mandoc/makewhatis.c - diff

Return to makewhatis.c CVS log

Up to [cvsweb.bsd.lv] / mandoc

Diff for /mandoc/Attic/makewhatis.c between version 1.4 and 1.14

version 1.4, 2011/06/21 13:13:15

version 1.14, 2011/07/01 10:46:32

Line 21

#include <sys/param.h>

#include <assert.h>

#ifdef __linux__

# include <db_185.h>

#else

# include <db.h>

#endif

#include <fcntl.h>

#include <getopt.h>

#include <stdio.h>

Line 33

Line 28

#include <stdlib.h>

#include <string.h>

#ifdef __linux__

# include <db_185.h>

#else

# include <db.h>

#endif

#include "man.h"

#include "mdoc.h"

#include "mandoc.h"

Line 42

Line 43

#define MANDOC_BUFSZ BUFSIZ

#define MANDOC_FLAGS O_CREAT|O_TRUNC|O_RDWR

#define TYPE_NAME 0x01

/* Bit-fields. See makewhatis.1. */

#define TYPE_FUNCTION 0x02

#define TYPE_UTILITY 0x04

#define TYPE_INCLUDES 0x08

#define TYPE_VARIABLE 0x10

#define TYPE_STANDARD 0x20

#define TYPE_AUTHOR 0x40

#define TYPE_CONFIG 0x80

#define TYPE__MAX TYPE_CONFIG

#define TYPE_NAME 0x01

#define TYPE_FUNCTION 0x02

#define TYPE_UTILITY 0x04

#define TYPE_INCLUDES 0x08

#define TYPE_VARIABLE 0x10

#define TYPE_STANDARD 0x20

#define TYPE_AUTHOR 0x40

#define TYPE_CONFIG 0x80

#define TYPE_DESC 0x100

#define TYPE_XREF 0x200

#define TYPE_PATH 0x400

/* Buffer for storing growable data. */

struct buf {

char *cp;

size_t len;

Line 60 struct buf {

Line 67 struct buf {

#define MAN_ARGS DB *hash, \

struct buf *buf, \

DBT *rval, size_t *rsz, \

struct buf *dbuf, \

const struct man_node *n

#define MDOC_ARGS DB *hash, \

struct buf *buf, \

DBT *rval, size_t *rsz, \

struct buf *dbuf, \

const struct mdoc_node *n, \

const struct mdoc_meta *m

static void dbt_append(DBT *, size_t *, const char *);

static void buf_appendmdoc(struct buf *,

static void dbt_appendb(DBT *, size_t *,

const struct mdoc_node *, int);

static void buf_append(struct buf *, const char *);

static void buf_appendb(struct buf *,

const void *, size_t);

static void dbt_init(DBT *, size_t *);

static void dbt_put(DB *, const char *, DBT *, DBT *);

static void hash_put(DB *, const struct buf *, int);

static void usage(void);

static int pman_node(MAN_ARGS);

static void pmdoc_node(MDOC_ARGS);

static void pmdoc_An(MDOC_ARGS);

Line 85 static void pmdoc_Fn(MDOC_ARGS);

Line 92 static void pmdoc_Fn(MDOC_ARGS);

static void pmdoc_Fo(MDOC_ARGS);

static void pmdoc_Nd(MDOC_ARGS);

static void pmdoc_Nm(MDOC_ARGS);

static void pmdoc_Pa(MDOC_ARGS);

static void pmdoc_St(MDOC_ARGS);

static void pmdoc_Vt(MDOC_ARGS);

static void pmdoc_Xr(MDOC_ARGS);

static void usage(void);

typedef void (*pmdoc_nf)(MDOC_ARGS);

static const char *progname;

static const pmdoc_nf mdocs[MDOC_MAX] = {

NULL, /* Ap */

NULL, /* Dd */

Line 128 static const pmdoc_nf mdocs[MDOC_MAX] = {

Line 136 static const pmdoc_nf mdocs[MDOC_MAX] = {

pmdoc_Nm, /* Nm */

NULL, /* Op */

NULL, /* Ot */

NULL, /* Pa */

pmdoc_Pa, /* Pa */

NULL, /* Rv */

pmdoc_St, /* St */

pmdoc_Vt, /* Va */

pmdoc_Vt, /* Vt */

NULL, /* Xr */

pmdoc_Xr, /* Xr */

NULL, /* %A */

NULL, /* %B */

NULL, /* %D */

Line 217 static const pmdoc_nf mdocs[MDOC_MAX] = {

Line 225 static const pmdoc_nf mdocs[MDOC_MAX] = {

NULL, /* Ta */

};

static const char *progname;

int

main(int argc, char *argv[])

{

Line 233 main(int argc, char *argv[])

Line 243 main(int argc, char *argv[])

fbuf[MAXPATHLEN], /* btree fname */

fbbuf[MAXPATHLEN], /* btree backup fname */

vbuf[8]; /* stringified record number */

int ch, seq;

int ch, seq, verb;

DB *idx, /* index database */

*db, /* keyword database */

*hash; /* temporary keyword hashtable */

DBT rkey, rval, /* recno entries */

DBT key, val;

key, val; /* persistent keyword entries */

size_t sv;

size_t sv, rsz;

BTREEINFO info; /* btree configuration */

recno_t rec; /* current record number */

struct buf buf; /* keyword buffer */

struct buf buf, /* keyword buffer */

dbuf; /* description buffer */

extern int optind;

extern char *optarg;

Line 253 main(int argc, char *argv[])

Line 263 main(int argc, char *argv[])

++progname;

dir = "";

verb = 0;

while (-1 != (ch = getopt(argc, argv, "d:")))

while (-1 != (ch = getopt(argc, argv, "d:v")))

switch (ch) {

case ('d'):

dir = optarg;

break;

case ('v'):

verb++;

break;

default:

usage();

return((int)MANDOCLEVEL_BADARG);

Line 295 main(int argc, char *argv[])

Line 309 main(int argc, char *argv[])

'\0' != fbbuf[MAXPATHLEN - 2] ||

'\0' != ibuf[MAXPATHLEN - 2] ||

'\0' != ibbuf[MAXPATHLEN - 2]) {

fprintf(stderr, "%s: Path too long\n", progname);

fprintf(stderr, "%s: Path too long\n", dir);

exit((int)MANDOCLEVEL_SYSERR);

}

Line 303 main(int argc, char *argv[])

Line 317 main(int argc, char *argv[])

* For the keyword database, open a BTREE database that allows

* duplicates.

* For the index database, use a standard RECNO database type.

* For the temporary keyword hashtable, use the HASH database

* type.

hash = dbopen(NULL, MANDOC_FLAGS, 0644, DB_HASH, NULL);

if (NULL == hash) {

perror("hash");

exit((int)MANDOCLEVEL_SYSERR);

}

memset(&info, 0, sizeof(BTREEINFO));

info.flags = R_DUP;

db = dbopen(fbbuf, MANDOC_FLAGS, 0644, DB_BTREE, &info);

if (NULL == db) {

perror(fbbuf);

(*hash->close)(hash);

exit((int)MANDOCLEVEL_SYSERR);

}

Line 328 main(int argc, char *argv[])

Line 333 main(int argc, char *argv[])

if (NULL == db) {

perror(ibbuf);

(*db->close)(db);

(*hash->close)(hash);

exit((int)MANDOCLEVEL_SYSERR);

}

* Try parsing the manuals given on the command line. If we

* Try parsing each manual given on the command line.

* totally fail, then just keep on going. Take resulting trees

* If we fail, then emit an error and keep on going.

* and push them down into the database code.

* Take resulting trees and push them down into the database code.

* Use the auto-parser and don't report any errors.

mp = mparse_alloc(MPARSE_AUTO, MANDOCLEVEL_FATAL, NULL, NULL);

memset(&key, 0, sizeof(DBT));

memset(&val, 0, sizeof(DBT));

memset(&rkey, 0, sizeof(DBT));

memset(&rval, 0, sizeof(DBT));

rkey.size = sizeof(recno_t);

rec = 1;

rsz = 0;

hash = NULL;

memset(&buf, 0, sizeof(struct buf));

memset(&dbuf, 0, sizeof(struct buf));

buf.size = MANDOC_BUFSZ;

buf.size = dbuf.size = MANDOC_BUFSZ;

buf.cp = mandoc_malloc(buf.size);

dbuf.cp = mandoc_malloc(dbuf.size);

while (NULL != (fn = *argv++)) {

mparse_reset(mp);

/* Initialise the in-memory hash of keywords. */

if (hash)

(*hash->close)(hash);

hash = dbopen(NULL, MANDOC_FLAGS, 0644, DB_HASH, NULL);

if (NULL == hash) {

perror("hash");

exit((int)MANDOCLEVEL_SYSERR);

}

/* Parse and get (non-empty) AST. */

if (mparse_readfd(mp, -1, fn) >= MANDOCLEVEL_FATAL) {

fprintf(stderr, "%s: Parse failure\n", fn);

continue;

}

mparse_result(mp, &mdoc, &man);

if (NULL == mdoc && NULL == man)

continue;

/* Manual section: can be empty string. */

msec = NULL != mdoc ?

mdoc_meta(mdoc)->msec :

man_meta(man)->msec;

Line 379 main(int argc, char *argv[])

Line 391 main(int argc, char *argv[])

man_meta(man)->title;

arch = NULL != mdoc ? mdoc_meta(mdoc)->arch : NULL;

assert(msec);

assert(mtitle);

* The index record value consists of a nil-terminated

* filename, a nil-terminated manual section, and a

Line 390 main(int argc, char *argv[])

Line 399 main(int argc, char *argv[])

* going to write a nil byte in its place.

dbt_init(&rval, &rsz);

dbuf.len = 0;

dbt_appendb(&rval, &rsz, fn, strlen(fn) + 1);

buf_appendb(&dbuf, fn, strlen(fn) + 1);

dbt_appendb(&rval, &rsz, msec, strlen(msec) + 1);

buf_appendb(&dbuf, msec, strlen(msec) + 1);

dbt_appendb(&rval, &rsz, mtitle, strlen(mtitle) + 1);

buf_appendb(&dbuf, mtitle, strlen(mtitle) + 1);

dbt_appendb(&rval, &rsz, arch ? arch : "",

buf_appendb(&dbuf, arch ? arch : "",

arch ? strlen(arch) + 1 : 1);

sv = rval.size;

sv = dbuf.len;

/* Fix the record number in the btree value. */

if (mdoc)

pmdoc_node(hash, &buf, &rval,

pmdoc_node(hash, &buf, &dbuf,

&rsz, mdoc_node(mdoc),

mdoc_node(mdoc), mdoc_meta(mdoc));

mdoc_meta(mdoc));

else

pman_node(hash, &buf, &rval,

pman_node(hash, &buf, &dbuf, man_node(man));

&rsz, man_node(man));

* Copy from the in-memory hashtable of pending keywords

Line 419 main(int argc, char *argv[])

Line 426 main(int argc, char *argv[])

seq = R_FIRST;

while (0 == (ch = (*hash->seq)(hash, &key, &val, seq))) {

seq = R_NEXT;

memcpy(vbuf, val.data, sizeof(uint32_t));

val.size = sizeof(vbuf);

val.data = vbuf;

if (verb > 1)

printf("%s: Keyword %s (%zu): 0x%x\n",

fn, (char *)key.data, key.size,

*(int *)val.data);

dbt_put(db, fbbuf, &key, &val);

/*fprintf(stderr, "Recording: %s (0x%x)\n",

(char *)key.data,

*(int *)val.data);*/

if ((*hash->del)(hash, &key, 0) < 0) {

perror("hash");

exit((int)MANDOCLEVEL_SYSERR);

}

seq = R_NEXT;

}

if (ch < 0) {

Line 443 main(int argc, char *argv[])

Line 451 main(int argc, char *argv[])

* set, put an empty one in now.

if (rval.size == sv)

if (dbuf.len == sv)

dbt_appendb(&rval, &rsz, "", 1);

buf_appendb(&dbuf, "", 1);

rkey.data = &rec;

key.data = &rec;

dbt_put(idx, ibbuf, &rkey, &rval);

key.size = sizeof(recno_t);

printf("Indexed: %s\n", fn);

val.data = dbuf.cp;

val.size = dbuf.len;

if (verb > 0)

printf("%s: Indexed\n", fn);

dbt_put(idx, ibbuf, &key, &val);

rec++;

}

(*db->close)(db);

(*idx->close)(idx);

(*hash->close)(hash);

if (hash)

(*hash->close)(hash);

mparse_free(mp);

free(rval.data);

free(buf.cp);

free(dbuf.cp);

/* Atomically replace the file with our temporary one. */

Line 473 main(int argc, char *argv[])

Line 489 main(int argc, char *argv[])

}

* Initialise the stored database key whose data buffer is shared

* Grow the buffer (if necessary) and copy in a binary string.

* between uses (as the key must sometimes be constructed from an array

* of

static void

dbt_init(DBT *key, size_t *ksz)

{

if (0 == *ksz) {

assert(0 == key->size);

assert(NULL == key->data);

key->data = mandoc_malloc(MANDOC_BUFSZ);

*ksz = MANDOC_BUFSZ;

}

key->size = 0;

}

* Append a binary value to a database entry. This can be invoked

* multiple times; the buffer is automatically resized.

static void

dbt_appendb(DBT *key, size_t *ksz, const void *cp, size_t sz)

{

assert(key->data);

/* Overshoot by MANDOC_BUFSZ. */

while (key->size + sz >= *ksz) {

*ksz = key->size + sz + MANDOC_BUFSZ;

key->data = mandoc_realloc(key->data, *ksz);

}

memcpy(key->data + (int)key->size, cp, sz);

key->size += sz;

}

static void

buf_appendb(struct buf *buf, const void *cp, size_t sz)

{

Line 528 buf_appendb(struct buf *buf, const void *cp, size_t sz

Line 507 buf_appendb(struct buf *buf, const void *cp, size_t sz

}

* Append a nil-terminated string to the database entry. This can be

* Append a nil-terminated string to the buffer.

* invoked multiple times. The database entry will be nil-terminated as

* This can be invoked multiple times.

* well; if invoked multiple times, a space is put between strings.

* The buffer string will be nil-terminated.

* If invoked multiple times, a space is put between strings.

static void

dbt_append(DBT *key, size_t *ksz, const char *cp)

buf_append(struct buf *buf, const char *cp)

{

size_t sz;

if (0 == (sz = strlen(cp)))

return;

assert(key->data);

if (buf->len)

buf->cp[(int)buf->len - 1] = ' ';

if (key->size)

buf_appendb(buf, cp, sz + 1);

((char *)key->data)[(int)key->size - 1] = ' ';

dbt_appendb(key, ksz, cp, sz + 1);

}

* Recursively add all text from a given node.

* This is optimised for general mdoc nodes in this context, which do

* not consist of subexpressions and having a recursive call for n->next

* would be wasteful.

* The "f" variable should be 0 unless called from pmdoc_Nd for the

* description buffer, which does not start at the beginning of the

* buffer.

static void

buf_append(struct buf *buf, const char *cp)

buf_appendmdoc(struct buf *buf, const struct mdoc_node *n, int f)

{

size_t sz;

if (0 == (sz = strlen(cp)))

for ( ; n; n = n->next) {

return;

if (n->child)

buf_appendmdoc(buf, n->child, f);

if (buf->len)

if (MDOC_TEXT == n->type && f) {

buf->cp[(int)buf->len - 1] = ' ';

f = 0;

buf_appendb(buf, n->string,

strlen(n->string) + 1);

} else if (MDOC_TEXT == n->type)

buf_append(buf, n->string);

buf_appendb(buf, cp, sz + 1);

}

/* ARGSUSED */

Line 570 pmdoc_An(MDOC_ARGS)

Line 561 pmdoc_An(MDOC_ARGS)

if (SEC_AUTHORS != n->sec)

return;

for (n = n->child; n; n = n->next)

buf_appendmdoc(buf, n->child, 0);

if (MDOC_TEXT == n->type)

buf_append(buf, n->string);

hash_put(hash, buf, TYPE_AUTHOR);

}

Line 631 pmdoc_Cd(MDOC_ARGS)

Line 619 pmdoc_Cd(MDOC_ARGS)

if (SEC_SYNOPSIS != n->sec)

return;

for (n = n->child; n; n = n->next)

buf_appendmdoc(buf, n->child, 0);

if (MDOC_TEXT == n->type)

buf_append(buf, n->string);

hash_put(hash, buf, TYPE_CONFIG);

}

Line 694 pmdoc_St(MDOC_ARGS)

Line 679 pmdoc_St(MDOC_ARGS)

/* ARGSUSED */

static void

pmdoc_Xr(MDOC_ARGS)

{

if (NULL == (n = n->child))

return;

buf_appendb(buf, n->string, strlen(n->string));

if (NULL != (n = n->next)) {

buf_appendb(buf, ".", 1);

buf_appendb(buf, n->string, strlen(n->string) + 1);

} else

buf_appendb(buf, ".", 2);

hash_put(hash, buf, TYPE_XREF);

}

/* ARGSUSED */

static void

pmdoc_Vt(MDOC_ARGS)

{

const char *start;

Line 748 pmdoc_Fo(MDOC_ARGS)

Line 752 pmdoc_Fo(MDOC_ARGS)

static void

pmdoc_Nd(MDOC_ARGS)

{

int first;

if (MDOC_BODY != n->type)

return;

buf_appendmdoc(dbuf, n->child, 1);

buf_appendmdoc(buf, n->child, 0);

hash_put(hash, buf, TYPE_DESC);

}

/* ARGSUSED */

static void

pmdoc_Pa(MDOC_ARGS)

{

if (SEC_FILES != n->sec)

return;

for (first = 1, n = n->child; n; n = n->next) {

buf_appendmdoc(buf, n->child, 0);

if (MDOC_TEXT != n->type)

hash_put(hash, buf, TYPE_PATH);

continue;

if (first)

dbt_appendb(rval, rsz, n->string, strlen(n->string) + 1);

else

dbt_append(rval, rsz, n->string);

first = 0;

}

/* ARGSUSED */

Line 767 pmdoc_Nm(MDOC_ARGS)

Line 780 pmdoc_Nm(MDOC_ARGS)

{

if (SEC_NAME == n->sec) {

for (n = n->child; n; n = n->next)

buf_appendmdoc(buf, n->child, 0);

if (MDOC_TEXT == n->type)

buf_append(buf, n->string);

hash_put(hash, buf, TYPE_NAME);

return;

} else if (SEC_SYNOPSIS != n->sec || MDOC_HEAD != n->type)

Line 778 pmdoc_Nm(MDOC_ARGS)

Line 789 pmdoc_Nm(MDOC_ARGS)

if (NULL == n->child)

buf_append(buf, m->name);

for (n = n->child; n; n = n->next)

buf_appendmdoc(buf, n->child, 0);

if (MDOC_TEXT == n->type)

buf_append(buf, n->string);

hash_put(hash, buf, TYPE_UTILITY);

}

Line 791 hash_put(DB *db, const struct buf *buf, int mask)

Line 799 hash_put(DB *db, const struct buf *buf, int mask)

DBT key, val;

int rc;

key.data = buf->cp;

if (buf->len < 2)

if ((key.size = buf->len) < 2)

return;

key.data = buf->cp;

key.size = buf->len;

if ((rc = (*db->get)(db, &key, &val, 0)) < 0) {

perror("hash");

exit((int)MANDOCLEVEL_SYSERR);

Line 805 hash_put(DB *db, const struct buf *buf, int mask)

Line 814 hash_put(DB *db, const struct buf *buf, int mask)

val.data = &mask;

val.size = sizeof(int);

/*fprintf(stderr, "Hashing: [%s] (0x%x)\n",

(char *)key.data, mask);*/

if ((rc = (*db->put)(db, &key, &val, 0)) < 0) {

perror("hash");

exit((int)MANDOCLEVEL_SYSERR);

Line 818 static void

Line 824 static void

dbt_put(DB *db, const char *dbn, DBT *key, DBT *val)

{

if (0 == key->size)

assert(key->size);

return;

assert(key->data);

assert(val->size);

assert(val->data);

if (0 == (*db->put)(db, key, val, 0))

return;

Line 858 pmdoc_node(MDOC_ARGS)

Line 860 pmdoc_node(MDOC_ARGS)

break;

buf->len = 0;

(*mdocs[n->tok])(hash, buf, rval, rsz, n, m);

(*mdocs[n->tok])(hash, buf, dbuf, n, m);

break;

default:

break;

}

pmdoc_node(hash, buf, rval, rsz, n->child, m);

pmdoc_node(hash, buf, dbuf, n->child, m);

pmdoc_node(hash, buf, rval, rsz, n->next, m);

pmdoc_node(hash, buf, dbuf, n->next, m);

}

static int

Line 952 pman_node(MAN_ARGS)

Line 954 pman_node(MAN_ARGS)

while (' ' == *start)

start++;

dbt_appendb(rval, rsz, start, strlen(start) + 1);

sz = strlen(start) + 1;

buf_appendb(dbuf, start, sz);

buf_appendb(buf, start, sz);

}

if (pman_node(hash, buf, rval, rsz, n->child))

if (pman_node(hash, buf, dbuf, n->child))

return(1);

if (pman_node(hash, buf, rval, rsz, n->next))

if (pman_node(hash, buf, dbuf, n->next))

return(1);

return(0);

Line 968 static void

Line 972 static void

usage(void)

{

fprintf(stderr, "usage: %s "

fprintf(stderr, "usage: %s [-v] [-d path] [file...]\n",

"[-d path] "

"[file...]\n",

progname);

}

CVSweb