How often are bugs reported to OpenCASCADE?

[lukasubo]

Because, as we all know, there are bugs. So many of them. Mainly related to fillets and boolean operations. However, I'd imagine they'd take note of them if there were many reports (and there seem to be many FreeCAD users who could make these).

I know the prevailing response here is that they only care about their commercial customers, but these bugs almost certainly affect them, too.

[ickby]

There are not many users that can make bug reports. This would require to build the latest opencascade master, rebuild freecad with this occ Version to verify the bug still exists and then learn the occ draw syntax to reproduce the bug for reporting in an appropriate format.

Shoogen did this regulary in the past, and quite a few of the reported bugs have been fixed, but then freecad did lag behindert in support of occ7. This is done now, it compiles again with the latest Version. Now we need to find someone who takes the responsibility to do this.

[tanderson69]

Another issue is that occt wants you to use their mantis bug tracker. Which is fine at face value, but in order to get permissions to the bug tracker you have to sign the CLA (contributors license agreement).

[yorik]

One thing we could do at FreeCAD level, provided of course we have people who can use drawexe and sign the cla, would be a bit like linux distributions do... Have a bug on our tracker, with an "upstream link" to a corresponding bug on their tracker. I think Mantis has some tools for that... Just an idea for later.

[lukasubo]

It would be nice, especially concerning fillets. I see so many people here who have issues with fillets, all of which are OCCT bugs. I'll look at the CLA sometime this summer after exams are over.

[ezzieyguywuf]

I have occ7 compiled and FreeCAD built against it. I don't really want to learn the draw.exe syntax, but I wouldn't mind verifying that bugs exist.

[Kunda1]

Please see [FreeCAD Volunteer Position] Liason between FreeCAD and OpenCascade

[Quaoar]

It would be nice, especially concerning fillets. I see so many people here who have issues with fillets, all of which are OCCT bugs. I'll look at the CLA sometime this summer after exams are over.

With fillets, you can do basically only one thing: rewrite them. This particular modeling feature is a pure legacy code inherited from Matra Datavision which is hardly maintainable. Unfortunately, there are no sponsors for this sort huge refactoring (i.e. restarting fillets), so do not expect too much from OCC in that direction. That's a good project, in fact, for a PhD student -)