Description of problem: I consistently get an exception in the Team Synchronizing perspective when doing a synchronization. Version-Release number of selected component (if applicable): eclipse-jdt-3.1.0_fc-2 How reproducible: Apparently always; I made a new workspace, did a cvs checkout of a project, then tried to synchronize.
Created attachment 116676 [details] backtrace
Well, it looks like this is probably a library bug since I removed org.eclipse.team.core_3.1.0.jar.so (and verified that it wasn't compiled on-the-fly in /tmp) and it still happened to me. I will investigate further tomorrow.
Update on How reproducible: The problem seems always repeatable with FC4-i386 distribution however I do NOT see this problem when using FC4-x86_64 (64-bit AMD64). On FC4-i386, I am seeing an exception and similar .log backtrace shown above, whilst synchronizing. Version: Eclipse Platform Version: 3.1.0 Build id: I20050401-1645 Using Fedora 4, eclipse CDT, using a pserver cvs connection.
This bug is still happening with 3.1.1 :( Is it a showstopper for anyone or can I go ahead with the FC4 update?
I'm actually getting an additional line at the top of my stack trace: org.eclipse.team.internal.ccvs.core.syncinfo.CVSResourceVariantTree.fetchMembers(org.eclipse.team.core.variants.IResourceVariant, org.eclipse.team.core.runtime.IProgressMonitor) (/usr/lib/gcj/eclipse/org.eclipse.team.cvs.core_3.1.1.jar.so) Just so it's all here and easy to see, this is the relevant code (poor formatting courtesy of bugzilla): /* (non-Javadoc) * @see org.eclipse.team.core.subscribers.RefreshOperation#getRemoteChildren(org.eclipse.team.core.subscribers.ISubscriberResource, org.eclipse.core.runtime.IProgressMonitor) */ protected IResourceVariant[] fetchMembers(IResourceVariant remote, IProgressMonitor progress) throws TeamException { ICVSRemoteResource[] children = remote != null ? (ICVSRemoteResource[])((RemoteResource)remote).members(progress) : new ICVSRemoteResource[0]; IResourceVariant[] result = new IResourceVariant[children.length]; for (int i = 0; i < children.length; i++) { result[i] = (IResourceVariant)children[i]; } return result; }
(In reply to comment #2) > Well, it looks like this is probably a library bug since I removed > org.eclipse.team.core_3.1.0.jar.so (and verified that it wasn't compiled > on-the-fly in /tmp) and it still happened to me. I will investigate further > tomorrow. I don't know what I did back then, but it is definitely a mis-compilation because I don't get it if I remove the corresponding .db (and rebuild-gcj-db).
This is fixed in rawhide. Andrew Haley suggested that it was caused by gcc PR 23182. gcc will probably be updated in FC4 at which point I will rebuild Eclipse with it and push another update that will fix this issue.
*** Bug 174567 has been marked as a duplicate of this bug. ***
Confirming this is fixed in updates-testing with the following package versions: gcc-4.0.2-8.fc4 eclipse-platform-3.1.1-1jpp_1fc.FC4.11