Georg Brandl
Georg Brandl
It's not only the `Vec`, you're building up `Vec`s and nested `struct`s everywhere in your solution. This is probably fine if you intend to build up an AST-like node tree...
> Would you feel up to the task to try and show correctness for the SingleCore impl and make a small write-up about it with motivation? Hehe, that's out of...
Sorry for taking a while. The biggest single increase is in a loop ``` for &byte in some_slice { queue.enqueue_unchecked(byte); } ``` It appears to be unrolled by the compiler...
Nice! I'm really loving the 0.6 changes, btw, it's a big improvement in ergonomics.
I've tried the current branch and it works fine for me. Thanks!
Sounds pretty specific. Are you sure it's worth it?
Tests are fine. Since there were tons of unrelated changes and conflicts in the branch, I manually cherry-picked the changes and pushed them in f1283e65e34483658a89bc48a885f840ce6c8a34.
I promise I'll have a look next week.
Yeah, `byteorder`'s WriteBytesExt and ReadBytesExt are not defined on `no_std` (in hindsight, makes sense, since they use `std::io` too).