OpenCascade Fillet Fiasco

[phoemur]

Man, this baffles me...

I'm back using FreeCAD after 10 years and I got really frustrated to see the fillet/chamfers still not working properly...
It crashes almost everytime if you need to make something a little more complex than a simple cube.
It seems this is one of the most common topics about it since forever. And here in this forum as well.

What is the point of adding new features if the basic things don't work, it there are bugs dating back to 2000's and still open ?

Shame on you OpenCascade.
I wish FreeCAD wasn't held hostage to this """wonderful""" piece of software.

[GlouGlou]

Please show to us your "fiasco".

[Chris_G]

Shame on you OpenCascade.

How did OpenCascade hurt you that you want to put shame on them ?
How much money did OpenCascade steal you that you want to put shame on them ?

Yes, some pieces of OCCT code are rather old.
Yes, there are some lacking features, or some non-optimal tools.

But we have an open-source geometric kernel.
Without OpenCascade, I don't know what FreeCAD would look like today.

I am not skilled enough to propose patches or fixes to OCCT.
I cannot put money into OCCT development.
So I am rather thankful we have what we have, even if it's not perfect.

[phoemur]

It hurts me when I spend some hours making a complex model. Then when it's done, you spend more time trying to get some decent finish on it than to design the whole piece. With segfaults on your face many times on a single session.
And yes, trying to redesign the whole thing in different ways just to make the fillets work.
The same issues from 10 years ago still not solved... The bugs still there...

And the FreeCAD community can't do anything about it because its a third party software.

And yes, I can do c++ and python myself, so I can really recognize the huge amount of effort involved in a project like OpenCascade.
But I don't see the point of arguments like " If you don't like it do it yourself better".

It's like saying you can't criticize some restaurant food just because you're unwilling to take the chef's job...

[phoemur]

Please show to us your "fiasco".

Take the ModelMania parts list for instance:
https://blogs.solidworks.com/tech/2020/ ... mania.html

In more than 50% of the parts I had fillet issues, meaning segfaults, program unexpectedly closing, and of course, redesigning the piece many times in different ways trying to make this "fiasco" feature work. Sometimes without success.
Many hours wasted.

Doesn't look like so functional to me.

[Kunda1]

What version of FreeCAD/OCC are you using ? Paste your About info

[phoemur]

What version of FreeCAD/OCC are you using ? Paste your About info

I'll try to build from the master branch and see if it helps
I am using 0.18.4 from Debian repo

[Kunda1]

What version of FreeCAD/OCC are you using ? Paste your About info

I'll try to build from the master branch and see if it helps
I am using 0.18.4 from Debian repo

You should use the PPA or the AppImage

[chrisb]

It's like saying you can't criticize some restaurant food just because you're unwilling to take the chef's job...

Usually you are not welcome in the kitchen and they are far far less to blame if they give you their food for free.

And you should definitely use 0.19. Much more stable! And sometimes better in Fillets.

[Quaoar]

Shame on you OpenCascade.

If you provide a simple scenario, we can prepare a bug report using the conventional bug reproducing practice (one shape, faces to blend, Tcl script). On behalf of OCC devs, I can assure you that everybody here is interested in making FreeCAD better, but you have to embrace reality:

1. FOSS development in OCC is driven mostly by paying customers and much less by the goodwill of the developers. The complexity of fillets (specifically) requires hundreds of man-days to maintain or (ideally) to rewrite from scratch (that's a legacy code inherited from Euclid). A small service company cannot afford such a huge R&D without clear value. That's a long shot, and we are not Siemens.

2. Rewriting fillets is still pretty possible, especially, with the help of the community. But the fact is that our community is very small and quite inactive in terms of contributions.

Treat OpenCascade as an open project where you can contribute to your own and other people's benefits. Do not take it as a private property of the Company. Anyone interested can make a change.