New crash with draw command tests on master

Hi,

I noticed that the undo/redo drawing regression test has started to
segfault after I cherry picked the patches in 1.0 to master. I haven’t
looked into it too deeply just yet, but I would certainly like to
confirm whether this is a regression or a problem with the test before
making a tarball for 1.0.1. I should also set up a second set of
dashboards for our release branch to catch things like this sooner.

http://my.cdash.org/testDetails.php?test=4830823&build=150381

I have closed quite a few other bugs, and was very pleased to get to
the bottom of a longstanding bug with surfaces. I will be merging in
some more of these bug fixes from Gerit tomorrow.

Marcus

Marcus D. Hanwell, Ph.D.
R&D Engineer, Kitware Inc.
(518) 881-4937

On Sun, Jan 30, 2011 at 10:00 PM, Marcus D. Hanwell
mhanwell@gmail.com wrote:

Hi,

I noticed that the undo/redo drawing regression test has started to
segfault after I cherry picked the patches in 1.0 to master. I haven’t
looked into it too deeply just yet, but I would certainly like to
confirm whether this is a regression or a problem with the test before
making a tarball for 1.0.1. I should also set up a second set of
dashboards for our release branch to catch things like this sooner.

http://my.cdash.org/testDetails.php?test=4830823&build=150381

I have closed quite a few other bugs, and was very pleased to get to
the bottom of a longstanding bug with surfaces. I will be merging in
some more of these bug fixes from Gerit tomorrow.

I have a proposed patch, this seems to maintain the fix (no crashes
for the two bugs referenced) and the draw command test no longer
segfaults.

http://review.source.kitware.com/#change,805

It reverts approximately half of the patch that introduced the
regression. I can’t spot any new regressions introduced and so plan to
merge this into 1.0 and master unless anyone spots new issues
introduced.

Marcus