Johan Forsberg

I see, I didn't think of that! It would be great though to not have to do this "carpet escaping". As far as I can tell from https://orgmode.org/worg/dev/org-syntax.html , mainly...

Maybe this doesn't address your problem, but "modifying" a namedtuple doesn't have to be too bad: ```python new_vec2 = vec2._replace(x=2.0) ```

I am seeing this right now too, I think. I noticed that my collapsed status buffer is behaving in a confusing way where TABbing to reveal diffs will sometimes reveal...

I can add that this does not seem to happen if you only use "n"/"p" to move between the headings. It may require some manual cursor movement at the beginning...

Hm, interesting @kyleam! That does seem related. However, I compiled a new emacs from master yesterday, and I still see this behavior, so I think there's something more going on.

Guess we can check off the "instancedCube" box due to #246. Also there is a "two-cubes" example already.

Does the "gpu" part necessarily depend on glfw?

I tried SSE using `sse-starlette` and `hx_sse` and it works fine as far as I can tell! Do you intend to integrate it further than that? `hx-sse` will go away...

Interesting! It was a little fiddly; adding the right HTMX attributes in the correct places, and creating SSE endpoints. But on the other hand it's flexible, and I found that...

Sure, here's a minimal example: ```python import asyncio import ludic.html as h from ludic.web import LudicApp from sse_starlette.sse import EventSourceResponse, ServerSentEvent app = LudicApp(debug=True) @app.get("/") async def index() -> h:...