[news.eclipse.tools] Re: Workbench shutdown does not complete

["Jeff McAffer" <jeff_mcaffer@xxxxxxx>]

Pat,

Stumped on this one.  Since you can get it to happen when running in debug
mode, perhaps you could pause the remaining threads and see what is
happening.  To do this, run the "target" Eclipse as you described (using the
debugger <aka 'the bug'>).  When you exit the target and it is left running,
select the remaining threads and use the debugger's pause button and review
the stacks.  Chances are that most of the threads are random things like
indexers or reconcilers etc. that don't cause much trouble. Look for threads
with "main" or "modal" in their name or with non-trivial stacks and some
sort of blocking call (semaphore acquisition, wait(), ...) at the top.  In
this case you might also look for "help" related threads.

Check this out and let us know what you find.  Its very curious...

Thanks

Jeff

"Pat McCarthy" <patmc@xxxxxxxxxx> wrote in message
news:3B4DC706.7FEC9CEA@xxxxxxxxxxxxx
> I'm having an intermittent problem during the shutdown of the workbench.
>
> The UI goes away and I can start the workbench again without trouble,
> but the first javaw.exe is still running in the background (not doing
> much).  I can get four or five of these hanging javaw.exe processes left
> running and consuming virtual memory.
>
> I thought I was alone with this problem as I could not get others to
> mimic my results on their machine.   I've now found several others that
> do have this problem but have not found anything consistent in the
> config that would help identify why this occurs.
>
> We do know that it seems to happen when the help perspective has been
> used/is left up when we shutdown.  If I close the help perspective and
> go through a shutdown/restart/shutdown cycle - the process is closed
> after the second shutdown (but the process from the first shutdown is
> still there).
>
> I've got IE 5.00.3103.nnnn installed.  It does not happen on a
> team-mates machine (he has IE 5.50.4134.nnn), but it does happen on
> another machine where IE is at 5.50.42nn.nnnn.  So there does not seem
> to be a relationship between the IE level and the hanging javaw.exe
> processes.
>
> I can get the same behavior if I start the workbench (run or debug) from
> a plugin in the PDE.  The task stays in the debug window after the
> launched workbench is shutdown (but closes nicely if I close the help
> perspective and shutdown/restart/shutdown).
>
> Is there a startup option I can use to find out why the javaw.exe
> process hangs?
>
> Thanks,
>     Pat Mc.
>