AWT feedback, again

[Norman Hendrich <hendrich at informatik dot uni-hamburg dot de>]

Hello all,

here are two more bug reports as requested by Tom...  

PR20014 - BufferedImage.getGraphics() aborts
PR20015 - JMenu stays open but should not

(Sorry, but I have not yet applied today's big GUI patch from to
my installation yet.)


Ranjit> Now if only we can address some of the more serious
Ranjit> bugs reported by Norman Hendrich in time for 4.0, it'll
Ranjig> be super cool.

Yep :-)

FWIW, even the 'less serious' bugs are show-stoppers for my apps.
Still, I guess that the mouse-button-2+3 thingy and perhaps the
Swing-JComponent-not-receiving-key-events-at-all bug are comparatively
easy to fix. The events themselves already are working, after all.

For fun, I attach a screenshot of jfig2 running in gij without Java2D and
with a few local fixes. (It's only 15 kbytes).
Rendering mostly works, as do left-mouse-button events, including mouse- 
dragging. The buttons don't show up due to the drawImage-with-negative-
coordinates-bug, but otherwise work as they should.


Mark> Tried it against a clean tree here with a --enable-java-awt=gtk
Mark> --enable-gtk-cairo (against cairo-0.3.0 snapshot) bootstrap which went
Mark> fine. The usual subject, GNU Classpath examples, jedit, freemind,
Mark> jfreechart, startup with this new patch (and actually show text now when
Mark> using -Dgnu.java.awt.peer.gtk.Graphics=Graphics2D !).

Aha! I knew I had forgotten something. Anyway, running my application(s)
or testcase(s) with gij or the-Dgnu.java.awt.peer.gtk.Graphics=Graphics2D
option only makes things worse. (I believe my system is setup correctly;
I do have cairo-0.3.0 and dependencies).

Using -Dgnu.java.awt.peer.gtk.Graphics=Graphics2D option makes *all*
accesses to BufferedImages or drawImage() fail. I will investigate...

- Norman

Attachment: screen-arrogant-gcj.png
Description: screen-arrogant-gcj.png