=================================================================== RCS file: /cvs/mandoc/Attic/makewhatis.c,v retrieving revision 1.6 retrieving revision 1.16 diff -u -p -r1.6 -r1.16 --- mandoc/Attic/makewhatis.c 2011/06/21 14:16:05 1.6 +++ mandoc/Attic/makewhatis.c 2011/07/01 13:46:39 1.16 @@ -1,4 +1,4 @@ -/* $Id: makewhatis.c,v 1.6 2011/06/21 14:16:05 kristaps Exp $ */ +/* $Id: makewhatis.c,v 1.16 2011/07/01 13:46:39 kristaps Exp $ */ /* * Copyright (c) 2011 Kristaps Dzonsons * @@ -21,11 +21,6 @@ #include #include -#ifdef __linux__ -# include -#else -# include -#endif #include #include #include @@ -33,6 +28,12 @@ #include #include +#ifdef __linux__ +# include +#else +# include +#endif + #include "man.h" #include "mdoc.h" #include "mandoc.h" @@ -44,15 +45,19 @@ /* Bit-fields. See makewhatis.1. */ -#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_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 +#define TYPE_ENV 0x800 +#define TYPE_ERR 0x1000 /* Buffer for storing growable data. */ @@ -72,6 +77,8 @@ struct buf { const struct mdoc_node *n, \ const struct mdoc_meta *m +static void buf_appendmdoc(struct buf *, + const struct mdoc_node *, int); static void buf_append(struct buf *, const char *); static void buf_appendb(struct buf *, const void *, size_t); @@ -81,14 +88,18 @@ static int pman_node(MAN_ARGS); static void pmdoc_node(MDOC_ARGS); static void pmdoc_An(MDOC_ARGS); static void pmdoc_Cd(MDOC_ARGS); +static void pmdoc_Er(MDOC_ARGS); +static void pmdoc_Ev(MDOC_ARGS); static void pmdoc_Fd(MDOC_ARGS); static void pmdoc_In(MDOC_ARGS); 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); @@ -114,8 +125,8 @@ static const pmdoc_nf mdocs[MDOC_MAX] = { pmdoc_Cd, /* Cd */ NULL, /* Cm */ NULL, /* Dv */ - NULL, /* Er */ - NULL, /* Ev */ + pmdoc_Er, /* Er */ + pmdoc_Ev, /* Ev */ NULL, /* Ex */ NULL, /* Fa */ pmdoc_Fd, /* Fd */ @@ -129,12 +140,12 @@ 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 */ @@ -236,12 +247,12 @@ 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 key, val; - size_t sv, rsz; + size_t sv; BTREEINFO info; /* btree configuration */ recno_t rec; /* current record number */ struct buf buf, /* keyword buffer */ @@ -256,12 +267,16 @@ 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); @@ -298,7 +313,7 @@ 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); } @@ -306,23 +321,14 @@ 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); } @@ -331,21 +337,20 @@ 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 - * totally fail, then just keep on going. Take resulting trees - * and push them down into the database code. + * Try parsing each manual given on the command line. + * If we fail, then emit an error and keep on going. + * 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); rec = 1; - rsz = 0; + hash = NULL; memset(&buf, 0, sizeof(struct buf)); memset(&dbuf, 0, sizeof(struct buf)); @@ -358,18 +363,30 @@ main(int argc, char *argv[]) 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; @@ -378,9 +395,6 @@ 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 @@ -416,16 +430,19 @@ 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); - if ((*hash->del)(hash, &key, 0) < 0) { - perror("hash"); - exit((int)MANDOCLEVEL_SYSERR); - } - seq = R_NEXT; } if (ch < 0) { @@ -447,14 +464,19 @@ main(int argc, char *argv[]) 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(buf.cp); @@ -508,6 +530,33 @@ buf_append(struct buf *buf, const char *cp) buf_appendb(buf, 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_appendmdoc(struct buf *buf, const struct mdoc_node *n, int f) +{ + + for ( ; n; n = n->next) { + if (n->child) + buf_appendmdoc(buf, n->child, f); + + if (MDOC_TEXT == n->type && f) { + f = 0; + buf_appendb(buf, n->string, + strlen(n->string) + 1); + } else if (MDOC_TEXT == n->type) + buf_append(buf, n->string); + + } +} + /* ARGSUSED */ static void pmdoc_An(MDOC_ARGS) @@ -516,10 +565,7 @@ pmdoc_An(MDOC_ARGS) if (SEC_AUTHORS != n->sec) return; - for (n = n->child; n; n = n->next) - if (MDOC_TEXT == n->type) - buf_append(buf, n->string); - + buf_appendmdoc(buf, n->child, 0); hash_put(hash, buf, TYPE_AUTHOR); } @@ -577,10 +623,7 @@ pmdoc_Cd(MDOC_ARGS) if (SEC_SYNOPSIS != n->sec) return; - for (n = n->child; n; n = n->next) - if (MDOC_TEXT == n->type) - buf_append(buf, n->string); - + buf_appendmdoc(buf, n->child, 0); hash_put(hash, buf, TYPE_CONFIG); } @@ -640,6 +683,25 @@ 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; @@ -694,37 +756,59 @@ pmdoc_Fo(MDOC_ARGS) static void pmdoc_Nd(MDOC_ARGS) { - int first; - size_t sz; - - for (first = 1, n = n->child; n; n = n->next) { - if (MDOC_TEXT != n->type) - continue; - if (first) { - sz = strlen(n->string) + 1; - buf_appendb(dbuf, n->string, sz); - buf_appendb(buf, n->string, sz); - } else { - buf_append(dbuf, n->string); - buf_append(buf, n->string); - } + if (MDOC_BODY != n->type) + return; - first = 0; - } + buf_appendmdoc(dbuf, n->child, 1); + buf_appendmdoc(buf, n->child, 0); hash_put(hash, buf, TYPE_DESC); } /* ARGSUSED */ static void +pmdoc_Er(MDOC_ARGS) +{ + + if (SEC_ERRORS != n->sec) + return; + + buf_appendmdoc(buf, n->child, 0); + hash_put(hash, buf, TYPE_ERR); +} + +/* ARGSUSED */ +static void +pmdoc_Ev(MDOC_ARGS) +{ + + if (SEC_ENVIRONMENT != n->sec) + return; + + buf_appendmdoc(buf, n->child, 0); + hash_put(hash, buf, TYPE_ENV); +} + +/* ARGSUSED */ +static void +pmdoc_Pa(MDOC_ARGS) +{ + + if (SEC_FILES != n->sec) + return; + + buf_appendmdoc(buf, n->child, 0); + hash_put(hash, buf, TYPE_PATH); +} + +/* ARGSUSED */ +static void pmdoc_Nm(MDOC_ARGS) { if (SEC_NAME == n->sec) { - for (n = n->child; n; n = n->next) - if (MDOC_TEXT == n->type) - buf_append(buf, n->string); + buf_appendmdoc(buf, n->child, 0); hash_put(hash, buf, TYPE_NAME); return; } else if (SEC_SYNOPSIS != n->sec || MDOC_HEAD != n->type) @@ -733,10 +817,7 @@ pmdoc_Nm(MDOC_ARGS) if (NULL == n->child) buf_append(buf, m->name); - for (n = n->child; n; n = n->next) - if (MDOC_TEXT == n->type) - buf_append(buf, n->string); - + buf_appendmdoc(buf, n->child, 0); hash_put(hash, buf, TYPE_UTILITY); } @@ -746,11 +827,12 @@ hash_put(DB *db, const struct buf *buf, int mask) DBT key, val; int rc; - key.data = buf->cp; - - if ((key.size = buf->len) < 2) + if (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); @@ -760,9 +842,6 @@ 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); @@ -773,9 +852,6 @@ static void dbt_put(DB *db, const char *dbn, DBT *key, DBT *val) { - if (0 == key->size) - return; - assert(key->size); assert(val->size); @@ -924,5 +1000,6 @@ static void usage(void) { - fprintf(stderr, "usage: %s [-d path] [file...]\n", progname); + fprintf(stderr, "usage: %s [-v] [-d path] [file...]\n", + progname); }