[icedtea-web] RFC: Patch to fix PR778

[Deepak Bhole]

* Omair Majid <omajid at redhat.com> [2011-10-20 17:10]:
> On 10/20/2011 04:43 PM, Deepak Bhole wrote:
> >* Omair Majid<omajid at redhat.com>  [2011-10-20 14:24]:
> >>On 10/20/2011 01:50 PM, Deepak Bhole wrote:
> >>>Hi,
> >>>
> >>>This is a patch to fix PR778 -
> >>>"Jar download and server certificate verification deadlock":
> >>>http://icedtea.classpath.org/bugzilla/show_bug.cgi?id=778
> >>>
> >>>ChangeLog:
> >>>2011-10-20  Deepak Bhole<dbhole at redhat.com>
> >>>
> >>>     PR778: Jar download and server certificate verification deadlock
> >>>     * netx/net/sourceforge/jnlp/GuiLaunchHandler.java (launchInitialized):
> >>>     Moved as much code as possible out of the invokeLater block.
> >>>
> >>>Unfortunately, the reproducer from the site has gone off-line and I was unable
> >>>to create one. I did however try this patch with the original reproducer and it
> >>>worked fine.
> >>>
> >>>Even without a reproducer, it makes sense from a logical standpoint to have as
> >>>little as possible in an invokeLater call.
> >>
> >>Yes, it does make sense. I do have a concern though.
> >>JNLPSplashScreen extends JDialog, so I am not sure if doing the
> >>equivalent of "new JDialog()" on a non-EDT thread is safe. The docs
> >>aren't very clear on this, but here is a blog that discusses the
> >>issue:
> >>http://bitguru.wordpress.com/2007/03/21/will-the-real-swing-single-threading-rule-please-stand-up/
> >>
> >
> >Interesting. I wasn't aware that constructor calls needed to happen in
> >the EDT. Nice catch!
> >
> >Modified patch attached.
> 
> Thanks for updating the patch. I have a minor concern: the patch
> swallows InvocationTargetExceptions. I think it would be better to
> at least print it before proceeding.
> 
> Otherwise, the patch looks good to me.
>

Pushed after adding a printStackTrace. Thanks!

Cheers,
Deepak