Tuna-i3-Plus VS crates.io

I did some similar things with my old libtuna for AVR, mainly with intsz and with how fixed-point values are handled.

As for crates, I haven't used https://github.com/nashaofu/screenshots-rs but it seems like it may do what you want? I found it by searching "rust wayland screenshot" on Google. It is apparently not the same library as the 'screenshots-rs' crate on crates.io, though

I've started improving my proof-of-concept Free Monad macro to make it ready for a proper release on crates.io.

The discoverability problem could be solved in other ways, like at least a community recommended libraries list, though curating it is difficult. Maybe something like most directly depended-upon crates is a good start, since transitive dependencies are by definition ones you didn't need to learn to use directly, but direct dependencies are ones that reflect how many people need to know a library exists. Someone did this analysis years ago, I guess it should be automated and made into a dynamic page. crates.io tracks the metadata for dependencies already, so that seems like a natural place for it.

you could just disallow code that uses this from entering the crates.io ecosystem. it would be used within end user projects.

I'm one of the core engineers on QuestDB. We're a time series database written in Java - with a twist: We don't do any garbage collection on the hot path by pooling memory and reusing objects. This approach makes us quite performance-competitive, but it does limit our access to other dependencies as most other Java libraries allocate memory which will eventually need to be collected later. In practice, this means that up until now we've been calling from Java into C and C++ libraries via JNI (Java Native Interface) for things like networking (epoll, io_uring, etc) and jitting our SQL queries to assembler. Enter Rust: Crates.io is pretty awesome. There's lots of crates out there that do all sorts and we'd like to make use of some of those. The Rust jni crate is pretty decent too.

For sure! There is not too much I can do about that right now sadly, the issue is that negative responses (404s) of the crates.io API aren't setting the right CORS headers, so diff.rs cannot know if a crate was not found (404) or if there was an actual error. The crates.io team has merged this: https://github.com/rust-lang/crates.io/commit/9e774556ec78f47d82a6ca467b25d3c096761381 which I think will let me fix this, but I have to do more testing. Thanks for the feedback!

It works by fetching a crate's metadata from crates.io, downloading the sources using reqwest, uncompressing them using flate2, extracting them (using tar), and finally rendering a diff (using similar). In the UI you can switch between different versions of the crate to diff against and it also has a search bar to enter the crate name.

Thank you for spotting this! Strange, it does work on Firefox for me. I'll investigate it and perhaps create some more merge requests for crates.io for the CORS stuff, have already had them fix something.

None of this is published on crates.io yet, since it can't be used with bevy's current stable release, and doesn't have the blessing of the related upstream parties re. naming adjacency.

https://crates.io/ homepage has new crates and just updated listed