<AWT Dev> Subject: <AWT dev>[11] Review request for JDK-8208125: Cannot input text into JOptionPane Text Input Dialog

Phil Race philip.race at oracle.com
Fri Aug 3 21:12:37 UTC 2018

So you are saying the fix here cumulatively backs out the fix for 
7158623 and 8204860
and reverts the product code to what it was before and re-introduces
7158623 which affected MacOS on all releases of JDK (7 and onwards) ?
I expect that is OK. We can live with a 6 year old bug in 11 as much better
than a stop-ship TCK failure.
However I'd prefer that you "exactly" reverted the fix

Originally we had

- final boolean resizable = isFrame ? ((Frame)target).isResizable() : 
(isDialog ? ((Dialog)target).isResizable() : false); you are proposing 
to keep the rewrite of that condition in the new fix. It is much easier 
for me to be sure of the equivalence if you revert to byte-for-byte what 
it was before we started. Ironically, the only other change to this file 
since you started on this was also backed out ! So once you are done we 
should be back to the same content as changeset e0b025915be8. That 
changeset was also touching maximise so I definitely think we want to 
make sure JCK passes.

Have you run the JCK tests on Mac to make sure there are no surprises ?
We don't have a lot of time left here.

Please enumerate all tests you have run on this "reversion" fix.
I expect it to include all automated jtreg + JCK tests, related manual 
tests, + SwingSet2
as well as the specific regression test.

Its a bit hard to tell what you are doing with the test, its not an 
exact reversion
You deleted a lot of machinery and I suppose you want to keep that part, 
just revert
the parts that verify the functionality you've backed out.
But you should make sure that if I as I expect the reverted test now 
fails on Mac,
and since it is automated, this fix should include problem listing it on 
If it does not fail on Mac .. I'd like to know why .. since this was the 
test that originally failed.
Also you should make sure it definitely still passes on Windows + Linux.

And can you file that new bug, with the full back story, and reference 
it here ?


On 08/03/2018 04:23 AM, Manajit Halder wrote:
> Hi All,
> Please review the fix for JDK11.

> <https://bugs.openjdk.java.net/browse/JDK-8208125>

> http://cr.openjdk.java.net/~mhalder/8208125/webrev.00/ 
> <http://cr.openjdk.java.net/%7Emhalder/8208125/webrev.00/>
>     This issue and JDK-8208127 was caused due to regression caused by 
> fix for the issue JDK-8204860.
> Fix:
>     I had a relook at the original issue fixed on Windows OS 
> (JDK-4980161 <https://bugs.openjdk.java.net/browse/JDK-4980161>) and 
> found that the fixes proposed by my earlier fixes were wrong w.r.t to 
> the behaviour on Windows and Ubuntu. The issue (JDK-4980161) on 
> Windows was that “non-focusable maximized frame was resizable” but a 
> focusable maximized frame was already non-resizable (working as 
> expected).
> Please read the problem description of issue JDK-4980161: which says 
> “Typically a maximized frame is not supposed to be resizable.”
> I verified the behaviour on Windows and Ubuntu and found that a 
> maximized Frame (frame.setExtendedState(Frame.MAXIMIZED_BOTH)) is not 
> resizable irrespective of Frame’s focusable behaviour. On Mac OS this 
> behaviour is not implemented i.e. the problem is “maximized frame is 
> resizable”.
> So the problem which needs to fixed on Mac OS is “Maximized frame 
> should not be resizable irrespective of its focusability”
> Therefor I have reverted my changes to the original state (before fix 
> of issue JDK-7158623 
> <https://bugs.openjdk.java.net/browse/JDK-7158623>) and propose to fix 
> the problem on Mac by creating a new issue.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/awt-dev/attachments/20180803/7ff2b92b/attachment.html>

More information about the awt-dev mailing list