From Kovarththanan Rajaratnam via bug 3702:

This patch optimizes the data source name processing in add_new_data_source()
by delaying it. We now simply store the constant string and lazily compute the
name when needed. This gives a performance boost because we only need the name
if we have multiple data sources.


git-svn-id: http://anonsvn.wireshark.org/wireshark/trunk@29066 f5534014-38df-0310-8fa8-9805f1628bb7

1 files changed, 2 insertions, 2 deletions

diff --git a/print.c b/print.c
index 580ba1ff88..d9178a842d 100644
--- a/print.c
+++ b/print.c
@@ -771,7 +771,7 @@ print_hex_data(print_stream_t *stream, epan_dissect_t *edt)
 	GSList *src_le;
 	data_source *src;
 	tvbuff_t *tvb;
-	char *name;
+	const char *name;
 	char *line;
 	const guchar *cp;
 	guint length;
@@ -789,7 +789,7 @@ print_hex_data(print_stream_t *stream, epan_dissect_t *edt)
 		src = src_le->data;
 		tvb = src->tvb;
 		if (multiple_sources) {
-			name = src->name;
+			name = get_data_source_name(src);
 			print_line(stream, 0, "");
 			line = g_strdup_printf("%s:", name);
 			print_line(stream, 0, line);