Eclipse hanging
Tom Davies
while rebuilding the workspace. I'm using 3.2.2 M20070212-1330 and the
1.4 CAL plugin.
This is the message from the log file:
!SESSION 2007-05-16 07:51:19.223
-----------------------------------------------
eclipse.buildId=M20070212-1330
java.version=1.5.0_07
java.vendor=Apple Computer, Inc.
BootLoader constants: OS=macosx, ARCH=x86, WS=carbon, NL=en_US
Framework arguments: -keyring /Users/tomd/.eclipse_keyring -
showlocation
Command-line arguments: -os macosx -ws carbon -keyring /Users/
tomd/.eclipse_keyring -consoleLog -showlocation
!ENTRY org.eclipse.osgi 4 0 2007-05-16 07:52:26.887
!MESSAGE While loading class
"org.openquark.cal.eclipse.core.CALModelManager$3$1", thread
"Worker-4" timed out waiting (5000ms) for thread "Worker-5" to finish
starting bundle "org.openquark.cal.eclipse.core". To avoid deadlock,
thread "Worker-4" is proceeding but
"org.openquark.cal.eclipse.core.CALModelManager$3$1" may not be fully
initialized.
!STACK 0
java.lang.Exception: Generated exception.
at
org.eclipse.core.runtime.internal.adaptor.EclipseLazyStarter.preFindLocalClass(EclipseLazyStarter.java:
77)
at
org.eclipse.osgi.baseadaptor.loader.ClasspathManager.findLocalClass(ClasspathManager.java:
412)
at
org.eclipse.osgi.internal.baseadaptor.DefaultClassLoader.findLocalClass(DefaultClassLoader.java:
189)
at
org.eclipse.osgi.framework.internal.core.BundleLoader.findLocalClass(BundleLoader.java:
334)
at
org.eclipse.osgi.framework.internal.core.BundleLoader.findClass(BundleLoader.java:
386)
at
org.eclipse.osgi.framework.internal.core.BundleLoader.findClass(BundleLoader.java:
347)
at
org.eclipse.osgi.internal.baseadaptor.DefaultClassLoader.loadClass(DefaultClassLoader.java:
83)
at java.lang.ClassLoader.loadClass(ClassLoader.java:251)
at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:319)
at org.openquark.cal.eclipse.core.CALModelManager
$3.run(CALModelManager.java:665)
at org.eclipse.core.internal.jobs.Worker.run(Worker.java:58)
Let me know if I can provide any other information which would help
(e.g. my .metadata directory and projects?)
Tom
Edward
Thanks for the bug, and especially for providing salient details.
Based on the stack trace, I have a suspicion that it might not really
be our fault, though it may (or may not) be something for which we can
code a workaround.
I haven't done extensive testing on the Mac though -- this is
something that we'll try.
Do you mind telling me:
- If you start Eclipse again, does the error go away?
- Does the error happen regularly? For instance, can you make it
happen if you start Eclipse a few times?
Of course, reproduction steps would be nice if you have them.
Thanks,
- Edward
Tom Davies
On May 16, 11:38 am, Edward <edward....@gmail.com> wrote:
> Do you mind telling me:
> - If you start Eclipse again, does the error go away?
No -- it happens every time.
> - Does the error happen regularly? For instance, can you make it
> happen if you start Eclipse a few times?
I can use a workspace happily for hours, with many restarts of
Eclipse. Then something happens and that workspace is no good any
more. This has happened twice.
I can email you my workspace if you like -- it's only 6MB as a tgz
Regards,
Tom
Edward
hung Eclipse.
I'll try to reproduce the problem.
Edward
https://bugs.eclipse.org/bugs/show_bug.cgi?id=106812
When the CAL Eclipse Core plugin starts up, some classloading is
taking too long. This causes activation of the plugin to timeout,
followed by a general badness.
Apart from voting for this bug on Eclipse, I'll try to code around the
problem, as it seems pretty nasty.
Thanks,
- Edward
Tom Davies
it at my old workspace, and then reinstalling the CAL plugin...
Tom
Edward
I wasn't able to reproduce the problem, as it's a threading / timing
issue.
But the code and stack trace provide a lot of hints -- I've checked in
something which should work around the problem.
The fix will be available in the next release.
I had a look at the suspect class in Eclipse 3.3 RC1
(org.eclipse.core.runtime.internal.adaptor.EclipseLazyStarter) -- the
changes here include moving the timeout to a different part of the
code (and a different part of the loading sequence). Hopefully that
makes the problem completely go away -- but please let me know if
you're still having problems.
Thanks,
- Edward
Tom Davies
I'm having this problem with Quark 1.5.0 and Eclipse 3.3 -- but I
think it may now be a different problem.
The hanging thread looks like this:
"Worker-3" prio=5 tid=0x0113cec0 nid=0x19cb600 runnable
[0xb092d000..0xb092ed10]
at java.lang.ClassLoader$NativeLibrary.load(Native Method)
at java.lang.ClassLoader.loadLibrary0(ClassLoader.java:1751)
- locked <0x24bb44b8> (a java.util.Vector)
- locked <0x24bb44d0> (a java.util.Vector)
at java.lang.ClassLoader.loadLibrary(ClassLoader.java:1668)
at java.lang.Runtime.loadLibrary0(Runtime.java:822)
- locked <0x24bb3800> (a java.lang.Runtime)
at java.lang.System.loadLibrary(System.java:992)
at
sun.security.action.LoadLibraryAction.run(LoadLibraryAction.java:50)
at java.security.AccessController.doPrivileged(Native Method)
at java.awt.Toolkit.loadLibraries(Toolkit.java:1509)
at java.awt.Toolkit.<clinit>(Toolkit.java:1530)
at java.awt.Color.<clinit>(Color.java:250)
at java.lang.Class.forName0(Native Method)
at java.lang.Class.forName(Class.java:164)
at
org.openquark.cal.eclipse.core.CALEclipseCorePlugin.start(CALEclipseCorePlugin.java:
107)
at org.eclipse.osgi.framework.internal.core.BundleContextImpl
$2.run(BundleContextImpl.java:999)
So it seems to be hanging when it tries to construct a Color instance,
which causes awt to be initialized.
I don't understand the problem, but it has been mentioned here:
https://bugs.eclipse.org/bugs/show_bug.cgi?id=67384#c29
What I discovered I needed to do was to give eclipse the -
Dapple.awt.usingSWT=true argument.
See https://bugs.eclipse.org/bugs/show_bug.cgi?id=67384#c127
I haven't figured out if this causes any other problems, so you might
want to look at not serialising Color objects?
Tom
Tom
Edward
on OS X")