Community
Participate
Working Groups
After every 2 hours or so of usage, my Eclipse UI freezes. Doesn't matter how long I wait -- CPU usage goes way up, and the UI is just unresponsive. I'm forced to kill it. The worst part of this for me is the workbench loss -- any customizations I've made, whatever windows were open, etc....all this is gone. So I've taken to restarting Eclipse myself periodically to "capture" something close to my current state. But it's not a fun way to work. What could be wrong? I've tried starting with the -clean flag, also with -debug; both no good.
Created attachment 52452 [details] jstack output from when Eclipse is frozen
Don't know if this is relevant, but 1.5.0_09-b01 was replaced with 1.5.0_09_b03 almost immediately afterwards: http://java.sun.com/j2se/1.5.0/ReleaseNotes.html "Special Note This update was originally released as 1.5.0_09-b01. After that release, it became necessary to provide an additional bug fix immediately. In order to simplify deployment, 1.5.0_09-b01 was replaced with 1.5.0_09-b03. This new release contains all the bug fixes contained in the old release, in addition to the new bug fix." The stacktrace shows 1.5.0_09-b01 in your VM. Would it be possible to verify that the same bug happens with 1.5.0_09-b03? PS You've gotta larf with JSR277 defining 'more powerful versioning' when the JDKs are the epitome for horrible version numbers ...
(In reply to comment #2) > Don't know if this is relevant, but 1.5.0_09-b01 was replaced with 1.5.0_09_b03 > almost immediately afterwards: > > http://java.sun.com/j2se/1.5.0/ReleaseNotes.html Wow, yeah, what a fantastic version numbering system. :-) OK, I've upgraded to the "real" 09 release and I'll post here again if I get another freeze.
Created attachment 52479 [details] jstack output from when Eclipse is frozen, using latest 1.5.0.9b3 JDK
(In reply to comment #4) > Created an attachment (id=52479) [edit] > jstack output from when Eclipse is frozen, using latest 1.5.0.9b3 JDK > The freezing is still happening, as reported before, even when using the "real" 1.5.0.9b3, as requested.
Created attachment 52492 [details] yet another jstack output
I have attached yet another jstack output. This time Eclipse froze in the middle of a "rename method" operation. Comparing this jstack trace to the other two, it looks like it's something different (but I know nothing about the Eclipse internals). Is there any other type of diagnostic output I can provide? Is there any reason that this would *not* be given priority P1? This is a terrible impediment to using Eclipse at all.
Not sure if it's also relevant, but Sun 6359295 discusses an issue with jstack trace generating a similar bug, which was claimed to be fixed in 1.5.0_08: http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6359295 It's possible that _09 has reintroduced the same bug, or it could be something else all together. The message in the bug report claimed that this was reproducible (the jstack output, not the hanging) when a remote file server over NFS went down and the file could no longer be accessed. Do you have the files on a network drive, or are they local? I merely ask though curiosity ... Since I don't have much knowledge of the Eclipse UI internals, I don't think I'm going to be of much help, but wanted to draw your attention to the newer VM and similar error message in case you (or anyone else) wasn't aware of it.
Could this a dupe of bug 153485?
I'm inclined to agree. If you can provide a standard stack dump using the kill -QUIT method that contradicts the other bug please re-open. *** This bug has been marked as a duplicate of 153485 ***
