[Bug swing/28061] JTree issues and regressions

[hendrich at informatik dot uni-hamburg dot de]

------- Comment #12 from hendrich at informatik dot uni-hamburg dot de  
2006-07-04 17:09 -------
(In reply to comment #11)
> Created an attachment (id=11816)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=11816&action=view) [edit]
> Partial fix (node expansion when the root node is not visible and vertical 
> line
> over root nodes)
> 
> I can reproduce only 3 and 4. This is not a regression, but stayed unnoticed
> until now. The fixes are simple.

Hello Audrius,
thanks for your fixes and comments. Even if 3 and 4 are not regressions, and
regardless of the bug priority level, they should be reported (and fixed)...

Concerning 1: the keyboard navigation actually does work in the free swing
demo, sometimes, initially, but it is very fragile. Resizing/maximizing the
internal frame with the tree seems to disable the keyboard navigation. When I
open the text demo, and then another tree demo, keyboard navigation in the tree
demo is dead on my system. This may be a keyboard focus issue, and it may well
be another bug related to my non-standard window-manager (ctwm).

Concerning 2: hmmm. I admit that I had some difficulties to create the attached
screenshot. Initially, everything worked fine. When I reported the bug earlier
today, I had played with SplitPaneDemo for some time, expanding an d collapsing
and scrolling for several minutes. The problem I observed seems to be related
to missing repaints after scrolling on heavy load; jamvm would keep up with
repainting, but the more nodes are expanded, the slower the repainting becomes.
As a result, many nodes are shown highlighted during scrolling, and finally a
result similar to the attached screenshot occurs. (Note: the JTree was
repainted correctly once I enforced a repaint via the window-manager.)

I initially thought this was a problem in JTree, but it may be another cliprect
problem (probably related to coalesced repaints). I know this sounds like a bad
joke, but could you try expanding thousands of subtrees/treenodes and keep the
VM busy with expanding/collapsing subtrees, scrolling and repainting for some
time? ;-)


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=28061