REGRESSION in ffmpeg support in transcode (intrepid)
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
transcode (Ubuntu) |
Fix Released
|
Undecided
|
Unassigned | ||
Intrepid |
Fix Released
|
Undecided
|
Unassigned | ||
Jaunty |
Fix Released
|
Undecided
|
Unassigned |
Bug Description
Hello,
Please consider this SRU for Intrepid.
This is entirely my fault, I'm afraid. I can only apologise and offer a fix now. A patch I applied to fix an FTBFS in the Intrepid cycle had a typo in it. This causes transcode to fail to encode using ffmpeg in some circumstances (I haven't pinned down exactly which, but I do have a simple testcase).
TEST CASE
Using the intrepid version of transcode, 2:1.0.2-
transcode -i path/to/file.avi -y ffmpeg -F mjpeg -N 0x2000
This should transcode the file into mjpeg video and AC3 audio, and send both to /dev/null. Observe the following error:
transcode: symbol lookup error: /usr/lib/
Install the proposed SRU version, 2:1.0.2-
[transcode] encoded 11708 frames (0 dropped, 0 cloned), clip length 390.66 s
REGRESSION POTENTIAL
The functionality is totally broken at the moment, so any regression potential is minimal, barirng some unexpected interaction with other parts of the program. The benefits (being able to encode in more cases) outweigh these miniscule risks.
The bug will be fixed in Jaunty by upgrading to the new upstream version (bug #311202), which includes the correct version of this same patch.
Related branches
Changed in transcode: | |
assignee: | nobody → mrpouit |
status: | Confirmed → In Progress |
This is a typo-bug in the Patch 06_libavcodec. dpatch: dpatch/ dpatch- run patch.dpatch by Iain Lane <email address hidden> bugs.sourcemage .org/show_ bug.cgi? id=13866>
+++++++++++++
#! /bin/sh /usr/share/
## 06_libavcodec.
##
## All lines beginning with `## DP:' are a description of the patch.
## DP: transcode FTBFS due to libavcodec changes without this patch,
## DP: which was found at <http://
@DPATCH@ 1.0.2~/ export/ aud_aux. c transcode- 1.0.2/export/ aud_aux. c 1.0.2~/ export/ aud_aux. c 2005-07-04 08:09:31.000000000 +0100 1.0.2/export/ aud_aux. c 2008-07-08 21:26:01.000000000 +0100
diff -urNad transcode-
--- transcode-
+++ transcode-
@@ -361,8 +361,7 @@
pthread_ mutex_lock( &init_avcodec_ lock); init(); avcodec( &ac3_encoder) ; avcodec( &mp2_encoder) ; regisiter_ all(); mutex_unlock( &init_avcodec_ lock);
avcodec_
- register_
- register_
+ avcodec_
pthread_
switch (o_codec) {
+++++++++++++++++++
According to http:// bugs.sourcemage .org/show_ bug.cgi? id=13866 it should read: avcodec( &ac3_encoder) ; avcodec( &mp2_encoder) ; register_ all();
++++++++++++++++
- register_
- register_
+ avcodec_
++++++++++++++++
I wonder, why no other has experienced this bug?