Repository navigation
Replies: 1 comment 6 replies
|
Please see: https://github.com/image-rs/image-extras#new-formats |
|
I think @fintelia meant that you publish your crate for JBIG2 and then implement the integration of your crate in |
|
I go back and forth on whether to have external crates or just tons of code in image-extras. There's tradeoffs both ways: more crates adds overhead to updates/publishing but having too much code in a single crate isn't optimal either |
|
I was sort of thinking just take the jbig2 repo and add it to the org, like as image-png exists today. Not sure what the rules are around this, but from an outsider perspective, the supply chain story is a lot cleaner when all the code |
|
Looking at the repository itself, the implementation seems to be roughly 15 thousand lines of LLM-generated code created over the course of a week. The image-rs org doesn't have a specific policy around LLM use, but I don't think that's the sort of crate that makes sense for us to adopt and promise to maintain long into the future |
|
no worries, I am not asking for you to maintain it, although I can see the implication there. i'm still interested in maintaining the code myself. why don't i submit a PR for image-extras and we'll see where it goes from there? |
Uh oh!
There was an error while loading. Please reload this page.
Hello - I recently put together a JBIG2 encoder and decoder in rust, and I was wondering what your process is for incorporating code into the library?
The project is at https://github.com/jmole/jbig2-rust
I haven't done extensive performance benchmarking on it yet, but we have a very complete set of conformance tests and validators for the format. The goal was to prioritize correctness before making it fast, given the spotty history of the JBIG2 format.
I'm happy to host it from my profile or incorporate it into the github org, or whatever else makes sense.
All reactions