fully integrated regex patch

Now have removed the previous method, as really nobody should call
it any longer (and now nobody does ;)). Also did some other cleanup.

4 files changed, 19 insertions, 35 deletions

diff --git a/ChangeLog b/ChangeLog
index 42fda5b3..fc9f807f 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -34,6 +34,8 @@ Version 4.1.6  [DEVEL] (rgerhards), 2009-03-??
   the full high availability features of rsyslog's engine
 - bugfix: fixed some segaults on Solaris, where vsprintf() does not
   check for NULL pointers
+- improved performance of regexp-based filters
+  Thanks to Arnaud Cornet for providing the idea and initial patch.
 ---------------------------------------------------------------------------
 Version 4.1.5  [DEVEL] (rgerhards), 2009-03-11
 - bugfix: parser did not correctly parse fields in UDP-received messages
diff --git a/runtime/stringbuf.c b/runtime/stringbuf.c
index 4a7cc4bd..35ec44c6 100644
--- a/runtime/stringbuf.c
+++ b/runtime/stringbuf.c
@@ -703,49 +703,34 @@ int rsCStrCaseInsensitveStartsWithSzStr(cstr_t *pCS1, uchar *psz, size_t iLenSz)
  * never is a \0 *inside* a property string.
  * Note that the function returns -1 if regexp functionality is not available.
  * rgerhards: 2009-03-04: ERE support added, via parameter iType: 0 - BRE, 1 - ERE
+ * Arnaud Cornet/rgerhards: 2009-04-02: performance improvement by caching compiled regex
+ * If a caller does not need the cached version, it must still provide memory for it
+ * and must call rsCStrRegexDestruct() afterwards.
  */
-rsRetVal rsCStrSzStrMatchRegex(cstr_t *pCS1, uchar *psz, int iType)
+rsRetVal rsCStrSzStrMatchRegex(cstr_t *pCS1, uchar *psz, int iType, void *rc)
 {
-	regex_t preq;
+	regex_t **cache = (regex_t**) rc;
 	int ret;
 	DEFiRet;
 
-	if(objUse(regexp, LM_REGEXP_FILENAME) == RS_RET_OK) {
-		regexp.regcomp(&preq, (char*) rsCStrGetSzStr(pCS1), (iType == 1 ? REG_EXTENDED : 0) | REG_NOSUB);
-		ret = regexp.regexec(&preq, (char*) psz, 0, NULL, 0);
-		regexp.regfree(&preq);
-		if(ret != 0)
-			ABORT_FINALIZE(RS_RET_NOT_FOUND);
-	} else {
-		ABORT_FINALIZE(RS_RET_NOT_FOUND);
-	}
-
-finalize_it:
-	RETiRet;
-}
-
-/* same as above, only not braindead */
-int rsCStrSzStrMatchRegexCache(cstr_t *pCS1, uchar *psz, void *rc)
-{
-	int ret;
-	regex_t **cache = (regex_t**) rc;
-
-	BEGINfunc
-
+	assert(pCS1 != NULL);
+	assert(psz != NULL);
 	assert(cache != NULL);
 
 	if(objUse(regexp, LM_REGEXP_FILENAME) == RS_RET_OK) {
 		if (*cache == NULL) {
 			*cache = calloc(sizeof(regex_t), 1);
-			regexp.regcomp(*cache, (char*) rsCStrGetSzStr(pCS1), 0);
+			regexp.regcomp(*cache, (char*) rsCStrGetSzStr(pCS1), (iType == 1 ? REG_EXTENDED : 0) | REG_NOSUB);
 		}
 		ret = regexp.regexec(*cache, (char*) psz, 0, NULL, 0);
+		if(ret != 0)
+			ABORT_FINALIZE(RS_RET_NOT_FOUND);
 	} else {
-		ret = 1; /* simulate "not found" */
+		ABORT_FINALIZE(RS_RET_NOT_FOUND);
 	}
 
-	ENDfunc
-	return ret;
+finalize_it:
+	RETiRet;
 }
 
 
diff --git a/runtime/stringbuf.h b/runtime/stringbuf.h
index 311d7f41..684133bb 100644
--- a/runtime/stringbuf.h
+++ b/runtime/stringbuf.h
@@ -136,8 +136,7 @@ int rsCStrCaseInsensitiveLocateInSzStr(cstr_t *pThis, uchar *sz);
 int rsCStrStartsWithSzStr(cstr_t *pCS1, uchar *psz, size_t iLenSz);
 int rsCStrCaseInsensitveStartsWithSzStr(cstr_t *pCS1, uchar *psz, size_t iLenSz);
 int rsCStrSzStrStartsWithCStr(cstr_t *pCS1, uchar *psz, size_t iLenSz);
-rsRetVal rsCStrSzStrMatchRegex(cstr_t *pCS1, uchar *psz, int iType);
-int rsCStrSzStrMatchRegexCache(cstr_t *pCS1, uchar *psz, void *cache);
+rsRetVal rsCStrSzStrMatchRegex(cstr_t *pCS1, uchar *psz, int iType, void *cache);
 void rsCStrRegexDestruct(void *rc);
 rsRetVal rsCStrConvertToNumber(cstr_t *pStr, number_t *pNumber);
 rsRetVal rsCStrConvertToBool(cstr_t *pStr, number_t *pBool);
diff --git a/tools/syslogd.c b/tools/syslogd.c
index d5429855..b23c12a7 100644
--- a/tools/syslogd.c
+++ b/tools/syslogd.c
@@ -1077,15 +1077,13 @@ static rsRetVal shouldProcessThisMessage(selector_t *f, msg_t *pMsg, int *bProce
 				bRet = 1; /* process message! */
 			break;
 		case FIOP_REGEX:
-			//TODO REGEX: this needs to be merged with new functionality below
-			//rgerhards, 2009-04-02
-			if(rsCStrSzStrMatchRegexCache(f->f_filterData.prop.pCSCompValue,
-					(unsigned char*) pszPropVal, &f->f_filterData.prop.regex_cache) == 0)
+			if(rsCStrSzStrMatchRegex(f->f_filterData.prop.pCSCompValue,
+					(unsigned char*) pszPropVal, 0, &f->f_filterData.prop.regex_cache) == RS_RET_OK)
 				bRet = 1;
 			break;
 		case FIOP_EREREGEX:
 			if(rsCStrSzStrMatchRegex(f->f_filterData.prop.pCSCompValue,
-					  (unsigned char*) pszPropVal, 1) == RS_RET_OK)
+					  (unsigned char*) pszPropVal, 1, &f->f_filterData.prop.regex_cache) == RS_RET_OK)
 				bRet = 1;
 			break;
 		default: