===================================================================
RCS file: /cvs/mandoc/cgi.c,v
retrieving revision 1.19
retrieving revision 1.47
diff -u -p -r1.19 -r1.47
--- mandoc/cgi.c	2011/12/08 22:47:09	1.19
+++ mandoc/cgi.c	2014/01/05 20:41:04	1.47
@@ -1,6 +1,6 @@
-/*	$Id: cgi.c,v 1.19 2011/12/08 22:47:09 kristaps Exp $ */
+/*	$Id: cgi.c,v 1.47 2014/01/05 20:41:04 schwarze Exp $ */
 /*
- * Copyright (c) 2011 Kristaps Dzonsons <kristaps@bsd.lv>
+ * Copyright (c) 2011, 2012 Kristaps Dzonsons <kristaps@bsd.lv>
  *
  * Permission to use, copy, modify, and distribute this software for any
  * purpose with or without fee is hereby granted, provided that the above
@@ -18,12 +18,12 @@
 #include "config.h"
 #endif
 
-#include <sys/param.h>
 #include <sys/wait.h>
 
 #include <assert.h>
 #include <ctype.h>
 #include <errno.h>
+#include <dirent.h>
 #include <fcntl.h>
 #include <limits.h>
 #include <regex.h>
@@ -34,14 +34,22 @@
 #include <string.h>
 #include <unistd.h>
 
+#if defined(__sun)
+/* for stat() */
+#include <fcntl.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#endif
+
 #include "apropos_db.h"
 #include "mandoc.h"
 #include "mdoc.h"
 #include "man.h"
 #include "main.h"
 #include "manpath.h"
+#include "mandocdb.h"
 
-#ifdef __linux__
+#if defined(__linux__) || defined(__sun)
 # include <db_185.h>
 #else
 # include <db.h>
@@ -54,32 +62,46 @@ enum	page {
 	PAGE__MAX
 };
 
-struct	kval {
-	char		*key;
-	char		*val;
+struct	paths {
+	char		*name;
+	char		*path;
 };
 
+/*
+ * A query as passed to the search function.
+ */
+struct	query {
+	const char	*arch; /* architecture */
+	const char	*sec; /* manual section */
+	const char	*expr; /* unparsed expression string */
+	int		 manroot; /* manroot index (or -1)*/
+	int		 legacy; /* whether legacy mode */
+};
+
 struct	req {
-	struct kval	*fields;
-	size_t		 fieldsz;
+	struct query	 q;
+	struct paths	*p;
+	size_t		 psz;
 	enum page	 page;
 };
 
 static	int		 atou(const char *, unsigned *);
-static	void		 catman(const char *);
+static	void		 catman(const struct req *, const char *);
 static	int	 	 cmp(const void *, const void *);
-static	void		 format(const char *);
+static	void		 format(const struct req *, const char *);
 static	void		 html_print(const char *);
+static	void		 html_printquery(const struct req *);
 static	void		 html_putchar(char);
-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 manpaths *,
-				const struct req *, char *);
-static	void		 pg_search(const struct manpaths *,
-				const struct req *, char *);
-static	void		 pg_show(const struct manpaths *,
-				const struct req *, char *);
+static	int 		 http_decode(char *);
+static	void		 http_parse(struct req *, char *);
+static	void		 http_print(const char *);
+static	void 		 http_putchar(char);
+static	void		 http_printquery(const struct req *);
+static	int		 pathstop(DIR *);
+static	void		 pathgen(DIR *, char *, struct req *);
+static	void		 pg_index(const struct req *, char *);
+static	void		 pg_search(const struct req *, char *);
+static	void		 pg_show(const struct req *, char *);
 static	void		 resp_bad(void);
 static	void		 resp_baddb(void);
 static	void		 resp_error400(void);
@@ -91,9 +113,10 @@ static	void		 resp_index(const struct req *);
 static	void		 resp_search(struct res *, size_t, void *);
 static	void		 resp_searchform(const struct req *);
 
-static	const char	 *progname;
-static	const char	 *cache;
-static	const char	 *host;
+static	const char	 *progname; /* cgi script name */
+static	const char	 *cache; /* cache directory */
+static	const char	 *css; /* css directory */
+static	const char	 *host; /* hostname */
 
 static	const char * const pages[PAGE__MAX] = {
 	"index", /* PAGE_INDEX */ 
@@ -117,13 +140,17 @@ atou(const char *buf, unsigned *v)
 		return(0);
 	if ((errno == ERANGE && (lval == LONG_MAX || 
 					lval == LONG_MIN)) ||
-			(lval > UINT_MAX || lval < 0))
+			(lval > INT_MAX || lval < 0))
 		return(0);
 
 	*v = (unsigned int)lval;
 	return(1);
 }
 
+/*
+ * Print a character, escaping HTML along the way.
+ * This will pass non-ASCII straight to output: be warned!
+ */
 static void
 html_putchar(char c)
 {
@@ -146,10 +173,44 @@ html_putchar(char c)
 		break;
 	}
 }
+static void
+http_printquery(const struct req *req)
+{
 
+	printf("&expr=");
+	http_print(req->q.expr ? req->q.expr : "");
+	printf("&sec=");
+	http_print(req->q.sec ? req->q.sec : "");
+	printf("&arch=");
+	http_print(req->q.arch ? req->q.arch : "");
+}
+
+
+static void
+html_printquery(const struct req *req)
+{
+
+	printf("&amp;expr=");
+	html_print(req->q.expr ? req->q.expr : "");
+	printf("&amp;sec=");
+	html_print(req->q.sec ? req->q.sec : "");
+	printf("&amp;arch=");
+	html_print(req->q.arch ? req->q.arch : "");
+}
+
+static void
+http_print(const char *p)
+{
+
+	if (NULL == p)
+		return;
+	while ('\0' != *p)
+		http_putchar(*p++);
+}
+
 /*
- * Print a word, escaping HTML along the way.
- * This will pass non-ASCII straight to output: be warned!
+ * Call through to html_putchar().
+ * Accepts NULL strings.
  */
 static void
 html_print(const char *p)
@@ -161,88 +222,105 @@ html_print(const char *p)
 		html_putchar(*p++);
 }
 
-static void
-kval_free(struct kval *p, size_t sz)
-{
-	int		 i;
-
-	for (i = 0; i < (int)sz; i++) {
-		free(p[i].key);
-		free(p[i].val);
-	}
-	free(p);
-}
-
 /*
  * Parse out key-value pairs from an HTTP request variable.
  * This can be either a cookie or a POST/GET string, although man.cgi
  * uses only GET for simplicity.
  */
 static void
-kval_parse(struct kval **kv, size_t *kvsz, char *p)
+http_parse(struct req *req, char *p)
 {
-	char            *key, *val;
-	size_t           sz, cur;
+	char            *key, *val, *manroot;
+	int		 i, legacy;
 
-	cur = 0;
+	memset(&req->q, 0, sizeof(struct query));
 
-	while (p && '\0' != *p) {
-		while (' ' == *p)
-			p++;
+	legacy = -1;
+	manroot = NULL;
 
+	while ('\0' != *p) {
 		key = p;
 		val = NULL;
 
-		if (NULL != (p = strchr(p, '='))) {
+		p += (int)strcspn(p, ";&");
+		if ('\0' != *p)
 			*p++ = '\0';
-			val = p;
+		if (NULL != (val = strchr(key, '=')))
+			*val++ = '\0';
 
-			sz = strcspn(p, ";&");
-			/* LINTED */
-			p += sz;
+		if ('\0' == *key || NULL == val || '\0' == *val)
+			continue;
 
-			if ('\0' != *p)
-				*p++ = '\0';
-		} else {
-			p = key;
-			sz = strcspn(p, ";&");
-			/* LINTED */
-			p += sz;
+		/* Just abort handling. */
 
-			if ('\0' != *p)
-				p++;
-			continue;
-		}
+		if ( ! http_decode(key))
+			break;
+		if (NULL != val && ! http_decode(val))
+			break;
 
-		if ('\0' == *key || '\0' == *val)
-			continue;
+		if (0 == strcmp(key, "expr"))
+			req->q.expr = val;
+		else if (0 == strcmp(key, "query"))
+			req->q.expr = val;
+		else if (0 == strcmp(key, "sec"))
+			req->q.sec = val;
+		else if (0 == strcmp(key, "sektion"))
+			req->q.sec = val;
+		else if (0 == strcmp(key, "arch"))
+			req->q.arch = val;
+		else if (0 == strcmp(key, "manpath"))
+			manroot = val;
+		else if (0 == strcmp(key, "apropos"))
+			legacy = 0 == strcmp(val, "0");
+	}
 
-		/* Just abort handling. */
+	/* Test for old man.cgi compatibility mode. */
 
-		if ( ! kval_decode(key))
-			return;
-		if ( ! kval_decode(val))
-			return;
+	req->q.legacy = legacy > 0;
 
-		if (*kvsz + 1 >= cur) {
-			cur++;
-			*kv = mandoc_realloc
-				(*kv, cur * sizeof(struct kval));
-		}
+	/* 
+	 * Section "0" means no section when in legacy mode.
+	 * For some man.cgi scripts, "default" arch is none.
+	 */
 
-		(*kv)[(int)*kvsz].key = mandoc_strdup(key);
-		(*kv)[(int)*kvsz].val = mandoc_strdup(val);
-		(*kvsz)++;
+	if (req->q.legacy && NULL != req->q.sec)
+		if (0 == strcmp(req->q.sec, "0"))
+			req->q.sec = NULL;
+	if (req->q.legacy && NULL != req->q.arch)
+		if (0 == strcmp(req->q.arch, "default"))
+			req->q.arch = NULL;
+
+	/* Default to first manroot. */
+
+	if (NULL != manroot) {
+		for (i = 0; i < (int)req->psz; i++)
+			if (0 == strcmp(req->p[i].name, manroot))
+				break;
+		req->q.manroot = i < (int)req->psz ? i : -1;
 	}
 }
 
+static void
+http_putchar(char c)
+{
+
+	if (isalnum((unsigned char)c)) {
+		putchar((unsigned char)c);
+		return;
+	} else if (' ' == c) {
+		putchar('+');
+		return;
+	}
+	printf("%%%.2x", c);
+}
+
 /*
  * HTTP-decode a string.  The standard explanation is that this turns
  * "%4e+foo" into "n foo" in the regular way.  This is done in-place
  * over the allocated string.
  */
 static int
-kval_decode(char *p)
+http_decode(char *p)
 {
 	char             hex[3];
 	int              c;
@@ -277,9 +355,9 @@ resp_begin_http(int code, const char *msg)
 	if (200 != code)
 		printf("Status: %d %s\n", code, msg);
 
-	puts("Content-Type: text/html; charset=utf-8"		"\n"
-	     "Cache-Control: no-cache"				"\n"
-	     "Pragma: no-cache"					"\n"
+	puts("Content-Type: text/html; charset=utf-8\n"
+	     "Cache-Control: no-cache\n"
+	     "Pragma: no-cache\n"
 	     "");
 
 	fflush(stdout);
@@ -291,76 +369,75 @@ resp_begin_html(int code, const char *msg)
 
 	resp_begin_http(code, msg);
 
-	puts("<!DOCTYPE HTML PUBLIC "				"\n"
-	     " \"-//W3C//DTD HTML 4.01//EN\""			"\n"
-	     " \"http://www.w3.org/TR/html4/strict.dtd\">"	"\n"
-	     "<HTML>"						"\n"
-	     " <HEAD>"						"\n"
-	     "  <META HTTP-EQUIV=\"Content-Type\" "		"\n"
-	     "        CONTENT=\"text/html; charset=utf-8\">"	"\n"
-	     "  <LINK REL=\"stylesheet\" HREF=\"/man.cgi.css\""	"\n"
-	     "        TYPE=\"text/css\" media=\"all\">"		"\n"
-	     "  <TITLE>System Manpage Reference</TITLE>"	"\n"
-	     " </HEAD>"						"\n"
-	     " <BODY>"						"\n"
-	     "<!-- Begin page content. //-->");
+	printf("<!DOCTYPE HTML PUBLIC "
+	       " \"-//W3C//DTD HTML 4.01//EN\""
+	       " \"http://www.w3.org/TR/html4/strict.dtd\">\n"
+	       "<HTML>\n"
+	       "<HEAD>\n"
+	       "<META HTTP-EQUIV=\"Content-Type\""
+	       " CONTENT=\"text/html; charset=utf-8\">\n"
+	       "<LINK REL=\"stylesheet\" HREF=\"%s/man-cgi.css\""
+	       " TYPE=\"text/css\" media=\"all\">\n"
+	       "<LINK REL=\"stylesheet\" HREF=\"%s/man.css\""
+	       " TYPE=\"text/css\" media=\"all\">\n"
+	       "<TITLE>System Manpage Reference</TITLE>\n"
+	       "</HEAD>\n"
+	       "<BODY>\n"
+	       "<!-- Begin page content. //-->\n", css, css);
 }
 
 static void
 resp_end_html(void)
 {
 
-	puts(" </BODY>\n</HTML>");
+	puts("</BODY>\n"
+	     "</HTML>");
 }
 
 static void
 resp_searchform(const struct req *req)
 {
-	int	 	 i;
-	const char	*expr, *sec, *arch;
+	int		 i;
 
-	expr = sec = arch = "";
-
-	for (i = 0; i < (int)req->fieldsz; i++)
-		if (0 == strcmp(req->fields[i].key, "expr"))
-			expr = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "query"))
-			expr = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "sec"))
-			sec = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "sektion"))
-			sec = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "arch"))
-			arch = req->fields[i].val;
-
-	if (NULL != sec && 0 == strcmp(sec, "0"))
-		sec = NULL;
-
 	puts("<!-- Begin search form. //-->");
-	printf("<FORM ACTION=\"");
-	html_print(progname);
-	printf("/search.html\" METHOD=\"get\">\n");
-	printf("<FIELDSET>\n"
+	printf("<DIV ID=\"mancgi\">\n"
+	       "<FORM ACTION=\"%s/search.html\" METHOD=\"get\">\n"
+	       "<FIELDSET>\n"
 	       "<LEGEND>Search Parameters</LEGEND>\n"
-	       "<INPUT TYPE=\"submit\" NAME=\"op\" "
-	        "VALUE=\"Whatis\"> or \n"
-	       "<INPUT TYPE=\"submit\" NAME=\"op\" "
-	        "VALUE=\"apropos\"> for manuals satisfying \n"
-	       "<INPUT TYPE=\"text\" NAME=\"expr\" VALUE=\"");
-	html_print(expr);
+	       "<INPUT TYPE=\"submit\" "
+	       " VALUE=\"Search\"> for manuals satisfying \n"
+	       "<INPUT TYPE=\"text\" NAME=\"expr\" VALUE=\"",
+	       progname);
+	html_print(req->q.expr ? req->q.expr : "");
 	printf("\">, section "
-	       "<INPUT TYPE=\"text\" "
-	        "SIZE=\"4\" NAME=\"sec\" VALUE=\"");
-	html_print(sec);
+	       "<INPUT TYPE=\"text\""
+	       " SIZE=\"4\" NAME=\"sec\" VALUE=\"");
+	html_print(req->q.sec ? req->q.sec : "");
 	printf("\">, arch "
-	       "<INPUT TYPE=\"text\" "
-	        "SIZE=\"8\" NAME=\"arch\" VALUE=\"");
-	html_print(arch);
-	puts("\">.\n"
+	       "<INPUT TYPE=\"text\""
+	       " SIZE=\"8\" NAME=\"arch\" VALUE=\"");
+	html_print(req->q.arch ? req->q.arch : "");
+	printf("\">");
+	if (req->psz > 1) {
+		puts(", <SELECT NAME=\"manpath\">");
+		for (i = 0; i < (int)req->psz; i++) {
+			printf("<OPTION %s VALUE=\"",
+				(i == req->q.manroot) ||
+				(0 == i && -1 == req->q.manroot) ?
+				"SELECTED=\"selected\"" : "");
+			html_print(req->p[i].name);
+			printf("\">");
+			html_print(req->p[i].name);
+			puts("</OPTION>");
+		}
+		puts("</SELECT>");
+	}
+	puts(".\n"
 	     "<INPUT TYPE=\"reset\" VALUE=\"Reset\">\n"
 	     "</FIELDSET>\n"
 	     "</FORM>\n"
-	     "<!-- End search form. //-->");
+	     "</DIV>");
+	puts("<!-- End search form. //-->");
 }
 
 static void
@@ -379,9 +456,9 @@ resp_error400(void)
 	resp_begin_html(400, "Query Malformed");
 	printf("<H1>Malformed Query</H1>\n"
 	       "<P>\n"
-	       "  The query your entered was malformed.\n"
-	       "  Try again from the\n"
-	       "  <A HREF=\"%s/index.html\">main page</A>\n"
+	       "The query your entered was malformed.\n"
+	       "Try again from the\n"
+	       "<A HREF=\"%s/index.html\">main page</A>.\n"
 	       "</P>", progname);
 	resp_end_html();
 }
@@ -393,13 +470,13 @@ resp_error404(const char *page)
 	resp_begin_html(404, "Not Found");
 	puts("<H1>Page Not Found</H1>\n"
 	     "<P>\n"
-	     "  The page you're looking for, ");
-	printf("  <B>");
+	     "The page you're looking for, ");
+	printf("<B>");
 	html_print(page);
 	printf("</B>,\n"
-	       "  could not be found.\n"
-	       "  Try searching from the\n"
-	       "  <A HREF=\"%s/index.html\">main page</A>\n"
+	       "could not be found.\n"
+	       "Try searching from the\n"
+	       "<A HREF=\"%s/index.html\">main page</A>.\n"
 	       "</P>", progname);
 	resp_end_html();
 }
@@ -424,77 +501,64 @@ resp_baddb(void)
 static void
 resp_search(struct res *r, size_t sz, void *arg)
 {
-	int		 i, whatis;
-	const char	*ep, *sec, *arch;
+	size_t		 i, matched;
 	const struct req *req;
 
-	whatis = 1;
-	ep = sec = arch = NULL;
+	req = (const struct req *)arg;
 
-	if (1 == sz) {
+	if (sz > 0)
+		assert(req->q.manroot >= 0);
+
+	for (matched = i = 0; i < sz; i++)
+		if (r[i].matched)
+			matched++;
+	
+	if (1 == matched) {
+		for (i = 0; i < sz; i++)
+			if (r[i].matched)
+				break;
 		/*
 		 * If we have just one result, then jump there now
 		 * without any delay.
 		 */
 		puts("Status: 303 See Other");
-		printf("Location: http://%s%s/show/%u/%u.html\n",
-				host, progname,
-				r[0].volume, r[0].rec);
-		puts("Content-Type: text/html; charset=utf-8\n");
+		printf("Location: http://%s%s/show/%d/%u/%u.html?",
+				host, progname, req->q.manroot,
+				r[i].volume, r[i].rec);
+		http_printquery(req);
+		puts("\n"
+		     "Content-Type: text/html; charset=utf-8\n");
 		return;
 	}
 
-	req = (const struct req *)arg;
-
-	for (i = 0; i < (int)req->fieldsz; i++)
-		if (0 == strcmp(req->fields[i].key, "expr"))
-			ep = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "query"))
-			ep = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "sec"))
-			sec = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "sektion"))
-			sec = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "arch"))
-			arch = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "apropos"))
-			whatis = 0 == strcmp
-				(req->fields[i].val, "0");
-		else if (0 == strcmp(req->fields[i].key, "op"))
-			whatis = 0 == strcasecmp
-				(req->fields[i].val, "whatis");
-
-	qsort(r, sz, sizeof(struct res), cmp);
-
 	resp_begin_html(200, NULL);
 	resp_searchform(req);
 
-	if (0 == sz) {
+	puts("<DIV CLASS=\"results\">");
+
+	if (0 == matched) {
 		puts("<P>\n"
-		     "No results found.");
-		if (whatis) {
-			printf("(Try <A HREF=\"");
-			html_print(progname);
-			printf("/search.html?op=apropos&amp;expr=");
-			html_print(ep ? ep : "");
-			printf("&amp;sec=");
-			html_print(sec ? sec : "");
-			printf("&amp;arch=");
-			html_print(arch ? arch : "");
-			puts("\">apropos</A>?)");
-		}
-		puts("</P>");
+		     "No results found.\n"
+		     "</P>\n"
+		     "</DIV>");
 		resp_end_html();
 		return;
 	}
 
-	puts("<P></P>\n"
-	     "<TABLE>");
+	qsort(r, sz, sizeof(struct res), cmp);
 
-	for (i = 0; i < (int)sz; i++) {
-		printf("<TR><TD CLASS=\"title\"><A HREF=\"");
-		html_print(progname);
-		printf("/show/%u/%u.html\">", r[i].volume, r[i].rec);
+	puts("<TABLE>");
+
+	for (i = 0; i < sz; i++) {
+		if ( ! r[i].matched)
+			continue;
+		printf("<TR>\n"
+		       "<TD CLASS=\"title\">\n"
+		       "<A HREF=\"%s/show/%d/%u/%u.html?", 
+				progname, req->q.manroot,
+				r[i].volume, r[i].rec);
+		html_printquery(req);
+		printf("\">");
 		html_print(r[i].title);
 		putchar('(');
 		html_print(r[i].cat);
@@ -502,26 +566,29 @@ resp_search(struct res *r, size_t sz, void *arg)
 			putchar('/');
 			html_print(r[i].arch);
 		}
-		printf(")</A></TD><TD CLASS=\"desc\">");
+		printf(")</A>\n"
+		       "</TD>\n"
+		       "<TD CLASS=\"desc\">");
 		html_print(r[i].desc);
-		puts("</TD></TR>");
+		puts("</TD>\n"
+		     "</TR>");
 	}
 
-	puts("</TABLE>");
-
+	puts("</TABLE>\n"
+	     "</DIV>");
 	resp_end_html();
 }
 
 /* ARGSUSED */
 static void
-pg_index(const struct manpaths *ps, const struct req *req, char *path)
+pg_index(const struct req *req, char *path)
 {
 
 	resp_index(req);
 }
 
 static void
-catman(const char *file)
+catman(const struct req *req, const char *file)
 {
 	FILE		*f;
 	size_t		 len;
@@ -534,22 +601,11 @@ catman(const char *file)
 		return;
 	}
 
-	resp_begin_http(200, NULL);
-	puts("<!DOCTYPE HTML PUBLIC "				"\n"
-	     " \"-//W3C//DTD HTML 4.01//EN\""			"\n"
-	     " \"http://www.w3.org/TR/html4/strict.dtd\">"	"\n"
-	     "<HTML>"						"\n"
-	     " <HEAD>"						"\n"
-	     "  <META HTTP-EQUIV=\"Content-Type\" "		"\n"
-	     "        CONTENT=\"text/html; charset=utf-8\">"	"\n"
-	     "  <LINK REL=\"stylesheet\" HREF=\"/catman.css\""	"\n"
-	     "        TYPE=\"text/css\" media=\"all\">"		"\n"
-	     "  <TITLE>System Manpage Reference</TITLE>"	"\n"
-	     " </HEAD>"						"\n"
-	     " <BODY>"						"\n"
-	     "<!-- Begin page content. //-->");
+	resp_begin_html(200, NULL);
+	resp_searchform(req);
+	puts("<DIV CLASS=\"catman\">\n"
+	     "<PRE>");
 
-	puts("<PRE>");
 	while (NULL != (p = fgetln(f, &len))) {
 		bold = italic = 0;
 		for (i = 0; i < (int)len - 1; i++) {
@@ -660,6 +716,7 @@ catman(const char *file)
 	}
 
 	puts("</PRE>\n"
+	     "</DIV>\n"
 	     "</BODY>\n"
 	     "</HTML>");
 
@@ -667,7 +724,7 @@ catman(const char *file)
 }
 
 static void
-format(const char *file)
+format(const struct req *req, const char *file)
 {
 	struct mparse	*mp;
 	int		 fd;
@@ -675,14 +732,14 @@ format(const char *file)
 	struct man	*man;
 	void		*vp;
 	enum mandoclevel rc;
-	char		 opts[MAXPATHLEN + 128];
+	char		 opts[PATH_MAX + 128];
 
 	if (-1 == (fd = open(file, O_RDONLY, 0))) {
 		resp_baddb();
 		return;
 	}
 
-	mp = mparse_alloc(MPARSE_AUTO, MANDOCLEVEL_FATAL, NULL, NULL);
+	mp = mparse_alloc(MPARSE_AUTO, MANDOCLEVEL_FATAL, NULL, NULL, 0);
 	rc = mparse_readfd(mp, fd, file);
 	close(fd);
 
@@ -691,64 +748,107 @@ format(const char *file)
 		return;
 	}
 
-	snprintf(opts, sizeof(opts), "style=/man.css,"
-			"man=%s/search.html?sec=%%S&expr=%%N,"
+	snprintf(opts, sizeof(opts), "fragment,"
+			"man=%s/search.html?sec=%%S&expr=Nm~^%%N$,"
 			/*"includes=/cgi-bin/man.cgi/usr/include/%%I"*/,
 			progname);
 
 	mparse_result(mp, &mdoc, &man);
+	if (NULL == man && NULL == mdoc) {
+		resp_baddb();
+		mparse_free(mp);
+		return;
+	}
+
+	resp_begin_html(200, NULL);
+	resp_searchform(req);
+
 	vp = html_alloc(opts);
 
-	if (NULL != mdoc) {
-		resp_begin_http(200, NULL);
+	if (NULL != mdoc)
 		html_mdoc(vp, mdoc);
-	} else if (NULL != man) {
-		resp_begin_http(200, NULL);
+	else
 		html_man(vp, man);
-	} else
-		resp_baddb();
 
+	puts("</BODY>\n"
+	     "</HTML>");
+
 	html_free(vp);
 	mparse_free(mp);
 }
 
 static void
-pg_show(const struct manpaths *ps, const struct req *req, char *path)
+pg_show(const struct req *req, char *path)
 {
+	struct manpaths	 ps;
+	size_t		 sz;
 	char		*sub;
-	char		 file[MAXPATHLEN];
-	const char	*fn, *cp;
-	int		 rc;
-	unsigned int	 vol, rec;
+	char		 file[PATH_MAX];
+	const char	*cp;
+	int		 rc, catm;
+	unsigned int	 vol, rec, mr;
 	DB		*idx;
 	DBT		 key, val;
 
-	if (NULL == path) {
+	idx = NULL;
+
+	/* Parse out mroot, volume, and record from the path. */
+
+	if (NULL == path || NULL == (sub = strchr(path, '/'))) {
 		resp_error400();
 		return;
-	} else if (NULL == (sub = strrchr(path, '/'))) {
+	} 
+	*sub++ = '\0';
+	if ( ! atou(path, &mr)) {
 		resp_error400();
 		return;
-	} else
-		*sub++ = '\0';
-
-	if ( ! (atou(path, &vol) && atou(sub, &rec))) {
+	}
+	path = sub;
+	if (NULL == (sub = strchr(path, '/'))) {
 		resp_error400();
 		return;
-	} else if (vol >= (unsigned int)ps->sz) {
+	}
+	*sub++ = '\0';
+	if ( ! atou(path, &vol) || ! atou(sub, &rec)) {
 		resp_error400();
 		return;
+	} else if (mr >= (unsigned int)req->psz) {
+		resp_error400();
+		return;
 	}
 
-	strlcpy(file, ps->paths[vol], MAXPATHLEN);
-	strlcat(file, "/mandoc.index", MAXPATHLEN);
+	/*
+	 * Begin by chdir()ing into the manroot.
+	 * This way we can pick up the database files, which are
+	 * relative to the manpath root.
+	 */
 
+	if (-1 == chdir(req->p[(int)mr].path)) {
+		perror(req->p[(int)mr].path);
+		resp_baddb();
+		return;
+	}
+
+	memset(&ps, 0, sizeof(struct manpaths));
+	manpath_manconf(&ps, "etc/catman.conf");
+
+	if (vol >= (unsigned int)ps.sz) {
+		resp_error400();
+		goto out;
+	}
+
+	sz = strlcpy(file, ps.paths[vol], PATH_MAX);
+	assert(sz < PATH_MAX);
+	strlcat(file, "/", PATH_MAX);
+	strlcat(file, MANDOC_IDX, PATH_MAX);
+
 	/* Open the index recno(3) database. */
 
 	idx = dbopen(file, O_RDONLY, 0, DB_RECNO, NULL);
 	if (NULL == idx) {
+		perror(file);
 		resp_baddb();
-		return;
+		goto out;
 	}
 
 	key.data = &rec;
@@ -757,71 +857,77 @@ pg_show(const struct manpaths *ps, const struct req *r
 	if (0 != (rc = (*idx->get)(idx, &key, &val, 0))) {
 		rc < 0 ? resp_baddb() : resp_error400();
 		goto out;
-	} 
+	} else if (0 == val.size) {
+		resp_baddb();
+		goto out;
+	}
 
 	cp = (char *)val.data;
+	catm = 'c' == *cp++;
 
-	if (NULL == (fn = memchr(cp, '\0', val.size)))
+	if (NULL == memchr(cp, '\0', val.size - 1)) 
 		resp_baddb();
-	else if (++fn - cp >= (int)val.size)
-		resp_baddb();
-	else if (NULL == memchr(fn, '\0', val.size - (fn - cp)))
-		resp_baddb();
 	else {
-		strlcpy(file, cache, MAXPATHLEN);
-		strlcat(file, "/", MAXPATHLEN);
-		strlcat(file, fn, MAXPATHLEN);
-		if (0 == strcmp(cp, "cat"))
-			catman(file);
+ 		file[(int)sz] = '\0';
+ 		strlcat(file, "/", PATH_MAX);
+ 		strlcat(file, cp, PATH_MAX);
+		if (catm) 
+			catman(req, file);
 		else
-			format(file);
+			format(req, file);
 	}
 out:
-	(*idx->close)(idx);
+	if (idx)
+		(*idx->close)(idx);
+	manpath_free(&ps);
 }
 
 static void
-pg_search(const struct manpaths *ps, const struct req *req, char *path)
+pg_search(const struct req *req, char *path)
 {
-	size_t		  tt;
-	int		  i, sz, rc, whatis;
+	size_t		  tt, ressz;
+	struct manpaths	  ps;
+	int		  i, sz, rc;
 	const char	 *ep, *start;
+	struct res	*res;
 	char		**cp;
 	struct opts	  opt;
 	struct expr	 *expr;
 
-	expr = NULL;
-	cp = NULL;
-	ep = NULL;
-	sz = 0;
-	whatis = 1;
+	if (req->q.manroot < 0 || 0 == req->psz) {
+		resp_search(NULL, 0, (void *)req);
+		return;
+	}
 
 	memset(&opt, 0, sizeof(struct opts));
 
-	for (sz = i = 0; i < (int)req->fieldsz; i++)
-		if (0 == strcmp(req->fields[i].key, "expr"))
-			ep = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "query"))
-			ep = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "sec"))
-			opt.cat = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "sektion"))
-			opt.cat = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "arch"))
-			opt.arch = req->fields[i].val;
-		else if (0 == strcmp(req->fields[i].key, "apropos"))
-			whatis = 0 == strcmp
-				(req->fields[i].val, "0");
-		else if (0 == strcmp(req->fields[i].key, "op"))
-			whatis = 0 == strcasecmp
-				(req->fields[i].val, "whatis");
+	ep 	 = req->q.expr;
+	opt.arch = req->q.arch;
+	opt.cat  = req->q.sec;
+	rc 	 = -1;
+	sz 	 = 0;
+	cp	 = NULL;
+	ressz	 = 0;
+	res	 = NULL;
 
-	if (NULL != opt.cat && 0 == strcmp(opt.cat, "0"))
-		opt.cat = NULL;
+	/*
+	 * Begin by chdir()ing into the root of the manpath.
+	 * This way we can pick up the database files, which are
+	 * relative to the manpath root.
+	 */
 
+	assert(req->q.manroot < (int)req->psz);
+	if (-1 == (chdir(req->p[req->q.manroot].path))) {
+		perror(req->p[req->q.manroot].path);
+		resp_search(NULL, 0, (void *)req);
+		return;
+	}
+
+	memset(&ps, 0, sizeof(struct manpaths));
+	manpath_manconf(&ps, "etc/catman.conf");
+
 	/*
-	 * Poor man's tokenisation.
-	 * Just break apart by spaces.
+	 * Poor man's tokenisation: just break apart by spaces.
 	 * Yes, this is half-ass.  But it works for now.
 	 */
 
@@ -840,68 +946,91 @@ pg_search(const struct manpaths *ps, const struct req 
 			ep++;
 	}
 
-	rc = -1;
-
 	/*
 	 * Pump down into apropos backend.
 	 * The resp_search() function is called with the results.
 	 */
 
-	expr = whatis ? termcomp(sz, cp, &tt) :
-		        exprcomp(sz, cp, &tt);
+	expr = req->q.legacy ? 
+		termcomp(sz, cp, &tt) : exprcomp(sz, cp, &tt);
 
 	if (NULL != expr)
 		rc = apropos_search
-			(ps->sz, ps->paths, &opt,
-			 expr, tt, (void *)req, resp_search);
+			(ps.sz, ps.paths, &opt, expr, tt, 
+			 (void *)req, &ressz, &res, resp_search);
 
 	/* ...unless errors occured. */
 
 	if (0 == rc)
 		resp_baddb();
 	else if (-1 == rc)
-		resp_search(NULL, 0, (void *)req);
+		resp_search(NULL, 0, NULL);
 
 	for (i = 0; i < sz; i++)
 		free(cp[i]);
 
 	free(cp);
+	resfree(res, ressz);
 	exprfree(expr);
+	manpath_free(&ps);
 }
 
 int
 main(void)
 {
 	int		 i;
+	char		 buf[PATH_MAX];
+	DIR		*cwd;
 	struct req	 req;
 	char		*p, *path, *subpath;
-	struct manpaths	 paths;
 
-	/* HTTP init: read and parse the query string. */
+	/* Scan our run-time environment. */
 
-	progname = getenv("SCRIPT_NAME");
-	if (NULL == progname)
+	if (NULL == (cache = getenv("CACHE_DIR")))
+		cache = "/cache/man.cgi";
+
+	if (NULL == (progname = getenv("SCRIPT_NAME")))
 		progname = "";
 
-	cache = getenv("CACHE_DIR");
-	if (NULL == cache)
-		cache = "/cache/man.cgi";
+	if (NULL == (css = getenv("CSS_DIR")))
+		css = "";
 
+	if (NULL == (host = getenv("HTTP_HOST")))
+		host = "localhost";
+
+	/*
+	 * First we change directory into the cache directory so that
+	 * subsequent scanning for manpath directories is rooted
+	 * relative to the same position.
+	 */
+
 	if (-1 == chdir(cache)) {
+		perror(cache);
 		resp_bad();
 		return(EXIT_FAILURE);
-	}
+	} else if (NULL == (cwd = opendir(cache))) {
+		perror(cache);
+		resp_bad();
+		return(EXIT_FAILURE);
+	} 
 
-	host = getenv("HTTP_HOST");
-	if (NULL == host)
-		host = "localhost";
-
 	memset(&req, 0, sizeof(struct req));
 
+	strlcpy(buf, ".", PATH_MAX);
+	pathgen(cwd, buf, &req);
+	closedir(cwd);
+
+	/* Next parse out the query string. */
+
 	if (NULL != (p = getenv("QUERY_STRING")))
-		kval_parse(&req.fields, &req.fieldsz, p);
+		http_parse(&req, p);
 
-	/* Resolve leading subpath component. */
+	/*
+	 * Now juggle paths to extract information.
+	 * We want to extract our filetype (the file suffix), the
+	 * initial path component, then the trailing component(s).
+	 * Start with leading subpath component. 
+	 */
 
 	subpath = path = NULL;
 	req.page = PAGE__MAX;
@@ -932,31 +1061,29 @@ main(void)
 				break;
 			}
 
-	/* Initialise MANPATH. */
-
-	memset(&paths, 0, sizeof(struct manpaths));
-	manpath_manconf("etc/catman.conf", &paths);
-
 	/* Route pages. */
 
 	switch (req.page) {
 	case (PAGE_INDEX):
-		pg_index(&paths, &req, subpath);
+		pg_index(&req, subpath);
 		break;
 	case (PAGE_SEARCH):
-		pg_search(&paths, &req, subpath);
+		pg_search(&req, subpath);
 		break;
 	case (PAGE_SHOW):
-		pg_show(&paths, &req, subpath);
+		pg_show(&req, subpath);
 		break;
 	default:
 		resp_error404(path);
 		break;
 	}
 
-	manpath_free(&paths);
-	kval_free(req.fields, req.fieldsz);
+	for (i = 0; i < (int)req.psz; i++) {
+		free(req.p[i].path);
+		free(req.p[i].name);
+	}
 
+	free(req.p);
 	return(EXIT_SUCCESS);
 }
 
@@ -968,3 +1095,139 @@ cmp(const void *p1, const void *p2)
 				((const struct res *)p2)->title));
 }
 
+/*
+ * Check to see if an "etc" path consists of a catman.conf file.  If it
+ * does, that means that the path contains a tree created by catman(8)
+ * and should be used for indexing.
+ */
+static int
+pathstop(DIR *dir)
+{
+	struct dirent	*d;
+#if defined(__sun)
+	struct stat	 sb;
+#endif
+
+	while (NULL != (d = readdir(dir))) {
+#if defined(__sun)
+		stat(d->d_name, &sb);
+		if (S_IFREG & sb.st_mode)
+#else
+		if (DT_REG == d->d_type)
+#endif
+			if (0 == strcmp(d->d_name, "catman.conf"))
+				return(1);
+  }
+
+	return(0);
+}
+
+/*
+ * Scan for indexable paths.
+ * This adds all paths with "etc/catman.conf" to the buffer.
+ */
+static void
+pathgen(DIR *dir, char *path, struct req *req)
+{
+	struct dirent	*d;
+	char		*cp;
+	DIR		*cd;
+	int		 rc;
+	size_t		 sz, ssz;
+#if defined(__sun)
+	struct stat	 sb;
+#endif
+
+	sz = strlcat(path, "/", PATH_MAX);
+	if (sz >= PATH_MAX) {
+		fprintf(stderr, "%s: Path too long", path);
+		return;
+	} 
+
+	/* 
+	 * First, scan for the "etc" directory.
+	 * If it's found, then see if it should cause us to stop.  This
+	 * happens when a catman.conf is found in the directory.
+	 */
+
+	rc = 0;
+	while (0 == rc && NULL != (d = readdir(dir))) {
+#if defined(__sun)
+		stat(d->d_name, &sb);
+		if (!(S_IFDIR & sb.st_mode)
+#else
+		if (DT_DIR != d->d_type
+#endif
+        || strcmp(d->d_name, "etc"))
+			continue;
+
+		path[(int)sz] = '\0';
+		ssz = strlcat(path, d->d_name, PATH_MAX);
+
+		if (ssz >= PATH_MAX) {
+			fprintf(stderr, "%s: Path too long", path);
+			return;
+		} else if (NULL == (cd = opendir(path))) {
+			perror(path);
+			return;
+		} 
+		
+		rc = pathstop(cd);
+		closedir(cd);
+	}
+
+	if (rc > 0) {
+		/* This also strips the trailing slash. */
+		path[(int)--sz] = '\0';
+		req->p = mandoc_realloc
+			(req->p, 
+			 (req->psz + 1) * sizeof(struct paths));
+		/*
+		 * Strip out the leading "./" unless we're just a ".",
+		 * in which case use an empty string as our name.
+		 */
+		req->p[(int)req->psz].path = mandoc_strdup(path);
+		req->p[(int)req->psz].name = 
+			cp = mandoc_strdup(path + (1 == sz ? 1 : 2));
+		req->psz++;
+		/* 
+		 * The name is just the path with all the slashes taken
+		 * out of it.  Simple but effective. 
+		 */
+		for ( ; '\0' != *cp; cp++) 
+			if ('/' == *cp)
+				*cp = ' ';
+		return;
+	} 
+
+	/*
+	 * If no etc/catman.conf was found, recursively enter child
+	 * directory and continue scanning.
+	 */
+
+	rewinddir(dir);
+	while (NULL != (d = readdir(dir))) {
+#if defined(__sun)
+		stat(d->d_name, &sb);
+		if (!(S_IFDIR & sb.st_mode)
+#else
+		if (DT_DIR != d->d_type
+#endif
+        || '.' == d->d_name[0])
+			continue;
+
+		path[(int)sz] = '\0';
+		ssz = strlcat(path, d->d_name, PATH_MAX);
+
+		if (ssz >= PATH_MAX) {
+			fprintf(stderr, "%s: Path too long", path);
+			return;
+		} else if (NULL == (cd = opendir(path))) {
+			perror(path);
+			return;
+		}
+
+		pathgen(cd, path, req);
+		closedir(cd);
+	}
+}