Community
Participate
Working Groups
I20070911-0833 I can't build my workspace anymore in the latest IBuild: The progress hangs at 61% and progress message cycles from cleaning, analyzing, compiling and back. It seams to reset always in the same package (eclipse/jdt/ui/wizards) I can reproduce with a 1.5 and a 1.6 vm. Nothing in the log. This happened after switching to the newest IBuild on start. I did turn off auto build, clean all, and I was able to build the workspace. If I close jdt/ui project with autobuild enabled it does hang again, I have to clean all again. This is a blocker for me I have to go back to the last IBuild. Mmm, interesting: When starting last weeks IBuild it says: !MESSAGE The workspace exited with unsaved changes in the previous session; refreshing workspace to recover changes. Although that's not true.
This must be caused by the change for bug 89301. Somehow the project's builder must think its been interrupted. I'll remove the change for now. Benno - how many projects were in your workspace ? Was it in sync with CVS before you switched to the new I build ?
(In reply to comment #1) > Benno - how many projects were in your workspace ? 67 > Was it in sync with CVS before you switched to the new I build ? No
All we're doing is checking to see if the project's builder believes it has been interrupted because of a change to one of its resources. John - what kind of resource changes could cause a project to constantly think its been interrupted?
Do you have a reproducible case? There is a .options file flag for tracking this down: # Reports the cause of autobuild interruption org.eclipse.core.resources/build/interrupt=true org.eclipse.core.resources/debug=true
Benno - please try your workspace with the latest I-build again with the options enabled. I've tried my JDT/UI workspace & its fine. We also haven't heard from any other developers that this is a problem for them.
(In reply to comment #4) > Do you have a reproducible case? Yes >There is a .options file flag for tracking this down: > > # Reports the cause of autobuild interruption > org.eclipse.core.resources/build/interrupt=true > org.eclipse.core.resources/debug=true > Starting application: 2311 Application Started: 14349 Autobuild was interrupted: java.lang.Exception at org.eclipse.core.internal.events.AutoBuildJob.interrupt(AutoBuildJob.java:187) at org.eclipse.core.internal.events.BuildManager.interrupt(BuildManager.java:792) at org.eclipse.core.internal.resources.Workspace.prepareOperation(Workspace.java:1742) at org.eclipse.core.internal.resources.Workspace.run(Workspace.java:1792) at org.polarion.team.svn.core.utility.ProgressMonitorUtility$1.run(ProgressMonitorUtility.java:59) at org.eclipse.core.internal.jobs.Worker.run(Worker.java:55) Scheduling rebuild due to interruption Autobuild was interrupted: java.lang.Exception at org.eclipse.core.internal.events.AutoBuildJob.interrupt(AutoBuildJob.java:187) at org.eclipse.core.internal.events.BuildManager.interrupt(BuildManager.java:792) at org.eclipse.core.internal.resources.Workspace.prepareOperation(Workspace.java:1742) at org.eclipse.core.internal.resources.Workspace.run(Workspace.java:1792) at org.polarion.team.svn.core.utility.ProgressMonitorUtility$1.run(ProgressMonitorUtility.java:59) at org.eclipse.core.internal.jobs.Worker.run(Worker.java:55) Scheduling rebuild due to interruption Autobuild was interrupted: and many more.
The workspace runnable for which prepareOperation is called resulting in the reset is: org.polarion.team.svn.core.utility.ProgressMonitorUtility$2@ec2bb1 they do pass in a none null scheduling rule: org.polarion.team.svn.core.operation.AbstractNonLockingOperation$NonLockingRule@16163b9
yep, removing the subversive plugin did help. I can build the workspace without it.
I suspect this operation has always been interrupting your autobuilds, but it's now more noticeable because the interruption has a much larger cost. Previously, the build would have been restarted, but each time it would successfully run one builder. Eventually all builders would run successfully, even with constant interruption. Now, it's not making any progress at all because the interruption causes the builder to restart from scratch. This is clearly a bug in the Polarion plugin you have installed, and it should be reported to them. However, if a misbehaving plugin can have such a drastic effect, I'm not sure we can fix bug 89301 after all.
I agree. If a misbehaving plugin can cause a builder to believe its interrupted then we cannot use it.
Removed the complete patch for bug 89301
(In reply to comment #11) > Removed the complete patch for bug 89301 > Sorry guys :-(
Verified for 3.4M2 using I20070917-1800
