version 1.58, 2013/06/05 02:00:26 |
version 1.235, 2017/01/10 21:59:47 |
|
|
/* $Id$ */ |
/* $Id$ */ |
/* |
/* |
* Copyright (c) 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv> |
* Copyright (c) 2011, 2012, 2013 Ingo Schwarze <schwarze@openbsd.org> |
* Copyright (c) 2011-2017 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. |
*/ |
*/ |
#ifdef HAVE_CONFIG_H |
|
#include "config.h" |
#include "config.h" |
#endif |
|
|
|
|
#include <sys/types.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 <stdint.h> |
#include <stdint.h> |
#include <stdlib.h> |
#include <stdlib.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 "manconf.h" |
#include "manpath.h" |
|
#include "mansearch.h" |
#include "mansearch.h" |
|
#include "dba_array.h" |
|
#include "dba.h" |
|
|
#define SQL_EXEC(_v) \ |
extern const char *const mansearch_keynames[]; |
if (SQLITE_OK != sqlite3_exec(db, (_v), NULL, NULL, NULL)) \ |
|
fprintf(stderr, "%s\n", sqlite3_errmsg(db)) |
|
#define SQL_BIND_TEXT(_s, _i, _v) \ |
|
if (SQLITE_OK != sqlite3_bind_text \ |
|
((_s), (_i)++, (_v), -1, SQLITE_STATIC)) \ |
|
fprintf(stderr, "%s\n", sqlite3_errmsg(db)) |
|
#define SQL_BIND_INT(_s, _i, _v) \ |
|
if (SQLITE_OK != sqlite3_bind_int \ |
|
((_s), (_i)++, (_v))) \ |
|
fprintf(stderr, "%s\n", sqlite3_errmsg(db)) |
|
#define SQL_BIND_INT64(_s, _i, _v) \ |
|
if (SQLITE_OK != sqlite3_bind_int64 \ |
|
((_s), (_i)++, (_v))) \ |
|
fprintf(stderr, "%s\n", sqlite3_errmsg(db)) |
|
#define SQL_STEP(_s) \ |
|
if (SQLITE_DONE != sqlite3_step((_s))) \ |
|
fprintf(stderr, "%s\n", 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_TEST /* change no databases, report potential problems */ |
OP_TEST /* change no databases, report potential problems */ |
}; |
}; |
|
|
enum form { |
|
FORM_SRC, /* format is -man or -mdoc */ |
|
FORM_CAT, /* format is cat */ |
|
FORM_NONE /* format is unknown */ |
|
}; |
|
|
|
struct str { |
struct str { |
char *utf8; /* key in UTF-8 form */ |
const struct mpage *mpage; /* if set, the owning parse */ |
const struct of *of; /* if set, the owning parse */ |
|
struct str *next; /* next in owning parse sequence */ |
|
uint64_t mask; /* bitmask in sequence */ |
uint64_t mask; /* bitmask in sequence */ |
char key[]; /* the string itself */ |
char key[]; /* rendered text */ |
}; |
}; |
|
|
struct id { |
struct inodev { |
ino_t ino; |
ino_t st_ino; |
dev_t dev; |
dev_t st_dev; |
}; |
}; |
|
|
struct of { |
struct mpage { |
struct id id; /* used for hashing routine */ |
struct inodev inodev; /* used for hashing routine */ |
struct of *next; /* next in ofs */ |
struct dba_array *dba; |
enum form dform; /* path-cued form */ |
char *sec; /* section from file content */ |
enum form sform; /* suffix-cued form */ |
char *arch; /* architecture from file content */ |
char file[PATH_MAX]; /* filename rel. to manpath */ |
char *title; /* title from file content */ |
const char *desc; /* parsed description */ |
char *desc; /* description from file content */ |
const char *sec; /* suffix-cued section (or empty) */ |
struct mpage *next; /* singly linked list */ |
const char *dsec; /* path-cued section (or empty) */ |
struct mlink *mlinks; /* singly linked list */ |
const char *arch; /* path-cued arch. (or empty) */ |
int name_head_done; |
const char *name; /* name (from filename) (not empty) */ |
enum form form; /* format from file content */ |
}; |
}; |
|
|
enum stmt { |
struct mlink { |
STMT_DELETE = 0, /* delete manpage */ |
char file[PATH_MAX]; /* filename rel. to manpath */ |
STMT_INSERT_DOC, /* insert manpage */ |
char *dsec; /* section from directory */ |
STMT_INSERT_KEY, /* insert parsed key */ |
char *arch; /* architecture from directory */ |
STMT__MAX |
char *name; /* name from file name (not empty) */ |
|
char *fsec; /* section from file name suffix */ |
|
struct mlink *next; /* singly linked list */ |
|
struct mpage *mpage; /* parent */ |
|
int gzip; /* filename has a .gz suffix */ |
|
enum form dform; /* format from directory */ |
|
enum form fform; /* format from file name suffix */ |
}; |
}; |
|
|
typedef int (*mdoc_fp)(struct of *, const struct mdoc_node *); |
typedef int (*mdoc_fp)(struct mpage *, const struct roff_meta *, |
|
const struct roff_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 flags; /* for use by pmdoc_node */ |
int taboo; /* node flags that must not be set */ |
#define MDOCF_CHILD 0x01 /* automatically index child nodes */ |
|
}; |
}; |
|
|
static void dbclose(const char *, int); |
|
static void dbindex(struct mchars *, int, |
int mandocdb(int, char *[]); |
const struct of *, const char *); |
|
static int dbopen(const char *, int); |
static void dbadd(struct dba *, struct mpage *); |
static void dbprune(const char *); |
static void dbadd_mlink(const struct mlink *mlink); |
static void fileadd(struct of *); |
static void dbprune(struct dba *); |
static int filecheck(const char *); |
static void dbwrite(struct dba *); |
static void filescan(const char *, const char *); |
static void filescan(const char *); |
static struct str *hashget(const char *, size_t); |
#if HAVE_FTS_COMPARE_CONST |
static void *hash_alloc(size_t, void *); |
static int fts_compare(const FTSENT *const *, const FTSENT *const *); |
static void hash_free(void *, size_t, void *); |
#else |
static void *hash_halloc(size_t, void *); |
static int fts_compare(const FTSENT **, const FTSENT **); |
static void inoadd(const struct stat *, struct of *); |
#endif |
static int inocheck(const struct stat *); |
static void mlink_add(struct mlink *, const struct stat *); |
static void ofadd(const char *, int, const char *, |
static void mlink_check(struct mpage *, struct mlink *); |
const char *, const char *, const char *, |
static void mlink_free(struct mlink *); |
const char *, const struct stat *); |
static void mlinks_undupe(struct mpage *); |
static void offree(void); |
static void mpages_free(void); |
static int ofmerge(struct mchars *, struct mparse *, const char *); |
static void mpages_merge(struct dba *, struct mparse *); |
static void parse_catpage(struct of *, const char *); |
static void parse_cat(struct mpage *, int); |
static int parse_man(struct of *, |
static void parse_man(struct mpage *, const struct roff_meta *, |
const struct man_node *); |
const struct roff_node *); |
static void parse_mdoc(struct of *, const struct mdoc_node *); |
static void parse_mdoc(struct mpage *, const struct roff_meta *, |
static int parse_mdoc_body(struct of *, const struct mdoc_node *); |
const struct roff_node *); |
static int parse_mdoc_head(struct of *, const struct mdoc_node *); |
static int parse_mdoc_head(struct mpage *, const struct roff_meta *, |
static int parse_mdoc_Fd(struct of *, const struct mdoc_node *); |
const struct roff_node *); |
static int parse_mdoc_Fn(struct of *, const struct mdoc_node *); |
static int parse_mdoc_Fd(struct mpage *, const struct roff_meta *, |
static int parse_mdoc_In(struct of *, const struct mdoc_node *); |
const struct roff_node *); |
static int parse_mdoc_Nd(struct of *, const struct mdoc_node *); |
static void parse_mdoc_fname(struct mpage *, const struct roff_node *); |
static int parse_mdoc_Nm(struct of *, const struct mdoc_node *); |
static int parse_mdoc_Fn(struct mpage *, const struct roff_meta *, |
static int parse_mdoc_Sh(struct of *, const struct mdoc_node *); |
const struct roff_node *); |
static int parse_mdoc_St(struct of *, const struct mdoc_node *); |
static int parse_mdoc_Fo(struct mpage *, const struct roff_meta *, |
static int parse_mdoc_Xr(struct of *, const struct mdoc_node *); |
const struct roff_node *); |
static int path_reset(const char *, int, const char *); |
static int parse_mdoc_Nd(struct mpage *, const struct roff_meta *, |
static void putkey(const struct of *, |
const struct roff_node *); |
const char *, uint64_t); |
static int parse_mdoc_Nm(struct mpage *, const struct roff_meta *, |
static void putkeys(const struct of *, |
const struct roff_node *); |
const char *, int, uint64_t); |
static int parse_mdoc_Sh(struct mpage *, const struct roff_meta *, |
static void putmdockey(const struct of *, |
const struct roff_node *); |
const struct mdoc_node *, uint64_t); |
static int parse_mdoc_Va(struct mpage *, const struct roff_meta *, |
static void say(const char *, const char *, const char *, ...); |
const struct roff_node *); |
static char *stradd(const char *); |
static int parse_mdoc_Xr(struct mpage *, const struct roff_meta *, |
static char *straddbuf(const char *, size_t); |
const struct roff_node *); |
static int treescan(const char *); |
static void putkey(const struct mpage *, char *, uint64_t); |
|
static void putkeys(const struct mpage *, char *, size_t, uint64_t); |
|
static void putmdockey(const struct mpage *, |
|
const struct roff_node *, uint64_t, int); |
|
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 size_t utf8(unsigned int, char [7]); |
static size_t utf8(unsigned int, char [7]); |
static void utf8key(struct mchars *, struct str *); |
|
static void wordaddbuf(const struct of *, |
|
const char *, size_t, uint64_t); |
|
|
|
static char *progname; |
|
static int use_all; /* use all found files */ |
|
static int nodb; /* no database changes */ |
static int nodb; /* no database changes */ |
static int verb; /* print what we're doing */ |
static int mparse_options; /* abort the parse early */ |
static int warnings; /* warn about crap */ |
static int use_all; /* use all found files */ |
static enum op op; /* operational mode */ |
static int debug; /* print what we're doing */ |
static struct ohash inos; /* table of inodes/devices */ |
static int warnings; /* warn about crap */ |
static struct ohash filenames; /* table of filenames */ |
static int write_utf8; /* write UTF-8 output; else ASCII */ |
|
static int exitcode; /* to be returned by main */ |
|
static enum op op; /* operational mode */ |
|
static char basedir[PATH_MAX]; /* current base directory */ |
|
static struct mpage *mpage_head; /* list of distinct manual pages */ |
|
static struct ohash mpages; /* table of distinct manual pages */ |
|
static struct ohash mlinks; /* table of directory entries */ |
|
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 struct of *ofs = NULL; /* vector of files to parse */ |
static uint64_t name_mask; |
static struct str *words = NULL; /* word list in current parse */ |
|
static sqlite3 *db = NULL; /* current database */ |
|
static sqlite3_stmt *stmts[STMT__MAX]; /* current statements */ |
|
|
|
static const struct mdoc_handler mdocs[MDOC_MAX] = { |
static const struct mdoc_handler mdocs[MDOC_MAX] = { |
{ NULL, 0, 0 }, /* Ap */ |
{ NULL, 0, 0 }, /* Ap */ |
{ NULL, 0, 0 }, /* Dd */ |
{ NULL, 0, NODE_NOPRT }, /* Dd */ |
{ NULL, 0, 0 }, /* Dt */ |
{ NULL, 0, NODE_NOPRT }, /* Dt */ |
{ NULL, 0, 0 }, /* Os */ |
{ NULL, 0, NODE_NOPRT }, /* Os */ |
{ parse_mdoc_Sh, TYPE_Sh, MDOCF_CHILD }, /* Sh */ |
{ parse_mdoc_Sh, TYPE_Sh, 0 }, /* Sh */ |
{ parse_mdoc_head, TYPE_Ss, MDOCF_CHILD }, /* Ss */ |
{ parse_mdoc_head, TYPE_Ss, 0 }, /* Ss */ |
{ NULL, 0, 0 }, /* Pp */ |
{ NULL, 0, 0 }, /* Pp */ |
{ NULL, 0, 0 }, /* D1 */ |
{ NULL, 0, 0 }, /* D1 */ |
{ NULL, 0, 0 }, /* Dl */ |
{ NULL, 0, 0 }, /* Dl */ |
Line 201 static const struct mdoc_handler mdocs[MDOC_MAX] = { |
|
Line 199 static const struct mdoc_handler mdocs[MDOC_MAX] = { |
|
{ NULL, 0, 0 }, /* El */ |
{ NULL, 0, 0 }, /* El */ |
{ NULL, 0, 0 }, /* It */ |
{ NULL, 0, 0 }, /* It */ |
{ NULL, 0, 0 }, /* Ad */ |
{ NULL, 0, 0 }, /* Ad */ |
{ NULL, TYPE_An, MDOCF_CHILD }, /* An */ |
{ NULL, TYPE_An, 0 }, /* An */ |
{ NULL, TYPE_Ar, MDOCF_CHILD }, /* Ar */ |
{ NULL, TYPE_Ar, 0 }, /* Ar */ |
{ NULL, TYPE_Cd, MDOCF_CHILD }, /* Cd */ |
{ NULL, TYPE_Cd, 0 }, /* Cd */ |
{ NULL, TYPE_Cm, MDOCF_CHILD }, /* Cm */ |
{ NULL, TYPE_Cm, 0 }, /* Cm */ |
{ NULL, TYPE_Dv, MDOCF_CHILD }, /* Dv */ |
{ NULL, TYPE_Dv, 0 }, /* Dv */ |
{ NULL, TYPE_Er, MDOCF_CHILD }, /* Er */ |
{ NULL, TYPE_Er, 0 }, /* Er */ |
{ NULL, TYPE_Ev, MDOCF_CHILD }, /* Ev */ |
{ NULL, TYPE_Ev, 0 }, /* Ev */ |
{ NULL, 0, 0 }, /* Ex */ |
{ NULL, 0, 0 }, /* Ex */ |
{ NULL, TYPE_Fa, MDOCF_CHILD }, /* Fa */ |
{ NULL, TYPE_Fa, 0 }, /* Fa */ |
{ parse_mdoc_Fd, TYPE_In, 0 }, /* Fd */ |
{ parse_mdoc_Fd, 0, 0 }, /* Fd */ |
{ NULL, TYPE_Fl, MDOCF_CHILD }, /* Fl */ |
{ NULL, TYPE_Fl, 0 }, /* Fl */ |
{ parse_mdoc_Fn, 0, 0 }, /* Fn */ |
{ parse_mdoc_Fn, 0, 0 }, /* Fn */ |
{ NULL, TYPE_Ft, MDOCF_CHILD }, /* Ft */ |
{ NULL, TYPE_Ft, 0 }, /* Ft */ |
{ NULL, TYPE_Ic, MDOCF_CHILD }, /* Ic */ |
{ NULL, TYPE_Ic, 0 }, /* Ic */ |
{ parse_mdoc_In, TYPE_In, MDOCF_CHILD }, /* In */ |
{ NULL, TYPE_In, 0 }, /* In */ |
{ NULL, TYPE_Li, MDOCF_CHILD }, /* Li */ |
{ NULL, TYPE_Li, 0 }, /* Li */ |
{ parse_mdoc_Nd, TYPE_Nd, MDOCF_CHILD }, /* Nd */ |
{ parse_mdoc_Nd, 0, 0 }, /* Nd */ |
{ parse_mdoc_Nm, TYPE_Nm, MDOCF_CHILD }, /* Nm */ |
{ parse_mdoc_Nm, 0, 0 }, /* Nm */ |
{ NULL, 0, 0 }, /* Op */ |
{ NULL, 0, 0 }, /* Op */ |
{ NULL, 0, 0 }, /* Ot */ |
{ NULL, 0, 0 }, /* Ot */ |
{ NULL, TYPE_Pa, MDOCF_CHILD }, /* Pa */ |
{ NULL, TYPE_Pa, NODE_NOSRC }, /* Pa */ |
{ NULL, 0, 0 }, /* Rv */ |
{ NULL, 0, 0 }, /* Rv */ |
{ parse_mdoc_St, TYPE_St, 0 }, /* St */ |
{ NULL, TYPE_St, 0 }, /* St */ |
{ NULL, TYPE_Va, MDOCF_CHILD }, /* Va */ |
{ parse_mdoc_Va, TYPE_Va, 0 }, /* Va */ |
{ parse_mdoc_body, TYPE_Va, MDOCF_CHILD }, /* Vt */ |
{ parse_mdoc_Va, TYPE_Vt, 0 }, /* Vt */ |
{ parse_mdoc_Xr, TYPE_Xr, 0 }, /* Xr */ |
{ parse_mdoc_Xr, 0, 0 }, /* Xr */ |
{ NULL, 0, 0 }, /* %A */ |
{ NULL, 0, 0 }, /* %A */ |
{ NULL, 0, 0 }, /* %B */ |
{ NULL, 0, 0 }, /* %B */ |
{ NULL, 0, 0 }, /* %D */ |
{ NULL, 0, 0 }, /* %D */ |
Line 241 static const struct mdoc_handler mdocs[MDOC_MAX] = { |
|
Line 239 static const struct mdoc_handler mdocs[MDOC_MAX] = { |
|
{ NULL, 0, 0 }, /* Ac */ |
{ NULL, 0, 0 }, /* Ac */ |
{ NULL, 0, 0 }, /* Ao */ |
{ NULL, 0, 0 }, /* Ao */ |
{ NULL, 0, 0 }, /* Aq */ |
{ NULL, 0, 0 }, /* Aq */ |
{ NULL, TYPE_At, MDOCF_CHILD }, /* At */ |
{ NULL, TYPE_At, 0 }, /* At */ |
{ NULL, 0, 0 }, /* Bc */ |
{ NULL, 0, 0 }, /* Bc */ |
{ NULL, 0, 0 }, /* Bf */ |
{ NULL, 0, 0 }, /* Bf */ |
{ NULL, 0, 0 }, /* Bo */ |
{ NULL, 0, 0 }, /* Bo */ |
{ NULL, 0, 0 }, /* Bq */ |
{ NULL, 0, 0 }, /* Bq */ |
{ NULL, TYPE_Bsx, MDOCF_CHILD }, /* Bsx */ |
{ NULL, TYPE_Bsx, NODE_NOSRC }, /* Bsx */ |
{ NULL, TYPE_Bx, MDOCF_CHILD }, /* Bx */ |
{ NULL, TYPE_Bx, 0 }, /* Bx */ |
{ NULL, 0, 0 }, /* Db */ |
{ NULL, 0, 0 }, /* Db */ |
{ NULL, 0, 0 }, /* Dc */ |
{ NULL, 0, 0 }, /* Dc */ |
{ NULL, 0, 0 }, /* Do */ |
{ NULL, 0, 0 }, /* Do */ |
{ NULL, 0, 0 }, /* Dq */ |
{ NULL, 0, 0 }, /* Dq */ |
{ NULL, 0, 0 }, /* Ec */ |
{ NULL, 0, 0 }, /* Ec */ |
{ NULL, 0, 0 }, /* Ef */ |
{ NULL, 0, 0 }, /* Ef */ |
{ NULL, TYPE_Em, MDOCF_CHILD }, /* Em */ |
{ NULL, TYPE_Em, 0 }, /* Em */ |
{ NULL, 0, 0 }, /* Eo */ |
{ NULL, 0, 0 }, /* Eo */ |
{ NULL, TYPE_Fx, MDOCF_CHILD }, /* Fx */ |
{ NULL, TYPE_Fx, NODE_NOSRC }, /* Fx */ |
{ NULL, TYPE_Ms, MDOCF_CHILD }, /* Ms */ |
{ NULL, TYPE_Ms, 0 }, /* Ms */ |
{ NULL, 0, 0 }, /* No */ |
{ NULL, 0, 0 }, /* No */ |
{ NULL, 0, 0 }, /* Ns */ |
{ NULL, 0, 0 }, /* Ns */ |
{ NULL, TYPE_Nx, MDOCF_CHILD }, /* Nx */ |
{ NULL, TYPE_Nx, NODE_NOSRC }, /* Nx */ |
{ NULL, TYPE_Ox, MDOCF_CHILD }, /* Ox */ |
{ NULL, TYPE_Ox, NODE_NOSRC }, /* Ox */ |
{ NULL, 0, 0 }, /* Pc */ |
{ NULL, 0, 0 }, /* Pc */ |
{ NULL, 0, 0 }, /* Pf */ |
{ NULL, 0, 0 }, /* Pf */ |
{ NULL, 0, 0 }, /* Po */ |
{ NULL, 0, 0 }, /* Po */ |
Line 277 static const struct mdoc_handler mdocs[MDOC_MAX] = { |
|
Line 275 static const struct mdoc_handler mdocs[MDOC_MAX] = { |
|
{ NULL, 0, 0 }, /* Sq */ |
{ NULL, 0, 0 }, /* Sq */ |
{ NULL, 0, 0 }, /* Sm */ |
{ NULL, 0, 0 }, /* Sm */ |
{ NULL, 0, 0 }, /* Sx */ |
{ NULL, 0, 0 }, /* Sx */ |
{ NULL, TYPE_Sy, MDOCF_CHILD }, /* Sy */ |
{ NULL, TYPE_Sy, 0 }, /* Sy */ |
{ NULL, TYPE_Tn, MDOCF_CHILD }, /* Tn */ |
{ NULL, TYPE_Tn, 0 }, /* Tn */ |
{ NULL, 0, 0 }, /* Ux */ |
{ NULL, 0, NODE_NOSRC }, /* Ux */ |
{ NULL, 0, 0 }, /* Xc */ |
{ NULL, 0, 0 }, /* Xc */ |
{ NULL, 0, 0 }, /* Xo */ |
{ NULL, 0, 0 }, /* Xo */ |
{ parse_mdoc_head, TYPE_Fn, 0 }, /* Fo */ |
{ parse_mdoc_Fo, 0, 0 }, /* Fo */ |
{ NULL, 0, 0 }, /* Fc */ |
{ NULL, 0, 0 }, /* Fc */ |
{ NULL, 0, 0 }, /* Oo */ |
{ NULL, 0, 0 }, /* Oo */ |
{ NULL, 0, 0 }, /* Oc */ |
{ NULL, 0, 0 }, /* Oc */ |
Line 292 static const struct mdoc_handler mdocs[MDOC_MAX] = { |
|
Line 290 static const struct mdoc_handler mdocs[MDOC_MAX] = { |
|
{ NULL, 0, 0 }, /* Hf */ |
{ NULL, 0, 0 }, /* Hf */ |
{ NULL, 0, 0 }, /* Fr */ |
{ NULL, 0, 0 }, /* Fr */ |
{ NULL, 0, 0 }, /* Ud */ |
{ NULL, 0, 0 }, /* Ud */ |
{ NULL, TYPE_Lb, MDOCF_CHILD }, /* Lb */ |
{ NULL, TYPE_Lb, 0 }, /* Lb */ |
{ NULL, 0, 0 }, /* Lp */ |
{ NULL, 0, 0 }, /* Lp */ |
{ NULL, TYPE_Lk, MDOCF_CHILD }, /* Lk */ |
{ NULL, TYPE_Lk, 0 }, /* Lk */ |
{ NULL, TYPE_Mt, MDOCF_CHILD }, /* Mt */ |
{ NULL, TYPE_Mt, NODE_NOSRC }, /* Mt */ |
{ NULL, 0, 0 }, /* Brq */ |
{ NULL, 0, 0 }, /* Brq */ |
{ NULL, 0, 0 }, /* Bro */ |
{ NULL, 0, 0 }, /* Bro */ |
{ NULL, 0, 0 }, /* Brc */ |
{ NULL, 0, 0 }, /* Brc */ |
{ NULL, 0, 0 }, /* %C */ |
{ NULL, 0, 0 }, /* %C */ |
{ NULL, 0, 0 }, /* Es */ |
{ NULL, 0, 0 }, /* Es */ |
{ NULL, 0, 0 }, /* En */ |
{ NULL, 0, 0 }, /* En */ |
{ NULL, TYPE_Dx, MDOCF_CHILD }, /* Dx */ |
{ NULL, TYPE_Dx, NODE_NOSRC }, /* Dx */ |
{ NULL, 0, 0 }, /* %Q */ |
{ NULL, 0, 0 }, /* %Q */ |
{ NULL, 0, 0 }, /* br */ |
{ NULL, 0, 0 }, /* br */ |
{ NULL, 0, 0 }, /* sp */ |
{ NULL, 0, 0 }, /* sp */ |
{ NULL, 0, 0 }, /* %U */ |
{ NULL, 0, 0 }, /* %U */ |
{ NULL, 0, 0 }, /* Ta */ |
{ NULL, 0, 0 }, /* Ta */ |
|
{ NULL, 0, 0 }, /* ll */ |
}; |
}; |
|
|
|
|
int |
int |
main(int argc, char *argv[]) |
mandocdb(int argc, char *argv[]) |
{ |
{ |
char cwd[PATH_MAX]; |
struct manconf conf; |
int ch, rc, fd, i; |
|
unsigned int index; |
|
size_t j, sz; |
|
const char *dir; |
|
struct str *s; |
|
struct mchars *mc; |
|
struct manpaths dirs; |
|
struct mparse *mp; |
struct mparse *mp; |
struct ohash_info ino_info, filename_info, str_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 fattr flock proc exec", NULL) == -1) { |
|
warn("pledge"); |
|
return (int)MANDOCLEVEL_SYSERR; |
|
} |
|
#endif |
|
|
ino_info.halloc = filename_info.halloc = |
#if HAVE_SANDBOX_INIT |
str_info.halloc = hash_halloc; |
if (sandbox_init(kSBXProfileNoInternet, SANDBOX_NAMED, NULL) == -1) { |
ino_info.hfree = filename_info.hfree = |
warnx("sandbox_init"); |
str_info.hfree = hash_free; |
return (int)MANDOCLEVEL_SYSERR; |
ino_info.alloc = filename_info.alloc = |
} |
str_info.alloc = hash_alloc; |
#endif |
|
|
ino_info.key_offset = offsetof(struct of, id); |
memset(&conf, 0, sizeof(conf)); |
filename_info.key_offset = offsetof(struct of, file); |
|
str_info.key_offset = offsetof(struct str, key); |
|
|
|
progname = strrchr(argv[0], '/'); |
|
if (progname == NULL) |
|
progname = argv[0]; |
|
else |
|
++progname; |
|
|
|
/* |
/* |
* Remember where we started by keeping a fd open to the origin |
* We accept a few different invocations. |
* path component: throughout this utility, we chdir() a lot to |
|
* handle relative paths, and by doing this, we can return to |
|
* the starting point. |
|
*/ |
|
if (NULL == getcwd(cwd, PATH_MAX)) { |
|
perror(NULL); |
|
return(EXIT_FAILURE); |
|
} else if (-1 == (fd = open(cwd, O_RDONLY, 0))) { |
|
perror(cwd); |
|
return(EXIT_FAILURE); |
|
} |
|
|
|
/* |
|
* 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_DEFAULT != (_op)) { \ |
fprintf(stderr, "-%c: Conflicting option\n", (_ch)); \ |
warnx("-%c: Conflicting option", (_ch)); \ |
goto usage; \ |
goto usage; \ |
} while (/*CONSTCOND*/0) |
} while (/*CONSTCOND*/0) |
|
|
dir = NULL; |
path_arg = NULL; |
op = OP_DEFAULT; |
op = OP_DEFAULT; |
|
|
while (-1 != (ch = getopt(argc, argv, "aC:d:ntu:vW"))) |
while (-1 != (ch = getopt(argc, argv, "aC:Dd:npQT:tu:v"))) |
switch (ch) { |
switch (ch) { |
case ('a'): |
case 'a': |
use_all = 1; |
use_all = 1; |
break; |
break; |
case ('C'): |
case 'C': |
CHECKOP(op, ch); |
CHECKOP(op, ch); |
dir = optarg; |
path_arg = optarg; |
op = OP_CONFFILE; |
op = OP_CONFFILE; |
break; |
break; |
case ('d'): |
case 'D': |
|
debug++; |
|
break; |
|
case 'd': |
CHECKOP(op, ch); |
CHECKOP(op, ch); |
dir = optarg; |
path_arg = optarg; |
op = OP_UPDATE; |
op = OP_UPDATE; |
break; |
break; |
case ('n'): |
case 'n': |
nodb = 1; |
nodb = 1; |
break; |
break; |
case ('t'): |
case 'p': |
|
warnings = 1; |
|
break; |
|
case 'Q': |
|
mparse_options |= MPARSE_QUICK; |
|
break; |
|
case 'T': |
|
if (strcmp(optarg, "utf8")) { |
|
warnx("-T%s: Unsupported output format", |
|
optarg); |
|
goto usage; |
|
} |
|
write_utf8 = 1; |
|
break; |
|
case 't': |
CHECKOP(op, ch); |
CHECKOP(op, ch); |
dup2(STDOUT_FILENO, STDERR_FILENO); |
dup2(STDOUT_FILENO, STDERR_FILENO); |
op = OP_TEST; |
op = OP_TEST; |
nodb = warnings = 1; |
nodb = warnings = 1; |
break; |
break; |
case ('u'): |
case 'u': |
CHECKOP(op, ch); |
CHECKOP(op, ch); |
dir = optarg; |
path_arg = optarg; |
op = OP_DELETE; |
op = OP_DELETE; |
break; |
break; |
case ('v'): |
case 'v': |
verb++; |
/* Compatibility with espie@'s makewhatis. */ |
break; |
break; |
case ('W'): |
|
warnings = 1; |
|
break; |
|
default: |
default: |
goto usage; |
goto usage; |
} |
} |
Line 414 main(int argc, char *argv[]) |
|
Line 406 main(int argc, char *argv[]) |
|
argc -= optind; |
argc -= optind; |
argv += optind; |
argv += optind; |
|
|
|
#if HAVE_PLEDGE |
|
if (nodb) { |
|
if (pledge("stdio rpath", NULL) == -1) { |
|
warn("pledge"); |
|
return (int)MANDOCLEVEL_SYSERR; |
|
} |
|
} |
|
#endif |
|
|
if (OP_CONFFILE == op && argc > 0) { |
if (OP_CONFFILE == op && argc > 0) { |
fprintf(stderr, "-C: Too many arguments\n"); |
warnx("-C: Too many arguments"); |
goto usage; |
goto usage; |
} |
} |
|
|
rc = 1; |
exitcode = (int)MANDOCLEVEL_OK; |
mp = mparse_alloc(MPARSE_AUTO, |
mchars_alloc(); |
MANDOCLEVEL_FATAL, NULL, NULL, NULL); |
mp = mparse_alloc(mparse_options, MANDOCLEVEL_BADARG, NULL, NULL); |
mc = mchars_alloc(); |
mandoc_ohash_init(&mpages, 6, offsetof(struct mpage, inodev)); |
|
mandoc_ohash_init(&mlinks, 6, offsetof(struct mlink, file)); |
|
|
ohash_init(&strings, 6, &str_info); |
|
ohash_init(&inos, 6, &ino_info); |
|
ohash_init(&filenames, 6, &filename_info); |
|
|
|
if (OP_UPDATE == op || OP_DELETE == op || OP_TEST == op) { |
if (OP_UPDATE == op || OP_DELETE == op || OP_TEST == op) { |
/* |
|
* Force processing all files. |
|
*/ |
|
use_all = 1; |
|
if (NULL == dir) |
|
dir = cwd; |
|
/* |
/* |
* All of these deal with a specific directory. |
* Most of these deal with a specific directory. |
* Jump into that directory then collect files specified |
* Jump into that directory first. |
* on the command-line. |
|
*/ |
*/ |
if (0 == path_reset(cwd, fd, dir)) |
if (OP_TEST != op && 0 == set_basedir(path_arg, 1)) |
goto out; |
goto out; |
for (i = 0; i < argc; i++) |
|
filescan(argv[i], dir); |
dba = nodb ? dba_new(128) : dba_read(MANDOC_DB); |
if (0 == dbopen(dir, 1)) |
if (dba != NULL) { |
goto out; |
/* |
if (OP_TEST != op) |
* The existing database is usable. Process |
dbprune(dir); |
* all files specified on the command-line. |
|
*/ |
|
#if HAVE_PLEDGE |
|
if (!nodb) { |
|
if (pledge("stdio rpath wpath cpath fattr flock", NULL) == -1) { |
|
warn("pledge"); |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
goto out; |
|
} |
|
} |
|
#endif |
|
use_all = 1; |
|
for (i = 0; i < argc; i++) |
|
filescan(argv[i]); |
|
if (nodb == 0) |
|
dbprune(dba); |
|
} else { |
|
/* Database missing or corrupt. */ |
|
if (op != OP_UPDATE || errno != ENOENT) |
|
say(MANDOC_DB, "%s: Automatically recreating" |
|
" from scratch", strerror(errno)); |
|
exitcode = (int)MANDOCLEVEL_OK; |
|
op = OP_DEFAULT; |
|
if (0 == treescan()) |
|
goto out; |
|
dba = dba_new(128); |
|
} |
if (OP_DELETE != op) |
if (OP_DELETE != op) |
rc = ofmerge(mc, mp, dir); |
mpages_merge(dba, mp); |
dbclose(dir, 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_calloc |
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, dir, NULL, NULL); |
manconf_parse(&conf, path_arg, NULL, NULL); |
|
|
|
if (conf.manpath.sz == 0) { |
|
exitcode = (int)MANDOCLEVEL_BADARG; |
|
say("", "Empty manpath"); |
|
} |
|
|
/* |
/* |
* First scan the tree rooted at a base directory. |
* First scan the tree rooted at a base directory, then |
* Then whak its database (if one exists), parse, and |
* build a new database and finally move it into place. |
* build up the database. |
|
* 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 (0 == sz) |
continue; |
continue; |
if (0 == path_reset(cwd, fd, dirs.paths[j])) |
|
goto out; |
|
if (0 == treescan(dirs.paths[j])) |
|
goto out; |
|
if (0 == path_reset(cwd, fd, dirs.paths[j])) |
|
goto out; |
|
if (0 == dbopen(dirs.paths[j], 0)) |
|
goto out; |
|
|
|
/* |
if (j) { |
* Since we're opening up a new database, we can |
mandoc_ohash_init(&mpages, 6, |
* turn off synchronous mode for much better |
offsetof(struct mpage, inodev)); |
* performance. |
mandoc_ohash_init(&mlinks, 6, |
*/ |
offsetof(struct mlink, file)); |
#ifndef __APPLE__ |
} |
SQL_EXEC("PRAGMA synchronous = OFF"); |
|
#endif |
|
|
|
if (0 == ofmerge(mc, mp, dirs.paths[j])) |
if ( ! set_basedir(conf.manpath.paths[j], argc > 0)) |
goto out; |
continue; |
dbclose(dirs.paths[j], 0); |
if (0 == treescan()) |
offree(); |
continue; |
ohash_delete(&inos); |
dba = dba_new(128); |
ohash_init(&inos, 6, &ino_info); |
mpages_merge(dba, mp); |
ohash_delete(&filenames); |
if (nodb == 0) |
ohash_init(&filenames, 6, &filename_info); |
dbwrite(dba); |
|
dba_free(dba); |
|
|
|
if (j + 1 < conf.manpath.sz) { |
|
mpages_free(); |
|
ohash_delete(&mpages); |
|
ohash_delete(&mlinks); |
|
} |
} |
} |
} |
} |
out: |
out: |
close(fd); |
manconf_free(&conf); |
manpath_free(&dirs); |
|
mchars_free(mc); |
|
mparse_free(mp); |
mparse_free(mp); |
for (s = ohash_first(&strings, &index); |
mchars_free(); |
NULL != s; s = ohash_next(&strings, &index)) { |
mpages_free(); |
if (s->utf8 != s->key) |
ohash_delete(&mpages); |
free(s->utf8); |
ohash_delete(&mlinks); |
free(s); |
return exitcode; |
} |
|
ohash_delete(&strings); |
|
ohash_delete(&inos); |
|
ohash_delete(&filenames); |
|
offree(); |
|
return(rc ? EXIT_SUCCESS : EXIT_FAILURE); |
|
usage: |
usage: |
fprintf(stderr, "usage: %s [-anvW] [-C file]\n" |
progname = getprogname(); |
" %s [-anvW] dir ...\n" |
fprintf(stderr, "usage: %s [-aDnpQ] [-C file] [-Tutf8]\n" |
" %s [-nvW] -d dir [file ...]\n" |
" %s [-aDnpQ] [-Tutf8] dir ...\n" |
" %s [-nvW] -u dir [file ...]\n" |
" %s [-DnpQ] [-Tutf8] -d dir [file ...]\n" |
" %s -t file ...\n", |
" %s [-Dnp] -u dir [file ...]\n" |
progname, progname, progname, |
" %s [-Q] -t file ...\n", |
progname, progname); |
progname, progname, progname, progname, progname); |
|
|
return(EXIT_FAILURE); |
return (int)MANDOCLEVEL_BADARG; |
} |
} |
|
|
/* |
/* |
* Scan a directory tree rooted at "base" for manpages. |
* 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. |
* 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. |
* |
* |
* If use_all has been specified, grok all files. |
* If use_all has been specified, grok all files. |
* If not, sanitise paths to the following: |
* If not, sanitise paths to the following: |
* |
* |
* [./]man*[/<arch>]/<name>.<section> |
* [./]man*[/<arch>]/<name>.<section> |
* 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(const char *base) |
treescan(void) |
{ |
{ |
|
char buf[PATH_MAX]; |
FTS *f; |
FTS *f; |
FTSENT *ff; |
FTSENT *ff; |
int dform; |
struct mlink *mlink; |
char *sec; |
int gzip; |
const char *dsec, *arch, *cp, *name, *path; |
enum form dform; |
|
char *dsec, *arch, *fsec, *cp; |
|
const char *path; |
const char *argv[2]; |
const char *argv[2]; |
|
|
argv[0] = "."; |
argv[0] = "."; |
argv[1] = (char *)NULL; |
argv[1] = (char *)NULL; |
|
|
/* |
f = fts_open((char * const *)argv, FTS_PHYSICAL | FTS_NOCHDIR, |
* Walk through all components under the directory, using the |
fts_compare); |
* logical descent of files. |
if (f == NULL) { |
*/ |
exitcode = (int)MANDOCLEVEL_SYSERR; |
f = fts_open((char * const *)argv, FTS_LOGICAL, NULL); |
say("", "&fts_open"); |
if (NULL == f) { |
return 0; |
perror(base); |
|
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) { |
|
|
/* |
/* |
* If we're a regular file, add an "of" by using the |
* Symbolic links require various sanity checks, |
* stored directory data and handling the filename. |
* then get handled just like regular files. |
* Disallow duplicate (hard-linked) files. |
|
*/ |
*/ |
if (FTS_F == ff->fts_info) { |
case FTS_SL: |
if ( ! use_all && ff->fts_level < 2) { |
if (realpath(path, buf) == NULL) { |
if (warnings) |
if (warnings) |
say(base, path, "Extraneous file"); |
say(path, "&realpath"); |
continue; |
continue; |
} else if (inocheck(ff->fts_statp)) { |
} |
|
if (strstr(buf, basedir) != buf |
|
#ifdef HOMEBREWDIR |
|
&& strstr(buf, HOMEBREWDIR) != buf |
|
#endif |
|
) { |
|
if (warnings) say("", |
|
"%s: outside base directory", buf); |
|
continue; |
|
} |
|
/* Use logical inode to avoid mpages dupe. */ |
|
if (stat(path, ff->fts_statp) == -1) { |
if (warnings) |
if (warnings) |
say(base, path, "Duplicate file"); |
say(path, "&stat"); |
continue; |
continue; |
} |
} |
|
/* FALLTHROUGH */ |
|
|
cp = ff->fts_name; |
/* |
|
* If we're a regular file, add an mlink by using the |
if (0 == strcmp(cp, "mandocdb.db")) { |
* stored directory data and handling the filename. |
|
*/ |
|
case FTS_F: |
|
if ( ! strcmp(path, MANDOC_DB)) |
|
continue; |
|
if ( ! use_all && ff->fts_level < 2) { |
if (warnings) |
if (warnings) |
say(base, path, "Skip database"); |
say(path, "Extraneous file"); |
continue; |
continue; |
} else if (NULL != (cp = strrchr(cp, '.'))) { |
} |
if (0 == strcmp(cp + 1, "html")) { |
gzip = 0; |
|
fsec = NULL; |
|
while (fsec == NULL) { |
|
fsec = strrchr(ff->fts_name, '.'); |
|
if (fsec == NULL || strcmp(fsec+1, "gz")) |
|
break; |
|
gzip = 1; |
|
*fsec = '\0'; |
|
fsec = NULL; |
|
} |
|
if (fsec == NULL) { |
|
if ( ! use_all) { |
if (warnings) |
if (warnings) |
say(base, path, "Skip html"); |
say(path, |
|
"No filename suffix"); |
continue; |
continue; |
} else if (0 == strcmp(cp + 1, "gz")) { |
|
if (warnings) |
|
say(base, path, "Skip gz"); |
|
continue; |
|
} else if (0 == strcmp(cp + 1, "ps")) { |
|
if (warnings) |
|
say(base, path, "Skip ps"); |
|
continue; |
|
} else if (0 == strcmp(cp + 1, "pdf")) { |
|
if (warnings) |
|
say(base, path, "Skip pdf"); |
|
continue; |
|
} |
} |
} |
} else if ( ! strcmp(++fsec, "html")) { |
|
if (warnings) |
|
say(path, "Skip html"); |
|
continue; |
|
} else if ( ! strcmp(fsec, "ps")) { |
|
if (warnings) |
|
say(path, "Skip ps"); |
|
continue; |
|
} else if ( ! strcmp(fsec, "pdf")) { |
|
if (warnings) |
|
say(path, "Skip pdf"); |
|
continue; |
|
} else if ( ! use_all && |
|
((dform == FORM_SRC && |
|
strncmp(fsec, dsec, strlen(dsec))) || |
|
(dform == FORM_CAT && strcmp(fsec, "0")))) { |
|
if (warnings) |
|
say(path, "Wrong filename suffix"); |
|
continue; |
|
} else |
|
fsec[-1] = '\0'; |
|
|
if (NULL != (sec = strrchr(ff->fts_name, '.'))) { |
mlink = mandoc_calloc(1, sizeof(struct mlink)); |
*sec = '\0'; |
if (strlcpy(mlink->file, path, |
sec = stradd(sec + 1); |
sizeof(mlink->file)) >= |
|
sizeof(mlink->file)) { |
|
say(path, "Filename too long"); |
|
free(mlink); |
|
continue; |
} |
} |
name = stradd(ff->fts_name); |
mlink->dform = dform; |
ofadd(base, dform, path, |
mlink->dsec = dsec; |
name, dsec, sec, arch, ff->fts_statp); |
mlink->arch = arch; |
|
mlink->name = ff->fts_name; |
|
mlink->fsec = fsec; |
|
mlink->gzip = gzip; |
|
mlink_add(mlink, ff->fts_statp); |
continue; |
continue; |
} else if (FTS_D != ff->fts_info && |
|
FTS_DP != ff->fts_info) |
case FTS_D: |
|
case FTS_DP: |
|
break; |
|
|
|
default: |
|
if (warnings) |
|
say(path, "Not a regular file"); |
continue; |
continue; |
|
} |
|
|
switch (ff->fts_level) { |
switch (ff->fts_level) { |
case (0): |
case 0: |
/* Ignore the root directory. */ |
/* Ignore the root directory. */ |
break; |
break; |
case (1): |
case 1: |
/* |
/* |
* This might contain manX/ or catX/. |
* This might contain manX/ or catX/. |
* Try to infer this from the name. |
* Try to infer this from the name. |
* If we're not in use_all, enforce it. |
* If we're not in use_all, enforce it. |
*/ |
*/ |
dsec = NULL; |
|
dform = FORM_NONE; |
|
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 = stradd(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 = stradd(cp + 3); |
dsec = cp + 3; |
|
} else { |
|
dform = FORM_NONE; |
|
dsec = NULL; |
} |
} |
|
|
if (NULL != dsec || use_all) |
if (dsec != NULL || use_all) |
break; |
break; |
|
|
if (warnings) |
if (warnings) |
say(base, path, "Unknown directory part"); |
say(path, "Unknown directory part"); |
fts_set(f, ff, FTS_SKIP); |
fts_set(f, ff, FTS_SKIP); |
break; |
break; |
case (2): |
case 2: |
/* |
/* |
* Possibly our architecture. |
* Possibly our architecture. |
* If we're descending, keep tabs on it. |
* If we're descending, keep tabs on it. |
*/ |
*/ |
arch = NULL; |
if (ff->fts_info != FTS_DP && dsec != NULL) |
if (FTS_DP != ff->fts_info && NULL != dsec) |
arch = ff->fts_name; |
arch = stradd(ff->fts_name); |
else |
|
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(base, path, "Extraneous directory part"); |
say(path, "Extraneous directory part"); |
fts_set(f, ff, FTS_SKIP); |
fts_set(f, ff, FTS_SKIP); |
break; |
break; |
} |
} |
} |
} |
|
|
fts_close(f); |
fts_close(f); |
return(1); |
return 1; |
} |
} |
|
|
/* |
/* |
* Add a file to the file vector. |
* Add a file to the mlinks table. |
* Do not verify that it's a "valid" looking manpage (we'll do that |
* Do not verify that it's a "valid" looking manpage (we'll do that |
* later). |
* later). |
* |
* |
* Try to infer the manual section, architecture, and page name from the |
* Try to infer the manual section, architecture, and page name from the |
* path, assuming it looks like |
* path, assuming it looks like |
* |
* |
* [./]man*[/<arch>]/<name>.<section> |
* [./]man*[/<arch>]/<name>.<section> |
* or |
* or |
* [./]cat<section>[/<arch>]/<name>.0 |
* [./]cat<section>[/<arch>]/<name>.0 |
* |
* |
* Stuff this information directly into the "of" vector. |
|
* 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, const char *base) |
filescan(const char *file) |
{ |
{ |
const char *sec, *arch, *name, *dsec; |
char buf[PATH_MAX]; |
char *p, *start, *buf; |
|
int dform; |
|
struct stat st; |
struct stat st; |
|
struct mlink *mlink; |
|
char *p, *start; |
|
|
assert(use_all); |
assert(use_all); |
|
|
if (0 == strncmp(file, "./", 2)) |
if (0 == strncmp(file, "./", 2)) |
file += 2; |
file += 2; |
|
|
if (-1 == stat(file, &st)) { |
/* |
if (warnings) |
* We have to do lstat(2) before realpath(3) loses |
say(base, file, "%s", strerror(errno)); |
* the information whether this is a symbolic link. |
|
* We need to know that because for symbolic links, |
|
* we want to use the orginal file name, while for |
|
* regular files, we want to use the real path. |
|
*/ |
|
if (-1 == lstat(file, &st)) { |
|
exitcode = (int)MANDOCLEVEL_BADARG; |
|
say(file, "&lstat"); |
return; |
return; |
} else if ( ! (S_IFREG & st.st_mode)) { |
} else if (0 == ((S_IFREG | S_IFLNK) & st.st_mode)) { |
if (warnings) |
exitcode = (int)MANDOCLEVEL_BADARG; |
say(base, file, "Not a regular file"); |
say(file, "Not a regular file"); |
return; |
return; |
} else if (inocheck(&st)) { |
} |
if (warnings) |
|
say(base, file, "Duplicate file"); |
/* |
|
* We have to resolve the file name to the real path |
|
* in any case for the base directory check. |
|
*/ |
|
if (NULL == realpath(file, buf)) { |
|
exitcode = (int)MANDOCLEVEL_BADARG; |
|
say(file, "&realpath"); |
return; |
return; |
} |
} |
|
|
buf = mandoc_strdup(file); |
if (OP_TEST == op) |
start = buf; |
start = buf; |
sec = arch = name = dsec = NULL; |
else if (strstr(buf, basedir) == buf) |
dform = FORM_NONE; |
start = buf + strlen(basedir); |
|
#ifdef HOMEBREWDIR |
|
else if (strstr(buf, HOMEBREWDIR) == buf) |
|
start = buf; |
|
#endif |
|
else { |
|
exitcode = (int)MANDOCLEVEL_BADARG; |
|
say("", "%s: outside base directory", buf); |
|
return; |
|
} |
|
|
/* |
/* |
|
* Now we are sure the file is inside our tree. |
|
* If it is a symbolic link, ignore the real path |
|
* 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) { |
|
if (-1 == stat(buf, &st)) { |
|
exitcode = (int)MANDOCLEVEL_BADARG; |
|
say(file, "&stat"); |
|
return; |
|
} |
|
if (strlcpy(buf, file, sizeof(buf)) >= sizeof(buf)) { |
|
say(file, "Filename too long"); |
|
return; |
|
} |
|
start = buf; |
|
if (OP_TEST != op && strstr(buf, basedir) == buf) |
|
start += strlen(basedir); |
|
} |
|
|
|
mlink = mandoc_calloc(1, sizeof(struct mlink)); |
|
mlink->dform = FORM_NONE; |
|
if (strlcpy(mlink->file, start, sizeof(mlink->file)) >= |
|
sizeof(mlink->file)) { |
|
say(start, "Filename too long"); |
|
free(mlink); |
|
return; |
|
} |
|
|
|
/* |
* 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, |
Line 741 filescan(const char *file, const char *base) |
|
Line 880 filescan(const char *file, const char *base) |
|
if (NULL != (p = strchr(start, '/'))) { |
if (NULL != (p = strchr(start, '/'))) { |
*p++ = '\0'; |
*p++ = '\0'; |
if (0 == strncmp(start, "man", 3)) { |
if (0 == strncmp(start, "man", 3)) { |
dform = FORM_SRC; |
mlink->dform = FORM_SRC; |
dsec = start + 3; |
mlink->dsec = start + 3; |
} else if (0 == strncmp(start, "cat", 3)) { |
} else if (0 == strncmp(start, "cat", 3)) { |
dform = FORM_CAT; |
mlink->dform = FORM_CAT; |
dsec = start + 3; |
mlink->dsec = start + 3; |
} |
} |
|
|
start = p; |
start = p; |
if (NULL != dsec && NULL != (p = strchr(start, '/'))) { |
if (NULL != mlink->dsec && NULL != (p = strchr(start, '/'))) { |
*p++ = '\0'; |
*p++ = '\0'; |
arch = start; |
mlink->arch = start; |
start = p; |
start = p; |
} |
} |
} |
} |
|
|
/* |
/* |
Line 766 filescan(const char *file, const char *base) |
|
Line 905 filescan(const char *file, const char *base) |
|
|
|
if ('.' == *p) { |
if ('.' == *p) { |
*p++ = '\0'; |
*p++ = '\0'; |
sec = p; |
mlink->fsec = p; |
} |
} |
|
|
/* |
/* |
* Now try to parse the name. |
* Now try to parse the name. |
* Use the filename portion of the path. |
* Use the filename portion of the path. |
*/ |
*/ |
name = start; |
mlink->name = start; |
if (NULL != (p = strrchr(start, '/'))) { |
if (NULL != (p = strrchr(start, '/'))) { |
name = p + 1; |
mlink->name = p + 1; |
*p = '\0'; |
*p = '\0'; |
} |
} |
|
mlink_add(mlink, &st); |
ofadd(base, dform, file, name, dsec, sec, arch, &st); |
|
free(buf); |
|
} |
} |
|
|
/* |
static void |
* See fileadd(). |
mlink_add(struct mlink *mlink, const struct stat *st) |
*/ |
|
static int |
|
filecheck(const char *name) |
|
{ |
{ |
unsigned int index; |
struct inodev inodev; |
|
struct mpage *mpage; |
|
unsigned int slot; |
|
|
index = ohash_qlookup(&filenames, name); |
assert(NULL != mlink->file); |
return(NULL != ohash_find(&filenames, index)); |
|
|
mlink->dsec = mandoc_strdup(mlink->dsec ? mlink->dsec : ""); |
|
mlink->arch = mandoc_strdup(mlink->arch ? mlink->arch : ""); |
|
mlink->name = mandoc_strdup(mlink->name ? mlink->name : ""); |
|
mlink->fsec = mandoc_strdup(mlink->fsec ? mlink->fsec : ""); |
|
|
|
if ('0' == *mlink->fsec) { |
|
free(mlink->fsec); |
|
mlink->fsec = mandoc_strdup(mlink->dsec); |
|
mlink->fform = FORM_CAT; |
|
} else if ('1' <= *mlink->fsec && '9' >= *mlink->fsec) |
|
mlink->fform = FORM_SRC; |
|
else |
|
mlink->fform = FORM_NONE; |
|
|
|
slot = ohash_qlookup(&mlinks, mlink->file); |
|
assert(NULL == ohash_find(&mlinks, slot)); |
|
ohash_insert(&mlinks, slot, mlink); |
|
|
|
memset(&inodev, 0, sizeof(inodev)); /* Clear padding. */ |
|
inodev.st_ino = st->st_ino; |
|
inodev.st_dev = st->st_dev; |
|
slot = ohash_lookup_memory(&mpages, (char *)&inodev, |
|
sizeof(struct inodev), inodev.st_ino); |
|
mpage = ohash_find(&mpages, slot); |
|
if (NULL == mpage) { |
|
mpage = mandoc_calloc(1, sizeof(struct mpage)); |
|
mpage->inodev.st_ino = inodev.st_ino; |
|
mpage->inodev.st_dev = inodev.st_dev; |
|
mpage->form = FORM_NONE; |
|
mpage->next = mpage_head; |
|
mpage_head = mpage; |
|
ohash_insert(&mpages, slot, mpage); |
|
} else |
|
mlink->next = mpage->mlinks; |
|
mpage->mlinks = mlink; |
|
mlink->mpage = mpage; |
} |
} |
|
|
/* |
|
* Use the standard hashing mechanism (K&R) to see if the given filename |
|
* already exists. |
|
*/ |
|
static void |
static void |
fileadd(struct of *of) |
mlink_free(struct mlink *mlink) |
{ |
{ |
unsigned int index; |
|
|
|
index = ohash_qlookup(&filenames, of->file); |
free(mlink->dsec); |
assert(NULL == ohash_find(&filenames, index)); |
free(mlink->arch); |
ohash_insert(&filenames, index, of); |
free(mlink->name); |
|
free(mlink->fsec); |
|
free(mlink); |
} |
} |
|
|
/* |
static void |
* See inoadd(). |
mpages_free(void) |
*/ |
|
static int |
|
inocheck(const struct stat *st) |
|
{ |
{ |
struct id id; |
struct mpage *mpage; |
uint32_t hash; |
struct mlink *mlink; |
unsigned int index; |
|
|
|
memset(&id, 0, sizeof(id)); |
while ((mpage = mpage_head) != NULL) { |
id.ino = hash = st->st_ino; |
while ((mlink = mpage->mlinks) != NULL) { |
id.dev = st->st_dev; |
mpage->mlinks = mlink->next; |
index = ohash_lookup_memory |
mlink_free(mlink); |
(&inos, (char *)&id, sizeof(id), hash); |
} |
|
mpage_head = mpage->next; |
return(NULL != ohash_find(&inos, index)); |
free(mpage->sec); |
|
free(mpage->arch); |
|
free(mpage->title); |
|
free(mpage->desc); |
|
free(mpage); |
|
} |
} |
} |
|
|
/* |
/* |
* The hashing function used here is quite simple: simply take the inode |
* For each mlink to the mpage, check whether the path looks like |
* and use uint32_t of its bits. |
* it is formatted, and if it does, check whether a source manual |
* Then when we do the lookup, use both the inode and device identifier. |
* exists by the same name, ignoring the suffix. |
|
* If both conditions hold, drop the mlink. |
*/ |
*/ |
static void |
static void |
inoadd(const struct stat *st, struct of *of) |
mlinks_undupe(struct mpage *mpage) |
{ |
{ |
uint32_t hash; |
char buf[PATH_MAX]; |
unsigned int index; |
struct mlink **prev; |
|
struct mlink *mlink; |
|
char *bufp; |
|
|
of->id.ino = hash = st->st_ino; |
mpage->form = FORM_CAT; |
of->id.dev = st->st_dev; |
prev = &mpage->mlinks; |
index = ohash_lookup_memory |
while (NULL != (mlink = *prev)) { |
(&inos, (char *)&of->id, sizeof(of->id), hash); |
if (FORM_CAT != mlink->dform) { |
|
mpage->form = FORM_NONE; |
assert(NULL == ohash_find(&inos, index)); |
goto nextlink; |
ohash_insert(&inos, index, of); |
} |
|
(void)strlcpy(buf, mlink->file, sizeof(buf)); |
|
bufp = strstr(buf, "cat"); |
|
assert(NULL != bufp); |
|
memcpy(bufp, "man", 3); |
|
if (NULL != (bufp = strrchr(buf, '.'))) |
|
*++bufp = '\0'; |
|
(void)strlcat(buf, mlink->dsec, sizeof(buf)); |
|
if (NULL == ohash_find(&mlinks, |
|
ohash_qlookup(&mlinks, buf))) |
|
goto nextlink; |
|
if (warnings) |
|
say(mlink->file, "Man source exists: %s", buf); |
|
if (use_all) |
|
goto nextlink; |
|
*prev = mlink->next; |
|
mlink_free(mlink); |
|
continue; |
|
nextlink: |
|
prev = &(*prev)->next; |
|
} |
} |
} |
|
|
static void |
static void |
ofadd(const char *base, int dform, const char *file, |
mlink_check(struct mpage *mpage, struct mlink *mlink) |
const char *name, const char *dsec, const char *sec, |
|
const char *arch, const struct stat *st) |
|
{ |
{ |
struct of *of; |
struct str *str; |
int sform; |
unsigned int slot; |
|
|
assert(NULL != file); |
/* |
|
* Check whether the manual section given in a file |
|
* agrees with the directory where the file is located. |
|
* Some manuals have suffixes like (3p) on their |
|
* section number either inside the file or in the |
|
* directory name, some are linked into more than one |
|
* section, like encrypt(1) = makekey(8). |
|
*/ |
|
|
if (NULL == name) |
if (FORM_SRC == mpage->form && |
name = ""; |
strcasecmp(mpage->sec, mlink->dsec)) |
if (NULL == sec) |
say(mlink->file, "Section \"%s\" manual in %s directory", |
sec = ""; |
mpage->sec, mlink->dsec); |
if (NULL == dsec) |
|
dsec = ""; |
|
if (NULL == arch) |
|
arch = ""; |
|
|
|
sform = FORM_NONE; |
/* |
if (NULL != sec && *sec <= '9' && *sec >= '1') |
* Manual page directories exist for each kernel |
sform = FORM_SRC; |
* architecture as returned by machine(1). |
else if (NULL != sec && *sec == '0') { |
* However, many manuals only depend on the |
sec = dsec; |
* application architecture as returned by arch(1). |
sform = FORM_CAT; |
* For example, some (2/ARM) manuals are shared |
} |
* across the "armish" and "zaurus" kernel |
|
* architectures. |
|
* A few manuals are even shared across completely |
|
* different architectures, for example fdformat(1) |
|
* on amd64, i386, and sparc64. |
|
*/ |
|
|
of = mandoc_calloc(1, sizeof(struct of)); |
if (strcasecmp(mpage->arch, mlink->arch)) |
strlcpy(of->file, file, PATH_MAX); |
say(mlink->file, "Architecture \"%s\" manual in " |
of->name = name; |
"\"%s\" directory", mpage->arch, mlink->arch); |
of->sec = sec; |
|
of->dsec = dsec; |
|
of->arch = arch; |
|
of->sform = sform; |
|
of->dform = dform; |
|
of->next = ofs; |
|
ofs = of; |
|
|
|
/* |
/* |
* Add to unique identifier hash. |
* XXX |
* Then if it's a source manual and we're going to use source in |
* parse_cat() doesn't set NAME_TITLE yet. |
* favour of catpages, add it to that hash. |
|
*/ |
*/ |
inoadd(st, of); |
|
fileadd(of); |
|
} |
|
|
|
static void |
if (FORM_CAT == mpage->form) |
offree(void) |
return; |
{ |
|
struct of *of; |
|
|
|
while (NULL != (of = ofs)) { |
/* |
ofs = of->next; |
* Check whether this mlink |
free(of); |
* appears as a name in the NAME section. |
} |
*/ |
|
|
|
slot = ohash_qlookup(&names, mlink->name); |
|
str = ohash_find(&names, slot); |
|
assert(NULL != str); |
|
if ( ! (NAME_TITLE & str->mask)) |
|
say(mlink->file, "Name missing in NAME section"); |
} |
} |
|
|
/* |
/* |
* Run through the files in the global vector "ofs" and add them to the |
* Run through the files in the global vector "mpages" |
* database specified in "base". |
* and add them to the database specified in "basedir". |
* |
* |
* This handles the parsing scheme itself, using the cues of directory |
* This handles the parsing scheme itself, using the cues of directory |
* and filename to determine whether the file is parsable or not. |
* and filename to determine whether the file is parsable or not. |
*/ |
*/ |
static int |
static void |
ofmerge(struct mchars *mc, struct mparse *mp, const char *base) |
mpages_merge(struct dba *dba, struct mparse *mp) |
{ |
{ |
int form; |
struct mpage *mpage, *mpage_dest; |
size_t sz; |
struct mlink *mlink, *mlink_dest; |
struct mdoc *mdoc; |
struct roff_man *man; |
struct man *man; |
char *sodest; |
char buf[PATH_MAX]; |
char *cp; |
char *bufp; |
int fd; |
const char *msec, *march, *mtitle, *cp; |
|
struct of *of; |
|
enum mandoclevel lvl; |
|
|
|
for (of = ofs; NULL != of; of = of->next) { |
for (mpage = mpage_head; mpage != NULL; mpage = mpage->next) { |
/* |
mlinks_undupe(mpage); |
* If we're a catpage (as defined by our path), then see |
if ((mlink = mpage->mlinks) == NULL) |
* if a manpage exists by the same name (ignoring the |
continue; |
* suffix). |
|
* If it does, then we want to use it instead of our |
|
* own. |
|
*/ |
|
if ( ! use_all && FORM_CAT == of->dform) { |
|
sz = strlcpy(buf, of->file, PATH_MAX); |
|
if (sz >= PATH_MAX) { |
|
if (warnings) |
|
say(base, of->file, |
|
"Filename too long"); |
|
continue; |
|
} |
|
bufp = strstr(buf, "cat"); |
|
assert(NULL != bufp); |
|
memcpy(bufp, "man", 3); |
|
if (NULL != (bufp = strrchr(buf, '.'))) |
|
*++bufp = '\0'; |
|
strlcat(buf, of->dsec, PATH_MAX); |
|
if (filecheck(buf)) { |
|
if (warnings) |
|
say(base, of->file, "Man " |
|
"source exists: %s", buf); |
|
continue; |
|
} |
|
} |
|
|
|
words = NULL; |
name_mask = NAME_MASK; |
|
mandoc_ohash_init(&names, 4, offsetof(struct str, key)); |
|
mandoc_ohash_init(&strings, 6, offsetof(struct str, key)); |
mparse_reset(mp); |
mparse_reset(mp); |
mdoc = NULL; |
|
man = NULL; |
man = NULL; |
form = 0; |
sodest = NULL; |
msec = of->dsec; |
|
march = of->arch; |
|
mtitle = of->name; |
|
|
|
|
if ((fd = mparse_open(mp, mlink->file)) == -1) { |
|
say(mlink->file, "&open"); |
|
goto nextpage; |
|
} |
|
|
/* |
/* |
* 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_SRC == of->dform || FORM_SRC == of->sform) { |
if (mlink->dform != FORM_CAT || mlink->fform != FORM_CAT) { |
lvl = mparse_readfd(mp, -1, of->file); |
mparse_readfd(mp, fd, mlink->file); |
if (lvl < MANDOCLEVEL_FATAL) |
close(fd); |
mparse_result(mp, &mdoc, &man); |
mparse_result(mp, &man, &sodest); |
} |
} |
|
|
if (NULL != mdoc) { |
if (sodest != NULL) { |
form = 1; |
mlink_dest = ohash_find(&mlinks, |
msec = mdoc_meta(mdoc)->msec; |
ohash_qlookup(&mlinks, sodest)); |
march = mdoc_meta(mdoc)->arch; |
if (mlink_dest == NULL) { |
mtitle = mdoc_meta(mdoc)->title; |
mandoc_asprintf(&cp, "%s.gz", sodest); |
} else if (NULL != man) { |
mlink_dest = ohash_find(&mlinks, |
form = 1; |
ohash_qlookup(&mlinks, cp)); |
msec = man_meta(man)->msec; |
free(cp); |
march = ""; |
} |
mtitle = man_meta(man)->title; |
if (mlink_dest != NULL) { |
} |
|
|
|
if (NULL == msec) |
/* The .so target exists. */ |
msec = ""; |
|
if (NULL == march) |
|
march = ""; |
|
if (NULL == mtitle) |
|
mtitle = ""; |
|
|
|
/* |
mpage_dest = mlink_dest->mpage; |
* Check whether the manual section given in a file |
while (1) { |
* agrees with the directory where the file is located. |
mlink->mpage = mpage_dest; |
* Some manuals have suffixes like (3p) on their |
|
* section number either inside the file or in the |
|
* directory name, some are linked into more than one |
|
* section, like encrypt(1) = makekey(8). Do not skip |
|
* manuals for such reasons. |
|
*/ |
|
if (warnings && !use_all && form && |
|
strcasecmp(msec, of->dsec)) |
|
say(base, of->file, "Section \"%s\" " |
|
"manual in %s directory", |
|
msec, of->dsec); |
|
|
|
/* |
/* |
* Manual page directories exist for each kernel |
* If the target was already |
* architecture as returned by machine(1). |
* processed, add the links |
* However, many manuals only depend on the |
* to the database now. |
* application architecture as returned by arch(1). |
* Otherwise, this will |
* For example, some (2/ARM) manuals are shared |
* happen when we come |
* across the "armish" and "zaurus" kernel |
* to the target. |
* architectures. |
*/ |
* A few manuals are even shared across completely |
|
* different architectures, for example fdformat(1) |
|
* on amd64, i386, sparc, and sparc64. |
|
* Thus, warn about architecture mismatches, |
|
* but don't skip manuals for this reason. |
|
*/ |
|
if (warnings && !use_all && strcasecmp(march, of->arch)) |
|
say(base, of->file, "Architecture \"%s\" " |
|
"manual in \"%s\" directory", |
|
march, of->arch); |
|
|
|
putkey(of, of->name, TYPE_Nm); |
if (mpage_dest->dba != NULL) |
|
dbadd_mlink(mlink); |
|
|
if (NULL != mdoc) { |
if (mlink->next == NULL) |
if (NULL != (cp = mdoc_meta(mdoc)->name)) |
break; |
putkey(of, cp, TYPE_Nm); |
mlink = mlink->next; |
parse_mdoc(of, mdoc_node(mdoc)); |
} |
} else if (NULL != man) |
|
parse_man(of, man_node(man)); |
/* Move all links to the target. */ |
|
|
|
mlink->next = mlink_dest->next; |
|
mlink_dest->next = mpage->mlinks; |
|
mpage->mlinks = NULL; |
|
} |
|
goto nextpage; |
|
} else if (man != NULL && man->macroset == MACROSET_MDOC) { |
|
mdoc_validate(man); |
|
mpage->form = FORM_SRC; |
|
mpage->sec = man->meta.msec; |
|
mpage->sec = mandoc_strdup( |
|
mpage->sec == NULL ? "" : mpage->sec); |
|
mpage->arch = man->meta.arch; |
|
mpage->arch = mandoc_strdup( |
|
mpage->arch == NULL ? "" : mpage->arch); |
|
mpage->title = mandoc_strdup(man->meta.title); |
|
} else if (man != NULL && man->macroset == MACROSET_MAN) { |
|
man_validate(man); |
|
mpage->form = FORM_SRC; |
|
mpage->sec = mandoc_strdup(man->meta.msec); |
|
mpage->arch = mandoc_strdup(mlink->arch); |
|
mpage->title = mandoc_strdup(man->meta.title); |
|
} else { |
|
mpage->form = FORM_CAT; |
|
mpage->sec = mandoc_strdup(mlink->dsec); |
|
mpage->arch = mandoc_strdup(mlink->arch); |
|
mpage->title = mandoc_strdup(mlink->name); |
|
} |
|
|
|
assert(mpage->desc == NULL); |
|
if (man != NULL && man->macroset == MACROSET_MDOC) |
|
parse_mdoc(mpage, &man->meta, man->first); |
|
else if (man != NULL) |
|
parse_man(mpage, &man->meta, man->first); |
else |
else |
parse_catpage(of, base); |
parse_cat(mpage, fd); |
|
if (mpage->desc == NULL) |
|
mpage->desc = mandoc_strdup(mpage->mlinks->name); |
|
|
dbindex(mc, form, of, base); |
if (warnings && !use_all) |
} |
for (mlink = mpage->mlinks; mlink; |
|
mlink = mlink->next) |
|
mlink_check(mpage, mlink); |
|
|
return(1); |
dbadd(dba, mpage); |
|
mlink = mpage->mlinks; |
|
|
|
nextpage: |
|
ohash_delete(&strings); |
|
ohash_delete(&names); |
|
} |
} |
} |
|
|
static void |
static void |
parse_catpage(struct of *of, const char *base) |
parse_cat(struct mpage *mpage, int fd) |
{ |
{ |
FILE *stream; |
FILE *stream; |
char *line, *p, *title; |
char *line, *p, *title; |
size_t len, plen, titlesz; |
size_t linesz, plen, titlesz; |
|
ssize_t len; |
|
int offs; |
|
|
if (NULL == (stream = fopen(of->file, "r"))) { |
stream = (-1 == fd) ? |
|
fopen(mpage->mlinks->file, "r") : |
|
fdopen(fd, "r"); |
|
if (NULL == stream) { |
|
if (-1 != fd) |
|
close(fd); |
if (warnings) |
if (warnings) |
say(base, of->file, "%s", strerror(errno)); |
say(mpage->mlinks->file, "&fopen"); |
return; |
return; |
} |
} |
|
|
|
line = NULL; |
|
linesz = 0; |
|
|
/* Skip to first blank line. */ |
/* Skip to first blank line. */ |
|
|
while (NULL != (line = fgetln(stream, &len))) |
while (getline(&line, &linesz, stream) != -1) |
if ('\n' == *line) |
if (*line == '\n') |
break; |
break; |
|
|
/* |
/* |
Line 1070 parse_catpage(struct of *of, const char *base) |
|
Line 1264 parse_catpage(struct of *of, const char *base) |
|
* 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; |
|
|
/* |
/* |
* Read up until the next section into a buffer. |
* Read up until the next section into a buffer. |
* Strip the leading and trailing newline from each read line, |
* Strip the leading and trailing newline from each read line, |
Line 1084 parse_catpage(struct of *of, const char *base) |
|
Line 1278 parse_catpage(struct of *of, const char *base) |
|
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 1108 parse_catpage(struct of *of, const char *base) |
|
Line 1302 parse_catpage(struct of *of, const char *base) |
|
|
|
if (NULL == title || '\0' == *title) { |
if (NULL == title || '\0' == *title) { |
if (warnings) |
if (warnings) |
say(base, of->file, "Cannot find NAME section"); |
say(mpage->mlinks->file, |
|
"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 1128 parse_catpage(struct of *of, const char *base) |
|
Line 1322 parse_catpage(struct of *of, const char *base) |
|
/* Skip to next word. */ ; |
/* Skip to next word. */ ; |
} else { |
} else { |
if (warnings) |
if (warnings) |
say(base, of->file, "No dash in title line"); |
say(mpage->mlinks->file, |
|
"No dash in title line"); |
p = title; |
p = title; |
} |
} |
|
|
Line 1141 parse_catpage(struct of *of, const char *base) |
|
Line 1336 parse_catpage(struct of *of, const char *base) |
|
if (0 == len) { |
if (0 == len) { |
memmove(line, line + 1, plen--); |
memmove(line, line + 1, plen--); |
continue; |
continue; |
} |
} |
memmove(line - 1, line + 1, plen - len); |
memmove(line - 1, line + 1, plen - len); |
plen -= 2; |
plen -= 2; |
} |
} |
|
|
of->desc = stradd(p); |
mpage->desc = mandoc_strdup(p); |
putkey(of, p, TYPE_Nd); |
|
fclose(stream); |
fclose(stream); |
free(title); |
free(title); |
} |
} |
Line 1156 parse_catpage(struct of *of, const char *base) |
|
Line 1350 parse_catpage(struct of *of, const char *base) |
|
* Put a type/word pair into the word database for this particular file. |
* Put a type/word pair into the word database for this particular file. |
*/ |
*/ |
static void |
static void |
putkey(const struct of *of, const char *value, uint64_t type) |
putkey(const struct mpage *mpage, char *value, uint64_t type) |
{ |
{ |
|
putkeys(mpage, value, strlen(value), type); |
assert(NULL != value); |
|
wordaddbuf(of, value, strlen(value), type); |
|
} |
} |
|
|
/* |
/* |
* Like putkey() but for unterminated strings. |
|
*/ |
|
static void |
|
putkeys(const struct of *of, const char *value, int sz, uint64_t type) |
|
{ |
|
|
|
wordaddbuf(of, value, sz, type); |
|
} |
|
|
|
/* |
|
* Grok all nodes at or below a certain mdoc node into putkey(). |
* Grok all nodes at or below a certain mdoc node into putkey(). |
*/ |
*/ |
static void |
static void |
putmdockey(const struct of *of, const struct mdoc_node *n, uint64_t m) |
putmdockey(const struct mpage *mpage, |
|
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(of, n->child, m); |
putmdockey(mpage, n->child, m, taboo); |
if (MDOC_TEXT == n->type) |
if (n->type == ROFFT_TEXT) |
putkey(of, n->string, m); |
putkey(mpage, n->string, m); |
} |
} |
} |
} |
|
|
static int |
static void |
parse_man(struct of *of, 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, *sv, *title; |
char *start, *title; |
char byte; |
char byte; |
size_t sz, titlesz; |
size_t sz; |
|
|
if (NULL == n) |
if (n == NULL) |
return(0); |
return; |
|
|
/* |
/* |
* We're only searching for one thing: the first text child in |
* We're only searching for one thing: the first text child in |
Line 1206 parse_man(struct of *of, const struct man_node *n) |
|
Line 1392 parse_man(struct of *of, const struct man_node *n) |
|
* 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 = body->child) && |
|
MAN_TEXT == body->type) { |
|
|
|
title = NULL; |
|
titlesz = 0; |
|
|
|
/* |
/* |
* Suck the entire NAME section into memory. |
* Suck the entire NAME section into memory. |
* Yes, we might run away. |
* Yes, we might run away. |
Line 1227 parse_man(struct of *of, const struct man_node *n) |
|
Line 1408 parse_man(struct of *of, const struct man_node *n) |
|
* NAME sections over many lines. |
* NAME sections over many lines. |
*/ |
*/ |
|
|
for ( ; NULL != body; body = body->next) { |
title = NULL; |
if (MAN_TEXT != body->type) |
deroff(&title, body); |
break; |
|
if (0 == (sz = strlen(body->string))) |
|
continue; |
|
title = mandoc_realloc |
|
(title, titlesz + sz + 1); |
|
memcpy(title + titlesz, body->string, sz); |
|
titlesz += sz + 1; |
|
title[titlesz - 1] = ' '; |
|
} |
|
if (NULL == title) |
if (NULL == title) |
return(1); |
return; |
|
|
title = mandoc_realloc(title, titlesz + 1); |
/* |
title[titlesz] = '\0'; |
|
|
|
/* Skip leading space. */ |
|
|
|
sv = title; |
|
while (isspace((unsigned char)*sv)) |
|
sv++; |
|
|
|
if (0 == (sz = strlen(sv))) { |
|
free(title); |
|
return(1); |
|
} |
|
|
|
/* Erase trailing space. */ |
|
|
|
start = &sv[sz - 1]; |
|
while (start > sv && isspace((unsigned char)*start)) |
|
*start-- = '\0'; |
|
|
|
if (start == sv) { |
|
free(title); |
|
return(1); |
|
} |
|
|
|
start = sv; |
|
|
|
/* |
|
* Go through a special heuristic dance here. |
* Go through a special heuristic dance here. |
* Conventionally, one or more manual names are |
* Conventionally, one or more manual names are |
* comma-specified prior to a whitespace, then a |
* comma-specified prior to a whitespace, then a |
Line 1276 parse_man(struct of *of, const struct man_node *n) |
|
Line 1421 parse_man(struct of *of, const struct man_node *n) |
|
* the name parts here. |
* the name parts here. |
*/ |
*/ |
|
|
|
start = title; |
for ( ;; ) { |
for ( ;; ) { |
sz = strcspn(start, " ,"); |
sz = strcspn(start, " ,"); |
if ('\0' == start[sz]) |
if ('\0' == start[sz]) |
Line 1284 parse_man(struct of *of, const struct man_node *n) |
|
Line 1430 parse_man(struct of *of, const struct man_node *n) |
|
byte = start[sz]; |
byte = start[sz]; |
start[sz] = '\0'; |
start[sz] = '\0'; |
|
|
putkey(of, start, TYPE_Nm); |
/* |
|
* Assume a stray trailing comma in the |
|
* name list if a name begins with a dash. |
|
*/ |
|
|
|
if ('-' == start[0] || |
|
('\\' == start[0] && '-' == start[1])) |
|
break; |
|
|
|
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; |
break; |
break; |
Line 1297 parse_man(struct of *of, const struct man_node *n) |
|
Line 1457 parse_man(struct of *of, const struct man_node *n) |
|
start++; |
start++; |
} |
} |
|
|
if (sv == start) { |
if (start == title) { |
putkey(of, start, TYPE_Nm); |
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(1); |
return; |
} |
} |
|
|
while (isspace((unsigned char)*start)) |
while (isspace((unsigned char)*start)) |
Line 1320 parse_man(struct of *of, const struct man_node *n) |
|
Line 1485 parse_man(struct of *of, const struct man_node *n) |
|
while (' ' == *start) |
while (' ' == *start) |
start++; |
start++; |
|
|
assert(NULL == of->desc); |
mpage->desc = mandoc_strdup(start); |
of->desc = stradd(start); |
|
putkey(of, start, TYPE_Nd); |
|
free(title); |
free(title); |
return(1); |
return; |
} |
} |
} |
} |
|
|
for (n = n->child; n; n = n->next) |
for (n = n->child; n; n = n->next) { |
if (parse_man(of, n)) |
if (NULL != mpage->desc) |
return(1); |
break; |
|
parse_man(mpage, meta, n); |
return(0); |
} |
} |
} |
|
|
static void |
static void |
parse_mdoc(struct of *of, const struct mdoc_node *n) |
parse_mdoc(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
assert(NULL != n); |
assert(NULL != n); |
for (n = n->child; NULL != n; n = n->next) { |
for (n = n->child; NULL != n; n = n->next) { |
|
if (n->flags & mdocs[n->tok].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 */ |
|
case (MDOC_BODY): |
|
/* FALLTHROUGH */ |
|
case (MDOC_TAIL): |
|
if (NULL != mdocs[n->tok].fp) |
if (NULL != mdocs[n->tok].fp) |
if (0 == (*mdocs[n->tok].fp)(of, n)) |
if (0 == (*mdocs[n->tok].fp)(mpage, meta, n)) |
break; |
break; |
|
if (mdocs[n->tok].mask) |
if (MDOCF_CHILD & mdocs[n->tok].flags) |
putmdockey(mpage, n->child, |
putmdockey(of, n->child, mdocs[n->tok].mask); |
mdocs[n->tok].mask, mdocs[n->tok].taboo); |
break; |
break; |
default: |
default: |
assert(MDOC_ROOT != n->type); |
assert(n->type != ROFFT_ROOT); |
continue; |
continue; |
} |
} |
if (NULL != n->child) |
if (NULL != n->child) |
parse_mdoc(of, n); |
parse_mdoc(mpage, meta, n); |
} |
} |
} |
} |
|
|
static int |
static int |
parse_mdoc_Fd(struct of *of, const struct mdoc_node *n) |
parse_mdoc_Fd(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
const char *start, *end; |
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 1384 parse_mdoc_Fd(struct of *of, const struct mdoc_node *n |
|
Line 1547 parse_mdoc_Fd(struct of *of, const struct mdoc_node *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 1399 parse_mdoc_Fd(struct of *of, const struct mdoc_node *n |
|
Line 1562 parse_mdoc_Fd(struct of *of, const struct mdoc_node *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) |
end--; |
end--; |
|
|
if (end > start) |
if (end > start) |
putkeys(of, start, end - start + 1, TYPE_In); |
putkeys(mpage, start, end - start + 1, TYPE_In); |
return(1); |
return 0; |
} |
} |
|
|
static int |
static void |
parse_mdoc_In(struct of *of, const struct mdoc_node *n) |
parse_mdoc_fname(struct mpage *mpage, const struct roff_node *n) |
{ |
{ |
|
char *cp; |
|
size_t sz; |
|
|
if (NULL != n->child && MDOC_TEXT == n->child->type) |
if (n->type != ROFFT_TEXT) |
return(0); |
return; |
|
|
putkey(of, n->child->string, TYPE_In); |
/* Skip function pointer punctuation. */ |
return(1); |
|
} |
|
|
|
static int |
cp = n->string; |
parse_mdoc_Fn(struct of *of, const struct mdoc_node *n) |
while (*cp == '(' || *cp == '*') |
{ |
|
const char *cp; |
|
|
|
if (NULL == (n = n->child) || MDOC_TEXT != n->type) |
|
return(0); |
|
|
|
/* |
|
* 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; |
|
|
|
while ('*' == *cp) |
|
cp++; |
cp++; |
|
sz = strcspn(cp, "()"); |
|
|
putkey(of, cp, TYPE_Fn); |
putkeys(mpage, cp, sz, TYPE_Fn); |
|
if (n->sec == SEC_SYNOPSIS) |
if (n->string < cp) |
putkeys(mpage, cp, sz, NAME_SYN); |
putkeys(of, n->string, cp - n->string, TYPE_Ft); |
|
|
|
for (n = n->next; NULL != n; n = n->next) |
|
if (MDOC_TEXT == n->type) |
|
putkey(of, n->string, TYPE_Fa); |
|
|
|
return(0); |
|
} |
} |
|
|
static int |
static int |
parse_mdoc_St(struct of *of, const struct mdoc_node *n) |
parse_mdoc_Fn(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
if (NULL == n->child || MDOC_TEXT != n->child->type) |
if (n->child == NULL) |
return(0); |
return 0; |
|
|
putkey(of, n->child->string, TYPE_St); |
parse_mdoc_fname(mpage, n->child); |
return(1); |
|
|
for (n = n->child->next; n != NULL; n = n->next) |
|
if (n->type == ROFFT_TEXT) |
|
putkey(mpage, n->string, TYPE_Fa); |
|
|
|
return 0; |
} |
} |
|
|
static int |
static int |
parse_mdoc_Xr(struct of *of, const struct mdoc_node *n) |
parse_mdoc_Fo(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
if (NULL == (n = n->child)) |
if (n->type != ROFFT_HEAD) |
return(0); |
return 1; |
|
|
putkey(of, n->string, TYPE_Xr); |
if (n->child != NULL) |
return(1); |
parse_mdoc_fname(mpage, n->child); |
|
|
|
return 0; |
} |
} |
|
|
static int |
static int |
parse_mdoc_Nd(struct of *of, const struct mdoc_node *n) |
parse_mdoc_Va(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
size_t sz; |
char *cp; |
char *sv, *desc; |
|
|
|
if (MDOC_BODY != n->type) |
if (n->type != ROFFT_ELEM && n->type != ROFFT_BODY) |
return(0); |
return 0; |
|
|
/* |
if (n->child != NULL && |
* Special-case the `Nd' because we need to put the description |
n->child->next == NULL && |
* into the document table. |
n->child->type == ROFFT_TEXT) |
*/ |
return 1; |
|
|
desc = NULL; |
cp = NULL; |
for (n = n->child; NULL != n; n = n->next) { |
deroff(&cp, n); |
if (MDOC_TEXT == n->type) { |
if (cp != NULL) { |
sz = strlen(n->string) + 1; |
putkey(mpage, cp, TYPE_Vt | (n->tok == MDOC_Va || |
if (NULL != (sv = desc)) |
n->type == ROFFT_BODY ? TYPE_Va : 0)); |
sz += strlen(desc) + 1; |
free(cp); |
desc = mandoc_realloc(desc, sz); |
|
if (NULL != sv) |
|
strlcat(desc, " ", sz); |
|
else |
|
*desc = '\0'; |
|
strlcat(desc, n->string, sz); |
|
} |
|
if (NULL != n->child) |
|
parse_mdoc_Nd(of, n); |
|
} |
} |
|
|
of->desc = NULL != desc ? stradd(desc) : NULL; |
return 0; |
free(desc); |
|
return(1); |
|
} |
} |
|
|
static int |
static int |
parse_mdoc_Nm(struct of *of, const struct mdoc_node *n) |
parse_mdoc_Xr(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
char *cp; |
|
|
if (SEC_NAME == n->sec) |
if (NULL == (n = n->child)) |
return(1); |
return 0; |
else if (SEC_SYNOPSIS != n->sec || MDOC_HEAD != n->type) |
|
return(0); |
|
|
|
return(1); |
if (NULL == n->next) { |
|
putkey(mpage, n->string, TYPE_Xr); |
|
return 0; |
|
} |
|
|
|
mandoc_asprintf(&cp, "%s(%s)", n->string, n->next->string); |
|
putkey(mpage, cp, TYPE_Xr); |
|
free(cp); |
|
return 0; |
} |
} |
|
|
static int |
static int |
parse_mdoc_Sh(struct of *of, const struct mdoc_node *n) |
parse_mdoc_Nd(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
return(SEC_CUSTOM == n->sec && MDOC_HEAD == n->type); |
if (n->type == ROFFT_BODY) |
|
deroff(&mpage->desc, n); |
|
return 0; |
} |
} |
|
|
static int |
static int |
parse_mdoc_head(struct of *of, const struct mdoc_node *n) |
parse_mdoc_Nm(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
return(MDOC_HEAD == n->type); |
if (SEC_NAME == n->sec) |
|
putmdockey(mpage, n->child, NAME_TITLE, 0); |
|
else if (n->sec == SEC_SYNOPSIS && n->type == ROFFT_HEAD) { |
|
if (n->child == NULL) |
|
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_body(struct of *of, const struct mdoc_node *n) |
parse_mdoc_Sh(struct mpage *mpage, const struct roff_meta *meta, |
|
const struct roff_node *n) |
{ |
{ |
|
|
return(MDOC_BODY == n->type); |
return n->sec == SEC_CUSTOM && n->type == ROFFT_HEAD; |
} |
} |
|
|
/* |
static int |
* See straddbuf(). |
parse_mdoc_head(struct mpage *mpage, const struct roff_meta *meta, |
*/ |
const struct roff_node *n) |
static char * |
|
stradd(const char *cp) |
|
{ |
{ |
|
|
return(straddbuf(cp, strlen(cp))); |
return n->type == ROFFT_HEAD; |
} |
} |
|
|
/* |
/* |
* This looks up or adds a string to the string table. |
* Add a string to the hash table for the current manual. |
* The string table is a table of all strings encountered during parse |
* Each string has a bitmask telling which macros it belongs to. |
* or file scan. |
* When we finish the manual, we'll dump the table. |
* In using it, we avoid having thousands of (e.g.) "cat1" string |
|
* allocations for the "of" table. |
|
* We also have a layer atop the string table for keeping track of words |
|
* in a parse sequence (see wordaddbuf()). |
|
*/ |
*/ |
static char * |
|
straddbuf(const char *cp, size_t sz) |
|
{ |
|
struct str *s; |
|
unsigned int index; |
|
const char *end; |
|
|
|
if (NULL != (s = hashget(cp, sz))) |
|
return(s->key); |
|
|
|
s = mandoc_calloc(sizeof(struct str) + sz + 1, 1); |
|
memcpy(s->key, cp, sz); |
|
|
|
end = cp + sz; |
|
index = ohash_qlookupi(&strings, cp, &end); |
|
assert(NULL == ohash_find(&strings, index)); |
|
ohash_insert(&strings, index, s); |
|
return(s->key); |
|
} |
|
|
|
static struct str * |
|
hashget(const char *cp, size_t sz) |
|
{ |
|
unsigned int index; |
|
const char *end; |
|
|
|
end = cp + sz; |
|
index = ohash_qlookupi(&strings, cp, &end); |
|
return(ohash_find(&strings, index)); |
|
} |
|
|
|
/* |
|
* Add a word to the current parse sequence. |
|
* Within the hashtable of strings, we maintain a list of strings that |
|
* are currently indexed. |
|
* Each of these ("words") has a bitmask modified within the parse. |
|
* When we finish a parse, we'll dump the list, then remove the head |
|
* entry -- since the next parse will have a new "of", it can keep track |
|
* of its entries without conflict. |
|
*/ |
|
static void |
static void |
wordaddbuf(const struct of *of, |
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 str *s; |
struct str *s; |
unsigned int index; |
|
const char *end; |
const char *end; |
|
unsigned int slot; |
|
int i, mustfree; |
|
|
if (0 == sz) |
if (0 == sz) |
return; |
return; |
|
|
s = hashget(cp, sz); |
mustfree = render_string(&cp, &sz); |
|
|
if (NULL != s && of == s->of) { |
if (TYPE_Nm & v) { |
|
htab = &names; |
|
v &= name_mask; |
|
if (v & NAME_FIRST) |
|
name_mask &= ~NAME_FIRST; |
|
if (debug > 1) |
|
say(mpage->mlinks->file, |
|
"Adding name %*s, bits=0x%llx", (int)sz, cp, |
|
(unsigned long long)v); |
|
} else { |
|
htab = &strings; |
|
if (debug > 1) |
|
for (i = 0; i < KEY_MAX; i++) |
|
if ((uint64_t)1 << i & v) |
|
say(mpage->mlinks->file, |
|
"Adding key %s=%*s", |
|
mansearch_keynames[i], (int)sz, cp); |
|
} |
|
|
|
end = cp + sz; |
|
slot = ohash_qlookupi(htab, cp, &end); |
|
s = ohash_find(htab, slot); |
|
|
|
if (NULL != s && mpage == s->mpage) { |
s->mask |= v; |
s->mask |= v; |
return; |
return; |
} else if (NULL == s) { |
} else if (NULL == s) { |
s = mandoc_calloc(sizeof(struct str) + sz + 1, 1); |
s = mandoc_calloc(1, sizeof(struct str) + sz + 1); |
memcpy(s->key, cp, sz); |
memcpy(s->key, cp, sz); |
end = cp + sz; |
ohash_insert(htab, slot, s); |
index = ohash_qlookupi(&strings, cp, &end); |
|
assert(NULL == ohash_find(&strings, index)); |
|
ohash_insert(&strings, index, s); |
|
} |
} |
|
s->mpage = mpage; |
s->next = words; |
|
s->of = of; |
|
s->mask = v; |
s->mask = v; |
words = s; |
|
|
if (mustfree) |
|
free(cp); |
} |
} |
|
|
/* |
/* |
Line 1681 utf8(unsigned int cp, char out[7]) |
|
Line 1821 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 UTF-8 version of a key, or alias the pointer if the key has |
* If the string contains escape sequences, |
* no UTF-8 transcription marks in it. |
* 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 |
utf8key(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 utfbuf[7], res[5]; |
char *dst; |
char *buf; |
size_t ssz, dsz, addsz; |
const char *seq, *cpp, *val; |
char utfbuf[7], res[6]; |
int len, u; |
int seqlen, unicode; |
enum mandoc_esc esc; |
|
|
|
assert(NULL == key->utf8); |
|
|
|
res[0] = '\\'; |
res[0] = '\\'; |
res[1] = '\t'; |
res[1] = '\t'; |
res[2] = ASCII_NBRSP; |
res[2] = ASCII_NBRSP; |
res[3] = ASCII_HYPH; |
res[3] = ASCII_HYPH; |
res[4] = '\0'; |
res[4] = ASCII_BREAK; |
|
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->utf8 = 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); |
} |
} |
|
|
if (ASCII_HYPH == *val) { |
/* Handle single-char special characters. */ |
buf[pos++] = '-'; |
|
val++; |
switch (*scp) { |
|
case '\\': |
|
break; |
|
case '\t': |
|
case ASCII_NBRSP: |
|
dst[dsz++] = ' '; |
|
scp++; |
continue; |
continue; |
} else if ('\t' == *val || ASCII_NBRSP == *val) { |
case ASCII_HYPH: |
buf[pos++] = ' '; |
dst[dsz++] = '-'; |
val++; |
/* FALLTHROUGH */ |
|
case ASCII_BREAK: |
|
scp++; |
continue; |
continue; |
} else if ('\\' != *val) |
default: |
break; |
abort(); |
|
} |
|
|
/* Read past the slash. */ |
|
|
|
val++; |
|
u = 0; |
|
|
|
/* |
/* |
* 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, &seq, &len); |
|
if (ESCAPE_ERROR == esc) |
|
break; |
|
|
|
if (ESCAPE_SPECIAL != esc) |
scp++; |
|
if (mandoc_escape(&scp, &seq, &seqlen) != ESCAPE_SPECIAL) |
continue; |
continue; |
if (0 == (u = mchars_spec2cp(mc, seq, len))) |
|
continue; |
|
|
|
/* |
/* |
* If we have a Unicode codepoint, try to convert that |
* Render the special character |
* to a UTF-8 byte string. |
* as either UTF-8 or ASCII. |
*/ |
*/ |
cpp = utfbuf; |
|
if (0 == (sz = utf8(u, utfbuf))) |
|
continue; |
|
|
|
|
if (write_utf8) { |
|
unicode = mchars_spec2cp(seq, seqlen); |
|
if (unicode <= 0) |
|
continue; |
|
addsz = utf8(unicode, utfbuf); |
|
if (addsz == 0) |
|
continue; |
|
addcp = utfbuf; |
|
} else { |
|
addcp = mchars_spec2str(seq, seqlen, &addsz); |
|
if (addcp == NULL) |
|
continue; |
|
if (*addcp == ASCII_NBRSP) { |
|
addcp = " "; |
|
addsz = 1; |
|
} |
|
} |
|
|
/* Copy the rendered glyph into the stream. */ |
/* Copy the rendered glyph into the stream. */ |
|
|
sz = strlen(cpp); |
ssz += addsz; |
bsz += sz; |
dst = mandoc_realloc(dst, ssz + 1); |
|
memcpy(dst + dsz, addcp, addsz); |
|
dsz += addsz; |
|
} |
|
if (dst != NULL) { |
|
*public = dst; |
|
*psz = dsz; |
|
} |
|
|
buf = mandoc_realloc(buf, bsz); |
/* Trim trailing whitespace and NUL-terminate. */ |
|
|
memcpy(&buf[pos], cpp, sz); |
while (*psz > 0 && (*public)[*psz - 1] == ' ') |
pos += sz; |
--*psz; |
} |
if (dst != NULL) { |
|
(*public)[*psz] = '\0'; |
|
return 1; |
|
} else |
|
return 0; |
|
} |
|
|
buf[pos] = '\0'; |
static void |
key->utf8 = buf; |
dbadd_mlink(const struct mlink *mlink) |
|
{ |
|
dba_page_alias(mlink->mpage->dba, mlink->name, NAME_FILE); |
|
dba_page_add(mlink->mpage->dba, DBP_SECT, mlink->dsec); |
|
dba_page_add(mlink->mpage->dba, DBP_SECT, mlink->fsec); |
|
dba_page_add(mlink->mpage->dba, DBP_ARCH, mlink->arch); |
|
dba_page_add(mlink->mpage->dba, DBP_FILE, mlink->file); |
} |
} |
|
|
/* |
/* |
* 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 |
* Also, handle escape sequences at the last possible moment. |
* little faster. |
|
* Also, UTF-8-encode the description at the last possible moment. |
|
*/ |
*/ |
static void |
static void |
dbindex(struct mchars *mc, int form, |
dbadd(struct dba *dba, struct mpage *mpage) |
const struct of *of, const char *base) |
|
{ |
{ |
|
struct mlink *mlink; |
struct str *key; |
struct str *key; |
const char *desc; |
char *cp; |
int64_t recno; |
uint64_t mask; |
size_t i; |
size_t i; |
|
unsigned int slot; |
|
int mustfree; |
|
|
if (verb) |
mlink = mpage->mlinks; |
say(base, of->file, "Adding to index"); |
|
|
|
if (nodb) |
if (nodb) { |
|
for (key = ohash_first(&names, &slot); NULL != key; |
|
key = ohash_next(&names, &slot)) |
|
free(key); |
|
for (key = ohash_first(&strings, &slot); NULL != key; |
|
key = ohash_next(&strings, &slot)) |
|
free(key); |
|
if (0 == debug) |
|
return; |
|
while (NULL != mlink) { |
|
fputs(mlink->name, stdout); |
|
if (NULL == mlink->next || |
|
strcmp(mlink->dsec, mlink->next->dsec) || |
|
strcmp(mlink->fsec, mlink->next->fsec) || |
|
strcmp(mlink->arch, mlink->next->arch)) { |
|
putchar('('); |
|
if ('\0' == *mlink->dsec) |
|
fputs(mlink->fsec, stdout); |
|
else |
|
fputs(mlink->dsec, stdout); |
|
if ('\0' != *mlink->arch) |
|
printf("/%s", mlink->arch); |
|
putchar(')'); |
|
} |
|
mlink = mlink->next; |
|
if (NULL != mlink) |
|
fputs(", ", stdout); |
|
} |
|
printf(" - %s\n", mpage->desc); |
return; |
return; |
|
|
desc = ""; |
|
if (NULL != of->desc) { |
|
key = hashget(of->desc, strlen(of->desc)); |
|
assert(NULL != key); |
|
if (NULL == key->utf8) |
|
utf8key(mc, key); |
|
desc = key->utf8; |
|
} |
} |
|
|
SQL_EXEC("BEGIN TRANSACTION"); |
if (debug) |
|
say(mlink->file, "Adding to database"); |
|
|
i = 1; |
cp = mpage->desc; |
SQL_BIND_TEXT(stmts[STMT_INSERT_DOC], i, of->file); |
i = strlen(cp); |
SQL_BIND_TEXT(stmts[STMT_INSERT_DOC], i, of->sec); |
mustfree = render_string(&cp, &i); |
SQL_BIND_TEXT(stmts[STMT_INSERT_DOC], i, of->arch); |
mpage->dba = dba_page_new(dba->pages, |
SQL_BIND_TEXT(stmts[STMT_INSERT_DOC], i, desc); |
*mpage->arch == '\0' ? mlink->arch : mpage->arch, |
SQL_BIND_INT(stmts[STMT_INSERT_DOC], i, form); |
cp, mlink->file, mpage->form); |
SQL_STEP(stmts[STMT_INSERT_DOC]); |
if (mustfree) |
recno = sqlite3_last_insert_rowid(db); |
free(cp); |
sqlite3_reset(stmts[STMT_INSERT_DOC]); |
dba_page_add(mpage->dba, DBP_SECT, mpage->sec); |
|
|
for (key = words; NULL != key; key = key->next) { |
while (mlink != NULL) { |
assert(key->of == of); |
dbadd_mlink(mlink); |
if (NULL == key->utf8) |
mlink = mlink->next; |
utf8key(mc, key); |
|
i = 1; |
|
SQL_BIND_INT64(stmts[STMT_INSERT_KEY], i, key->mask); |
|
SQL_BIND_TEXT(stmts[STMT_INSERT_KEY], i, key->utf8); |
|
SQL_BIND_INT64(stmts[STMT_INSERT_KEY], i, recno); |
|
SQL_STEP(stmts[STMT_INSERT_KEY]); |
|
sqlite3_reset(stmts[STMT_INSERT_KEY]); |
|
} |
} |
|
|
SQL_EXEC("END TRANSACTION"); |
for (key = ohash_first(&names, &slot); NULL != key; |
|
key = ohash_next(&names, &slot)) { |
|
assert(key->mpage == mpage); |
|
dba_page_alias(mpage->dba, key->key, key->mask); |
|
free(key); |
|
} |
|
for (key = ohash_first(&strings, &slot); NULL != key; |
|
key = ohash_next(&strings, &slot)) { |
|
assert(key->mpage == mpage); |
|
i = 0; |
|
for (mask = TYPE_Xr; mask <= TYPE_Lb; mask *= 2) { |
|
if (key->mask & mask) |
|
dba_macro_add(dba->macros, i, |
|
key->key, mpage->dba); |
|
i++; |
|
} |
|
free(key); |
|
} |
} |
} |
|
|
static void |
static void |
dbprune(const char *base) |
dbprune(struct dba *dba) |
{ |
{ |
struct of *of; |
struct dba_array *page, *files; |
size_t i; |
char *file; |
|
|
if (nodb) |
dba_array_FOREACH(dba->pages, page) { |
return; |
files = dba_array_get(page, DBP_FILE); |
|
dba_array_FOREACH(files, file) { |
for (of = ofs; NULL != of; of = of->next) { |
if (*file < ' ') |
i = 1; |
file++; |
SQL_BIND_TEXT(stmts[STMT_DELETE], i, of->file); |
if (ohash_find(&mlinks, ohash_qlookup(&mlinks, |
SQL_STEP(stmts[STMT_DELETE]); |
file)) != NULL) { |
sqlite3_reset(stmts[STMT_DELETE]); |
if (debug) |
if (verb) |
say(file, "Deleting from database"); |
say(base, of->file, "Deleted from index"); |
dba_array_del(dba->pages); |
|
break; |
|
} |
|
} |
} |
} |
} |
} |
|
|
/* |
/* |
* 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(const char *base, int real) |
dbwrite(struct dba *dba) |
{ |
{ |
size_t i; |
char tfn[32]; |
char file[PATH_MAX]; |
int status; |
|
pid_t child; |
|
|
if (nodb) |
if (dba_write(MANDOC_DB "~", dba) != -1) { |
|
if (rename(MANDOC_DB "~", MANDOC_DB) == -1) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(MANDOC_DB, "&rename"); |
|
unlink(MANDOC_DB "~"); |
|
} |
return; |
return; |
|
} |
|
|
for (i = 0; i < STMT__MAX; i++) { |
(void)strlcpy(tfn, "/tmp/mandocdb.XXXXXXXX", sizeof(tfn)); |
sqlite3_finalize(stmts[i]); |
if (mkdtemp(tfn) == NULL) { |
stmts[i] = NULL; |
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "&%s", tfn); |
|
return; |
} |
} |
|
|
sqlite3_close(db); |
(void)strlcat(tfn, "/" MANDOC_DB, sizeof(tfn)); |
db = NULL; |
if (dba_write(tfn, dba) == -1) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say(tfn, "&dba_write"); |
|
goto out; |
|
} |
|
|
if (real) |
switch (child = fork()) { |
|
case -1: |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "&fork cmp"); |
return; |
return; |
|
case 0: |
|
execlp("cmp", "cmp", "-s", tfn, MANDOC_DB, (char *)NULL); |
|
say("", "&exec cmp"); |
|
exit(0); |
|
default: |
|
break; |
|
} |
|
if (waitpid(child, &status, 0) == -1) { |
|
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"); |
|
} |
|
|
strlcpy(file, MANDOC_DB, PATH_MAX); |
out: |
strlcat(file, "~", PATH_MAX); |
*strrchr(tfn, '/') = '\0'; |
if (-1 == rename(file, MANDOC_DB)) |
switch (child = fork()) { |
perror(MANDOC_DB); |
case -1: |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "&fork rm"); |
|
return; |
|
case 0: |
|
execlp("rm", "rm", "-rf", tfn, (char *)NULL); |
|
say("", "&exec rm"); |
|
exit((int)MANDOCLEVEL_SYSERR); |
|
default: |
|
break; |
|
} |
|
if (waitpid(child, &status, 0) == -1) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "&wait rm"); |
|
} else if (WIFSIGNALED(status) || WEXITSTATUS(status)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "%s: Cannot remove temporary directory", tfn); |
|
} |
} |
} |
|
|
/* |
|
* 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 |
static int |
dbopen(const char *base, int real) |
set_basedir(const char *targetdir, int report_baddir) |
{ |
{ |
char file[PATH_MAX]; |
static char startdir[PATH_MAX]; |
const char *sql; |
static int getcwd_status; /* 1 = ok, 2 = failure */ |
int rc, ofl; |
static int chdir_status; /* 1 = changed directory */ |
size_t sz; |
char *cp; |
|
|
if (nodb) |
/* |
return(1); |
* Remember the original working directory, if possible. |
|
* This will be needed if the second or a later directory |
sz = strlcpy(file, MANDOC_DB, PATH_MAX); |
* on the command line is given as a relative path. |
if ( ! real) |
* Do not error out if the current directory is not |
sz = strlcat(file, "~", PATH_MAX); |
* searchable: Maybe it won't be needed after all. |
|
*/ |
if (sz >= PATH_MAX) { |
if (0 == getcwd_status) { |
fprintf(stderr, "%s: Path too long\n", file); |
if (NULL == getcwd(startdir, sizeof(startdir))) { |
return(0); |
getcwd_status = 2; |
|
(void)strlcpy(startdir, strerror(errno), |
|
sizeof(startdir)); |
|
} else |
|
getcwd_status = 1; |
} |
} |
|
|
if ( ! real) |
/* |
remove(file); |
* We are leaving the old base directory. |
|
* Do not use it any longer, not even for messages. |
|
*/ |
|
*basedir = '\0'; |
|
|
ofl = SQLITE_OPEN_READWRITE | |
/* |
(0 == real ? SQLITE_OPEN_EXCLUSIVE : 0); |
* If and only if the directory was changed earlier and |
|
* the next directory to process is given as a relative path, |
rc = sqlite3_open_v2(file, &db, ofl, NULL); |
* first go back, or bail out if that is impossible. |
if (SQLITE_OK == rc) |
*/ |
goto prepare_statements; |
if (chdir_status && '/' != *targetdir) { |
if (SQLITE_CANTOPEN != rc) { |
if (2 == getcwd_status) { |
perror(file); |
exitcode = (int)MANDOCLEVEL_SYSERR; |
return(0); |
say("", "getcwd: %s", startdir); |
|
return 0; |
|
} |
|
if (-1 == chdir(startdir)) { |
|
exitcode = (int)MANDOCLEVEL_SYSERR; |
|
say("", "&chdir %s", startdir); |
|
return 0; |
|
} |
} |
} |
|
|
sqlite3_close(db); |
/* |
db = NULL; |
* Always resolve basedir to the canonicalized absolute |
|
* pathname and append a trailing slash, such that |
if (SQLITE_OK != (rc = sqlite3_open(file, &db))) { |
* we can reliably check whether files are inside. |
perror(file); |
*/ |
return(0); |
if (NULL == realpath(targetdir, basedir)) { |
|
if (report_baddir || errno != ENOENT) { |
|
exitcode = (int)MANDOCLEVEL_BADARG; |
|
say("", "&%s: realpath", targetdir); |
|
} |
|
return 0; |
|
} else if (-1 == chdir(basedir)) { |
|
if (report_baddir || errno != ENOENT) { |
|
exitcode = (int)MANDOCLEVEL_BADARG; |
|
say("", "&chdir"); |
|
} |
|
return 0; |
} |
} |
|
chdir_status = 1; |
sql = "CREATE TABLE \"docs\" (\n" |
cp = strchr(basedir, '\0'); |
" \"file\" TEXT NOT NULL,\n" |
if ('/' != cp[-1]) { |
" \"sec\" TEXT NOT NULL,\n" |
if (cp - basedir >= PATH_MAX - 1) { |
" \"arch\" TEXT NOT NULL,\n" |
exitcode = (int)MANDOCLEVEL_SYSERR; |
" \"desc\" TEXT NOT NULL,\n" |
say("", "Filename too long"); |
" \"form\" INTEGER NOT NULL,\n" |
return 0; |
" \"id\" INTEGER PRIMARY KEY AUTOINCREMENT NOT NULL\n" |
} |
");\n" |
*cp++ = '/'; |
"\n" |
*cp = '\0'; |
"CREATE TABLE \"keys\" (\n" |
|
" \"bits\" INTEGER NOT NULL,\n" |
|
" \"key\" TEXT NOT NULL,\n" |
|
" \"docid\" INTEGER NOT NULL REFERENCES docs(id) " |
|
"ON DELETE CASCADE,\n" |
|
" \"id\" INTEGER PRIMARY KEY AUTOINCREMENT NOT NULL\n" |
|
");\n" |
|
"\n" |
|
"CREATE INDEX \"key_index\" ON keys (key);\n"; |
|
|
|
if (SQLITE_OK != sqlite3_exec(db, sql, NULL, NULL, NULL)) { |
|
perror(sqlite3_errmsg(db)); |
|
return(0); |
|
} |
} |
|
return 1; |
prepare_statements: |
|
SQL_EXEC("PRAGMA foreign_keys = ON"); |
|
sql = "DELETE FROM docs where file=?"; |
|
sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_DELETE], NULL); |
|
sql = "INSERT INTO docs " |
|
"(file,sec,arch,desc,form) VALUES (?,?,?,?,?)"; |
|
sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_DOC], NULL); |
|
sql = "INSERT INTO keys " |
|
"(bits,key,docid) VALUES (?,?,?)"; |
|
sqlite3_prepare_v2(db, sql, -1, &stmts[STMT_INSERT_KEY], NULL); |
|
return(1); |
|
} |
} |
|
|
static void * |
|
hash_halloc(size_t sz, void *arg) |
|
{ |
|
|
|
return(mandoc_calloc(sz, 1)); |
|
} |
|
|
|
static void * |
|
hash_alloc(size_t sz, void *arg) |
|
{ |
|
|
|
return(mandoc_malloc(sz)); |
|
} |
|
|
|
static void |
static void |
hash_free(void *p, size_t sz, void *arg) |
say(const char *file, const char *format, ...) |
{ |
{ |
|
|
free(p); |
|
} |
|
|
|
static int |
|
path_reset(const char *cwd, int fd, const char *base) |
|
{ |
|
|
|
if (-1 == fchdir(fd)) { |
|
perror(cwd); |
|
return(0); |
|
} else if (-1 == chdir(base)) { |
|
perror(base); |
|
return(0); |
|
} |
|
return(1); |
|
} |
|
|
|
static void |
|
say(const char *dir, const char *file, const char *format, ...) |
|
{ |
|
va_list ap; |
va_list ap; |
|
int use_errno; |
|
|
fprintf(stderr, "%s", dir); |
if ('\0' != *basedir) |
|
fprintf(stderr, "%s", basedir); |
|
if ('\0' != *basedir && '\0' != *file) |
|
fputc('/', stderr); |
if ('\0' != *file) |
if ('\0' != *file) |
fprintf(stderr, "//%s", file); |
fprintf(stderr, "%s", file); |
fputs(": ", stderr); |
|
|
|
va_start(ap, format); |
use_errno = 1; |
vfprintf(stderr, format, ap); |
if (NULL != format) { |
va_end(ap); |
switch (*format) { |
|
case '&': |
fputc('\n', stderr); |
format++; |
|
break; |
|
case '\0': |
|
format = NULL; |
|
break; |
|
default: |
|
use_errno = 0; |
|
break; |
|
} |
|
} |
|
if (NULL != format) { |
|
if ('\0' != *basedir || '\0' != *file) |
|
fputs(": ", stderr); |
|
va_start(ap, format); |
|
vfprintf(stderr, format, ap); |
|
va_end(ap); |
|
} |
|
if (use_errno) { |
|
if ('\0' != *basedir || '\0' != *file || NULL != format) |
|
fputs(": ", stderr); |
|
perror(NULL); |
|
} else |
|
fputc('\n', stderr); |
} |
} |