I reopened the puzzle to see if it was reproducible (knew it was but this time grabbed more log data)
Looks like 2 separate errors causing problems… Segmentation Fault and a memory allocation error (double free?)
(these are the most common errors on the monomer puzzles now for me as well.)
<.b>Check this feedback thread:</b>
http://fold.it/portal/node/2003770#comment-34763
Thanks for the feedback on this one everyone.
We are working hard to debug it, but it is proving very difficult to track down. It only seems to happen on Windows, and also seems to disappear when we run the 'debug' version of the game - a 'Heisenbug' - https://en.wikipedia.org/wiki/Heisenbug
We are really excited to have players get working on symmetric designs, especially in light of the recent success of a player made structure, so any further information you might have on the nature of this bug would be greatly appreciated.
Are you able to cause the crash reliably at UW using the non-debug version of the game?
Yes.
The issue is that the bug crashes in different parts of the code every time. This likely means that the actual problem is happening somewhere else in a way that doesn't cause things to crash immediately. We have tried a number of tools that are designed to help in these situations, but haven't gotten anything so far.
I'm hoping that with more specific data on the nature of the bug, I might be able to narrow things down to a portion of code that I can review carefully by hand and discover the bug.
There seems to have been improvement with the new build in June 2017. Of the three symmetry puzzles I tested, two are still crashing quickly and reliably with my simple remix test script, but one has stopped crashing with the script. That's not to say it wouldn't crash if I played with it more, but a little progress is better than none. I do now have log files with crashes and log files for the same puzzle, same actions, no crashes.
build 20170620-a0c2bbb6d0-win_x86-devprev
Devprev Symmetric Dimer Design puzzle, remix test script (see above), from reset puzzle
build 20170620-a0c2bbb6d0-win_x86-devprev
Beginner Puzzle <150 Symmetry Design, remix test script, from reset puzzle
build 20170620-a0c2bbb6d0-win_x86-devprev
Devprev Symmetric Dimer: Buried Hydrogen Bond Networks, remix test script, from reset puzzle
Just to be sure there was a real improvement in the HBN puzzle, I re-tested it with the previous devprev build. The log files are confusing because I had other clients running in that directory at the same time, but you can see the final call to ActionBuild.
build 20170503-ce4d057080-win_x86-devprev
Devprev Symmetric Dimer: Buried Hydrogen Bond Networks, remix test script, from reset puzzle