BUG : foamToEnsight -faceZones still crashes sometimes in develop

Possibly related to #317 (closed), foamToEnsight -faceZones still crashes sometimes. I tested with the latest develop commit.

The example model is attached and is a simple pipe with a few faceZones and a cellZone created during meshing. When foamToEnsight -faceZones is run at the end, it crashes and goes into a large loop of errors.

tube_test_inter.tar.gz

Reassigned to @mark

Mentioned in commit 9a1f460723fb2499c8aab963a1c7e1eda9deb96a

Hi Justin, thank you very much for finding this regression. I am very grateful for your additional testing. I believe that I have correctly isolated the problem and have pushed the commit accordingly.

/mark

Added ~15 label

Milestone changed to %Version v1612

Hi Mark, thanks once again for your quick fix! foamToEnsight -faceZones no longer crashes for this model.

However, when I read the converted model into EnSight, the faceZones are not drawn properly. I also tested it in Paraview's EnSight reader and it shows the same thing (see below)...

By Justin Graupman on 2016-12-14T18:37:04 (imported from GitLab project)

Hi Justin, I'm really sorry about all these iterations. I've located a format error in ASCII, but that is presumably not your problem. This isn't the same geometry that you posted earlier, is it?

Hi Mark, it's all good, I just want to make sure the issues are worked out before the release.

Yes, it's the same model attached to this report. Run Allrun and open the resulting EnSight model in EnSight or Paraview, then isolate the cellZone faceZone (cz) and the faceZones (inter_1 and inter_2). You'll see what I see. Maybe this is an ordering problem for the faceZones?

Just a note, I did pull only your commit 9a1f4607 to test foamToEnsight as I didn't want to pull all the recent foundation merge changes yet. If your fix for foamToEnsight was dependent on other commits, let me know.

By Justin Graupman on 2016-12-14T19:32:20 (imported from GitLab project)

At least two good things: the addressing is also off in serial (ie, not a parallel-only issue), and foamToEnsightParts is delivering proper looking geometry.

Mentioned in commit 4d6847c6795254f43f2fffbbe251d80df87bf60b

Thanks Mark for your diligence in resolving these issues. Your most recent commit (4d6847c6) resolved all the issues with foamToEnsight that I'm aware of. As always, I will let you know if I find more. :) If you have no other comments, you may close this ticket.

By Justin Graupman on 2016-12-15T23:44:40 (imported from GitLab project)

Closing since this was fixed by 4d6847c6

By Justin Graupman on 2016-12-19T18:43:09 (imported from GitLab project)

Status changed to closed

By Justin Graupman on 2016-12-19T18:43:09 (imported from GitLab project)