Maxim Zaks

`04 00 00 00 F4 FF FF FF 04 00 00 00 00 00 00 00 06 00 08 00 04 00` - flatcc, doublevec 0 len a vector no...

Ah, @mikkelfj I think you are correct, sorry. It is `i32` because of the vTable reuse feature. I forgot about it.

> Your next analysis is correct from a cursory look. That is why I earlier said the second set of buffers appear correct, but the first do not make any...

HTTP/3 RFC 9114 is now officially out https://datatracker.ietf.org/doc/rfc9114/ I would be interested in contributing a pure Swift implementation.

Actually the reading of a variable length offset can be quite efficient as we have a simplified problem on our hands. The runtime value can be UInt64, the buffer is...

@mikkelfj most variable length encodings rely on loops and conditions for decoding. My proposal is branch less. It is still definitely slower than just direct four byte read, but as...

Totally agree regarding testing. And I am also not so sure regarding lookup performance. That sad, the mask can also be computed purely arithmetically: ``` mask = 0xffffffffffffffff >> ((8...

> Do you or @mzaks already have a generic sorting algorithm on AnyType/some refined trait such as the ordering one? If so, would very much welcome a PR into the...

And another point, there is a class of sorting algorithms specific for strings. https://github.com/rantala/string-sorting is a treasure trove when it comes to string sorting. I ported multi key sorting algorithm...

@JoeLoser how do we want to start? You will opensource the sorting package? I can add the `ComparableCollectionElement` trait which is a `CollectionElement` and has the compare methods. And implement...