Olivier Chafik

@ShelbyJenkins Oh this one is weirding me out indeed. A workaround seems to be `root ::= "\n" item{0,7}`, but it feels like the totally empty case is broken in grammar-constrained...

@phymbert great to read you've already mitigated the issue, I'd missed that commit in the flow. > Things get messy when talking about Windows and MacOS. In fact, I have...

> I would prefer to have a classical npm build then. @phymbert possibly inevitable in the long run, would unlock new horizons... > Also what about: https://stackoverflow.com/questions/11813271/embed-resources-eg-shader-code-images-into-executable-library-with-cmake Hah, nifty! Might...

> Nomic's Kompute fork builds xxd from source (it's a single .c source file) unconditionally: https://github.com/nomic-ai/kompute/blob/d1e3b0953cf66acc94b2e29693e221427b2c1f3f/CMakeLists.txt#L187 @cebtenzzre hah great, so we kinda always have xxd after all... (assuming one does...

> I think this kind of break rocm for cmake in windows. > [debugbuild.txt](https://github.com/ggerganov/llama.cpp/files/15062347/debugbuild.txt) @sorasoras glancing through it, think it might related to https://github.com/ggerganov/llama.cpp/pull/6797 @jart FYI

Alongside an optional cap, I think we should make the server stop generating when the connection is closed for whatever reason (clients may well have a timeout or interrupt things...

@skoulik thanks so much for testing this out, and for reporting this stringification issue! The `expects_stringified_function_arguments` thing was me trying to work around some templates hard-coding that expectation of stringified...

(btw, congratulations for this very nice project : had thought about doing something like this for quite a while, but you did it much better than I would have hoped...

> Using Manifold, this should be possible, as demoed by https://manifoldcad.org/#Tetrahedron%20Puzzle. This feature is likely a generalization of #5047 (**edit**: incorrect statement, see reply below) ~~Manifold supports per-vertex colors, not...

I took a stab at using Manifold's OriginalID mechanism: https://github.com/openscad/openscad/pull/5180 (thanks @elalish for the explanation! also... could this be supported in 2D? would be a good incentive to switch OpenSCAD's...