Modified to only match text MIME parts in the message rather than the

2000-09-13  Jeffrey Stedfast  <fejj@helixcode.com>

	* filter-message-search.c (body_contains): Modified to only match
	text MIME parts in the message rather than the entire MIME message
	(header_contains): Added regex matching

svn path=/trunk/; revision=5406

2 files changed, 131 insertions, 36 deletions

diff --git a/filter/ChangeLog b/filter/ChangeLog
index 06284d16e5..0cdd6d20fd 100644
--- a/filter/ChangeLog
+++ b/filter/ChangeLog
@@ -1,3 +1,9 @@
+2000-09-13  Jeffrey Stedfast  <fejj@helixcode.com>
+
+	* filter-message-search.c (body_contains): Modified to only match
+	text MIME parts in the message rather than the entire MIME message
+	(header_contains): Added regex matching
+
 2000-09-12  Dan Winship  <danw@helixcode.com>
 
 	* rule-context.c (d): 
diff --git a/filter/filter-message-search.c b/filter/filter-message-search.c
index 9730deb61e..9dc051dee4 100644
--- a/filter/filter-message-search.c
+++ b/filter/filter-message-search.c
@@ -69,11 +69,38 @@ header_contains (struct _ESExp *f, int argc, struct _ESExpResult **argv, FilterM
 	if (argc == 2) {
 		char *header = (argv[0])->value.string;
 		char *match = (argv[1])->value.string;
+		regex_t regexpat;        /* regex patern */
+		regmatch_t *fltmatch;
+		gint regerr = 0;
+		size_t reglen = 0;
+		gchar *regmsg;
 		const char *contents;
 		
 		contents = camel_medium_get_header (CAMEL_MEDIUM (fms->message), header);
-		if (contents && e_strstrcase (contents, match))
-			matched = TRUE;
+		
+		regerr = regcomp (&regexpat, match, REG_EXTENDED | REG_NEWLINE | REG_ICASE);
+		if (regerr) {
+			/* regerror gets called twice to get the full error string 
+			   length to do proper posix error reporting */
+			reglen = regerror (regerr, &regexpat, 0, 0);
+			regmsg = g_malloc0 (reglen + 1);
+			regerror (regerr, &regexpat, regmsg, reglen);
+			camel_exception_setv (fms->ex, CAMEL_EXCEPTION_SYSTEM,
+					      "Failed to perform regex search on message body: %s",
+					      regmsg);
+			g_free (regmsg);
+			regfree (&regexpat);
+		} else {
+			if (contents) {
+				fltmatch = g_new0 (regmatch_t, regexpat.re_nsub);
+				
+				if (!regexec (&regexpat, contents, regexpat.re_nsub, fltmatch, 0))
+					matched = TRUE;
+				
+				g_free (fltmatch);
+				regfree (&regexpat);
+			}
+		}
 	}
 	
 	r = e_sexp_result_new (ESEXP_RES_BOOL);
@@ -85,7 +112,6 @@ header_contains (struct _ESExp *f, int argc, struct _ESExpResult **argv, FilterM
 static ESExpResult *
 match_all (struct _ESExp *f, int argc, struct _ESExpTerm **argv, FilterMessageSearch *fms)
 {
-	/* FIXME: is this right? I dunno */
 	/* match-all: when dealing with single messages is a no-op */
 	ESExpResult *r;
 	
@@ -98,6 +124,90 @@ match_all (struct _ESExp *f, int argc, struct _ESExpTerm **argv, FilterMessageSe
 	return r;
 }
 
+static gboolean
+mime_part_matches (CamelMimePart *mime_part, const char *match, CamelException *ex)
+{
+	CamelStream *stream;
+	GByteArray *array;
+	char *text;
+	regex_t regexpat;        /* regex patern */
+	regmatch_t *fltmatch;
+	gint regerr = 0;
+	size_t reglen = 0;
+	gchar *regmsg;
+	gboolean matched = FALSE;
+	
+	array = g_byte_array_new ();
+	stream = camel_stream_mem_new_with_byte_array (array);
+	camel_data_wrapper_write_to_stream (CAMEL_DATA_WRAPPER (mime_part), stream);
+	camel_stream_reset (stream);
+	g_byte_array_append (array, "", 1);
+	
+	text = array->data;
+	
+	regerr = regcomp (&regexpat, match, REG_EXTENDED | REG_NEWLINE | REG_ICASE);
+	if (regerr) {
+		/* regerror gets called twice to get the full error string 
+		   length to do proper posix error reporting */
+		reglen = regerror (regerr, &regexpat, 0, 0);
+		regmsg = g_malloc0 (reglen + 1);
+		regerror (regerr, &regexpat, regmsg, reglen);
+		camel_exception_setv (ex, CAMEL_EXCEPTION_SYSTEM,
+				      "Failed to perform regex search on message body: %s",
+				      regmsg);
+		g_free (regmsg);
+		regfree (&regexpat);
+	} else {
+		fltmatch = g_new0 (regmatch_t, regexpat.re_nsub);
+		
+		if (!regexec (&regexpat, text, regexpat.re_nsub, fltmatch, 0))
+			matched = TRUE;
+		
+		g_free (fltmatch);
+		regfree (&regexpat);
+	}
+	
+	return matched;
+}
+
+static gboolean
+handle_multipart (CamelMultipart *multipart, const char *match, CamelException *ex)
+{
+	gboolean matched = FALSE;
+	int i, nparts;
+	
+	nparts = camel_multipart_get_number (multipart);
+	
+	for (i = 0; i < nparts && !matched; i++) {
+		GMimeContentField *content;
+		CamelMimePart *mime_part;
+		
+		mime_part = camel_multipart_get_part (multipart, i);
+		content = camel_mime_part_get_content_type (mime_part);
+		
+		if (gmime_content_field_is_type (content, "text", "*")) {
+			/* we only want to match text parts */
+			matched = mime_part_matches (mime_part, match, ex);
+			
+			if (camel_exception_is_set (ex))
+				break;
+		} else if (gmime_content_field_is_type (content, "multipart", "*")) {
+			CamelDataWrapper *wrapper;
+			CamelMultipart *mpart;
+			
+			wrapper = camel_medium_get_content_object (CAMEL_MEDIUM (mime_part));
+			mpart = CAMEL_MULTIPART (wrapper);
+			
+			matched = handle_multipart (mpart, match, ex);
+			
+			if (camel_exception_is_set (ex))
+				break;
+		}
+	}
+	
+	return matched;
+}
+
 static ESExpResult *
 body_contains (struct _ESExp *f, int argc, struct _ESExpResult **argv, FilterMessageSearch *fms)
 {
@@ -105,45 +215,24 @@ body_contains (struct _ESExp *f, int argc, struct _ESExpResult **argv, FilterMes
 	ESExpResult *r;
 	
 	if (argc > 0) {
-		CamelStream *stream;
-		GByteArray *array;
-		char *match, *body;
-		regex_t regexpat;        /* regex patern */
-		regmatch_t *fltmatch;
-		gint regerr = 0;
-		size_t reglen = 0;
-		gchar *regmsg;
+		GMimeContentField *content;
+		char *match;
 		
 		match = (*argv)->value.string;
 		
-		array = g_byte_array_new ();
-		stream = camel_stream_mem_new_with_byte_array (array);
-		camel_data_wrapper_write_to_stream (CAMEL_DATA_WRAPPER (fms->message), stream);
-		camel_stream_reset (stream);
-		g_byte_array_append (array, "", 1);
+		content = camel_mime_part_get_content_type (CAMEL_MIME_PART (fms->message));
 		
-		body = array->data;
-		
-		regerr = regcomp (&regexpat, match, REG_EXTENDED | REG_NEWLINE | REG_ICASE);
-		if (regerr) {
-			/* regerror gets called twice to get the full error string 
-			   length to do proper posix error reporting */
-			reglen = regerror (regerr, &regexpat, 0, 0);
-			regmsg = g_malloc0 (reglen + 1);
-			regerror (regerr, &regexpat, regmsg, reglen);
-			camel_exception_setv (fms->ex, CAMEL_EXCEPTION_SYSTEM,
-					      "Failed to perform regex search on message body: %s",
-					      regmsg);
-			g_free (regmsg);
-			regfree (&regexpat);
-		} else {
-			fltmatch = g_new0 (regmatch_t, regexpat.re_nsub);
+		if (gmime_content_field_is_type (content, "multipart", "*")) {
+			CamelDataWrapper *wrapper;
+			CamelMultipart *multipart;
 			
-			if (!regexec (&regexpat, body, regexpat.re_nsub, fltmatch, 0))
-				matched = TRUE;
+			wrapper = camel_medium_get_content_object (CAMEL_MEDIUM (CAMEL_MIME_PART (fms->message)));
+			multipart = CAMEL_MULTIPART (wrapper);
 			
-			g_free (fltmatch);
-			regfree (&regexpat);
+			matched = handle_multipart (multipart, match, fms->ex);
+		} else {
+			/* single-part message so just search the entire message */
+			matched = mime_part_matches (CAMEL_MIME_PART (fms->message), match, fms->ex);
 		}
 	}