Possible Fix for Crash-After-Draw-Delete Bug

I’ve had a reproducible bug – if you use the draw tool and right-click an atom to delete it, then resume drawing, you’ll consistently get a crash.

I think I’ve fixed this in the 1.0 branch:
http://review.source.kitware.com/#change,1340

In debugging, I was noticing that the m_beginAtom and m_endAtom were pointing to completely bogus id()'s. Making sure to clear these variables after the mouseRelease for delete seems to solve the bug.

Any other remaining crashers for 1.0.2?

-Geoff

On Wed, Apr 6, 2011 at 11:13 PM, Geoffrey Hutchison geoff.hutchison@gmail.com wrote:

I’ve had a reproducible bug – if you use the draw tool and right-click an atom to delete it, then resume drawing, you’ll consistently get a crash.

I think I’ve fixed this in the 1.0 branch:
http://review.source.kitware.com/#change,1340

In debugging, I was noticing that the m_beginAtom and m_endAtom were pointing to completely bogus id()'s. Making sure to clear these variables after the mouseRelease for delete seems to solve the bug.

Any other remaining crashers for 1.0.2?

Glad you got to the bottom of that one, I have a very small fix for
MOPAC that I have solved locally. I will upload a patch tomorrow, but
I have it fixed locally (change in behavior in the latest MOPAC).

Then I think we can finally make the release.

Marcus

On Thu, Apr 7, 2011 at 3:26 AM, Marcus D. Hanwell
mhanwell@gmail.com wrote:

On Wed, Apr 6, 2011 at 11:13 PM, Geoffrey Hutchison geoff.hutchison@gmail.com wrote:

I’ve had a reproducible bug – if you use the draw tool and right-click an atom to delete it, then resume drawing, you’ll consistently get a crash.

I think I’ve fixed this in the 1.0 branch:
http://review.source.kitware.com/#change,1340

In debugging, I was noticing that the m_beginAtom and m_endAtom were pointing to completely bogus id()'s. Making sure to clear these variables after the mouseRelease for delete seems to solve the bug.

Any other remaining crashers for 1.0.2?

Glad you got to the bottom of that one, I have a very small fix for
MOPAC that I have solved locally. I will upload a patch tomorrow, but
I have it fixed locally (change in behavior in the latest MOPAC).

Then I think we can finally make the release.

For whatever reason I could not reproduce your crash here on Mac or
Linux. I also uploaded a couple of new patches to fix bugs pointed out
in the bug tracker. I have a clean build of the 1.0 tree, so I am not
sure why it is not crashing here. If there are no other blockers I
would like to get the release out.

Marcus

On Apr 7, 2011, at 9:32 AM, Marcus D. Hanwell wrote:

For whatever reason I could not reproduce your crash here on Mac or
Linux.

A few of these bugs seem to be compiler/optimizer dependent. I can reproduce it on my laptop using Clang, but not on my desktop using GCC.

I also have a pending fix for Konstantin’s “change-to-hydrogen” crash I’ll upload to Gerrit in a few minutes.

-Geoff