Created attachment 371957 [details] [review]
patch

Review of attachment 371957 [details] [review]:

::: gst/subparse/gstsubparse.c
@@ +759,3 @@
 {
   gchar *cur, *next_tag;
+  GSList *open_tags = NULL;

GSlist sounds a bit suboptimal here, you kind of know that there shouldn't be so many elements (the 32 limit seems to have passed unnoticed for a long time :-)), so avoiding allocating all the time might be better. In practice it might not matter much at all :-) (If it does a GPtrArray could be better).

@@ +819,3 @@
         if (num_open_tags == 0
+            || !g_slist_find_custom (open_tags, tag_name,
+                (GCompareFunc) g_strcmp0)) {

Looks like you are not ignoring the case anymore?

@@ -823,3 +824,3 @@
           next_tag -= strlen (end_tag);
         } else {
-          --num_open_tags;
+          GSList *first = g_slist_nth (open_tags, 0);

Well, that is open_tags itself :-)

Created attachment 371963 [details] [review]
updated patch

Review of attachment 371963 [details] [review]:

::: gst/subparse/gstsubparse.c
@@ -818,3 +821,3 @@
         if (num_open_tags == 0
-            || g_ascii_strncasecmp (end_tag - 1, open_tags[num_open_tags - 1],
-                strlen (open_tags[num_open_tags - 1]))) {
+            || !g_ptr_array_find_with_equal_func (open_tags, tag_name,
+                g_str_equal, &index)) {

I always have doubt about the use of g_str_equal outside GHashTable, otherwise looks good I think.

Created attachment 371964 [details] [review]
updated patch

Review of attachment 371964 [details] [review]:

Better, Go :-)

Review of attachment 371964 [details] [review]:

Better, Go :-)