===================================================================
RCS file: /cvs/mandoc/Attic/makewhatis.c,v
retrieving revision 1.10
retrieving revision 1.18
diff -u -p -r1.10 -r1.18
--- mandoc/Attic/makewhatis.c	2011/06/22 10:36:36	1.10
+++ mandoc/Attic/makewhatis.c	2011/07/11 09:36:15	1.18
@@ -1,4 +1,4 @@
-/*	$Id: makewhatis.c,v 1.10 2011/06/22 10:36:36 kristaps Exp $ */
+/*	$Id: makewhatis.c,v 1.18 2011/07/11 09:36:15 kristaps Exp $ */
 /*
  * Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -54,6 +54,10 @@
 #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. */
 
@@ -73,23 +77,30 @@ 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);
 static	void		  dbt_put(DB *, const char *, DBT *, DBT *);
 static	void		  hash_put(DB *, const struct buf *, int);
+static	void		  hash_reset(DB **);
 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);
@@ -115,8 +126,8 @@ 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 */
@@ -130,12 +141,12 @@ 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 */
@@ -233,15 +244,14 @@ main(int argc, char *argv[])
 			*arch, /* manual architecture */
 	      		*dir; /* result dir (default: cwd) */
 	char		 ibuf[MAXPATHLEN], /* index fname */
-			 ibbuf[MAXPATHLEN], /* index backup fname */
 			 fbuf[MAXPATHLEN],  /* btree fname */
-			 fbbuf[MAXPATHLEN], /* btree backup fname */
 			 vbuf[8]; /* stringified record number */
 	int		 ch, seq, verb;
 	DB		*idx, /* index database */
 			*db, /* keyword database */
 			*hash; /* temporary keyword hashtable */
 	DBT		 key, val;
+	enum mandocerr	 ec;
 	size_t		 sv;
 	BTREEINFO	 info; /* btree configuration */
 	recno_t		 rec; /* current record number */
@@ -258,7 +268,14 @@ main(int argc, char *argv[])
 
 	dir = "";
 	verb = 0;
+	db = idx = NULL;
+	mp = NULL;
+	hash = NULL;
+	ec = MANDOCLEVEL_SYSERR;
 
+	memset(&buf, 0, sizeof(struct buf));
+	memset(&dbuf, 0, sizeof(struct buf));
+
 	while (-1 != (ch = getopt(argc, argv, "d:v")))
 		switch (ch) {
 		case ('d'):
@@ -275,36 +292,19 @@ main(int argc, char *argv[])
 	argc -= optind;
 	argv += optind;
 
-	/*
-	 * Set up temporary file-names into which we're going to write
-	 * all of our data (both for the index and database).  These
-	 * will be securely renamed to the real file-names after we've
-	 * written all of our data.
-	 */
-
 	ibuf[0] = ibuf[MAXPATHLEN - 2] =
-		ibbuf[0] = ibbuf[MAXPATHLEN - 2] = 
-		fbuf[0] = fbuf[MAXPATHLEN - 2] = 
-		fbbuf[0] = fbbuf[MAXPATHLEN - 2] = '\0';
+		fbuf[0] = fbuf[MAXPATHLEN - 2] = '\0';
 
 	strlcat(fbuf, dir, MAXPATHLEN);
 	strlcat(fbuf, MANDOC_DB, MAXPATHLEN);
 
-	strlcat(fbbuf, fbuf, MAXPATHLEN);
-	strlcat(fbbuf, "~", MAXPATHLEN);
-
 	strlcat(ibuf, dir, MAXPATHLEN);
 	strlcat(ibuf, MANDOC_IDX, MAXPATHLEN);
 
-	strlcat(ibbuf, ibuf, MAXPATHLEN);
-	strlcat(ibbuf, "~", MAXPATHLEN);
-
 	if ('\0' != fbuf[MAXPATHLEN - 2] ||
-			'\0' != fbbuf[MAXPATHLEN - 2] ||
-			'\0' != ibuf[MAXPATHLEN - 2] ||
-			'\0' != ibbuf[MAXPATHLEN - 2]) {
+			'\0' != ibuf[MAXPATHLEN - 2]) {
 		fprintf(stderr, "%s: Path too long\n", dir);
-		exit((int)MANDOCLEVEL_SYSERR);
+		goto out;
 	}
 
 	/*
@@ -315,19 +315,16 @@ main(int argc, char *argv[])
 
 	memset(&info, 0, sizeof(BTREEINFO));
 	info.flags = R_DUP;
-	db = dbopen(fbbuf, MANDOC_FLAGS, 0644, DB_BTREE, &info);
 
-	if (NULL == db) {
-		perror(fbbuf);
-		exit((int)MANDOCLEVEL_SYSERR);
-	}
+	db = dbopen(fbuf, MANDOC_FLAGS, 0644, DB_BTREE, &info);
+	idx = dbopen(ibuf, MANDOC_FLAGS, 0644, DB_RECNO, NULL);
 
-	idx = dbopen(ibbuf, MANDOC_FLAGS, 0644, DB_RECNO, NULL);
-
 	if (NULL == db) {
-		perror(ibbuf);
-		(*db->close)(db);
-		exit((int)MANDOCLEVEL_SYSERR);
+		perror(fbuf);
+		goto out;
+	} else if (NULL == db) {
+		perror(ibuf);
+		goto out;
 	}
 
 	/*
@@ -339,51 +336,31 @@ main(int argc, char *argv[])
 
 	mp = mparse_alloc(MPARSE_AUTO, MANDOCLEVEL_FATAL, NULL, NULL);
 
-	rec = 1;
-	hash = NULL;
-
-	memset(&buf, 0, sizeof(struct buf));
-	memset(&dbuf, 0, sizeof(struct buf));
-
 	buf.size = dbuf.size = MANDOC_BUFSZ;
-
 	buf.cp = mandoc_malloc(buf.size);
 	dbuf.cp = mandoc_malloc(dbuf.size);
 
+	rec = 1;
+
 	while (NULL != (fn = *argv++)) {
 		mparse_reset(mp);
+		hash_reset(&hash);
 
-		/* 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;
 
 		msec = NULL != mdoc ? 
-			mdoc_meta(mdoc)->msec :
-			man_meta(man)->msec;
+			mdoc_meta(mdoc)->msec : man_meta(man)->msec;
 		mtitle = NULL != mdoc ? 
-			mdoc_meta(mdoc)->title :
-			man_meta(man)->title;
-		arch = NULL != mdoc ? mdoc_meta(mdoc)->arch : NULL;
+			mdoc_meta(mdoc)->title : man_meta(man)->title;
+		arch = NULL != mdoc ? 
+			mdoc_meta(mdoc)->arch : "";
 
 		/* 
 		 * The index record value consists of a nil-terminated
@@ -397,8 +374,7 @@ main(int argc, char *argv[])
 		buf_appendb(&dbuf, fn, strlen(fn) + 1);
 		buf_appendb(&dbuf, msec, strlen(msec) + 1);
 		buf_appendb(&dbuf, mtitle, strlen(mtitle) + 1);
-		buf_appendb(&dbuf, arch ? arch : "", 
-				arch ? strlen(arch) + 1 : 1);
+		buf_appendb(&dbuf, arch, strlen(arch) + 1);
 
 		sv = dbuf.len;
 
@@ -427,14 +403,13 @@ main(int argc, char *argv[])
 			val.data = vbuf;
 
 			if (verb > 1)
-				printf("%s: Keyword %s (%zu): 0x%x\n", 
-					fn, (char *)key.data, key.size, 
+				printf("%s: Keyword %s: 0x%x\n", 
+					fn, (char *)key.data, 
 					*(int *)val.data);
 
-			dbt_put(db, fbbuf, &key, &val);
+			dbt_put(db, fbuf, &key, &val);
 
 		}
-
 		if (ch < 0) {
 			perror("hash");
 			exit((int)MANDOCLEVEL_SYSERR);
@@ -457,29 +432,26 @@ main(int argc, char *argv[])
 		if (verb > 0)
 			printf("%s: Indexed\n", fn);
 
-		dbt_put(idx, ibbuf, &key, &val);
+		dbt_put(idx, ibuf, &key, &val);
 		rec++;
 	}
 
-	(*db->close)(db);
-	(*idx->close)(idx);
+	ec = MANDOCLEVEL_OK;
 
+out:
+	if (db)
+		(*db->close)(db);
+	if (idx)
+		(*idx->close)(idx);
 	if (hash)
 		(*hash->close)(hash);
+	if (mp)
+		mparse_free(mp);
 
-	mparse_free(mp);
-
 	free(buf.cp);
 	free(dbuf.cp);
 
-	/* Atomically replace the file with our temporary one. */
-
-	if (-1 == rename(fbbuf, fbuf))
-		perror(fbuf);
-	if (-1 == rename(ibbuf, ibuf))
-		perror(fbuf);
-
-	return((int)MANDOCLEVEL_OK);
+	return((int)ec);
 }
 
 /*
@@ -520,6 +492,33 @@ 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)
@@ -528,13 +527,25 @@ 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);
 }
 
+static void
+hash_reset(DB **db)
+{
+	DB		*hash;
+
+	if (NULL != (hash = *db))
+		(*hash->close)(hash);
+
+	*db = dbopen(NULL, MANDOC_FLAGS, 0644, DB_HASH, NULL);
+	if (NULL == *db) {
+		perror("hash");
+		exit((int)MANDOCLEVEL_SYSERR);
+	}
+}
+
 /* ARGSUSED */
 static void
 pmdoc_Fd(MDOC_ARGS)
@@ -589,10 +600,7 @@ 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);
 }
 
@@ -652,6 +660,25 @@ 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;
@@ -706,37 +733,59 @@ 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)
@@ -745,10 +794,7 @@ 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);
 }
 
@@ -892,8 +938,9 @@ pman_node(MAN_ARGS)
 					start++;
 			}
 
+			buf->len = 0;
+
 			if (sv == start) {
-				buf->len = 0;
 				buf_append(buf, start);
 				return(1);
 			}
@@ -916,6 +963,8 @@ pman_node(MAN_ARGS)
 			sz = strlen(start) + 1;
 			buf_appendb(dbuf, start, sz);
 			buf_appendb(buf, start, sz);
+
+			hash_put(hash, buf, TYPE_DESC);
 		}
 	}