jimy byerley

For the moment, I think using the main branch of moderngl is fine. I will tell the madcad users to do so as well if they need to run on...

of course it can, but less ergonomic and computationnaly efficient than just supporting tuples

Sure it would be nice ! But we will have hard-times to make compatible our dependencies (pyglm, arrex, cython, moderngl) work in wasm and adapt our graphical pipeline to webGL...

This feature is added in [the plan](https://github.com/jimy-byerley/pymadcad-refactor-rendering) to improve the rendering pipeline

@bourbonut is working on a refactoring of the rendering pipeline that could address this feature, It will follow the architecture discussed on [this repo](https://github.com/jimy-byerley/pymadcad-refactor-rendering)

Did you tried to run it in gdb to get how it crashed on the C side ? ```bash $ sudo apt install python3-dbg $ gdb python gdb) run -m...

Hello @GlennWSo, any news on this ? did you experienced similar crashes since ?

I didn't tested your example yet, I have to I can see on the wireframe that the bottom flat face has none of the subdivisions of the side extruded surface,...

This is a floating point issue I can avoid it by adding these lines after your `randsurf()` ```python surf1 = randsurf(res=res, seed=1) for i, p in enumerate(surf1.points): if abs(p[1]) <...

Ah, I finally found the cause of this oversimplification of the bottom face: it comes from the ngon retriangulation after intersection `triangulation_outline()` is processing triplets of points in the outline...