Mac client: crash running recipe "Banded worm pairs infinite" overnight. Seems to be in server update code.

Started by NinjaGreg

bkoep Staff Lv 1

Thanks @NinjaGreg! This looks related to another crash that has cropped up recently. The crash occurs randomly, so unfortunately I have no tips on avoiding it. But I do have an idea about the underlying bug, so hopefully we'll have a bugfix soon!

NinjaGreg Lv 1

Hmmm, it would seem that in puzzle 2199, where I have a two-helix four(including ligand)-sheet design, that running the recipe "iwdn ufreeze SC 1/0". unfreezing the core sidechains, then running "iwdn HBNW 1.3" seems to trigger this bug pretty reliably. FYI. I posted a share of this design for the group if you'd like a copy. I run it on a track, if that makes a difference.

bkoep Staff Lv 1

@NinjaGreg Today's devprev release includes a change that we hope will fix this crash. If you get a chance, please try it out.

Unfortunately we were unable to reproduce the same crash. But your crash report was helpful enough that I believe we can now catch the issue before it crashes the application. After running the latest update, it would be helpful if you could check the log.txt for a messages like "Error posting solution: …", and let us know the exact message and frequency of these messages.

NinjaGreg Lv 1

@bkoep: Aha! Got a crash. No instances of "Error posting solution". I cut the log.txt down to the last 500 lines, and it's posted here:

log.txt

There wasn't a debug.txt, oddly enough. The crashlog that MacOS displayed seemed a bit unusually lengthy, so I've included that also:

crashlog.txt

HTH.

(further note: I was running two clients to try to aggravate things; only one crashed)