378 private links
Scramjet is an experimental interception based web proxy designed to evade internet censorship, bypass arbitrary web browser restrictions and innovate web proxy technologies. This project strives to maintain security, developer friendliness and performance unlike many other web proxies regardless of its open source nature.
It's more than a proxy
It's hard to set zero in the RAM actually
It's absolutely possible to beat even the best sort implementations with domain specific knowledge, careful benchmarking and an understanding of CPU micro-architectures. At the same time, assumptions will become invalid, mistakes can creep in silently and good sort implementations can be surprisingly fast even without prior domain knowledge. If you have access to a high-quality sort implementation, think twice about replacing it with something home-grown.
Help to build microservices as WASM components in Rust.
It can be used to handle HTTP requests for example.
There is a demonstration: https://www.youtube.com/watch?v=UoRfr3Q2R8A
An extremely fast PHP linter, formatter, and static analyzer, written in Rust.
That's interesting
KISS for maintainability: "Nothing in Rust forces us to get fancy. You can write straightforward code in Rust just like in any other language. But in code reviews, I often see people trying to outsmart themselves and stumble over their own shoelaces. They use all the advanced features at their disposal without thinking much about maintainability."
Here an real life example.
But if simplicity is so obviously “better,” why isn’t it the norm? Because achieving simplicity is hard!
Even in Rust: abstractions are never zero cost for developers.
Often, simple code can be optimized by the compiler more easily and runs faster on CPUs. That’s because CPUs are optimized for basic data structures and predictable access patterns. And parallelizing work is also easier when that is the case. All of that works in our favor when our code is simple.
Most of the code you’ll write for companies will be application code, not library code. That’s because most companies don’t make money writing libraries, but business logic. There’s no need to get fancy here. Application code should be straightforward to keep your fellow developers in mind.
Tips:
- start small
- avoid optimization early
- delay refactoring: we have limited information at the time of writing our first prototype.
- write code for humans
- The right abstractions guide you to do the right thing
A phishing attack is running on crates.io
En intégrant SIMD, les performances semblent meilleures que les outils. Ici l'outil base64 passe de 4,9 secondes à 3,1.
Bound checking costs 0.3%
It would be great to have a crate registry independent from github (and Microsoft).
Competition would also lead to crates.io improving or the two registries inspiring each other.
My honest assessment is that the Rust / Cargo leadership need to be bold, deprecate packages publishing to crates.io and move to a decentralized package distribution architecture.
There is no other way around. Rust needs to copy Go, it's as simple as that.
The second part of the solution, way harder and more expensive to implement is to release an extended standard library. We need to reduce the amount of third-party developers that we need to trust to release our software.
A workaround from now is to import rust crates from git itself, such as ring = { git = "https://github.com/briansmith/ring", version = "0.12" }
Optimizing some endpoints in Rust inside a go app.
The results shows nearly 2x performance.