Regression or OCC?

[chrisb]

I have this (heavily stripped down) file which I was able to recompute with these versions without errors:
OS: macOS 10.13
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.18.16146 (Git)
Build type: Release
Branch: (HEAD detached at 0.18.4)
Hash: 980bf9060e28555fecd9e3462f68ca74007b70f8
Python version: 3.6.7
Qt version: 5.6.2
Coin version: 4.0.0a
OCC version: 7.3.0
Locale: English/UnitedStates (en_US)

and
OS: macOS High Sierra (10.13)
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.19.17798 (Git)
Build type: Release
Branch: master
Hash: 7e60631239109c632a8f6cca83f6e7e5502e43a1
Python version: 3.7.3
Qt version: 5.9.7
Coin version: 4.0.0a
OCC version: 7.3.0
Locale: English/Germany (en_DE)

But not with the current. Here the Fillet shows an error:
OS: macOS High Sierra (10.13)
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.19.20802 (Git)
Build type: Release
Branch: master
Hash: f4279f1bea8377a7ef63f7f4c69b67159d498965
Python version: 3.8.2
Qt version: 5.12.5
Coin version: 4.0.0
OCC version: 7.4.0
Locale: C/Default (C)

Is this a regression or is it OCC 7.3 vs. 7.4?

[kisolre]

Recomputes ok with:

OS: Windows 8.1 (6.3)
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.19.20738 (Git)
Build type: Release
Branch: master
Hash: 0c96f77184edc0a6747483b2ff1462fc6dae46ff
Python version: 3.6.8
Qt version: 5.12.1
Coin version: 4.0.0a
OCC version: 7.3.0
Locale: Bulgarian/Bulgaria (bg_BG)
and
OS: Windows 8.1 (6.3)
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.19.20839 (Git)
Build type: Release
Branch: master
Hash: 09005766f10539dfe32314c33c2cd3efe9141d0b
Python version: 3.6.8
Qt version: 5.12.1
Coin version: 4.0.0a
OCC version: 7.3.0
Locale: Bulgarian/Bulgaria (bg_BG)
With
OS: Windows 8.1 (6.3)
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.19.20802 (Git)
Build type: Release
Branch: master
Hash: f4279f1bea8377a7ef63f7f4c69b67159d498965
Python version: 3.8.2
Qt version: 5.12.5
Coin version: 4.0.0
OCC version: 7.4.0
Locale: Bulgarian/Bulgaria (bg_BG)
fillet fails because of TopoNaming
Fixing the edges fixes the fillet.
.

[chrisb]

fillet fails because of TopoNaming
Fixing the edges fixes the fillet.

I know how to fix it, but in the real model it was more than this. And where does the toponaming come from. Usually it occurs if shapes change, but here nothing was changed.

So it's more a reliability question. If errors occur it is sort of good, because they cannot slip through unattended, but there were some changes without noticing in the first place.

[kisolre]

The edge/face/vertex order was never guaranteed as far as I understand it. And that is where the whole naming fail comes from. They have no names. They have indices, generated from algorithms. And as long as the algorithms produce proper geometrical result they dont care about the order of things. And you can not expect that improving/fixing/extending/... algorithms will keep the ordering. So I find it perfectly normal if OCC version changes results to also change - hopefully for the better.

[chrisb]

Thanks!