[FreeCAD Volunteer Position] Liason between FreeCAD and OpenCascade

[Kunda1]

FreeCAD uses OpenCascade (OCC) as it's CAD engine. Naturally there is a need to keep a open rapport with the upstream OCC issues that FC encounters. This is a volunteer position for the FreeCAD project to improve, streamline, report, and monitor upstream OCC issues. There are some requirements by the OCC devs for one to be able to have access to their bugtracker.

Requirements:

• In order to get permissions to report on the OCC bugtracker you have to sign the CLA (contributors license agreement)
• Ability to build the latest OpenCascade master
• Ability to rebuild FreeCAD with said OCC master version to verify the bug still exists
• Learn the OCC draw syntax to reproduce the bug for reporting in an appropriate format in the OCC bugtracker

References:

• A need to have an open dialogue with upstream OCC team
• How often are bugs reported to OpenCASCADE?

[looo]

In my opinion it would be much easier to work with oce, as they are on github, and I think contributions are accepted faster. In the long run we should use oce-sources again. Therefore oce should move to occt7.x. I already asked something in this direction:
https://github.com/tpaviot/smesh/issues/46

[NormandC]

Therefore oce should move to occt7.x.

You have to face the sad truth that the OCE project is dead. The project's pillars have apparently moved on to other things. Their last update from May 2017 was based on OCC 6.9.1 which itself came out on 28 September 2015 (!!!), and based on work tanderson69 (blobfish) did more than a year before.

Thomas admits what I consider is the failing of OCE:

The difference between oce and occt is just the timeline, i.e. oce has a delay over occt release cycle.

A delay of two years (and still counting) is simply unacceptable. FreeCAD has benefited from huge gains in going from OCE 0.17.1/OCC 6.8.0 to OCC 7.1.0. Countless geometry bugs encountered by end users are now gone. Proper projections and HLR in TechDraw would be impossible with the current OCE.

Thinking that OCE can be brought to OCCT 7.1.0 in a timely manner is wishful thinking. OCCT will probably already be at 7.5.0 when/if that happens.

[looo]

FreeCAD has benefited from huge gains in going from OCE 0.17.1/OCC 6.8.0 to OCC 7.1.0.

That is true. My suggestion to oce is quite simple: Adopt latest occt7.1, try to apply some of the patches and update the dependencies. This is more like a restart and shouldn't be that much work. Patches can be applied one by one...

[NormandC]

My suggestion to oce is quite simple: Adopt latest occt7.1, try to apply some of the patches and update the dependencies. This is more like a restart and shouldn't be that much work.

Are you not aware that there is nobody left anymore to do even little work (I doubt what you're proposing represents little work but anyway)? You can hear crickets on the oce-dev mailing list.

[looo]

at least I got some fast replies on github.

[NormandC]

Yes, I'm actually quite surprised by that. If you look at OCE's commits, the latest upgrade and occasional commit notwithstanding, there is no actual work being done anymore and hasn't been for the past year or two.

[DeepSOIC]

Can we (FreeCAD) pay OCCT to have some sweep across FreeCAD bugs marked as OCC bugs, I wonder?

[Kunda1]

Can we (FreeCAD) pay OCCT to have some sweep across FreeCAD bugs marked as OCC bugs, I wonder?

I actually think this is pretty good idea. Case in point, Defeaturing WB.