NULL; |
break; |
break; |
|
} |
|
|
if (0 == strncmp(cp, "man", 3)) { |
if ( ! strncmp(cp, "man", 3)) { |
dform = FORM_SRC; |
dform = FORM_SRC; |
dsec = cp + 3; |
dsec = cp + 3; |
} else if (0 == strncmp(cp, "cat", 3)) { |
} else if ( ! strncmp(cp, "cat", 3)) { |
dform = FORM_CAT; |
dform = FORM_CAT; |
dsec = cp + 3; |
dsec = cp + 3; |
} else { |
} else { |
|
|
dsec = NULL; |
dsec = NULL; |
} |
} |
|
|
if (NULL != dsec || use_all) |
if (dsec != NULL || use_all) |
break; |
break; |
|
|
if (warnings) |
if (warnings) |
|
|
* Possibly our architecture. |
* Possibly our architecture. |
* If we're descending, keep tabs on it. |
* If we're descending, keep tabs on it. |
*/ |
*/ |
if (FTS_DP != ff->fts_info && NULL != dsec) |
if (ff->fts_info != FTS_DP && dsec != NULL) |
arch = ff->fts_name; |
arch = ff->fts_name; |
else |
else |
arch = NULL; |
arch = NULL; |
break; |
break; |
default: |
default: |
if (FTS_DP == ff->fts_info || use_all) |
if (ff->fts_info == FTS_DP || use_all) |
break; |
break; |
if (warnings) |
if (warnings) |
say(path, "Extraneous directory part"); |
say(path, "Extraneous directory part"); |
|
|
} |
} |
|
|
fts_close(f); |
fts_close(f); |
return(1); |
return 1; |
} |
} |
|
|
/* |
/* |
|
|
* See treescan() for the fts(3) version of this. |
* See treescan() for the fts(3) version of this. |
*/ |
*/ |
static void |
static void |
filescan(const char *file) |
filescan(const char *infile) |
{ |
{ |
char buf[PATH_MAX]; |
|
struct stat st; |
struct stat st; |
struct mlink *mlink; |
struct mlink *mlink; |
char *p, *start; |
char *linkfile, *p, *realdir, *start, *usefile; |
|
size_t realdir_len; |
|
|
assert(use_all); |
assert(use_all); |
|
|
if (0 == strncmp(file, "./", 2)) |
if (strncmp(infile, "./", 2) == 0) |
file += 2; |
infile += 2; |
|
|
/* |
/* |
* We have to do lstat(2) before realpath(3) loses |
* We have to do lstat(2) before realpath(3) loses |
Line 780 filescan(const char *file) |
|
Line 807 filescan(const char *file) |
|
* we want to use the orginal file name, while for |
* we want to use the orginal file name, while for |
* regular files, we want to use the real path. |
* regular files, we want to use the real path. |
*/ |
*/ |
if (-1 == lstat(file, &st)) { |
if (lstat(infile, &st) == -1) { |
exitcode = (int)MANDOCLEVEL_BADARG; |
exitcode = (int)MANDOCLEVEL_BADARG; |
say(file, "&lstat"); |
say(infile, "&lstat"); |
return; |
return; |
} else if (0 == ((S_IFREG | S_IFLNK) & st.st_mode)) { |
} else if (S_ISREG(st.st_mode) == 0 && S_ISLNK(st.st_mode) == 0) { |
exitcode = (int)MANDOCLEVEL_BADARG; |
exitcode = (int)MANDOCLEVEL_BADARG; |
say(file, "Not a regular file"); |
say(infile, "Not a regular file"); |
return; |
return; |
} |
} |
|
|
Line 794 filescan(const char *file) |
|
Line 821 filescan(const char *file) |
|
* We have to resolve the file name to the real path |
* We have to resolve the file name to the real path |
* in any case for the base directory check. |
* in any case for the base directory check. |
*/ |
*/ |
if (NULL == realpath(file, buf)) { |
if ((usefile = realpath(infile, NULL)) == NULL) { |
exitcode = (int)MANDOCLEVEL_BADARG; |
exitcode = (int)MANDOCLEVEL_BADARG; |
say(file, "&realpath"); |
say(infile, "&realpath"); |
return; |
return; |
} |
} |
|
|
if (OP_TEST == op) |
if (op == OP_TEST) |
start = buf; |
start = usefile; |
else if (strstr(buf, basedir) == buf) |
else if (strncmp(usefile, basedir, basedir_len) == 0) |
start = buf + strlen(basedir); |
start = usefile + basedir_len; |
|
#ifdef READ_ALLOWED_PATH |
|
else if (read_allowed(usefile)) |
|
start = usefile; |
|
#endif |
else { |
else { |
exitcode = (int)MANDOCLEVEL_BADARG; |
exitcode = (int)MANDOCLEVEL_BADARG; |
say("", "%s: outside base directory", buf); |
say("", "%s: outside base directory", infile); |
|
free(usefile); |
return; |
return; |
} |
} |
|
|
Line 814 filescan(const char *file) |
|
Line 846 filescan(const char *file) |
|
* Now we are sure the file is inside our tree. |
* Now we are sure the file is inside our tree. |
* If it is a symbolic link, ignore the real path |
* If it is a symbolic link, ignore the real path |
* and use the original name. |
* and use the original name. |
* This implies passing stuff like "cat1/../man1/foo.1" |
|
* on the command line won't work. So don't do that. |
|
* Note the stat(2) can still fail if the link target |
|
* doesn't exist. |
|
*/ |
*/ |
if (S_IFLNK & st.st_mode) { |
do { |
if (-1 == stat(buf, &st)) { |
if (S_ISLNK(st.st_mode) == 0) |
|
break; |
|
|
|
/* |
|
* Some implementations of realpath(3) may succeed |
|
* even if the target of the link does not exist, |
|
* so check again for extra safety. |
|
*/ |
|
if (stat(usefile, &st) == -1) { |
exitcode = (int)MANDOCLEVEL_BADARG; |
exitcode = (int)MANDOCLEVEL_BADARG; |
say(file, "&stat"); |
say(infile, "&stat"); |
|
free(usefile); |
return; |
return; |
} |
} |
if (strlcpy(buf, file, sizeof(buf)) >= sizeof(buf)) { |
linkfile = mandoc_strdup(infile); |
say(file, "Filename too long"); |
if (op == OP_TEST) { |
return; |
free(usefile); |
|
start = usefile = linkfile; |
|
break; |
} |
} |
start = buf; |
if (strncmp(infile, basedir, basedir_len) == 0) { |
if (OP_TEST != op && strstr(buf, basedir) == buf) |
free(usefile); |
start += strlen(basedir); |
usefile = linkfile; |
} |
start = usefile + basedir_len; |
|
break; |
|
} |
|
|
|
/* |
|
* This symbolic link points into the basedir |
|
* from the outside. Let's see whether any of |
|
* the parent directories resolve to the basedir. |
|
*/ |
|
p = strchr(linkfile, '\0'); |
|
do { |
|
while (*--p != '/') |
|
continue; |
|
*p = '\0'; |
|
if ((realdir = realpath(linkfile, NULL)) == NULL) { |
|
exitcode = (int)MANDOCLEVEL_BADARG; |
|
say(infile, "&realpath"); |
|
free(linkfile); |
|
free(usefile); |
|
return; |
|
} |
|
realdir_len = strlen(realdir) + 1; |
|
free(realdir); |
|
*p = '/'; |
|
} while (realdir_len > basedir_len); |
|
|
|
/* |
|
* If one of the directories resolves to the basedir, |
|
* use the rest of the original name. |
|
* Otherwise, the best we can do |
|
* is to use the filename pointed to. |
|
*/ |
|
if (realdir_len == basedir_len) { |
|
free(usefile); |
|
usefile = linkfile; |
|
start = p + 1; |
|
} else { |
|
free(linkfile); |
|
start = usefile + basedir_len; |
|
} |
|
} while (/* CONSTCOND */ 0); |
|
|
mlink = mandoc_calloc(1, sizeof(struct mlink)); |
mlink = mandoc_calloc(1, sizeof(struct mlink)); |
|
mlink->dform = FORM_NONE; |
if (strlcpy(mlink->file, start, sizeof(mlink->file)) >= |
if (strlcpy(mlink->file, start, sizeof(mlink->file)) >= |
sizeof(mlink->file)) { |
sizeof(mlink->file)) { |
say(start, "Filename too long"); |
say(start, "Filename too long"); |
|
free(mlink); |
|
free(usefile); |
return; |
return; |
} |
} |
|
|
/* |
/* |
|
* In test mode or when the original name is absolute |
|
* but outside our tree, guess the base directory. |
|
*/ |
|
|
|
if (op == OP_TEST || (start == usefile && *start == '/')) { |
|
if (strncmp(usefile, "man/", 4) == 0) |
|
start = usefile + 4; |
|
else if ((start = strstr(usefile, "/man/")) != NULL) |
|
start += 5; |
|
else |
|
start = usefile; |
|
} |
|
|
|
/* |
* First try to guess our directory structure. |
* First try to guess our directory structure. |
* If we find a separator, try to look for man* or cat*. |
* If we find a separator, try to look for man* or cat*. |
* If we find one of these and what's underneath is a directory, |
* If we find one of these and what's underneath is a directory, |
* assume it's an architecture. |
* assume it's an architecture. |
*/ |
*/ |
if (NULL != (p = strchr(start, '/'))) { |
if ((p = strchr(start, '/')) != NULL) { |
*p++ = '\0'; |
*p++ = '\0'; |
if (0 == strncmp(start, "man", 3)) { |
if (strncmp(start, "man", 3) == 0) { |
mlink->dform = FORM_SRC; |
mlink->dform = FORM_SRC; |
mlink->dsec = start + 3; |
mlink->dsec = start + 3; |
} else if (0 == strncmp(start, "cat", 3)) { |
} else if (strncmp(start, "cat", 3) == 0) { |
mlink->dform = FORM_CAT; |
mlink->dform = FORM_CAT; |
mlink->dsec = start + 3; |
mlink->dsec = start + 3; |
} |
} |
|
|
start = p; |
start = p; |
if (NULL != mlink->dsec && NULL != (p = strchr(start, '/'))) { |
if (mlink->dsec != NULL && (p = strchr(start, '/')) != NULL) { |
*p++ = '\0'; |
*p++ = '\0'; |
mlink->arch = start; |
mlink->arch = start; |
start = p; |
start = p; |
Line 870 filescan(const char *file) |
|
Line 966 filescan(const char *file) |
|
* Suffix of `.0' indicates a catpage, `.1-9' is a manpage. |
* Suffix of `.0' indicates a catpage, `.1-9' is a manpage. |
*/ |
*/ |
p = strrchr(start, '\0'); |
p = strrchr(start, '\0'); |
while (p-- > start && '/' != *p && '.' != *p) |
while (p-- > start && *p != '/' && *p != '.') |
/* Loop. */ ; |
continue; |
|
|
if ('.' == *p) { |
if (*p == '.') { |
*p++ = '\0'; |
*p++ = '\0'; |
mlink->fsec = p; |
mlink->fsec = p; |
} |
} |
Line 883 filescan(const char *file) |
|
Line 979 filescan(const char *file) |
|
* Use the filename portion of the path. |
* Use the filename portion of the path. |
*/ |
*/ |
mlink->name = start; |
mlink->name = start; |
if (NULL != (p = strrchr(start, '/'))) { |
if ((p = strrchr(start, '/')) != NULL) { |
mlink->name = p + 1; |
mlink->name = p + 1; |
*p = '\0'; |
*p = '\0'; |
} |
} |
mlink_add(mlink, &st); |
mlink_add(mlink, &st); |
|
free(usefile); |
} |
} |
|
|
static void |
static void |
Line 917 mlink_add(struct mlink *mlink, const struct stat *st) |
|
Line 1014 mlink_add(struct mlink *mlink, const struct stat *st) |
|
assert(NULL == ohash_find(&mlinks, slot)); |
assert(NULL == ohash_find(&mlinks, slot)); |
ohash_insert(&mlinks, slot, mlink); |
ohash_insert(&mlinks, slot, mlink); |
|
|
|
memset(&inodev, 0, sizeof(inodev)); /* Clear padding. */ |
inodev.st_ino = st->st_ino; |
inodev.st_ino = st->st_ino; |
inodev.st_dev = st->st_dev; |
inodev.st_dev = st->st_dev; |
slot = ohash_lookup_memory(&mpages, (char *)&inodev, |
slot = ohash_lookup_memory(&mpages, (char *)&inodev, |
Line 926 mlink_add(struct mlink *mlink, const struct stat *st) |
|
Line 1024 mlink_add(struct mlink *mlink, const struct stat *st) |
|
mpage = mandoc_calloc(1, sizeof(struct mpage)); |
mpage = mandoc_calloc(1, sizeof(struct mpage)); |
mpage->inodev.st_ino = inodev.st_ino; |
mpage->inodev.st_ino = inodev.st_ino; |
mpage->inodev.st_dev = inodev.st_dev; |
mpage->inodev.st_dev = inodev.st_dev; |
|
mpage->form = FORM_NONE; |
|
mpage->next = mpage_head; |
|
mpage_head = mpage; |
ohash_insert(&mpages, slot, mpage); |
ohash_insert(&mpages, slot, mpage); |
} else |
} else |
mlink->next = mpage->mlinks; |
mlink->next = mpage->mlinks; |
Line 949 mpages_free(void) |
|
Line 1050 mpages_free(void) |
|
{ |
{ |
struct mpage *mpage; |
struct mpage *mpage; |
struct mlink *mlink; |
struct mlink *mlink; |
unsigned int slot; |
|
|
|
mpage = ohash_first(&mpages, &slot); |
while ((mpage = mpage_head) != NULL) { |
while (NULL != mpage) { |
while ((mlink = mpage->mlinks) != NULL) { |
while (NULL != (mlink = mpage->mlinks)) { |
|
mpage->mlinks = mlink->next; |
mpage->mlinks = mlink->next; |
mlink_free(mlink); |
mlink_free(mlink); |
} |
} |
|
mpage_head = mpage->next; |
free(mpage->sec); |
free(mpage->sec); |
free(mpage->arch); |
free(mpage->arch); |
free(mpage->title); |
free(mpage->title); |
free(mpage->desc); |
free(mpage->desc); |
free(mpage); |
free(mpage); |
mpage = ohash_next(&mpages, &slot); |
|
} |
} |
} |
} |
|
|
Line 1039 mlink_check(struct mpage *mpage, struct mlink *mlink) |
|
Line 1138 mlink_check(struct mpage *mpage, struct mlink *mlink) |
|
* architectures. |
* architectures. |
* A few manuals are even shared across completely |
* A few manuals are even shared across completely |
* different architectures, for example fdformat(1) |
* different architectures, for example fdformat(1) |
* on amd64, i386, sparc, and sparc64. |
* on amd64, i386, and sparc64. |
*/ |
*/ |
|
|
if (strcasecmp(mpage->arch, mlink->arch)) |
if (strcasecmp(mpage->arch, mlink->arch)) |
Line 1074 mlink_check(struct mpage *mpage, struct mlink *mlink) |
|
Line 1173 mlink_check(struct mpage *mpage, struct mlink *mlink) |
|
* and filename to determine whether the file is parsable or not. |
* and filename to determine whether the file is parsable or not. |
*/ |
*/ |
static void |
static void |
mpages_merge(struct mchars *mc, struct mparse *mp) |
mpages_merge(struct dba *dba, struct mparse *mp) |
{ |
{ |
char any[] = "any"; |
|
struct ohash_info str_info; |
|
int fd[2]; |
|
struct mpage *mpage, *mpage_dest; |
struct mpage *mpage, *mpage_dest; |
struct mlink *mlink, *mlink_dest; |
struct mlink *mlink, *mlink_dest; |
struct mdoc *mdoc; |
struct roff_meta *meta; |
struct man *man; |
|
char *sodest; |
|
char *cp; |
char *cp; |
pid_t child_pid; |
int fd; |
int status; |
|
unsigned int pslot; |
|
enum mandoclevel lvl; |
|
|
|
str_info.alloc = hash_alloc; |
for (mpage = mpage_head; mpage != NULL; mpage = mpage->next) { |
str_info.calloc = hash_calloc; |
|
str_info.free = hash_free; |
|
str_info.key_offset = offsetof(struct str, key); |
|
|
|
if (0 == nodb) |
|
SQL_EXEC("BEGIN TRANSACTION"); |
|
|
|
mpage = ohash_first(&mpages, &pslot); |
|
while (NULL != mpage) { |
|
mlinks_undupe(mpage); |
mlinks_undupe(mpage); |
if (NULL == mpage->mlinks) { |
if ((mlink = mpage->mlinks) == NULL) |
mpage = ohash_next(&mpages, &pslot); |
|
continue; |
continue; |
} |
|
|
|
name_mask = NAME_MASK; |
name_mask = NAME_MASK; |
ohash_init(&names, 4, &str_info); |
mandoc_ohash_init(&names, 4, offsetof(struct str, key)); |
ohash_init(&strings, 6, &str_info); |
mandoc_ohash_init(&strings, 6, offsetof(struct str, key)); |
mparse_reset(mp); |
mparse_reset(mp); |
mdoc = NULL; |
meta = NULL; |
man = NULL; |
|
sodest = NULL; |
|
child_pid = 0; |
|
fd[0] = -1; |
|
fd[1] = -1; |
|
|
|
if (mpage->mlinks->gzip) { |
if ((fd = mparse_open(mp, mlink->file)) == -1) { |
if (-1 == pipe(fd)) { |
say(mlink->file, "&open"); |
exitcode = (int)MANDOCLEVEL_SYSERR; |
goto nextpage; |
say(mpage->mlinks->file, "&pipe gunzip"); |
|
goto nextpage; |
|
} |
|
switch (child_pid = fork()) { |
|
case -1: |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(mpage->mlinks->file, "&fork gunzip"); |
|
child_pid = 0; |
|
close(fd[1]); |
|
close(fd[0]); |
|
goto nextpage; |
|
case 0: |
|
close(fd[0]); |
|
if (-1 == dup2(fd[1], STDOUT_FILENO)) { |
|
say(mpage->mlinks->file, |
|
"&dup gunzip"); |
|
exit(1); |
|
} |
|
execlp("gunzip", "gunzip", "-c", |
|
mpage->mlinks->file, NULL); |
|
say(mpage->mlinks->file, "&exec gunzip"); |
|
exit(1); |
|
default: |
|
close(fd[1]); |
|
break; |
|
} |
|
} |
} |
|
|
/* |
/* |
* Try interpreting the file as mdoc(7) or man(7) |
* Interpret the file as mdoc(7) or man(7) source |
* source code, unless it is already known to be |
* code, unless it is known to be formatted. |
* formatted. Fall back to formatted mode. |
|
*/ |
*/ |
if (FORM_CAT != mpage->mlinks->dform || |
if (mlink->dform != FORM_CAT || mlink->fform != FORM_CAT) { |
FORM_CAT != mpage->mlinks->fform) { |
mparse_readfd(mp, fd, mlink->file); |
lvl = mparse_readfd(mp, fd[0], mpage->mlinks->file); |
close(fd); |
if (lvl < MANDOCLEVEL_FATAL) |
fd = -1; |
mparse_result(mp, &mdoc, &man, &sodest); |
meta = mparse_result(mp); |
} |
} |
|
|
if (NULL != sodest) { |
if (meta != NULL && meta->sodest != NULL) { |
mlink_dest = ohash_find(&mlinks, |
mlink_dest = ohash_find(&mlinks, |
ohash_qlookup(&mlinks, sodest)); |
ohash_qlookup(&mlinks, meta->sodest)); |
if (NULL != mlink_dest) { |
if (mlink_dest == NULL) { |
|
mandoc_asprintf(&cp, "%s.gz", meta->sodest); |
|
mlink_dest = ohash_find(&mlinks, |
|
ohash_qlookup(&mlinks, cp)); |
|
free(cp); |
|
} |
|
if (mlink_dest != NULL) { |
|
|
/* The .so target exists. */ |
/* The .so target exists. */ |
|
|
mpage_dest = mlink_dest->mpage; |
mpage_dest = mlink_dest->mpage; |
mlink = mpage->mlinks; |
|
while (1) { |
while (1) { |
mlink->mpage = mpage_dest; |
mlink->mpage = mpage_dest; |
|
|
Line 1181 mpages_merge(struct mchars *mc, struct mparse *mp) |
|
Line 1234 mpages_merge(struct mchars *mc, struct mparse *mp) |
|
* to the target. |
* to the target. |
*/ |
*/ |
|
|
if (mpage_dest->pageid) |
if (mpage_dest->dba != NULL) |
dbadd_mlink(mlink); |
dbadd_mlink(mlink); |
|
|
if (NULL == mlink->next) |
if (mlink->next == NULL) |
break; |
break; |
mlink = mlink->next; |
mlink = mlink->next; |
} |
} |
Line 1194 mpages_merge(struct mchars *mc, struct mparse *mp) |
|
Line 1247 mpages_merge(struct mchars *mc, struct mparse *mp) |
|
mlink->next = mlink_dest->next; |
mlink->next = mlink_dest->next; |
mlink_dest->next = mpage->mlinks; |
mlink_dest->next = mpage->mlinks; |
mpage->mlinks = NULL; |
mpage->mlinks = NULL; |
|
goto nextpage; |
} |
} |
goto nextpage; |
meta->macroset = MACROSET_NONE; |
} else if (NULL != mdoc) { |
} |
|
if (meta != NULL && meta->macroset == MACROSET_MDOC) { |
mpage->form = FORM_SRC; |
mpage->form = FORM_SRC; |
mpage->sec = |
mpage->sec = meta->msec; |
mandoc_strdup(mdoc_meta(mdoc)->msec); |
mpage->sec = mandoc_strdup( |
mpage->arch = mdoc_meta(mdoc)->arch; |
mpage->sec == NULL ? "" : mpage->sec); |
|
mpage->arch = meta->arch; |
mpage->arch = mandoc_strdup( |
mpage->arch = mandoc_strdup( |
NULL == mpage->arch ? "" : mpage->arch); |
mpage->arch == NULL ? "" : mpage->arch); |
mpage->title = |
mpage->title = mandoc_strdup(meta->title); |
mandoc_strdup(mdoc_meta(mdoc)->title); |
} else if (meta != NULL && meta->macroset == MACROSET_MAN) { |
} else if (NULL != man) { |
if (*meta->msec != '\0' || *meta->title != '\0') { |
mpage->form = FORM_SRC; |
mpage->form = FORM_SRC; |
mpage->sec = |
mpage->sec = mandoc_strdup(meta->msec); |
mandoc_strdup(man_meta(man)->msec); |
mpage->arch = mandoc_strdup(mlink->arch); |
mpage->arch = |
mpage->title = mandoc_strdup(meta->title); |
mandoc_strdup(mpage->mlinks->arch); |
} else |
mpage->title = |
meta = NULL; |
mandoc_strdup(man_meta(man)->title); |
|
} else { |
|
mpage->form = FORM_CAT; |
|
mpage->sec = |
|
mandoc_strdup(mpage->mlinks->dsec); |
|
mpage->arch = |
|
mandoc_strdup(mpage->mlinks->arch); |
|
mpage->title = |
|
mandoc_strdup(mpage->mlinks->name); |
|
} |
} |
putkey(mpage, mpage->sec, TYPE_sec); |
|
putkey(mpage, '\0' == *mpage->arch ? |
|
any : mpage->arch, TYPE_arch); |
|
|
|
for (mlink = mpage->mlinks; mlink; mlink = mlink->next) { |
assert(mpage->desc == NULL); |
if ('\0' != *mlink->dsec) |
if (meta == NULL || meta->sodest != NULL) { |
putkey(mpage, mlink->dsec, TYPE_sec); |
mpage->sec = mandoc_strdup(mlink->dsec); |
if ('\0' != *mlink->fsec) |
mpage->arch = mandoc_strdup(mlink->arch); |
putkey(mpage, mlink->fsec, TYPE_sec); |
mpage->title = mandoc_strdup(mlink->name); |
putkey(mpage, '\0' == *mlink->arch ? |
if (meta == NULL) { |
any : mlink->arch, TYPE_arch); |
mpage->form = FORM_CAT; |
putkey(mpage, mlink->name, NAME_FILE); |
parse_cat(mpage, fd); |
|
} else |
|
mpage->form = FORM_SRC; |
|
} else if (meta->macroset == MACROSET_MDOC) |
|
parse_mdoc(mpage, meta, meta->first); |
|
else |
|
parse_man(mpage, meta, meta->first); |
|
if (mpage->desc == NULL) { |
|
mpage->desc = mandoc_strdup(mlink->name); |
|
if (warnings) |
|
say(mlink->file, "No one-line description, " |
|
"using filename \"%s\"", mlink->name); |
} |
} |
|
|
assert(NULL == mpage->desc); |
for (mlink = mpage->mlinks; |
if (NULL != mdoc) { |
mlink != NULL; |
if (NULL != (cp = mdoc_meta(mdoc)->name)) |
mlink = mlink->next) { |
putkey(mpage, cp, NAME_HEAD); |
putkey(mpage, mlink->name, NAME_FILE); |
parse_mdoc(mpage, mdoc_node(mdoc)); |
if (warnings && !use_all) |
} else if (NULL != man) |
|
parse_man(mpage, man_node(man)); |
|
else |
|
parse_cat(mpage, fd[0]); |
|
if (NULL == mpage->desc) |
|
mpage->desc = mandoc_strdup(mpage->mlinks->name); |
|
|
|
if (warnings && !use_all) |
|
for (mlink = mpage->mlinks; mlink; |
|
mlink = mlink->next) |
|
mlink_check(mpage, mlink); |
mlink_check(mpage, mlink); |
|
} |
|
|
dbadd(mpage, mc); |
dbadd(dba, mpage); |
|
|
nextpage: |
nextpage: |
if (child_pid) { |
|
if (-1 == waitpid(child_pid, &status, 0)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(mpage->mlinks->file, "&wait gunzip"); |
|
} else if (WIFSIGNALED(status)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(mpage->mlinks->file, |
|
"gunzip died from signal %d", |
|
WTERMSIG(status)); |
|
} else if (WEXITSTATUS(status)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(mpage->mlinks->file, |
|
"gunzip failed with code %d", |
|
WEXITSTATUS(status)); |
|
} |
|
} |
|
ohash_delete(&strings); |
ohash_delete(&strings); |
ohash_delete(&names); |
ohash_delete(&names); |
mpage = ohash_next(&mpages, &pslot); |
|
} |
} |
|
|
if (0 == nodb) |
|
SQL_EXEC("END TRANSACTION"); |
|
} |
} |
|
|
static void |
static void |
names_check(void) |
|
{ |
|
sqlite3_stmt *stmt; |
|
const char *name, *sec, *arch, *key; |
|
int irc; |
|
|
|
sqlite3_prepare_v2(db, |
|
"SELECT name, sec, arch, key FROM (" |
|
"SELECT name AS key, pageid FROM names " |
|
"WHERE bits & ? AND NOT EXISTS (" |
|
"SELECT pageid FROM mlinks " |
|
"WHERE mlinks.pageid == names.pageid " |
|
"AND mlinks.name == names.name" |
|
")" |
|
") JOIN (" |
|
"SELECT sec, arch, name, pageid FROM mlinks " |
|
"GROUP BY pageid" |
|
") USING (pageid);", |
|
-1, &stmt, NULL); |
|
|
|
if (SQLITE_OK != sqlite3_bind_int64(stmt, 1, NAME_TITLE)) |
|
say("", "%s", sqlite3_errmsg(db)); |
|
|
|
while (SQLITE_ROW == (irc = sqlite3_step(stmt))) { |
|
name = sqlite3_column_text(stmt, 0); |
|
sec = sqlite3_column_text(stmt, 1); |
|
arch = sqlite3_column_text(stmt, 2); |
|
key = sqlite3_column_text(stmt, 3); |
|
say("", "%s(%s%s%s) lacks mlink \"%s\"", name, sec, |
|
'\0' == *arch ? "" : "/", |
|
'\0' == *arch ? "" : arch, key); |
|
} |
|
sqlite3_finalize(stmt); |
|
} |
|
|
|
static void |
|
parse_cat(struct mpage *mpage, int fd) |
parse_cat(struct mpage *mpage, int fd) |
{ |
{ |
FILE *stream; |
FILE *stream; |
char *line, *p, *title; |
struct mlink *mlink; |
size_t len, plen, titlesz; |
char *line, *p, *title, *sec; |
|
size_t linesz, plen, titlesz; |
|
ssize_t len; |
|
int offs; |
|
|
stream = (-1 == fd) ? |
mlink = mpage->mlinks; |
fopen(mpage->mlinks->file, "r") : |
stream = fd == -1 ? fopen(mlink->file, "r") : fdopen(fd, "r"); |
fdopen(fd, "r"); |
if (stream == NULL) { |
if (NULL == stream) { |
if (fd != -1) |
|
close(fd); |
if (warnings) |
if (warnings) |
say(mpage->mlinks->file, "&fopen"); |
say(mlink->file, "&fopen"); |
return; |
return; |
} |
} |
|
|
|
line = NULL; |
|
linesz = 0; |
|
|
|
/* Parse the section number from the header line. */ |
|
|
|
while (getline(&line, &linesz, stream) != -1) { |
|
if (*line == '\n') |
|
continue; |
|
if ((sec = strchr(line, '(')) == NULL) |
|
break; |
|
if ((p = strchr(++sec, ')')) == NULL) |
|
break; |
|
free(mpage->sec); |
|
mpage->sec = mandoc_strndup(sec, p - sec); |
|
if (warnings && *mlink->dsec != '\0' && |
|
strcasecmp(mpage->sec, mlink->dsec)) |
|
say(mlink->file, |
|
"Section \"%s\" manual in %s directory", |
|
mpage->sec, mlink->dsec); |
|
break; |
|
} |
|
|
/* Skip to first blank line. */ |
/* Skip to first blank line. */ |
|
|
while (NULL != (line = fgetln(stream, &len))) |
while (line == NULL || *line != '\n') |
if ('\n' == *line) |
if (getline(&line, &linesz, stream) == -1) |
break; |
break; |
|
|
/* |
/* |
Line 1344 parse_cat(struct mpage *mpage, int fd) |
|
Line 1360 parse_cat(struct mpage *mpage, int fd) |
|
* is the first section header. Skip to it. |
* is the first section header. Skip to it. |
*/ |
*/ |
|
|
while (NULL != (line = fgetln(stream, &len))) |
while (getline(&line, &linesz, stream) != -1) |
if ('\n' != *line && ' ' != *line) |
if (*line != '\n' && *line != ' ') |
break; |
break; |
|
|
/* |
/* |
Line 1358 parse_cat(struct mpage *mpage, int fd) |
|
Line 1374 parse_cat(struct mpage *mpage, int fd) |
|
titlesz = 0; |
titlesz = 0; |
title = NULL; |
title = NULL; |
|
|
while (NULL != (line = fgetln(stream, &len))) { |
while ((len = getline(&line, &linesz, stream)) != -1) { |
if (' ' != *line || '\n' != line[len - 1]) |
if (*line != ' ') |
break; |
break; |
while (len > 0 && isspace((unsigned char)*line)) { |
offs = 0; |
line++; |
while (isspace((unsigned char)line[offs])) |
len--; |
offs++; |
} |
if (line[offs] == '\0') |
if (1 == len) |
|
continue; |
continue; |
title = mandoc_realloc(title, titlesz + len); |
title = mandoc_realloc(title, titlesz + len - offs); |
memcpy(title + titlesz, line, len); |
memcpy(title + titlesz, line + offs, len - offs); |
titlesz += len; |
titlesz += len - offs; |
title[titlesz - 1] = ' '; |
title[titlesz - 1] = ' '; |
} |
} |
|
free(line); |
|
|
/* |
/* |
* If no page content can be found, or the input line |
* If no page content can be found, or the input line |
Line 1382 parse_cat(struct mpage *mpage, int fd) |
|
Line 1398 parse_cat(struct mpage *mpage, int fd) |
|
|
|
if (NULL == title || '\0' == *title) { |
if (NULL == title || '\0' == *title) { |
if (warnings) |
if (warnings) |
say(mpage->mlinks->file, |
say(mlink->file, "Cannot find NAME section"); |
"Cannot find NAME section"); |
|
fclose(stream); |
fclose(stream); |
free(title); |
free(title); |
return; |
return; |
} |
} |
|
|
title = mandoc_realloc(title, titlesz + 1); |
title[titlesz - 1] = '\0'; |
title[titlesz] = '\0'; |
|
|
|
/* |
/* |
* Skip to the first dash. |
* Skip to the first dash. |
Line 1403 parse_cat(struct mpage *mpage, int fd) |
|
Line 1417 parse_cat(struct mpage *mpage, int fd) |
|
/* Skip to next word. */ ; |
/* Skip to next word. */ ; |
} else { |
} else { |
if (warnings) |
if (warnings) |
say(mpage->mlinks->file, |
say(mlink->file, "No dash in title line, " |
"No dash in title line"); |
"reusing \"%s\" as one-line description", title); |
p = title; |
p = title; |
} |
} |
|
|
Line 1422 parse_cat(struct mpage *mpage, int fd) |
|
Line 1436 parse_cat(struct mpage *mpage, int fd) |
|
plen -= 2; |
plen -= 2; |
} |
} |
|
|
mpage->desc = mandoc_strdup(p); |
/* |
|
* Cut off excessive one-line descriptions. |
|
* Bad pages are not worth better heuristics. |
|
*/ |
|
|
|
mpage->desc = mandoc_strndup(p, 150); |
fclose(stream); |
fclose(stream); |
free(title); |
free(title); |
} |
} |
Line 1433 parse_cat(struct mpage *mpage, int fd) |
|
Line 1452 parse_cat(struct mpage *mpage, int fd) |
|
static void |
static void |
putkey(const struct mpage *mpage, char *value, uint64_t type) |
putkey(const struct mpage *mpage, char *value, uint64_t type) |
{ |
{ |
char *cp; |
|
|
|
assert(NULL != value); |
|
if (TYPE_arch == type) |
|
for (cp = value; *cp; cp++) |
|
if (isupper((unsigned char)*cp)) |
|
*cp = _tolower((unsigned char)*cp); |
|
putkeys(mpage, value, strlen(value), type); |
putkeys(mpage, value, strlen(value), type); |
} |
} |
|
|
Line 1448 putkey(const struct mpage *mpage, char *value, uint64_ |
|
Line 1460 putkey(const struct mpage *mpage, char *value, uint64_ |
|
*/ |
*/ |
static void |
static void |
putmdockey(const struct mpage *mpage, |
putmdockey(const struct mpage *mpage, |
const struct mdoc_node *n, uint64_t m) |
const struct roff_node *n, uint64_t m, int taboo) |
{ |
{ |
|
|
for ( ; NULL != n; n = n->next) { |
for ( ; NULL != n; n = n->next) { |
|
if (n->flags & taboo) |
|
continue; |
if (NULL != n->child) |
if (NULL != n->child) |
putmdockey(mpage, n->child, m); |
putmdockey(mpage, n->child, m, taboo); |
if (MDOC_TEXT == n->type) |
if (n->type == ROFFT_TEXT) |
putkey(mpage, n->string, m); |
putkey(mpage, n->string, m); |
} |
} |
} |
} |
|
|
static void |
static void |
parse_man(struct mpage *mpage, const struct man_node *n) |
parse_man(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
const struct man_node *head, *body; |
const struct roff_node *head, *body; |
char *start, *title; |
char *start, *title; |
char byte; |
char byte; |
size_t sz; |
size_t sz; |
|
|
if (NULL == n) |
if (n == NULL) |
return; |
return; |
|
|
/* |
/* |
Line 1477 parse_man(struct mpage *mpage, const struct man_node * |
|
Line 1492 parse_man(struct mpage *mpage, const struct man_node * |
|
* the correct section or not. |
* the correct section or not. |
*/ |
*/ |
|
|
if (MAN_BODY == n->type && MAN_SH == n->tok) { |
if (n->type == ROFFT_BODY && n->tok == MAN_SH) { |
body = n; |
body = n; |
assert(body->parent); |
if ((head = body->parent->head) != NULL && |
if (NULL != (head = body->parent->head) && |
(head = head->child) != NULL && |
1 == head->nchild && |
head->next == NULL && |
NULL != (head = (head->child)) && |
head->type == ROFFT_TEXT && |
MAN_TEXT == head->type && |
strcmp(head->string, "NAME") == 0 && |
0 == strcmp(head->string, "NAME") && |
body->child != NULL) { |
NULL != body->child) { |
|
|
|
/* |
/* |
* Suck the entire NAME section into memory. |
* Suck the entire NAME section into memory. |
Line 1495 parse_man(struct mpage *mpage, const struct man_node * |
|
Line 1509 parse_man(struct mpage *mpage, const struct man_node * |
|
*/ |
*/ |
|
|
title = NULL; |
title = NULL; |
man_deroff(&title, body); |
deroff(&title, body); |
if (NULL == title) |
if (NULL == title) |
return; |
return; |
|
|
Line 1526 parse_man(struct mpage *mpage, const struct man_node * |
|
Line 1540 parse_man(struct mpage *mpage, const struct man_node * |
|
break; |
break; |
|
|
putkey(mpage, start, NAME_TITLE); |
putkey(mpage, start, NAME_TITLE); |
|
if ( ! (mpage->name_head_done || |
|
strcasecmp(start, meta->title))) { |
|
putkey(mpage, start, NAME_HEAD); |
|
mpage->name_head_done = 1; |
|
} |
|
|
if (' ' == byte) { |
if (' ' == byte) { |
start += sz + 1; |
start += sz + 1; |
Line 1540 parse_man(struct mpage *mpage, const struct man_node * |
|
Line 1559 parse_man(struct mpage *mpage, const struct man_node * |
|
|
|
if (start == title) { |
if (start == title) { |
putkey(mpage, start, NAME_TITLE); |
putkey(mpage, start, NAME_TITLE); |
|
if ( ! (mpage->name_head_done || |
|
strcasecmp(start, meta->title))) { |
|
putkey(mpage, start, NAME_HEAD); |
|
mpage->name_head_done = 1; |
|
} |
free(title); |
free(title); |
return; |
return; |
} |
} |
Line 1561 parse_man(struct mpage *mpage, const struct man_node * |
|
Line 1585 parse_man(struct mpage *mpage, const struct man_node * |
|
while (' ' == *start) |
while (' ' == *start) |
start++; |
start++; |
|
|
mpage->desc = mandoc_strdup(start); |
/* |
|
* Cut off excessive one-line descriptions. |
|
* Bad pages are not worth better heuristics. |
|
*/ |
|
|
|
mpage->desc = mandoc_strndup(start, 150); |
free(title); |
free(title); |
return; |
return; |
} |
} |
Line 1570 parse_man(struct mpage *mpage, const struct man_node * |
|
Line 1599 parse_man(struct mpage *mpage, const struct man_node * |
|
for (n = n->child; n; n = n->next) { |
for (n = n->child; n; n = n->next) { |
if (NULL != mpage->desc) |
if (NULL != mpage->desc) |
break; |
break; |
parse_man(mpage, n); |
parse_man(mpage, meta, n); |
} |
} |
} |
} |
|
|
static void |
static void |
parse_mdoc(struct mpage *mpage, const struct mdoc_node *n) |
parse_mdoc(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
const struct mdoc_handler *handler; |
|
|
assert(NULL != n); |
for (n = n->child; n != NULL; n = n->next) { |
for (n = n->child; NULL != n; n = n->next) { |
if (n->tok == TOKEN_NONE || n->tok < ROFF_MAX) |
|
continue; |
|
assert(n->tok >= MDOC_Dd && n->tok < MDOC_MAX); |
|
handler = mdoc_handlers + (n->tok - MDOC_Dd); |
|
if (n->flags & handler->taboo) |
|
continue; |
|
|
switch (n->type) { |
switch (n->type) { |
case MDOC_ELEM: |
case ROFFT_ELEM: |
/* FALLTHROUGH */ |
case ROFFT_BLOCK: |
case MDOC_BLOCK: |
case ROFFT_HEAD: |
/* FALLTHROUGH */ |
case ROFFT_BODY: |
case MDOC_HEAD: |
case ROFFT_TAIL: |
/* FALLTHROUGH */ |
if (handler->fp != NULL && |
case MDOC_BODY: |
(*handler->fp)(mpage, meta, n) == 0) |
/* FALLTHROUGH */ |
break; |
case MDOC_TAIL: |
if (handler->mask) |
if (NULL != mdocs[n->tok].fp) |
|
if (0 == (*mdocs[n->tok].fp)(mpage, n)) |
|
break; |
|
if (mdocs[n->tok].mask) |
|
putmdockey(mpage, n->child, |
putmdockey(mpage, n->child, |
mdocs[n->tok].mask); |
handler->mask, handler->taboo); |
break; |
break; |
default: |
default: |
assert(MDOC_ROOT != n->type); |
|
continue; |
continue; |
} |
} |
if (NULL != n->child) |
if (NULL != n->child) |
parse_mdoc(mpage, n); |
parse_mdoc(mpage, meta, n); |
} |
} |
} |
} |
|
|
static int |
static int |
parse_mdoc_Fd(struct mpage *mpage, const struct mdoc_node *n) |
parse_mdoc_Fa(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
const char *start, *end; |
uint64_t mask; |
|
|
|
mask = TYPE_Fa; |
|
if (n->sec == SEC_SYNOPSIS) |
|
mask |= TYPE_Vt; |
|
|
|
putmdockey(mpage, n->child, mask, 0); |
|
return 0; |
|
} |
|
|
|
static int |
|
parse_mdoc_Fd(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
|
{ |
|
char *start, *end; |
size_t sz; |
size_t sz; |
|
|
if (SEC_SYNOPSIS != n->sec || |
if (SEC_SYNOPSIS != n->sec || |
NULL == (n = n->child) || |
NULL == (n = n->child) || |
MDOC_TEXT != n->type) |
n->type != ROFFT_TEXT) |
return(0); |
return 0; |
|
|
/* |
/* |
* Only consider those `Fd' macro fields that begin with an |
* Only consider those `Fd' macro fields that begin with an |
Line 1623 parse_mdoc_Fd(struct mpage *mpage, const struct mdoc_n |
|
Line 1670 parse_mdoc_Fd(struct mpage *mpage, const struct mdoc_n |
|
*/ |
*/ |
|
|
if (strcmp("#include", n->string)) |
if (strcmp("#include", n->string)) |
return(0); |
return 0; |
|
|
if (NULL == (n = n->next) || MDOC_TEXT != n->type) |
if ((n = n->next) == NULL || n->type != ROFFT_TEXT) |
return(0); |
return 0; |
|
|
/* |
/* |
* Strip away the enclosing angle brackets and make sure we're |
* Strip away the enclosing angle brackets and make sure we're |
Line 1638 parse_mdoc_Fd(struct mpage *mpage, const struct mdoc_n |
|
Line 1685 parse_mdoc_Fd(struct mpage *mpage, const struct mdoc_n |
|
start++; |
start++; |
|
|
if (0 == (sz = strlen(start))) |
if (0 == (sz = strlen(start))) |
return(0); |
return 0; |
|
|
end = &start[(int)sz - 1]; |
end = &start[(int)sz - 1]; |
if ('>' == *end || '"' == *end) |
if ('>' == *end || '"' == *end) |
Line 1646 parse_mdoc_Fd(struct mpage *mpage, const struct mdoc_n |
|
Line 1693 parse_mdoc_Fd(struct mpage *mpage, const struct mdoc_n |
|
|
|
if (end > start) |
if (end > start) |
putkeys(mpage, start, end - start + 1, TYPE_In); |
putkeys(mpage, start, end - start + 1, TYPE_In); |
return(0); |
return 0; |
} |
} |
|
|
static int |
static void |
parse_mdoc_Fn(struct mpage *mpage, const struct mdoc_node *n) |
parse_mdoc_fname(struct mpage *mpage, const struct roff_node *n) |
{ |
{ |
char *cp; |
char *cp; |
|
size_t sz; |
|
|
if (NULL == (n = n->child) || MDOC_TEXT != n->type) |
if (n->type != ROFFT_TEXT) |
return(0); |
return; |
|
|
/* |
/* Skip function pointer punctuation. */ |
* Parse: .Fn "struct type *name" "char *arg". |
|
* First strip away pointer symbol. |
|
* Then store the function name, then type. |
|
* Finally, store the arguments. |
|
*/ |
|
|
|
if (NULL == (cp = strrchr(n->string, ' '))) |
cp = n->string; |
cp = n->string; |
while (*cp == '(' || *cp == '*') |
|
|
while ('*' == *cp) |
|
cp++; |
cp++; |
|
sz = strcspn(cp, "()"); |
|
|
putkey(mpage, cp, TYPE_Fn); |
putkeys(mpage, cp, sz, TYPE_Fn); |
|
if (n->sec == SEC_SYNOPSIS) |
|
putkeys(mpage, cp, sz, NAME_SYN); |
|
} |
|
|
if (n->string < cp) |
static int |
putkeys(mpage, n->string, cp - n->string, TYPE_Ft); |
parse_mdoc_Fn(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
|
{ |
|
uint64_t mask; |
|
|
for (n = n->next; NULL != n; n = n->next) |
if (n->child == NULL) |
if (MDOC_TEXT == n->type) |
return 0; |
putkey(mpage, n->string, TYPE_Fa); |
|
|
|
return(0); |
parse_mdoc_fname(mpage, n->child); |
|
|
|
n = n->child->next; |
|
if (n != NULL && n->type == ROFFT_TEXT) { |
|
mask = TYPE_Fa; |
|
if (n->sec == SEC_SYNOPSIS) |
|
mask |= TYPE_Vt; |
|
putmdockey(mpage, n, mask, 0); |
|
} |
|
|
|
return 0; |
} |
} |
|
|
static int |
static int |
parse_mdoc_Xr(struct mpage *mpage, const struct mdoc_node *n) |
parse_mdoc_Fo(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
|
if (n->type != ROFFT_HEAD) |
|
return 1; |
|
|
|
if (n->child != NULL) |
|
parse_mdoc_fname(mpage, n->child); |
|
|
|
return 0; |
|
} |
|
|
|
static int |
|
parse_mdoc_Va(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
|
{ |
|
char *cp; |
|
|
|
if (n->type != ROFFT_ELEM && n->type != ROFFT_BODY) |
|
return 0; |
|
|
|
if (n->child != NULL && |
|
n->child->next == NULL && |
|
n->child->type == ROFFT_TEXT) |
|
return 1; |
|
|
|
cp = NULL; |
|
deroff(&cp, n); |
|
if (cp != NULL) { |
|
putkey(mpage, cp, TYPE_Vt | (n->tok == MDOC_Va || |
|
n->type == ROFFT_BODY ? TYPE_Va : 0)); |
|
free(cp); |
|
} |
|
|
|
return 0; |
|
} |
|
|
|
static int |
|
parse_mdoc_Xr(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
|
{ |
char *cp; |
char *cp; |
|
|
if (NULL == (n = n->child)) |
if (NULL == (n = n->child)) |
return(0); |
return 0; |
|
|
if (NULL == n->next) { |
if (NULL == n->next) { |
putkey(mpage, n->string, TYPE_Xr); |
putkey(mpage, n->string, TYPE_Xr); |
return(0); |
return 0; |
} |
} |
|
|
mandoc_asprintf(&cp, "%s(%s)", n->string, n->next->string); |
mandoc_asprintf(&cp, "%s(%s)", n->string, n->next->string); |
putkey(mpage, cp, TYPE_Xr); |
putkey(mpage, cp, TYPE_Xr); |
free(cp); |
free(cp); |
return(0); |
return 0; |
} |
} |
|
|
static int |
static int |
parse_mdoc_Nd(struct mpage *mpage, const struct mdoc_node *n) |
parse_mdoc_Nd(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
if (MDOC_BODY == n->type) |
if (n->type == ROFFT_BODY) |
mdoc_deroff(&mpage->desc, n); |
deroff(&mpage->desc, n); |
return(0); |
return 0; |
} |
} |
|
|
static int |
static int |
parse_mdoc_Nm(struct mpage *mpage, const struct mdoc_node *n) |
parse_mdoc_Nm(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
if (SEC_NAME == n->sec) |
if (SEC_NAME == n->sec) |
putmdockey(mpage, n->child, NAME_TITLE); |
putmdockey(mpage, n->child, NAME_TITLE, 0); |
else if (SEC_SYNOPSIS == n->sec && MDOC_HEAD == n->type) |
else if (n->sec == SEC_SYNOPSIS && n->type == ROFFT_HEAD) { |
putmdockey(mpage, n->child, NAME_SYN); |
if (n->child == NULL) |
return(0); |
putkey(mpage, meta->name, NAME_SYN); |
|
else |
|
putmdockey(mpage, n->child, NAME_SYN, 0); |
|
} |
|
if ( ! (mpage->name_head_done || |
|
n->child == NULL || n->child->string == NULL || |
|
strcasecmp(n->child->string, meta->title))) { |
|
putkey(mpage, n->child->string, NAME_HEAD); |
|
mpage->name_head_done = 1; |
|
} |
|
return 0; |
} |
} |
|
|
static int |
static int |
parse_mdoc_Sh(struct mpage *mpage, const struct mdoc_node *n) |
parse_mdoc_Sh(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
return(SEC_CUSTOM == n->sec && MDOC_HEAD == n->type); |
return n->sec == SEC_CUSTOM && n->type == ROFFT_HEAD; |
} |
} |
|
|
static int |
static int |
parse_mdoc_head(struct mpage *mpage, const struct mdoc_node *n) |
parse_mdoc_head(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
return(MDOC_HEAD == n->type); |
return n->type == ROFFT_HEAD; |
} |
} |
|
|
static int |
|
parse_mdoc_body(struct mpage *mpage, const struct mdoc_node *n) |
|
{ |
|
|
|
return(MDOC_BODY == n->type); |
|
} |
|
|
|
/* |
/* |
* Add a string to the hash table for the current manual. |
* Add a string to the hash table for the current manual. |
* Each string has a bitmask telling which macros it belongs to. |
* Each string has a bitmask telling which macros it belongs to. |
* When we finish the manual, we'll dump the table. |
* When we finish the manual, we'll dump the table. |
*/ |
*/ |
static void |
static void |
putkeys(const struct mpage *mpage, |
putkeys(const struct mpage *mpage, char *cp, size_t sz, uint64_t v) |
const char *cp, size_t sz, uint64_t v) |
|
{ |
{ |
struct ohash *htab; |
struct ohash *htab; |
struct str *s; |
struct str *s; |
const char *end; |
const char *end; |
unsigned int slot; |
unsigned int slot; |
int i; |
int i, mustfree; |
|
|
if (0 == sz) |
if (0 == sz) |
return; |
return; |
|
|
|
mustfree = render_string(&cp, &sz); |
|
|
if (TYPE_Nm & v) { |
if (TYPE_Nm & v) { |
htab = &names; |
htab = &names; |
v &= name_mask; |
v &= name_mask; |
name_mask &= ~NAME_FIRST; |
if (v & NAME_FIRST) |
|
name_mask &= ~NAME_FIRST; |
if (debug > 1) |
if (debug > 1) |
say(mpage->mlinks->file, |
say(mpage->mlinks->file, |
"Adding name %*s", sz, cp); |
"Adding name %*s, bits=0x%llx", (int)sz, cp, |
|
(unsigned long long)v); |
} else { |
} else { |
htab = &strings; |
htab = &strings; |
if (debug > 1) |
if (debug > 1) |
for (i = 0; i < mansearch_keymax; i++) |
for (i = 0; i < KEY_MAX; i++) |
if (1 << i & v) |
if ((uint64_t)1 << i & v) |
say(mpage->mlinks->file, |
say(mpage->mlinks->file, |
"Adding key %s=%*s", |
"Adding key %s=%*s", |
mansearch_keynames[i], sz, cp); |
mansearch_keynames[i], (int)sz, cp); |
} |
} |
|
|
end = cp + sz; |
end = cp + sz; |
Line 1791 putkeys(const struct mpage *mpage, |
|
Line 1898 putkeys(const struct mpage *mpage, |
|
} |
} |
s->mpage = mpage; |
s->mpage = mpage; |
s->mask = v; |
s->mask = v; |
|
|
|
if (mustfree) |
|
free(cp); |
} |
} |
|
|
/* |
/* |
Line 1839 utf8(unsigned int cp, char out[7]) |
|
Line 1949 utf8(unsigned int cp, char out[7]) |
|
out[4] = (cp >> 6 & 63) | 128; |
out[4] = (cp >> 6 & 63) | 128; |
out[5] = (cp & 63) | 128; |
out[5] = (cp & 63) | 128; |
} else |
} else |
return(0); |
return 0; |
|
|
out[rc] = '\0'; |
out[rc] = '\0'; |
return(rc); |
return rc; |
} |
} |
|
|
/* |
/* |
* Store the rendered version of a key, or alias the pointer |
* If the string contains escape sequences, |
* if the key contains no escape sequences. |
* replace it with an allocated rendering and return 1, |
|
* such that the caller can free it after use. |
|
* Otherwise, do nothing and return 0. |
*/ |
*/ |
static void |
static int |
render_key(struct mchars *mc, struct str *key) |
render_string(char **public, size_t *psz) |
{ |
{ |
size_t sz, bsz, pos; |
const char *src, *scp, *addcp, *seq; |
|
char *dst; |
|
size_t ssz, dsz, addsz; |
char utfbuf[7], res[6]; |
char utfbuf[7], res[6]; |
char *buf; |
int seqlen, unicode; |
const char *seq, *cpp, *val; |
|
int len, u; |
|
enum mandoc_esc esc; |
|
|
|
assert(NULL == key->rendered); |
|
|
|
res[0] = '\\'; |
res[0] = '\\'; |
res[1] = '\t'; |
res[1] = '\t'; |
res[2] = ASCII_NBRSP; |
res[2] = ASCII_NBRSP; |
Line 1868 render_key(struct mchars *mc, struct str *key) |
|
Line 1977 render_key(struct mchars *mc, struct str *key) |
|
res[4] = ASCII_BREAK; |
res[4] = ASCII_BREAK; |
res[5] = '\0'; |
res[5] = '\0'; |
|
|
val = key->key; |
src = scp = *public; |
bsz = strlen(val); |
ssz = *psz; |
|
dst = NULL; |
|
dsz = 0; |
|
|
/* |
while (scp < src + *psz) { |
* Pre-check: if we have no stop-characters, then set the |
|
* pointer as ourselvse and get out of here. |
|
*/ |
|
if (strcspn(val, res) == bsz) { |
|
key->rendered = key->key; |
|
return; |
|
} |
|
|
|
/* Pre-allocate by the length of the input */ |
/* Leave normal characters unchanged. */ |
|
|
buf = mandoc_malloc(++bsz); |
if (strchr(res, *scp) == NULL) { |
pos = 0; |
if (dst != NULL) |
|
dst[dsz++] = *scp; |
|
scp++; |
|
continue; |
|
} |
|
|
while ('\0' != *val) { |
|
/* |
/* |
* Halt on the first escape sequence. |
* Found something that requires replacing, |
* This also halts on the end of string, in which case |
* make sure we have a destination buffer. |
* we just copy, fallthrough, and exit the loop. |
|
*/ |
*/ |
if ((sz = strcspn(val, res)) > 0) { |
|
memcpy(&buf[pos], val, sz); |
if (dst == NULL) { |
pos += sz; |
dst = mandoc_malloc(ssz + 1); |
val += sz; |
dsz = scp - src; |
|
memcpy(dst, src, dsz); |
} |
} |
|
|
switch (*val) { |
/* Handle single-char special characters. */ |
case ASCII_HYPH: |
|
buf[pos++] = '-'; |
switch (*scp) { |
val++; |
case '\\': |
continue; |
break; |
case '\t': |
case '\t': |
/* FALLTHROUGH */ |
|
case ASCII_NBRSP: |
case ASCII_NBRSP: |
buf[pos++] = ' '; |
dst[dsz++] = ' '; |
val++; |
scp++; |
|
continue; |
|
case ASCII_HYPH: |
|
dst[dsz++] = '-'; |
/* FALLTHROUGH */ |
/* FALLTHROUGH */ |
case ASCII_BREAK: |
case ASCII_BREAK: |
|
scp++; |
continue; |
continue; |
default: |
default: |
break; |
abort(); |
} |
} |
if ('\\' != *val) |
|
break; |
|
|
|
/* Read past the slash. */ |
|
|
|
val++; |
|
|
|
/* |
/* |
* Parse the escape sequence and see if it's a |
* Found an escape sequence. |
* predefined character or special character. |
* Read past the slash, then parse it. |
|
* Ignore everything except characters. |
*/ |
*/ |
|
|
esc = mandoc_escape((const char **)&val, |
scp++; |
&seq, &len); |
if (mandoc_escape(&scp, &seq, &seqlen) != ESCAPE_SPECIAL) |
if (ESCAPE_ERROR == esc) |
|
break; |
|
if (ESCAPE_SPECIAL != esc) |
|
continue; |
continue; |
|
|
/* |
/* |
Line 1938 render_key(struct mchars *mc, struct str *key) |
|
Line 2040 render_key(struct mchars *mc, struct str *key) |
|
*/ |
*/ |
|
|
if (write_utf8) { |
if (write_utf8) { |
if (0 == (u = mchars_spec2cp(mc, seq, len))) |
unicode = mchars_spec2cp(seq, seqlen); |
|
if (unicode <= 0) |
continue; |
continue; |
cpp = utfbuf; |
addsz = utf8(unicode, utfbuf); |
if (0 == (sz = utf8(u, utfbuf))) |
if (addsz == 0) |
continue; |
continue; |
sz = strlen(cpp); |
addcp = utfbuf; |
} else { |
} else { |
cpp = mchars_spec2str(mc, seq, len, &sz); |
addcp = mchars_spec2str(seq, seqlen, &addsz); |
if (NULL == cpp) |
if (addcp == NULL) |
continue; |
continue; |
if (ASCII_NBRSP == *cpp) { |
if (*addcp == ASCII_NBRSP) { |
cpp = " "; |
addcp = " "; |
sz = 1; |
addsz = 1; |
} |
} |
} |
} |
|
|
/* Copy the rendered glyph into the stream. */ |
/* Copy the rendered glyph into the stream. */ |
|
|
bsz += sz; |
ssz += addsz; |
buf = mandoc_realloc(buf, bsz); |
dst = mandoc_realloc(dst, ssz + 1); |
memcpy(&buf[pos], cpp, sz); |
memcpy(dst + dsz, addcp, addsz); |
pos += sz; |
dsz += addsz; |
} |
} |
|
if (dst != NULL) { |
|
*public = dst; |
|
*psz = dsz; |
|
} |
|
|
buf[pos] = '\0'; |
/* Trim trailing whitespace and NUL-terminate. */ |
key->rendered = buf; |
|
|
while (*psz > 0 && (*public)[*psz - 1] == ' ') |
|
--*psz; |
|
if (dst != NULL) { |
|
(*public)[*psz] = '\0'; |
|
return 1; |
|
} else |
|
return 0; |
} |
} |
|
|
static void |
static void |
dbadd_mlink(const struct mlink *mlink) |
dbadd_mlink(const struct mlink *mlink) |
{ |
{ |
size_t i; |
dba_page_alias(mlink->mpage->dba, mlink->name, NAME_FILE); |
|
dba_page_add(mlink->mpage->dba, DBP_SECT, mlink->dsec); |
i = 1; |
dba_page_add(mlink->mpage->dba, DBP_SECT, mlink->fsec); |
SQL_BIND_TEXT(stmts[STMT_INSERT_LINK], i, mlink->dsec); |
dba_page_add(mlink->mpage->dba, DBP_ARCH, mlink->arch); |
SQL_BIND_TEXT(stmts[STMT_INSERT_LINK], i, mlink->arch); |
dba_page_add(mlink->mpage->dba, DBP_FILE, mlink->file); |
SQL_BIND_TEXT(stmts[STMT_INSERT_LINK], i, mlink->name); |
|
SQL_BIND_INT64(stmts[STMT_INSERT_LINK], i, mlink->mpage->pageid); |
|
SQL_STEP(stmts[STMT_INSERT_LINK]); |
|
sqlite3_reset(stmts[STMT_INSERT_LINK]); |
|
} |
} |
|
|
/* |
/* |
* Flush the current page's terms (and their bits) into the database. |
* Flush the current page's terms (and their bits) into the database. |
* Wrap the entire set of additions in a transaction to make sqlite be a |
|
* little faster. |
|
* Also, handle escape sequences at the last possible moment. |
* Also, handle escape sequences at the last possible moment. |
*/ |
*/ |
static void |
static void |
dbadd(struct mpage *mpage, struct mchars *mc) |
dbadd(struct dba *dba, struct mpage *mpage) |
{ |
{ |
struct mlink *mlink; |
struct mlink *mlink; |
struct str *key; |
struct str *key; |
|
char *cp; |
|
uint64_t mask; |
size_t i; |
size_t i; |
unsigned int slot; |
unsigned int slot; |
|
int mustfree; |
|
|
mlink = mpage->mlinks; |
mlink = mpage->mlinks; |
|
|
if (nodb) { |
if (nodb) { |
for (key = ohash_first(&names, &slot); NULL != key; |
for (key = ohash_first(&names, &slot); NULL != key; |
key = ohash_next(&names, &slot)) { |
key = ohash_next(&names, &slot)) |
if (key->rendered != key->key) |
|
free(key->rendered); |
|
free(key); |
free(key); |
} |
|
for (key = ohash_first(&strings, &slot); NULL != key; |
for (key = ohash_first(&strings, &slot); NULL != key; |
key = ohash_next(&strings, &slot)) { |
key = ohash_next(&strings, &slot)) |
if (key->rendered != key->key) |
|
free(key->rendered); |
|
free(key); |
free(key); |
} |
|
if (0 == debug) |
if (0 == debug) |
return; |
return; |
while (NULL != mlink) { |
while (NULL != mlink) { |
Line 2037 dbadd(struct mpage *mpage, struct mchars *mc) |
|
Line 2142 dbadd(struct mpage *mpage, struct mchars *mc) |
|
if (debug) |
if (debug) |
say(mlink->file, "Adding to database"); |
say(mlink->file, "Adding to database"); |
|
|
i = strlen(mpage->desc) + 1; |
cp = mpage->desc; |
key = mandoc_calloc(1, sizeof(struct str) + i); |
i = strlen(cp); |
memcpy(key->key, mpage->desc, i); |
mustfree = render_string(&cp, &i); |
render_key(mc, key); |
mpage->dba = dba_page_new(dba->pages, |
|
*mpage->arch == '\0' ? mlink->arch : mpage->arch, |
|
cp, mlink->file, mpage->form); |
|
if (mustfree) |
|
free(cp); |
|
dba_page_add(mpage->dba, DBP_SECT, mpage->sec); |
|
|
i = 1; |
while (mlink != NULL) { |
SQL_BIND_TEXT(stmts[STMT_INSERT_PAGE], i, key->rendered); |
|
SQL_BIND_INT(stmts[STMT_INSERT_PAGE], i, FORM_SRC == mpage->form); |
|
SQL_STEP(stmts[STMT_INSERT_PAGE]); |
|
mpage->pageid = sqlite3_last_insert_rowid(db); |
|
sqlite3_reset(stmts[STMT_INSERT_PAGE]); |
|
|
|
if (key->rendered != key->key) |
|
free(key->rendered); |
|
free(key); |
|
|
|
while (NULL != mlink) { |
|
dbadd_mlink(mlink); |
dbadd_mlink(mlink); |
mlink = mlink->next; |
mlink = mlink->next; |
} |
} |
mlink = mpage->mlinks; |
|
|
|
for (key = ohash_first(&names, &slot); NULL != key; |
for (key = ohash_first(&names, &slot); NULL != key; |
key = ohash_next(&names, &slot)) { |
key = ohash_next(&names, &slot)) { |
assert(key->mpage == mpage); |
assert(key->mpage == mpage); |
if (NULL == key->rendered) |
dba_page_alias(mpage->dba, key->key, key->mask); |
render_key(mc, key); |
|
i = 1; |
|
SQL_BIND_INT64(stmts[STMT_INSERT_NAME], i, key->mask); |
|
SQL_BIND_TEXT(stmts[STMT_INSERT_NAME], i, key->rendered); |
|
SQL_BIND_INT64(stmts[STMT_INSERT_NAME], i, mpage->pageid); |
|
SQL_STEP(stmts[STMT_INSERT_NAME]); |
|
sqlite3_reset(stmts[STMT_INSERT_NAME]); |
|
if (key->rendered != key->key) |
|
free(key->rendered); |
|
free(key); |
free(key); |
} |
} |
for (key = ohash_first(&strings, &slot); NULL != key; |
for (key = ohash_first(&strings, &slot); NULL != key; |
key = ohash_next(&strings, &slot)) { |
key = ohash_next(&strings, &slot)) { |
assert(key->mpage == mpage); |
assert(key->mpage == mpage); |
if (NULL == key->rendered) |
i = 0; |
render_key(mc, key); |
for (mask = TYPE_Xr; mask <= TYPE_Lb; mask *= 2) { |
i = 1; |
if (key->mask & mask) |
SQL_BIND_INT64(stmts[STMT_INSERT_KEY], i, key->mask); |
dba_macro_add(dba->macros, i, |
SQL_BIND_TEXT(stmts[STMT_INSERT_KEY], i, key->rendered); |
key->key, mpage->dba); |
SQL_BIND_INT64(stmts[STMT_INSERT_KEY], i, mpage->pageid); |
i++; |
SQL_STEP(stmts[STMT_INSERT_KEY]); |
} |
sqlite3_reset(stmts[STMT_INSERT_KEY]); |
|
if (key->rendered != key->key) |
|
free(key->rendered); |
|
free(key); |
free(key); |
} |
} |
} |
} |
|
|
static void |
static void |
dbprune(void) |
dbprune(struct dba *dba) |
{ |
{ |
struct mpage *mpage; |
struct dba_array *page, *files; |
struct mlink *mlink; |
char *file; |
size_t i; |
|
unsigned int slot; |
|
|
|
if (0 == nodb) |
dba_array_FOREACH(dba->pages, page) { |
SQL_EXEC("BEGIN TRANSACTION"); |
files = dba_array_get(page, DBP_FILE); |
|
dba_array_FOREACH(files, file) { |
for (mpage = ohash_first(&mpages, &slot); NULL != mpage; |
if (*file < ' ') |
mpage = ohash_next(&mpages, &slot)) { |
file++; |
mlink = mpage->mlinks; |
if (ohash_find(&mlinks, ohash_qlookup(&mlinks, |
if (debug) |
file)) != NULL) { |
say(mlink->file, "Deleting from database"); |
if (debug) |
if (nodb) |
say(file, "Deleting from database"); |
continue; |
dba_array_del(dba->pages); |
for ( ; NULL != mlink; mlink = mlink->next) { |
break; |
i = 1; |
} |
SQL_BIND_TEXT(stmts[STMT_DELETE_PAGE], |
|
i, mlink->dsec); |
|
SQL_BIND_TEXT(stmts[STMT_DELETE_PAGE], |
|
i, mlink->arch); |
|
SQL_BIND_TEXT(stmts[STMT_DELETE_PAGE], |
|
i, mlink->name); |
|
SQL_STEP(stmts[STMT_DELETE_PAGE]); |
|
sqlite3_reset(stmts[STMT_DELETE_PAGE]); |
|
} |
} |
} |
} |
|
|
if (0 == nodb) |
|
SQL_EXEC("END TRANSACTION"); |
|
} |
} |
|
|
/* |
/* |
* Close an existing database and its prepared statements. |
* Write the database from memory to disk. |
* If "real" is not set, rename the temporary file into the real one. |
|
*/ |
*/ |
static void |
static void |
dbclose(int real) |
dbwrite(struct dba *dba) |
{ |
{ |
size_t i; |
struct stat sb1, sb2; |
int status; |
char tfn[33], *cp1, *cp2; |
pid_t child; |
off_t i; |
|
int fd1, fd2; |
|
|
if (nodb) |
/* |
return; |
* Do not write empty databases, and delete existing ones |
|
* when makewhatis -u causes them to become empty. |
|
*/ |
|
|
for (i = 0; i < STMT__MAX; i++) { |
dba_array_start(dba->pages); |
sqlite3_finalize(stmts[i]); |
if (dba_array_next(dba->pages) == NULL) { |
stmts[i] = NULL; |
if (unlink(MANDOC_DB) == -1 && errno != ENOENT) |
|
say(MANDOC_DB, "&unlink"); |
|
return; |
} |
} |
|
|
sqlite3_close(db); |
/* |
db = NULL; |
* Build the database in a temporary file, |
|
* then atomically move it into place. |
|
*/ |
|
|
if (real) |
if (dba_write(MANDOC_DB "~", dba) != -1) { |
return; |
if (rename(MANDOC_DB "~", MANDOC_DB) == -1) { |
|
|
if ('\0' == *tempfilename) { |
|
if (-1 == rename(MANDOC_DB "~", MANDOC_DB)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
exitcode = (int)MANDOCLEVEL_SYSERR; |
say(MANDOC_DB, "&rename"); |
say(MANDOC_DB, "&rename"); |
|
unlink(MANDOC_DB "~"); |
} |
} |
return; |
return; |
} |
} |
|
|
switch (child = fork()) { |
/* |
case -1: |
* We lack write permission and cannot replace the database |
exitcode = (int)MANDOCLEVEL_SYSERR; |
* file, but let's at least check whether the data changed. |
say("", "&fork cmp"); |
*/ |
return; |
|
case 0: |
|
execlp("cmp", "cmp", "-s", |
|
tempfilename, MANDOC_DB, NULL); |
|
say("", "&exec cmp"); |
|
exit(0); |
|
default: |
|
break; |
|
} |
|
if (-1 == waitpid(child, &status, 0)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "&wait cmp"); |
|
} else if (WIFSIGNALED(status)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "cmp died from signal %d", WTERMSIG(status)); |
|
} else if (WEXITSTATUS(status)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(MANDOC_DB, |
|
"Data changed, but cannot replace database"); |
|
} |
|
|
|
*strrchr(tempfilename, '/') = '\0'; |
(void)strlcpy(tfn, "/tmp/mandocdb.XXXXXXXX", sizeof(tfn)); |
switch (child = fork()) { |
if (mkdtemp(tfn) == NULL) { |
case -1: |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
exitcode = (int)MANDOCLEVEL_SYSERR; |
say("", "&fork rm"); |
say("", "&%s", tfn); |
return; |
return; |
case 0: |
|
execlp("rm", "rm", "-rf", tempfilename, NULL); |
|
say("", "&exec rm"); |
|
exit((int)MANDOCLEVEL_SYSERR); |
|
default: |
|
break; |
|
} |
} |
if (-1 == waitpid(child, &status, 0)) { |
cp1 = cp2 = MAP_FAILED; |
exitcode = (int)MANDOCLEVEL_SYSERR; |
fd1 = fd2 = -1; |
say("", "&wait rm"); |
(void)strlcat(tfn, "/" MANDOC_DB, sizeof(tfn)); |
} else if (WIFSIGNALED(status) || WEXITSTATUS(status)) { |
if (dba_write(tfn, dba) == -1) { |
exitcode = (int)MANDOCLEVEL_SYSERR; |
say(tfn, "&dba_write"); |
say("", "%s: Cannot remove temporary directory", |
goto err; |
tempfilename); |
|
} |
} |
} |
if ((fd1 = open(MANDOC_DB, O_RDONLY)) == -1) { |
|
say(MANDOC_DB, "&open"); |
/* |
goto err; |
* This is straightforward stuff. |
|
* Open a database connection to a "temporary" database, then open a set |
|
* of prepared statements we'll use over and over again. |
|
* If "real" is set, we use the existing database; if not, we truncate a |
|
* temporary one. |
|
* Must be matched by dbclose(). |
|
*/ |
|
static int |
|
dbopen(int real) |
|
{ |
|
const char *sql; |
|
int rc, ofl; |
|
|
|
if (nodb) |
|
return(1); |
|
|
|
*tempfilename = '\0'; |
|
ofl = SQLITE_OPEN_READWRITE; |
|
|
|
if (real) { |
|
rc = sqlite3_open_v2(MANDOC_DB, &db, ofl, NULL); |
|
if (SQLITE_OK != rc) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
if (SQLITE_CANTOPEN != rc) |
|
say(MANDOC_DB, "%s", sqlite3_errstr(rc)); |
|
return(0); |
|
} |
|
goto prepare_statements; |
|
} |
} |
|
if ((fd2 = open(tfn, O_RDONLY)) == -1) { |
ofl |= SQLITE_OPEN_CREATE | SQLITE_OPEN_EXCLUSIVE; |
say(tfn, "&open"); |
|
goto err; |
remove(MANDOC_DB "~"); |
|
rc = sqlite3_open_v2(MANDOC_DB "~", &db, ofl, NULL); |
|
if (SQLITE_OK == rc) |
|
goto create_tables; |
|
if (MPARSE_QUICK & mparse_options) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(MANDOC_DB "~", "%s", sqlite3_errstr(rc)); |
|
return(0); |
|
} |
} |
|
if (fstat(fd1, &sb1) == -1) { |
(void)strlcpy(tempfilename, "/tmp/mandocdb.XXXXXX", |
say(MANDOC_DB, "&fstat"); |
sizeof(tempfilename)); |
goto err; |
if (NULL == mkdtemp(tempfilename)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "&%s", tempfilename); |
|
return(0); |
|
} |
} |
(void)strlcat(tempfilename, "/" MANDOC_DB, |
if (fstat(fd2, &sb2) == -1) { |
sizeof(tempfilename)); |
say(tfn, "&fstat"); |
rc = sqlite3_open_v2(tempfilename, &db, ofl, NULL); |
goto err; |
if (SQLITE_OK != rc) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "%s: %s", tempfilename, sqlite3_errstr(rc)); |
|
return(0); |
|
} |
} |
|
if (sb1.st_size != sb2.st_size) |
create_tables: |
goto err; |
sql = "CREATE TABLE \"mpages\" (\n" |
if ((cp1 = mmap(NULL, sb1.st_size, PROT_READ, MAP_PRIVATE, |
" \"desc\" TEXT NOT NULL,\n" |
fd1, 0)) == MAP_FAILED) { |
" \"form\" INTEGER NOT NULL,\n" |
say(MANDOC_DB, "&mmap"); |
" \"pageid\" INTEGER PRIMARY KEY AUTOINCREMENT NOT NULL\n" |
goto err; |
");\n" |
|
"\n" |
|
"CREATE TABLE \"mlinks\" (\n" |
|
" \"sec\" TEXT NOT NULL,\n" |
|
" \"arch\" TEXT NOT NULL,\n" |
|
" \"name\" TEXT NOT NULL,\n" |
|
" \"pageid\" INTEGER NOT NULL REFERENCES mpages(pageid) " |
|
"ON DELETE CASCADE\n" |
|
");\n" |
|
"CREATE INDEX mlinks_pageid_idx ON mlinks (pageid);\n" |
|
"\n" |
|
"CREATE TABLE \"names\" (\n" |
|
" \"bits\" INTEGER NOT NULL,\n" |
|
" \"name\" TEXT NOT NULL,\n" |
|
" \"pageid\" INTEGER NOT NULL REFERENCES mpages(pageid) " |
|
"ON DELETE CASCADE\n" |
|
");\n" |
|
"\n" |
|
"CREATE TABLE \"keys\" (\n" |
|
" \"bits\" INTEGER NOT NULL,\n" |
|
" \"key\" TEXT NOT NULL,\n" |
|
" \"pageid\" INTEGER NOT NULL REFERENCES mpages(pageid) " |
|
"ON DELETE CASCADE\n" |
|
");\n" |
|
"CREATE INDEX keys_pageid_idx ON keys (pageid);\n"; |
|
|
|
if (SQLITE_OK != sqlite3_exec(db, sql, NULL, NULL, NULL)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(MANDOC_DB, "%s", sqlite3_errmsg(db)); |
|
sqlite3_close(db); |
|
return(0); |
|
} |
} |
|
if ((cp2 = mmap(NULL, sb2.st_size, PROT_READ, MAP_PRIVATE, |
prepare_statements: |
fd2, 0)) == MAP_FAILED) { |
if (SQLITE_OK != sqlite3_exec(db, |
say(tfn, "&mmap"); |
"PRAGMA foreign_keys = ON", NULL, NULL, NULL)) { |
goto err; |
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(MANDOC_DB, "PRAGMA foreign_keys: %s", |
|
sqlite3_errmsg(db)); |
|
sqlite3_close(db); |
|
return(0); |
|
} |
} |
|
for (i = 0; i < sb1.st_size; i++) |
|
if (cp1[i] != cp2[i]) |
|
goto err; |
|
goto out; |
|
|
sql = "DELETE FROM mpages WHERE pageid IN " |
err: |
"(SELECT pageid FROM mlinks WHERE " |
exitcode = (int)MANDOCLEVEL_SYSERR; |
"sec=? AND arch=? AND name=?)"; |
say(MANDOC_DB, "Data changed, but cannot replace database"); |
sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_DELETE_PAGE], NULL); |
|
sql = "INSERT INTO mpages " |
|
"(desc,form) VALUES (?,?)"; |
|
sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_PAGE], NULL); |
|
sql = "INSERT INTO mlinks " |
|
"(sec,arch,name,pageid) VALUES (?,?,?,?)"; |
|
sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_LINK], NULL); |
|
sql = "INSERT INTO names " |
|
"(bits,name,pageid) VALUES (?,?,?)"; |
|
sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_NAME], NULL); |
|
sql = "INSERT INTO keys " |
|
"(bits,key,pageid) VALUES (?,?,?)"; |
|
sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_KEY], NULL); |
|
|
|
#ifndef __APPLE__ |
out: |
/* |
if (cp1 != MAP_FAILED) |
* When opening a new database, we can turn off |
munmap(cp1, sb1.st_size); |
* synchronous mode for much better performance. |
if (cp2 != MAP_FAILED) |
*/ |
munmap(cp2, sb2.st_size); |
|
if (fd1 != -1) |
if (real && SQLITE_OK != sqlite3_exec(db, |
close(fd1); |
"PRAGMA synchronous = OFF", NULL, NULL, NULL)) { |
if (fd2 != -1) |
exitcode = (int)MANDOCLEVEL_SYSERR; |
close(fd2); |
say(MANDOC_DB, "PRAGMA synchronous: %s", |
unlink(tfn); |
sqlite3_errmsg(db)); |
*strrchr(tfn, '/') = '\0'; |
sqlite3_close(db); |
rmdir(tfn); |
return(0); |
|
} |
|
#endif |
|
|
|
return(1); |
|
} |
} |
|
|
static void * |
|
hash_calloc(size_t n, size_t sz, void *arg) |
|
{ |
|
|
|
return(mandoc_calloc(n, sz)); |
|
} |
|
|
|
static void * |
|
hash_alloc(size_t sz, void *arg) |
|
{ |
|
|
|
return(mandoc_malloc(sz)); |
|
} |
|
|
|
static void |
|
hash_free(void *p, void *arg) |
|
{ |
|
|
|
free(p); |
|
} |
|
|
|
static int |
static int |
set_basedir(const char *targetdir) |
set_basedir(const char *targetdir, int report_baddir) |
{ |
{ |
static char startdir[PATH_MAX]; |
static char startdir[PATH_MAX]; |
static int getcwd_status; /* 1 = ok, 2 = failure */ |
static int getcwd_status; /* 1 = ok, 2 = failure */ |
static int chdir_status; /* 1 = changed directory */ |
static int chdir_status; /* 1 = changed directory */ |
char *cp; |
|
|
|
/* |
/* |
* Remember the original working directory, if possible. |
* Remember the original working directory, if possible. |
Line 2386 set_basedir(const char *targetdir) |
|
Line 2319 set_basedir(const char *targetdir) |
|
* Do not error out if the current directory is not |
* Do not error out if the current directory is not |
* searchable: Maybe it won't be needed after all. |
* searchable: Maybe it won't be needed after all. |
*/ |
*/ |
if (0 == getcwd_status) { |
if (getcwd_status == 0) { |
if (NULL == getcwd(startdir, sizeof(startdir))) { |
if (getcwd(startdir, sizeof(startdir)) == NULL) { |
getcwd_status = 2; |
getcwd_status = 2; |
(void)strlcpy(startdir, strerror(errno), |
(void)strlcpy(startdir, strerror(errno), |
sizeof(startdir)); |
sizeof(startdir)); |
Line 2400 set_basedir(const char *targetdir) |
|
Line 2333 set_basedir(const char *targetdir) |
|
* Do not use it any longer, not even for messages. |
* Do not use it any longer, not even for messages. |
*/ |
*/ |
*basedir = '\0'; |
*basedir = '\0'; |
|
basedir_len = 0; |
|
|
/* |
/* |
* If and only if the directory was changed earlier and |
* If and only if the directory was changed earlier and |
* the next directory to process is given as a relative path, |
* the next directory to process is given as a relative path, |
* first go back, or bail out if that is impossible. |
* first go back, or bail out if that is impossible. |
*/ |
*/ |
if (chdir_status && '/' != *targetdir) { |
if (chdir_status && *targetdir != '/') { |
if (2 == getcwd_status) { |
if (getcwd_status == 2) { |
exitcode = (int)MANDOCLEVEL_SYSERR; |
exitcode = (int)MANDOCLEVEL_SYSERR; |
say("", "getcwd: %s", startdir); |
say("", "getcwd: %s", startdir); |
return(0); |
return 0; |
} |
} |
if (-1 == chdir(startdir)) { |
if (chdir(startdir) == -1) { |
exitcode = (int)MANDOCLEVEL_SYSERR; |
exitcode = (int)MANDOCLEVEL_SYSERR; |
say("", "&chdir %s", startdir); |
say("", "&chdir %s", startdir); |
return(0); |
return 0; |
} |
} |
} |
} |
|
|
Line 2424 set_basedir(const char *targetdir) |
|
Line 2358 set_basedir(const char *targetdir) |
|
* pathname and append a trailing slash, such that |
* pathname and append a trailing slash, such that |
* we can reliably check whether files are inside. |
* we can reliably check whether files are inside. |
*/ |
*/ |
if (NULL == realpath(targetdir, basedir)) { |
if (realpath(targetdir, basedir) == NULL) { |
exitcode = (int)MANDOCLEVEL_BADARG; |
if (report_baddir || errno != ENOENT) { |
say("", "&%s: realpath", targetdir); |
exitcode = (int)MANDOCLEVEL_BADARG; |
return(0); |
say("", "&%s: realpath", targetdir); |
} else if (-1 == chdir(basedir)) { |
} |
exitcode = (int)MANDOCLEVEL_BADARG; |
*basedir = '\0'; |
say("", "&chdir"); |
return 0; |
return(0); |
} else if (chdir(basedir) == -1) { |
|
if (report_baddir || errno != ENOENT) { |
|
exitcode = (int)MANDOCLEVEL_BADARG; |
|
say("", "&chdir"); |
|
} |
|
*basedir = '\0'; |
|
return 0; |
} |
} |
chdir_status = 1; |
chdir_status = 1; |
cp = strchr(basedir, '\0'); |
basedir_len = strlen(basedir); |
if ('/' != cp[-1]) { |
if (basedir[basedir_len - 1] != '/') { |
if (cp - basedir >= PATH_MAX - 1) { |
if (basedir_len >= PATH_MAX - 1) { |
exitcode = (int)MANDOCLEVEL_SYSERR; |
exitcode = (int)MANDOCLEVEL_SYSERR; |
say("", "Filename too long"); |
say("", "Filename too long"); |
return(0); |
*basedir = '\0'; |
|
basedir_len = 0; |
|
return 0; |
} |
} |
*cp++ = '/'; |
basedir[basedir_len++] = '/'; |
*cp = '\0'; |
basedir[basedir_len] = '\0'; |
} |
} |
return(1); |
return 1; |
} |
} |
|
|
|
#ifdef READ_ALLOWED_PATH |
|
static int |
|
read_allowed(const char *candidate) |
|
{ |
|
const char *cp; |
|
size_t len; |
|
|
|
for (cp = READ_ALLOWED_PATH;; cp += len) { |
|
while (*cp == ':') |
|
cp++; |
|
if (*cp == '\0') |
|
return 0; |
|
len = strcspn(cp, ":"); |
|
if (strncmp(candidate, cp, len) == 0) |
|
return 1; |
|
} |
|
} |
|
#endif |
|
|
static void |
static void |
say(const char *file, const char *format, ...) |
say(const char *file, const char *format, ...) |
{ |
{ |
va_list ap; |
va_list ap; |
int use_errno; |
int use_errno; |
|
|
if ('\0' != *basedir) |
if (*basedir != '\0') |
fprintf(stderr, "%s", basedir); |
fprintf(stderr, "%s", basedir); |
if ('\0' != *basedir && '\0' != *file) |
if (*basedir != '\0' && *file != '\0') |
fputc('/', stderr); |
fputc('/', stderr); |
if ('\0' != *file) |
if (*file != '\0') |
fprintf(stderr, "%s", file); |
fprintf(stderr, "%s", file); |
|
|
use_errno = 1; |
use_errno = 1; |
if (NULL != format) { |
if (format != NULL) { |
switch (*format) { |
switch (*format) { |
case '&': |
case '&': |
format++; |
format++; |
Line 2474 say(const char *file, const char *format, ...) |
|
Line 2435 say(const char *file, const char *format, ...) |
|
break; |
break; |
} |
} |
} |
} |
if (NULL != format) { |
if (format != NULL) { |
if ('\0' != *basedir || '\0' != *file) |
if (*basedir != '\0' || *file != '\0') |
fputs(": ", stderr); |
fputs(": ", stderr); |
va_start(ap, format); |
va_start(ap, format); |
vfprintf(stderr, format, ap); |
vfprintf(stderr, format, ap); |
va_end(ap); |
va_end(ap); |
} |
} |
if (use_errno) { |
if (use_errno) { |
if ('\0' != *basedir || '\0' != *file || NULL != format) |
if (*basedir != '\0' || *file != '\0' || format != NULL) |
fputs(": ", stderr); |
fputs(": ", stderr); |
perror(NULL); |
perror(NULL); |
} else |
} else |