[pacman-dev] [PATCH] _alpm_db_search-cleanup: remove needless regex compilation, string duplication and code cleanup

[Jürgen Hötzel]

Hi,

* avoid repeated regex compilations (regex for search string do not
  change while scanning the package database)

* remove needless string duplication (regex function do not change target
  string nor free them)

* code cleanup

This patch improves search performance:

bash-3.2$ time ./src/pacman/pacman.static.old -Ss "(database|web).*server" >/dev/null 

real    0m1.026s
user    0m0.544s
sys     0m0.208s
bash-3.2$ time ./src/pacman/pacman.static -Ss "(database|web).*server" >/dev/null 

real    0m0.777s
user    0m0.456s
sys     0m0.128s
bash-3.2$ 

Jürgen
-------------- next part --------------
Index: lib/libalpm/db.c
===================================================================
RCS file: /home/cvs-pacman/pacman-lib/lib/libalpm/db.c,v
retrieving revision 1.52
diff -u -r1.52 db.c
--- lib/libalpm/db.c	21 Dec 2006 01:53:41 -0000	1.52
+++ lib/libalpm/db.c	16 Jan 2007 22:39:42 -0000
@@ -37,6 +37,7 @@
 #include <sys/stat.h>
 #include <dirent.h>
 #include <libintl.h>
+#include <regex.h>
 #ifdef CYGWIN
 #include <limits.h> /* PATH_MAX */
 #endif
@@ -101,68 +102,47 @@
 
 	for(i = needles; i; i = i->next) {
 		char *targ;
-		int retval;
 
 		if(i->data == NULL) {
 			continue;
 		}
-		targ = strdup(i->data);
+		targ = i->data;
 		_alpm_log(PM_LOG_DEBUG, "searching for target '%s'", targ);
 
 		for(j = _alpm_db_get_pkgcache(db, INFRQ_DESC|INFRQ_DEPENDS); j; j = j->next) {
 			pmpkg_t *pkg = j->data;
-			char *haystack;
-			int match = 0;
-
-			/* check name */
-			haystack = strdup(pkg->name);
-			retval = _alpm_reg_match(haystack, targ);
-			if(retval < 0) {
+			char *matched = NULL;
+			regex_t reg;
+			
+			if(regcomp(&reg, targ, REG_EXTENDED | REG_NOSUB | REG_ICASE) != 0) {
 				/* bad regexp */
-				FREE(haystack);
-				return(NULL);
-			} else if(retval) {
-				_alpm_log(PM_LOG_DEBUG, "    search target '%s' matched '%s'", targ, haystack);
-				match = 1;
-			}
-			FREE(haystack);
-
-			/* check description */
-			if(!match) {
-				haystack = strdup(pkg->desc);
-				retval = _alpm_reg_match(haystack, targ);
-				if(retval < 0) {
-					/* bad regexp */
-					FREE(haystack);
-					return(NULL);
-				} else if(retval) {
-					match = 1;
-				}
-				FREE(haystack);
+				RET_ERR(PM_ERR_INVALID_REGEX, NULL);
 			}
 
+			/* check name */
+			if (regexec(&reg, pkg->name, 0, 0, 0) == 0) {
+				matched = pkg->name;
+			}
+			/* check desc */
+			else if (regexec(&reg, pkg->desc, 0, 0, 0) == 0) {
+				matched = pkg->desc;
+			}
 			/* check provides */
-			if(!match) {
+			else {
 				for(k = pkg->provides; k; k = k->next) {
-					haystack = strdup(k->data);
-					retval = _alpm_reg_match(haystack, targ);
-					if(retval < 0) {
-						/* bad regexp */
-						FREE(haystack);
-						return(NULL);
-					} else if(retval) {
-						match = 1;
+					if (regexec(&reg, k->data, 0, 0, 0) == 0) {
+						matched = k->data;
+						break;
 					}
-					FREE(haystack);
 				}
 			}
+			regfree(&reg);
 
-			if(match) {
+			if(matched != NULL) {
+				_alpm_log(PM_LOG_DEBUG, "    search target '%s' matched '%s'", targ, matched);
 				ret = _alpm_list_add(ret, pkg);
 			}
 		}
-
-		FREE(targ);
 	}
 
 	return(ret);
Index: lib/libalpm/util.c
===================================================================
RCS file: /home/cvs-pacman/pacman-lib/lib/libalpm/util.c,v
retrieving revision 1.35
diff -u -r1.35 util.c
--- lib/libalpm/util.c	22 Nov 2006 09:03:42 -0000	1.35
+++ lib/libalpm/util.c	16 Jan 2007 22:39:43 -0000
@@ -54,7 +54,6 @@
 #ifndef __sun__
 #include <mntent.h>
 #endif
-#include <regex.h>
 
 /* pacman */
 #include "log.h"
@@ -619,20 +618,6 @@
 }
 #endif
 
-/* match a string against a regular expression */
-int _alpm_reg_match(char *string, char *pattern)
-{
-	int result;
-	regex_t reg;
-
-	if(regcomp(&reg, pattern, REG_EXTENDED | REG_NOSUB | REG_ICASE) != 0) {
-		RET_ERR(PM_ERR_INVALID_REGEX, -1);
-	}
-	result = regexec(&reg, string, 0, 0, 0);
-	regfree(&reg);
-	return(!(result));
-}
-
 /* convert a time_t to a string - buffer MUST be large enough for
  * YYYYMMDDHHMMSS - 15 chars */
 void _alpm_time2string(time_t t, char *buffer)
Index: lib/libalpm/util.h
===================================================================
RCS file: /home/cvs-pacman/pacman-lib/lib/libalpm/util.h,v
retrieving revision 1.15
diff -u -r1.15 util.h
--- lib/libalpm/util.h	22 Nov 2006 09:03:42 -0000	1.15
+++ lib/libalpm/util.h	16 Jan 2007 22:39:43 -0000
@@ -70,7 +70,6 @@
 int _alpm_check_freespace(pmtrans_t *trans, pmlist_t **data);
 #endif
 #endif
-int _alpm_reg_match(char *string, char *pattern);
 void _alpm_time2string(time_t t, char *buffer);
 #ifdef __sun__
 char* strsep(char** str, const char* delims);
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://archlinux.org/pipermail/pacman-dev/attachments/20070117/f9bab49e/attachment.pgp>