lylythechosenone
lylythechosenone
Actually @Wumpf hold before you merge this. I may have (perhaps annoyingly) found a way to entirely eliminate it. If you'd like, you can move this to a draft. Edit:...
Alright. I have to figure out what causes that ICE. Once I do that, I'll probably remove the `dynamic` module and move it into core, with minor changes. I'll illustrate...
Alright @Wumpf if this check passes, this should be fine to merge.
Wait, never mind, back to draft I go! EDIT: this was a mistake, please ignore. EDIT 2: I was actually right.
I think that should do it! #5124: all of the hal steps should now be done. For now, I just patched `HalApi` to ignore these changes, but I'll start the...
I think we can do it without that change [`A` as an associated type], and as such, we should avoid that change in an effort to reduce breaking changes.
Currently my work on core is very much theoretical. The basic idea is that we slowly incorporate non-concrete Apis into core resources, using nical's DST ideas. It will require some...
> I'm not sure why we need to remove the associated type from the hal traits. Creating the associated type was my own previous change. I'm reverting it because it's...
@jimblandy the implementation in core is going to be a bit more complicated than this, because we want to keep resources and their metadata inline. This means turning e.g. `Texture`...
P.S.: the reason that as_any is needed is that downcast_* are methods on dyn Any, not Any itself (that's how it stays object safe). With the new trait upcasting feature,...