Page MenuHomePhabricator

OggHandler Cortado stop button doesn't work right, makes java go out to lunch.
Closed, ResolvedPublic

Description

Author: gmaxwell

Description:
The stop button does not appear to work correctly and causes misbehavior in Firefox with the cortado player using Sun JVM 1.5 or 1.6. (confirmed on Linux with JVM 1.5 and windows with JVM 1.6)

I'll give detailed reproduction instructions but you can reproduce this with pretty much anything:

Go to http://en.wikipedia.org/wiki/Bassoon#Audio_examples Play the first example. When it finishes. Hit stop. Notice that java goes out to lunch and no player work for 10 seconds or so. Scrolling the screen makes the player iframes go blank.

Now go to
http://tools.wikimedia.de/~gmaxwell/jorbis/commonsJOrbisPlayer.php?path=Bassoon-technical-range.ogg

Observe that the same behavior doesn't happen.

I think this is also related to the player intermittently going out to lunch before it starts but thats harder to reproduce.


Version: unspecified
Severity: major

Details

Reference
bz11245

Event Timeline

bzimport raised the priority of this task from to Medium.Nov 21 2014, 9:53 PM
bzimport set Reference to bz11245.

Can't reproduce. What sort of lunch? Sandwich?

gmaxwell wrote:

Mmm.. What kind of lunch? Something stale no doubt....

The issue is 100% reproducible for me, and apparently some other folks... Shall I send you a VMware player image with a Linux install that experiences this? ;) I'm not sure what to say.. if I use the patched jar that OggHander uses on my external player it gives the same behavior. http://tools.wikimedia.de/~gmaxwell/jorbis/commonsJOrbisPlayer-ohjar.php?path=Bassoon-technical-range.ogg

I also get peroidic clicking in the output audio when I use the OggHandler jar file. The clicking doesn't happen with the jar I've been using.

What did you compile that jar with? I used Sun javac 1.6.0 with -source and -target set to um.. I think 1.2.. whichever was the lowest version number that didn't have issues compiling the code.

gmaxwell wrote:

My primary test case is w/ Java(TM) Plug-in 1.5.0_09-b03 on "Mozilla/5.0 (X11; U; Linux i686 (x86_64); en-US; rv:1.8.0.10) Gecko/20070510 Fedora/1.5.0.10-6.fc6 Firefox/1.5.0.10". Also had someone else test with FF2 on Windows XP with Java(TM) Platform SE 6 U1.

gmaxwell wrote:

I recompiled the plugin with our patch using sun javac 1.6 and ant. http://tools.wikimedia.de/~gmaxwell/jorbis/cortado-new.jar It doesn't have the problem. I've asked tim to replace the copy on our site.

gmaxwell wrote:

The new jar was not displaying the status bar when status height = applet height. This was as coded and should have never worked. I'm guessing it worked at all because slight differences between swing and awt behavior exposed by building against the wrong target version.

I changed the code to allow them to be equal.. since thats actually harmless. The jar compiled with this change is http://tools.wikimedia.de/~gmaxwell/jorbis/cortado-new2.jar

  • ./src/com/fluendo/player/Cortado.java 2007-09-09 03:05:45.693764000 +0000

+++ ./src/com/fluendo/player/Cortado.java~ 2007-09-09 03:05:45.678859000 +0000
@@ -302,7 +302,7 @@

int dheight = getSize().height;
 
/* sometimes dimension is wrong */
  • if (dwidth <= 0 || dheight < statusHeight) {

+ if (dwidth <= 0 || dheight <= statusHeight) {

  appletDimension = null;
  return;
}

gmaxwell wrote:

(above patch is backwards ;) )

gmaxwell wrote:

Domas pushed out the new jar. All is good.