Hi Denis,
The code that works with the NET protocol should be run under the
awtLock() since it accesses the Xlib APIs.
Otherwise the fix looks good to me. Thanks.
--
best regards,
Anthony
On 1/23/2012 8:16 PM, Denis S. Fokin wrote:
Anthony, Artem,
please look at this iteration
http://cr.openjdk.java.net/~denis/7117011/webrev.02
Thank you,
Denis.
On 1/23/2012 6:34 PM, Denis S. Fokin wrote:
Anthony, Artem, thank you for your input.
Could you take another look at the fix.
http://cr.openjdk.java.net/~denis/7078460/webrev.01/
I have run Window, Dialog, Frame JTreg tests. I have not found any new
failures.
Thank you,
Denis.
On 1/20/2012 6:57 PM, Anthony Petrov wrote:
+1 as long as the fix doesn't cause any regressions. The simple window
case needs to be fixed, obviously.
--
best regards,
Anthony
On 1/20/2012 6:54 PM, Artem Ananiev wrote:
Hi, Anthony,
I personally find the current code quite complicated because of:
1. SKIP_TASKBAR should only be set (or not set) once, because it
depends on the window's owner, that cannot be changed on the fly.
Having SKIP_TASKBAR code in setBounds() doesn't make any sense to me.
2. XDecoratedPeer.setBounds() not calling super.setBounds() is not a
typical pattern, so it's not trivial to understand that SKIP_TASKBAR
is not really set for dialogs. This is yet another reason to move this
code to postInit().
So I like the proposed change. The only missed part is that we should
set SKIP_TASKBAR on simple windows (not frames/dialogs) regardless of
their owner.
Thanks,
Artem
On 1/20/2012 4:03 PM, Anthony Petrov wrote:
Hi Denis,
The bug you're fixing is about a JDialog which is a Dialog descendant,
and as such uses the XDecoratedPeer. Therefore, the
XWindowPeer.setBounds() is never really invoked for JDialogs.
Also, the lines 482-489 which you're removing from this method come
from
JDK6. This piece of code resets the SKIP_TASKBAR state on every
setBounds() call, and please note that it's unconditional (i.e.
there's
no if (owner!=null) check).
I'd suggest to not remove the lines 482-489 unless this is strictly
necessary and doesn't introduce any regressions.
--
best regards,
Anthony
On 1/19/2012 7:57 PM, Denis S. Fokin wrote:
Hi AWT team,
Please review a fix for the CR 7078460 at
http://cr.openjdk.java.net/~denis/7078460/webrev/
CR URL:
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=7078460
Thank you,
Denis.
