Bug 43312 - Switching perspectives while debugging causes segfault
Summary: Switching perspectives while debugging causes segfault
Status: RESOLVED FIXED
Alias: None
Product: Platform
Classification: Eclipse Project
Component: SWT (show other bugs)
Version: 3.0   Edit
Hardware: PC Linux-GTK
: P3 critical (vote)
Target Milestone: ---   Edit
Assignee: Felipe Heidrich CLA
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2003-09-18 14:31 EDT by Douglas Pollock CLA
Modified: 2004-06-02 17:09 EDT (History)
0 users

See Also:

Attachments
core dump (22.91 KB, text/plain)
2003-09-18 14:32 EDT, Douglas Pollock CLA 		no flags Details
Another core dump (23.26 KB, text/plain)
2003-09-19 14:56 EDT, Douglas Pollock CLA 		no flags Details
core dump #3 (running UiTestSuite in platform-ui) (21.24 KB, text/plain)
2003-09-29 10:58 EDT, Douglas Pollock CLA 		no flags Details
core dump while jumping to test suite failure (27.46 KB, text/plain)
2003-10-06 15:03 EDT, Douglas Pollock CLA 		no flags Details
while running ShowViewTest.test1() (19.23 KB, text/plain)
2003-10-08 12:09 EDT, Douglas Pollock CLA 		no flags Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Douglas Pollock CLA 2003-09-18 14:31:43 EDT 
I'm not sure how reproducible this bug is.  I was debugging a self-hosted
workbench through a JUnit plug-in test.  A patch was neeeded, so I clicked on
the team synchronize perspective.  A segfault occurred before any perspective
change occurred -- killing both the self-hosted windows and the Eclipse windows.

The machine config is: RedHat 9, Eclipse-GTK-I20030917, GNOME, gtk2-2.2.1-4.
Comment 1 Douglas Pollock CLA 2003-09-18 14:32:45 EDT 
Created attachment 6155 [details]
core dump
Comment 2 Douglas Pollock CLA 2003-09-19 14:56:14 EDT 
Created attachment 6173 [details]
Another core dump

Another core dump.  This time while trying to double-click on an item in the
Team Synchronizing perspective to get it do a compare.	Eclipse-GTK-I20030917.

Not easily reproducible, but this seems to be more than a rare occurrence.

Running UiTestSuite sometimes crashes in the same way.
Comment 3 Grant Gayed CLA 2003-09-22 15:39:19 EDT 
Doug, if you find more scenarios where this happens then please add them to 
this report.
Comment 4 Douglas Pollock CLA 2003-09-29 10:58:21 EDT 
Created attachment 6271 [details]
core dump #3 (running UiTestSuite in platform-ui)

This should hopefully make things a bit more reproducible for you.  This
occurred while running the UiTestSuite for platform-ui.  It is available in CVS
in the project "org.eclipse.ui.tests".	The failure occurred on consecutive
runs, on tests #192 and #186, respectively.

Using Eclipse-GTK-I20030925
Comment 5 Douglas Pollock CLA 2003-10-06 15:03:48 EDT 
Created attachment 6340 [details]
core dump while jumping to test suite failure

This looks different than the other three.  This might be unrelated.  If it is,
could you file another bug and CC me.  Thanks.
Comment 6 Douglas Pollock CLA 2003-10-08 12:09:18 EDT 
Created attachment 6369 [details]
while running ShowViewTest.test1()

Another failure while running the UiTestSuite.
Comment 7 Douglas Pollock CLA 2003-10-08 12:56:16 EDT 
I've run the test suite twice without using x2vnc, and the tests will then
complete normally.  This is enough to make me feel strongly that this is related
to Bug 44344.  Check out that bug for a more verbose description of where things
might be going wrong.  Basically, I think some pieces of either GTK or SWT don't
deal very well with the possibility of the X Windows session having serviced a
request for XGrabPointer or XGrabKeyboard.
Comment 8 Grant Gayed CLA 2003-10-08 15:36:59 EDT 
Moving since Felipe has bug 44344.
Comment 9 Felipe Heidrich CLA 2004-05-25 17:30:07 EDT 
Does this still happen ?
I think SSQ changes for Clipboard and Shell#setVisible() got this problem fixed.
Please, retry. let me know if this is fixed.
Comment 10 Douglas Pollock CLA 2004-05-28 17:52:36 EDT 
I'm not sure that Silenio's fix had anything to do with this though, as I 
haven't seen this in quite a while.  But it does appear to be fixed.
Comment 11 Felipe Heidrich CLA 2004-06-02 17:09:42 EDT 
closing