sled

sled copied to clipboard

The README of the `main` branch advertises zstd compression, but https://github.com/spacejam/sled/pull/1426 removed it entirely. While the plan is to reintroduce this on top of marble, semantics on the `main` branch...

flokli

Updates the requirements on [zerocopy](https://github.com/google/zerocopy) to permit the latest version. Commits See full diff in compare view Dependabot will resolve any conflicts with this PR as long as you don't...

dependabot[bot]

Updates the requirements on [env_logger](https://github.com/rust-cli/env_logger) to permit the latest version. Changelog Sourced from env_logger's changelog. 0.10.1 - 2023-11-10 0.10.0 - 2022-11-24 MSRV changed to 1.60 to hide optional dependencies Fixes...

dependabot[bot]

Removes unsafe implementations where in theory it should not affect perfoemance. I checked all implmenetations in compiler explorer, they give the equivalent code except for `slice.fill()` which is one zero...

Wicpar

Storage issues when using high count of trees

1

Hey there, I have a need of using really a lot of trees to make difference/versioning of some values. Unfortunately, looks like this cause lot of storage issues as the...

Slixe

`IoBuf::get_mut_range` unsoundly extends arbitrary lifetime to 'static

2

This method takes `&self` and returns a `'static` bounded reference. https://github.com/spacejam/sled/blob/69294e59c718289ab3cb6bd03ac3b9e1e072a1e7/src/pagecache/iobuf.rs#L84-L98 This can cause a use-after-free. The following example fails Miri: ```rust use std::{ alloc::{alloc, dealloc, Layout}, sync::Arc, cell::UnsafeCell };...

vikramnitin9

Less unsafe, in more controlled instances

2

#### Use Case: Unsafe is used almost everywhere in the code. No benchmarks are there to make sure that they give a true benefit over safe alternatives. One instance if...

Wicpar

impl for T allows also use for &T and &&T when doing T.trait_fn, else it can just be dereferenced, it's better than potentially hazardous code duplication

Wicpar

fix(benchmarks): fixed criterion benchmark not compiling on windows

1

fixed criterion benchmark not compiling on windows due to jemalloc

Wicpar

I don't know if there is a PR specification, but I believe it is a simple enough PR. **from: https://github.com/tokio-rs/tokio/issues/6085**

shuyanzhu