I had tested with the almost latest 0.19 and as far as I know nothing has changed in sketcher since.
Jim, you use the Levenberg-Marquardt solver. As this isn't the default, could you retry with the dogleg solver? I get the message
Code: Select all
19:06:04 Updating geometry: Error build geometry(0):
19:06:04 Invalid solution from DogLeg solver.
There was an issue with Levenberg-Marquardt not being able to solve some pretty simple sketches, but I'm not sure if it was fixed or not.
I had uploaded a file with an ellipse, just to make sure that different ways of creation don't play a role, I link it below. With that file I experience the following behaviour:
- opening the file and trimming the right lower quarter and then the left lower yields the error described above with the invalid constraints
- starting with trimming the right upper quarter yields different results, undo and trim again is usually not identical to first trim
- - it can remove the arc and create conflicting constraints

- Snip macro screenshot-ffcc23.png (17.19 KiB) Viewed 551 times
- - it can collapse to a point

- Snip macro screenshot-9c249e.png (820 Bytes) Viewed 551 times
- - it can collapse to a line

- Snip macro screenshot-3b64cf.png (1.21 KiB) Viewed 551 times
- - it can collapse to a rather small (correct) variant of the trimmed ellipse

- Snip macro screenshot-7ee228.png (1.31 KiB) Viewed 551 times
- - it can look correct but is in fact unsolved

- Snip macro screenshot-daaa6b.png (20.01 KiB) Viewed 551 times
OS: macOS High Sierra (10.13)
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.19.22246 (Git)
Build type: Release
Branch: master
Hash: abd18ff88f7682e9b1160c36bb785b14a33473ef
Python version: 3.8.5
Qt version: 5.12.5
Coin version: 4.0.0
OCC version: 7.4.0
Locale: C/Default (C)
ellipseTrim.FCStd