gst/subparse/gstsubparse.c: For SubRip (.srt) subtitles, ignore all markup tags we...

Original commit message from CVS:
* gst/subparse/gstsubparse.c: (subrip_remove_unhandled_tag),
(subrip_remove_unhandled_tags), (parse_subrip):
For SubRip (.srt) subtitles, ignore all markup tags we don't
handle (like font tags, for example).
* tests/check/elements/subparse.c:
Add test for this.

diff --git a/ChangeLog b/ChangeLog
index c37f870..3be8e44 100644 (file)
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,13 @@
+2007-02-10  Tim-Philipp Müller  <tim at centricular dot net>
+
+       * gst/subparse/gstsubparse.c: (subrip_remove_unhandled_tag),
+       (subrip_remove_unhandled_tags), (parse_subrip):
+         For SubRip (.srt) subtitles, ignore all markup tags we don't
+         handle (like font tags, for example).
+
+       * tests/check/elements/subparse.c:
+         Add test for this.
+
 2007-02-09  Tim-Philipp Müller  <tim at centricular dot net>
 
        * gst/playback/gstdecodebin.c: (add_fakesink),

diff --git a/gst/subparse/gstsubparse.c b/gst/subparse/gstsubparse.c
index 779e728..138ef46 100644 (file)
--- a/gst/subparse/gstsubparse.c
+++ b/gst/subparse/gstsubparse.c
@@ -547,6 +547,42 @@ subrip_unescape_formatting (gchar * txt)
   }
 }
 
+
+static gboolean
+subrip_remove_unhandled_tag (gchar * start, gchar * stop)
+{
+  gchar *tag, saved;
+
+  tag = start + strlen ("<");
+  if (*tag == '/')
+    ++tag;
+
+  if (g_ascii_tolower (*tag) < 'a' || g_ascii_tolower (*tag) > 'z')
+    return FALSE;
+
+  saved = *stop;
+  *stop = '\0';
+  GST_LOG ("removing unhandled tag '%s'", start);
+  *stop = saved;
+  g_memmove (start, stop, strlen (stop) + 1);
+  return TRUE;
+}
+
+/* remove tags we haven't explicitly allowed earlier on, like font tags
+ * for example */
+static void
+subrip_remove_unhandled_tags (gchar * txt)
+{
+  gchar *pos, *gt;
+
+  for (pos = txt; pos != NULL && *pos != '\0'; ++pos) {
+    if (strncmp (pos, "&lt;", 4) == 0 && (gt = strstr (pos + 4, "&gt;"))) {
+      if (subrip_remove_unhandled_tag (pos, gt + strlen ("&gt;")))
+        --pos;
+    }
+  }
+}
+
 /* we only allow <i>, <u> and <b>, so let's take a simple approach. This code
  * assumes the input has been escaped and subrip_unescape_formatting() has then
  * been run over the input! This function adds missing closing markup tags and
@@ -670,6 +706,7 @@ parse_subrip (ParserState * state, const gchar * line)
         g_string_truncate (state->buf, 0);
         state->state = 0;
         subrip_unescape_formatting (ret);
+        subrip_remove_unhandled_tags (ret);
         strip_trailing_newlines (ret);
         subrip_fix_up_markup (&ret);
         return ret;

diff --git a/tests/check/elements/subparse.c b/tests/check/elements/subparse.c
index 3cd7d3b..809f8d7 100644 (file)
--- a/tests/check/elements/subparse.c
+++ b/tests/check/elements/subparse.c
@@ -100,7 +100,10 @@ static SubParseInputChunk srt_input[] = {
         "26\n00:04:00,000 --> 00:05:00,000\n<i>Rock & Roll</i>\n\n",
       240 * GST_SECOND, 300 * GST_SECOND, "<i>Rock &amp; Roll</i>"}, {
         "27\n00:06:00,000 --> 00:08:00,000\nRock & Roll\n\n",
-      360 * GST_SECOND, 480 * GST_SECOND, "Rock &amp; Roll"}
+      360 * GST_SECOND, 480 * GST_SECOND, "Rock &amp; Roll"}, {
+        "28\n00:10:00,000 --> 00:11:00,000\n"
+        "<font \"#0000FF\"><joj>This is </xxx>in blue but <5</font>\n\n",
+      600 * GST_SECOND, 660 * GST_SECOND, "This is in blue but &lt;5"}
 };
 
 static void