Add support for subtitle files with UTF-8 BOM at the beginning by simple stripping...

Original commit message from CVS:
* gst/subparse/gstsubparse.c:
(gst_sub_parse_data_format_autodetect), (handle_buffer),
(gst_sub_parse_change_state):
* gst/subparse/gstsubparse.h:
* tests/check/elements/subparse.c: (GST_START_TEST):
Add support for subtitle files with UTF-8 BOM at the beginning
by simple stripping it from the first line before passing it
to any parsing code. Fixes bug #555257 and playback of files
created by Gnome Subtitles.

diff --git a/ChangeLog b/ChangeLog
index fc26d8f..4bd31b1 100644 (file)
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,15 @@
+2008-10-10  Sebastian Dröge  <sebastian.droege@collabora.co.uk>
+
+       * gst/subparse/gstsubparse.c:
+       (gst_sub_parse_data_format_autodetect), (handle_buffer),
+       (gst_sub_parse_change_state):
+       * gst/subparse/gstsubparse.h:
+       * tests/check/elements/subparse.c: (GST_START_TEST):
+       Add support for subtitle files with UTF-8 BOM at the beginning
+       by simple stripping it from the first line before passing it
+       to any parsing code. Fixes bug #555257 and playback of files
+       created by Gnome Subtitles.
+
 2008-10-10  Wim Taymans  <wim.taymans@collabora.co.uk>
 
        * gst/audiotestsrc/gstaudiotestsrc.c:

diff --git a/gst/subparse/gstsubparse.c b/gst/subparse/gstsubparse.c
index dc2a345..5fbcfdd 100644 (file)
--- a/gst/subparse/gstsubparse.c
+++ b/gst/subparse/gstsubparse.c
@@ -922,6 +922,11 @@ gst_sub_parse_data_format_autodetect (gchar * match_str)
     }
   }
 
+  /* If the string contains a UTF-8 BOM drop it */
+  if ((guint8) match_str[0] == 0xEF && (guint8) match_str[1] == 0xBB
+      && (guint8) match_str[2] == 0xBF)
+    match_str += 3;
+
   if (regexec (&mdvd_rx, match_str, 0, NULL, 0) == 0) {
     GST_LOG ("MicroDVD (frame based) format detected");
     return GST_SUB_PARSE_FORMAT_MDVDSUB;
@@ -1073,11 +1078,22 @@ handle_buffer (GstSubParse * self, GstBuffer * buf)
   }
 
   while ((line = get_next_line (self)) && !self->flushing) {
+    guint offset = 0;
+
+    /* If this is the first line and it contains a UTF-8 BOM drop it */
+    if (self->first_line && strlen (line) >= 3 &&
+        (guint8) line[0] == 0xEF && (guint8) line[1] == 0xBB
+        && (guint8) line[2] == 0xBF) {
+      offset = 3;
+    }
+
+    self->first_line = FALSE;
+
     /* Set segment on our parser state machine */
     self->state.segment = &self->segment;
     /* Now parse the line, out of segment lines will just return NULL */
-    GST_LOG_OBJECT (self, "Parsing line '%s'", line);
-    subtitle = self->parse_line (&self->state, line);
+    GST_LOG_OBJECT (self, "Parsing line '%s'", line + offset);
+    subtitle = self->parse_line (&self->state, line + offset);
     g_free (line);
 
     if (subtitle) {
@@ -1252,6 +1268,7 @@ gst_sub_parse_change_state (GstElement * element, GstStateChange transition)
       self->next_offset = 0;
       self->parser_type = GST_SUB_PARSE_FORMAT_UNKNOWN;
       self->valid_utf8 = TRUE;
+      self->first_line = TRUE;
       g_string_truncate (self->textbuf, 0);
       break;
     default:

diff --git a/gst/subparse/gstsubparse.h b/gst/subparse/gstsubparse.h
index 86bf301..510c3e8 100644 (file)
--- a/gst/subparse/gstsubparse.h
+++ b/gst/subparse/gstsubparse.h
@@ -93,6 +93,8 @@ struct _GstSubParse {
   gboolean flushing;
   gboolean valid_utf8;
   gchar   *encoding;
+
+  gboolean first_line;
 };
 
 struct _GstSubParseClass {

diff --git a/tests/check/elements/subparse.c b/tests/check/elements/subparse.c
index 05e3619..56beab4 100644 (file)
--- a/tests/check/elements/subparse.c
+++ b/tests/check/elements/subparse.c
@@ -132,6 +132,13 @@ static SubParseInputChunk srt_input1[] = {
       44 * GST_SECOND + 44 * GST_MSECOND, "I still cant see anything."}
 };
 
+/* has UTF-8 BOM at the start */
+static SubParseInputChunk srt_input2[] = {
+  {
+        "\xef\xbb\xbf" "1\n00:00:00,000 --> 00:00:03,50\nJust testing.\n\n",
+      0, 3 * GST_SECOND + 50 * GST_MSECOND, "Just testing."}
+};
+
 static void
 setup_subparse (void)
 {
@@ -237,6 +244,9 @@ GST_START_TEST (test_srt)
 
   /* try with spaces instead of doubled zeroes (which is not exactly according to spec) */
   test_srt_do_test (srt_input1, 0, G_N_ELEMENTS (srt_input1));
+
+  /* try with UTF-8 BOM at the start */
+  test_srt_do_test (srt_input1, 0, G_N_ELEMENTS (srt_input2));
 }
 
 GST_END_TEST;