Page MenuHomePhabricator
Create Task
Maniphest T16610

Cortado is broken on Java 6 with the server VM
Closed, ResolvedPublic
Actions

Description

Author: helmutpozimski

Description:
I compiled the latest svn-sourcecode of cortado using the Java 6 SDK supplied by the Debian repositories. Compiling works fine but after that it is not possible to play a video. Cortado buffers, starts playing and stops with messages like that:

java.lang.ArrayIndexOutOfBoundsException: 64

at com.fluendo.jheora.DCTDecode.ExpandToken(DCTDecode.java:531)
at com.fluendo.jheora.Decode.unpackAndExpandToken(Decode.java:455)
at com.fluendo.jheora.Decode.unPackVideo(Decode.java:598)
at com.fluendo.jheora.Decode.loadAndDecode(Decode.java:650)
at com.fluendo.jheora.State.decodePacketin(State.java:74)
at com.fluendo.plugin.TheoraDec$2.chainFunc(TheoraDec.java:211)
at com.fluendo.jst.Pad.chain(Pad.java:257)
at com.fluendo.jst.Pad.push(Pad.java:271)
at com.fluendo.plugin.Queue$1.taskFunc(Queue.java:135)
at com.fluendo.jst.Pad.run(Pad.java:339)
at java.lang.Thread.run(Unknown Source)

[INFO] [Message]: Pad: videodec:sink type: ERROR, 64
[INFO] got EOS Pad: videodec:sink
[INFO] Pad: videosink:sink got EOS
[INFO] Element: [pipeline] got EOS from sink: Element: [videosink]
[INFO] ogg: got EOS
[INFO] got EOS: Pad: v_queue:sink
[INFO] got EOS: Pad: a_queue:sink
[INFO] video pad removed Pad: :serial_1118493027
[INFO] audio pad removed Pad: :serial_1145241550
[INFO] cleanup
[INFO] Status: Show
[INFO] [Message]: Pad: v_queue:src type: STREAM_STATUS, stop, reason: error, flow stopped
[INFO] [Message]: Pad: buffer:src type: STREAM_STATUS, stop, reason: error, flow stopped
[INFO] [Message]: Pad: a_queue:src type: STREAM_STATUS, stop, reason: wrong-state, stopping
[INFO] [Message]: Pad: a_queue:src type: STREAM_STATUS, stop, reason: wrong-state, flow stopped
[INFO] [Message]: Pad: v_queue:src type: STREAM_STATUS, stop, reason: wrong-state, stopping
[INFO] [Message]: Pad: buffer:src type: STREAM_STATUS, stop, reason: wrong-state, stopping
[INFO] [Message]: Pad: httpsrc:src type: STREAM_STATUS, stop, reason: wrong-state, reason: wrong-state
[INFO] [Message]: Pad: httpsrc:src type: STREAM_STATUS, stop, reason: wrong-state, stopping

Again, with Java 5 (aka 1.5) everything works fine.

Im am using a Debian lenny system, amd64 architecture and Java 6 update 6

Version: unspecified
Severity: major
OS: Linux
Platform: PC

Details

Reference
bz14610

Event Timeline

• bzimport raised the priority of this task from to Medium.Nov 21 2014, 10:14 PM
• bzimport added a project: MediaWiki-extensions-OggHandler.
• bzimport set Reference to bz14610.
• bzimport created this task.Jun 22 2008, 3:32 PM
• bzimport added a comment.Jun 22 2008, 3:34 PM

helmutpozimski wrote:

Oh, I forgot something. Sample files:

http://pozimski.eu/itheora/data/bd1_pp.ogg
http://pozimski.eu/itheora/data/Parteitag.ogg
http://pozimski.eu/itheora/data/RedeDirk.ogg

tstarling added a comment.Jul 2 2008, 4:46 AM

bd1_pp.ogg works just fine on my 32-bit system. If someone could confirm this on a 64-bit system, that would be handy.

Have you tried using the precompiled jar file that we distribute with OggHandler?

http://svn.wikimedia.org/svnroot/mediawiki/trunk/extensions/OggHandler/cortado-ovt-stripped-wm_r31776.jar

• bzimport added a comment.Jul 2 2008, 12:52 PM

helmutpozimski wrote:

Yes, I tried it as well and it shows the same messages and doesn't work either on my system.

tstarling added a comment.Jul 3 2008, 12:47 AM

Confirmed under QEMU.

tstarling added a comment.Jul 3 2008, 5:55 AM

The problem appears to be the background compilation feature of JVM, not the architecture. I reproduced it in 32-bit mode and found it to be dependent on JVM options:

works: appletviewer -J-client test.html
broken: appletviewer -J-server test.html
works: appletviewer -J-server -J-Xbatch test.html

Apparently if you're on a 64-bit architecture, it always uses the server VM.

I've submitted this to the Java bug tracker, now waiting for review ID 1284931.

• bzimport added a comment.Jul 31 2008, 5:02 PM

matt.ingenthron wrote:

A bug against the compiler has been filed under 6732194. This should be posted to http://bugs.sun.com shortly. Any updates on the bug should be visible there. It's currently in the queue for evaluation.

tstarling added a comment.Aug 6 2008, 2:49 PM

I've applied the workaround suggested on the Sun bug report, updated the jar in OggHandler SVN, and deployed it on Wikimedia. It works for me with -J-server, I assume it works on 64-bit too. My thanks to Matt for getting this to the right people.

• bzimport added a comment.Aug 6 2008, 3:15 PM

helmutpozimski wrote:

I compiled it myself and I can confirm it works now even on 64-bit. Thank you very much for your great work.

• bzimport added a comment.Sep 12 2008, 5:15 PM

matt.ingenthron wrote:

The bug has now been fixed in OpenJDK/JDK 7 development, so once it's mainstream, the workaround should no longer be required. It shouldn't hurt anything to have the workaround in there though.

Content licensed under Creative Commons Attribution-ShareAlike (CC BY-SA) 4.0 unless otherwise noted; code licensed under GNU General Public License (GPL) 2.0 or later and other open source licenses. By using this site, you agree to the Terms of Use, Privacy Policy, and Code of Conduct. · Wikimedia Foundation · Privacy Policy · Code of Conduct · Terms of Use · Disclaimer · CC-BY-SA · GPL