OpenJDK6 regression, javax.swing.UIDefaults broken

[Dr Andrew John Hughes]

On 10:02 Mon 25 Oct     , Joe Darcy wrote:
> Hello.
> 
> Alexander Potochkin wrote:
> > Hello Phil
> > Hello Clement
> >
> > Indeed, it was an urgent fix,
> > to make it work CR #6795356 has to be backported to 6-open
> >
> > when it is there, the problem will be gone
> >
> > I am not sure who takes care of 6-open,
> > Phil could you help me to pass this information to the 6-open team?
> 
> I approve 6795356 being backported to OpenJDK 6 b21 if the affected 
> files are branded appropriately with Oracle copyrights.
> 
> Alexander, if you'd like you're welcome do to the backport.
> 

I've pushed this with the amended copyrights.

I have to admit I'd feel a little better about OpenJDK6 if it was being
considered as a matter of course when fixing such regressions, rather than
us relying on the issue being caught like this.

> -Joe
> 
> >
> > Thanks
> > alexp
> >
> >> SFAIK bugzilla is still officially just for tracking patches.
> >> I've cc'd the engineer who created the 6 open fix. Hopefully that
> >> should be all that's needed.
> >> The fix looks identical to the 6u22 one but the problem may be
> >> that 6-open was forked off an early jdk7 build so that might mean
> >> there's an issue resulting from that divergence, but that's just a 
> >> guess.
> >>
> >> BTW there is a specific 6-dev list and I think any further discussion
> >> ought to move there
> >> after this reply: So I've cc'd jdk6-dev at openjdk.java.net
> >>
> >> -phil.
> >>
> >> Clemens Eisserer wrote:
> >>> Hi,
> >>>
> >>> After updating OpenJDK6 recently an application I've written is
> >>> broken, when running it as Applet in an AppletViewer.
> >>> The app displays boolean values in a JTable, using
> >>> JTable.getDefaultRenderer(Class) to get the CellRenderer.
> >>> After digging a bit deeper, I found the problem is the
> >>> checkAccess()-call when loading "javax.swing.JTable$BooleanRenderer"
> >>> in javax.swing.UIDefaults.java : 1103.
> >>>
> >>> The code there looks like this:
> >>>> c = Class.forName(className, true, (ClassLoader)cl);
> >>>> checkAccess(c.getModifiers()); //Fails, unfourtunatly
> >>>> Exception-Handler kills Exception, so no idea what it is.
> >>>> if (methodName != null) {
> >>>
> >>> In a 2-3 weeks old OpenJDK7 repo, I didn't find the
> >>> checkAccess()-Call, it simply says:
> >>>> c = Class.forName(className, true, (ClassLoader)cl);
> >>>> if (methodName != null) {
> >>>
> >>> Looks like some kind of emergency security fix ;)
> >>> Oracle's JDK6u22 works fine however.
> >>>
> >>> Should I report this against bugs.sun.com or the OpenJDK bugzilla?
> >>> I know for now only OpenJDK seems affected, but I wonder if the oracle
> >>> guys look at bugzilla at all?
> >>>
> >>> Thanks, Clemens
> >>>
> >>> PS: Tetscase attached.
> >>
> >
> 

-- 
Andrew :)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and the OpenJDK
http://www.gnu.org/software/classpath
http://openjdk.java.net
PGP Key: 94EFD9D8 (http://subkeys.pgp.net)
Fingerprint = F8EF F1EA 401E 2E60 15FA  7927 142C 2591 94EF D9D8