molikto

> > 2 ms (even faster than Scala?!!!) > > A bit suspicious to me, maybe you are not doing the same amount of work. For forcing, you could also...

@AndrasKovacs it seems cubicaltt use cbv and renormalize after context restriction (although it is in haskell, a lazy langauge, don't know if this makes any difference?) mlang instead always works...

https://github.com/mortberg/cubicaltt/blob/master/Eval.hs#L106 here, after a nominal operation (act, swap, etc.), neutral values needs to be eliminated again (compare the VPair and `Fst` case), because context restriction can makes neutral value evaluate...

@jozefg hello, are you implementing XTT? :-)

@AndrasKovacs Mortberg suggested me to checkout pi4s3_nobug branch in cubicaltt when I am working on mlang. This is the branch with most updated computation rules I think.

> LGTM modulo comments, but I am a bit confused: is it actually possible for us to end up with `OpCompositeInsert`s with multiple indices? Do we have a pass that...

> > I suspect it's if a closure captures an storage buffer (`&[u32]`) and then another acceleration structure, the closure capturing generate a top level composite, then the array is...

> Does something like this not satisfy your use-case? > > fn setup(mut commands: Commands) { > commands.spawn((Name::new("parent"), optional_children())); > } > > fn optional_children() -> impl Bundle { >...

> Merging discussion into [#19715](https://github.com/bevyengine/bevy/issues/19715). Not sure they are exactly duplicate. What I want is allow multiple `children` in a `Bundle` by perform merging bundle content at insertion time. It...

tree sitter?