Make apropos's lookup use a find(1)-like expression. I'll write more on

this when it completes; this is to keep it in-tree.

Right now this uses prefix notation.  Ignore it.  I'll make this into
infix notation real soon.

The goal of this (exprcomp and exprexec) is to have arbitrary logical
expressions.

4 files changed, 150 insertions, 88 deletions

diff --git a/apropos.c b/apropos.c
index 4e438e99..6bb5b7a9 100644
--- a/apropos.c
+++ b/apropos.c
@@ -1,4 +1,4 @@
-/*	$Id: apropos.c,v 1.9 2011/11/09 10:53:48 kristaps Exp $ */
+/*	$Id: apropos.c,v 1.10 2011/11/09 22:05:56 kristaps Exp $ */
 /*
  * Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -24,29 +24,6 @@
 #include "apropos.h"
 #include "mandoc.h"
 
-struct	type {
-	int		 mask;
-	const char	*name; /* command-line type name */
-};
-
-static	const struct type types[] = {
-	{ TYPE_NAME, "name" },
-	{ TYPE_FUNCTION, "func" },
-	{ TYPE_UTILITY, "utility" },
-	{ TYPE_INCLUDES, "incl" },
-	{ TYPE_VARIABLE, "var" },
-	{ TYPE_STANDARD, "stand" },
-	{ TYPE_AUTHOR, "auth" },
-	{ TYPE_CONFIG, "conf" },
-	{ TYPE_DESC, "desc" },
-	{ TYPE_XREF, "xref" },
-	{ TYPE_PATH, "path" },
-	{ TYPE_ENV, "env" },
-	{ TYPE_ERR, "err" },
-	{ INT_MAX, "all" },
-	{ 0, NULL }
-};
-
 static	int	 cmp(const void *, const void *);
 static	void	 list(struct rec *, size_t, void *);
 static	void	 usage(void);
@@ -56,23 +33,23 @@ static	char	*progname;
 int
 main(int argc, char *argv[])
 {
-	int		 ch, i;
-	char		*q, *v;
+	int		 ch, cs;
 	struct opts	 opts;
+	struct expr	*e;
 	extern int	 optind;
 	extern char	*optarg;
 
 	memset(&opts, 0, sizeof(struct opts));
 
-	q = NULL;
-
 	progname = strrchr(argv[0], '/');
 	if (progname == NULL)
 		progname = argv[0];
 	else
 		++progname;
 
-	while (-1 != (ch = getopt(argc, argv, "S:s:It:"))) 
+	cs = 0;
+
+	while (-1 != (ch = getopt(argc, argv, "S:s:I"))) 
 		switch (ch) {
 		case ('S'):
 			opts.arch = optarg;
@@ -81,26 +58,8 @@ main(int argc, char *argv[])
 			opts.cat = optarg;
 			break;
 		case ('I'):
-			opts.flags |= OPTS_INSENS;
+			cs = 1;
 			break;
-		case ('t'):
-			while (NULL != (v = strsep(&optarg, ","))) {
-				if ('\0' == *v)
-					continue;
-				for (i = 0; types[i].mask; i++) {
-					if (strcmp(types[i].name, v))
-						continue;
-					break;
-				}
-				if (0 == types[i].mask)
-					break;
-				opts.types |= types[i].mask;
-			}
-			if (NULL == v)
-				break;
-			
-			fprintf(stderr, "%s: Bad type\n", v);
-			return(EXIT_FAILURE);
 		default:
 			usage();
 			return(EXIT_FAILURE);
@@ -109,14 +68,13 @@ main(int argc, char *argv[])
 	argc -= optind;
 	argv += optind;
 
-	if (0 == argc || '\0' == **argv) {
-		usage();
+	if (0 == argc) 
 		return(EXIT_SUCCESS);
-	} else
-		q = *argv;
 
-	if (0 == opts.types)
-		opts.types = TYPE_NAME | TYPE_DESC;
+	if (NULL == (e = exprcomp(cs, argv, argc))) {
+		fprintf(stderr, "Bad expression\n");
+		return(EXIT_FAILURE);
+	}
 
 	/*
 	 * Configure databases.
@@ -125,7 +83,8 @@ main(int argc, char *argv[])
 	 * The index database is a recno.
 	 */
 
-	apropos_search(&opts, q, NULL, list);
+	apropos_search(&opts, e, NULL, list);
+	exprfree(e);
 	return(EXIT_SUCCESS);
 }
 
@@ -161,6 +120,6 @@ usage(void)
 			"[-I] "
 			"[-S arch] "
 			"[-s section] "
-			"[-t type[,...]] "
-			"key\n", progname);
+			"EXPR\n", 
+			progname);
 }
diff --git a/apropos.h b/apropos.h
index 4a06db87..16ebcda2 100644
--- a/apropos.h
+++ b/apropos.h
@@ -1,4 +1,4 @@
-/*	$Id: apropos.h,v 1.1 2011/11/09 01:24:23 kristaps Exp $ */
+/*	$Id: apropos.h,v 1.2 2011/11/09 22:05:56 kristaps Exp $ */
 /*
  * Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -53,15 +53,18 @@ struct	rec {
 struct	opts {
 	const char	*arch; /* restrict to architecture */
 	const char	*cat; /* restrict to manual section */
-	int		 types; /* only types in bitmask */
-	int		 flags;
-#define	OPTS_INSENS	(0x01) /* case-insensitive match */
 };
 
 __BEGIN_DECLS
 
-void	 apropos_search(const struct opts *, const char *, 
-		void *, void (*)(struct rec *, size_t, void *));
+struct	expr;
+
+void	 	 apropos_search(const struct opts *, 
+			const struct expr *, void *, 
+			void (*)(struct rec *, size_t, void *));
+
+struct	expr	*exprcomp(int, char *[], int);
+void		 exprfree(struct expr *);
 
 __END_DECLS
 
diff --git a/cgi.c b/cgi.c
index edf63f8d..a596254a 100644
--- a/cgi.c
+++ b/cgi.c
@@ -1,4 +1,4 @@
-/*	$Id: cgi.c,v 1.1 2011/11/09 01:24:23 kristaps Exp $ */
+/*	$Id: cgi.c,v 1.2 2011/11/09 22:05:56 kristaps Exp $ */
 #include <assert.h>
 #include <fcntl.h>
 #include <regex.h>
@@ -47,13 +47,17 @@ struct	req {
 	enum page	 page;
 };
 
+#if 0
 static	void		 html_printtext(const char *);
+#endif
 static	int 		 kval_decode(char *);
 static	void		 kval_parse(struct kval **, size_t *, char *);
 static	void		 kval_free(struct kval *, size_t);
 static	void		 pg_index(const struct req *, char *);
 static	void		 pg_search(const struct req *, char *);
+#if 0
 static	void		 pg_searchres(struct rec *, size_t, void *);
+#endif
 
 static	const char * const pages[PAGE__MAX] = {
 	"index", /* PAGE_INDEX */ 
@@ -64,6 +68,7 @@ static	const char * const medias[MEDIA__MAX] = {
 	"html", /* MEDIA_HTML */
 };
 
+#if 0
 static void
 html_printtext(const char *p)
 {
@@ -88,6 +93,7 @@ html_printtext(const char *p)
 			break;
 		}
 }
+#endif
 
 static void
 kval_free(struct kval *p, size_t sz)
@@ -205,6 +211,7 @@ pg_index(const struct req *req, char *path)
 
 }
 
+#if 0
 static void
 pg_searchres(struct rec *recs, size_t sz, void *arg)
 {
@@ -223,6 +230,7 @@ pg_searchres(struct rec *recs, size_t sz, void *arg)
 		puts(")</A>");
 	}
 }
+#endif
 
 static void
 pg_search(const struct req *req, char *path)
@@ -238,8 +246,8 @@ pg_search(const struct req *req, char *path)
 		return;
 
 	memset(&opt, 0, sizeof(struct opts));
-	opt.types = TYPE_NAME | TYPE_DESC;
-	apropos_search(&opt, req->fields[i].val, NULL, pg_searchres);
+	/*opt.types = TYPE_NAME | TYPE_DESC;
+	apropos_search(&opt, req->fields[i].val, NULL, pg_searchres);*/
 }
 
 int
diff --git a/db.c b/db.c
index adba3eb1..512c48aa 100644
--- a/db.c
+++ b/db.c
@@ -1,4 +1,4 @@
-/*	$Id: db.c,v 1.1 2011/11/09 01:24:23 kristaps Exp $ */
+/*	$Id: db.c,v 1.2 2011/11/09 22:05:56 kristaps Exp $ */
 /*
  * Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
  *
@@ -30,8 +30,46 @@
 #include "apropos.h"
 #include "mandoc.h"
 
+enum	match {
+	MATCH_REGEX,
+	MATCH_REGEXCASE,
+	MATCH_STR,
+	MATCH_STRCASE
+};
+
+struct	expr {
+	enum match	 match;
+	int	 	 mask;
+	char		*v;
+	regex_t	 	 re;
+};
+
+struct	type {
+	int		 mask;
+	const char	*name;
+};
+
+static	const struct type types[] = {
+	{ TYPE_NAME, "name" },
+	{ TYPE_FUNCTION, "func" },
+	{ TYPE_UTILITY, "utility" },
+	{ TYPE_INCLUDES, "incl" },
+	{ TYPE_VARIABLE, "var" },
+	{ TYPE_STANDARD, "stand" },
+	{ TYPE_AUTHOR, "auth" },
+	{ TYPE_CONFIG, "conf" },
+	{ TYPE_DESC, "desc" },
+	{ TYPE_XREF, "xref" },
+	{ TYPE_PATH, "path" },
+	{ TYPE_ENV, "env" },
+	{ TYPE_ERR, "err" },
+	{ INT_MAX, "all" },
+	{ 0, NULL }
+};
+
 static	DB	*btree_open(void);
 static	int	 btree_read(const DBT *, const struct mchars *, char **);
+static	int	 exprexec(const struct expr *, char *);
 static	DB	*index_open(void);
 static	int	 index_read(const DBT *, const DBT *, 
 			const struct mchars *, struct rec *);
@@ -277,16 +315,15 @@ index_read(const DBT *key, const DBT *val,
 }
 
 /*
- * Search the mandocdb database for the regular expression "q".
+ * Search the mandocdb database for the expression "expr".
  * Filter out by "opts".
  * Call "res" with the results, which may be zero.
  */
 void
-apropos_search(const struct opts *opts, const char *q, void *arg, 
-		void (*res)(struct rec *, size_t, void *))
+apropos_search(const struct opts *opts, const struct expr *expr,
+		void *arg, void (*res)(struct rec *, size_t, void *))
 {
 	int		 i, len, root, leaf;
-	regex_t		 reg;
 	DBT		 key, val;
 	DB		*btree, *idx;
 	struct mchars	*mc;
@@ -307,19 +344,8 @@ apropos_search(const struct opts *opts, const char *q, void *arg,
 
 	memset(&srec, 0, sizeof(struct rec));
 
-	if (NULL != q && '\0' == *q)
-		q = NULL;
-
-	ch = REG_EXTENDED | REG_NOSUB | 
-		(OPTS_INSENS & opts->flags ? REG_ICASE : 0);
-
 	/* XXX: error out with bad regexp? */
 
-	if (NULL == q || regcomp(&reg, q, ch)) {
-		(*res)(NULL, 0, arg);
-		return;
-	}
-
 	mc = mchars_alloc();
 
 	/* XXX: return fact that we've errored? */
@@ -337,13 +363,10 @@ apropos_search(const struct opts *opts, const char *q, void *arg,
 		 */
 		if (key.size < 2 || 8 != val.size) 
 			break;
-
-		if ( ! (*(int32_t *)val.data & opts->types))
-			continue;
-
 		if ( ! btree_read(&key, mc, &buf))
 			break;
-		if (regexec(&reg, buf, 0, NULL, 0))
+
+		if ( ! exprexec(expr, buf))
 			continue;
 
 		memcpy(&rec, val.data + 4, sizeof(recno_t));
@@ -432,5 +455,74 @@ out:
 
 	free(buf);
 	free(recs);
-	regfree(&reg);
+}
+
+struct expr *
+exprcomp(int cs, char *argv[], int argc)
+{
+	struct expr	*p;
+	struct expr	 e;
+	int		 i, ch;
+
+	if (3 != argc)
+		return(NULL);
+
+	if (0 == strcmp("-eq", argv[0]))
+		e.match = cs ? MATCH_STRCASE : MATCH_STR;
+	else if (0 == strcmp("-ieq", argv[0]))
+		e.match = MATCH_STRCASE;
+	else if (0 == strcmp("-re", argv[0]))
+		e.match = cs ? MATCH_REGEXCASE : MATCH_REGEX;
+	else if (0 == strcmp("-ire", argv[0]))
+		e.match = MATCH_REGEXCASE;
+	else
+		return(NULL);
+
+	for (i = 0; 0 != types[i].mask; i++)
+		if (0 == strcmp(types[i].name, argv[1]))
+			break;
+
+	if (0 == (e.mask = types[i].mask))
+		return(NULL);
+
+	e.v = mandoc_strdup(argv[2]);
+
+	if (MATCH_REGEX == e.match || MATCH_REGEXCASE == e.match) {
+		ch = REG_EXTENDED | REG_NOSUB;
+		if (MATCH_REGEXCASE == e.match)
+			ch |= REG_ICASE;
+		if (regcomp(&e.re, e.v, ch))
+			return(NULL);
+	}
+
+	p = mandoc_calloc(1, sizeof(struct expr));
+	memcpy(p, &e, sizeof(struct expr));
+	return(p);
+}
+
+void
+exprfree(struct expr *p)
+{
+
+	if (NULL == p)
+		return;
+
+	if (MATCH_REGEX == p->match)
+		regfree(&p->re);
+
+	free(p->v);
+	free(p);
+}
+
+static int
+exprexec(const struct expr *p, char *cp)
+{
+
+	if (MATCH_STR == p->match)
+		return(0 == strcmp(p->v, cp));
+	else if (MATCH_STRCASE == p->match)
+		return(0 == strcasecmp(p->v, cp));
+
+	assert(MATCH_REGEX == p->match);
+	return(0 == regexec(&p->re, cp, 0, NULL, 0));
 }