Re: Gtk+ peer broken?

On Thu, Mar 10, 2005 at 11:41:21PM -0500, Chris Pickett wrote:
> Grzegorz B. Prokopski wrote:
> >On Thu, 2005-10-03 at 23:00 -0500, Chris Pickett wrote:
> >>Grzegorz B. Prokopski wrote:
> >>>On Thu, 2005-10-03 at 23:45 +0100, Rene Wagner wrote:
> >>>>does the Gtk+ AWT peer in sablevm(-classpath) 1.1.10 work for anybody? I'm testing with a simple test class that just pops up a window
> >>>>with a "quit" button in it.
> >>>Yes, the AWT/Swing in 1.1.10 and current staging are horribly broken,
> >>>see http://sablevm.org/bugs/109
> >>>Any help with locating the problem is greatly appreciated,
> >>Um, isn't that a _regression_ and sort of against the mandate of what staging is supposed to be?  Isn't it better to rollback staging until it's fixed (assuming a non-trivial fix is needed), and then when it 
> >>does get resolved release 1.1.10-1?  Or is there too much overhead involved in that?
> >You can't rollback a release.
> 
> I meant to rollback staging and issue 1.1.10-1 as a bugfix release ASAP (I think that's better than 1.1.11 personally, but okay, that works too).
> 
> I'll leave it up to you, I just think it's not good if people don't have *anything* working to use besides 1.1.9.  Perhaps even just minimally identifying the revision where things broke and announcing it would 
> be useful.
> 
> Cheers,
> Chris
> 
> P.S.  Sorry if I'm overly pessimistic.
> 
> >But in any case, this issue has to be fixed.  I'd rather see developers
> >being hit by the same problems as the users, so that the problems got
> >fixed.  Getting 2 months back with classpath is not an improvement.
> >OTOH a 1.1.11 release with this issue fixed would be one.
> >				GBP

Etienne and I talked about it on #classpath and it seemed to be some
stack corruption in sablevm. Then same testcases work fine with current
kaffe and jamvm/classpath 0.14.


Michael
-- 
Java Trap: http://www.gnu.org/philosophy/java-trap.html