[news.eclipse.platform.swt] Re: SWT standalone application crashing on M

["Grant Gayed" <grant_gayed@xxxxxxxxxx>]

swt is only meant to be run in the main thread on all platforms.  What did
your event loop code look like that was being run in the other thread?  I
ask because SWT should throw an error (not a crash like you saw) whenever an
attempt is made to use it in a non-UI thread, so I wonder if SWT is missing
a check somewhere.

Grant


"Elfi Heck" <eheck@xxxxxxxxxxx> wrote in message
news:el68bt$qcq$1@xxxxxxxxxxxxxxxxxxxx
> I found out more in the meantime. Part of the problem seems to be that I
> run (ran) the event loop in another than the main thread (the
> application was originally developed on windows where this is not a
> problem at all). After I moved the event loop to the main thread the
> crashes went away. I couldn't find out however if what I did is
> something that *should* work. I does (or seems to) at least on the
> MacMini. And even on the MacBook there were a few times the application
> did not crash.
> I don't have the MacBook available right now but when I get it back I
> will try to reproduce the crashes with a simple window and two threads.
>
> Elfi
>
>
> Grant Gayed wrote:
> > Hi Elfi,
> >
> > Yes, SWT's libraries are now compiled as universal binaries.
> >
> > You seem to be hitting a bug in swt, but it's difficult to know much
about
> > it based on what you've said so far.  It would be ideal if you could
find a
> > pattern of steps that make the crash reproducable for you, then log a
bug
> > with SWT (
https://bugs.eclipse.org/bugs/enter_bug.cgi?product=Platform )
> > that describes the steps and attach any additional code that's needed to
> > reproduce it.  This is sometimes easier said than done though, so if you
> > don't succeed with this then you can still log a report, include
> > the -Xcheck:jni info, and it may be possible for someone to debug it
with
> > you remotely.
> >
> > Grant