musli

musli copied to clipboard

This allows the `System` allocator to be statically re-used.

udoprog

udoprog

Document no_std, no_alloc capabilities.

1

It would be nice if the readme or the doc would make it clear which crates/formats are well suited for no_std or even no_alloc environments. I see that you support...

t-moe

zerocopy: deriving the total size in bytes including `Ref` fields

4

Would it make sense to include in `ZeroCopy` a function that returns the total size in bytes of the struct, taking into account the length of `Ref`s? One use case...

duarten

This is intended to improve support for Option #4 in this blog post. The idea here is to have two parallel data models, and the ability to automatically convert between...

udoprog

Macros should use Result and Option through fully qualified names

2

The first example on https://docs.rs/musli/0.0.49/musli/derives/index.html does not build. I ran `cargo test --doc` locally and noticed there were three `ignored` tests which included this failing one, so I presume there...

AaronKutch

Untagged enum support

1

In the doc the section is missing so I guess it's unimplemented. It would be handy to have it similar to serde: https://serde.rs/enum-representations.html#untagged

l4l

Floating point variable-length encoding is optimized for values like 0.000000000000000000000000000000000000000000014

5

I'm coming at this from the perspective of writing a musli-wire format for a very float-heavy data structure. As a primary observation, floats in musli-wire are encoded by bitcasting them...

cormac-ainc

Feature request: variant of encode_bytes that lets you avoid allocating your own buffer for conversion

1

Continuing on from #24 but off-topic for that issue: Most of the floats in this float-heavy format will be done using custom Encode/Decode impls on newtyped `Vec` or `Vec`. Doing...

cormac-ainc

Musli-wire forces variable-length integer encoding everywhere inside #[musli(packed)]

1

I was looking into why a `#[musli(packed)] struct Point3 { x: f32, y: f32, z: f32 }` was so small when it was all-zeros despite setting the integer encoding to...

cormac-ainc