Skip to content

Conversation

@Vizonex
Copy link

@Vizonex Vizonex commented Sep 8, 2025

This pull request will add a Cython implementation to HPack to make it run faster as well as allow users to utilize this entire library inside of cython to increase performance.

I had decided to go a step above what the aiohttp developers wanted out of this library to achieve a faster experience which should be really exciting the only restraint I'm going to have is how much time I will get to work on something like this since I have a parttime job that runs in the afternoons until evenings. However I am more than happy to have the aiohttp developers and h2 developers review my changes to see what would be the best outcome.

as of right now my changes are incomplete so that's just warning. This shouldn't be too difficult to maintain since I could always come in and maintain portions of this library if my assistance is requested.

fixes #282

@Kriechi
Copy link
Member

Kriechi commented Sep 8, 2025

Thanks for raising this and the effort in this PR. Unfortunately, as it is right now, it would result in a major API break and change that is backwards incompatible. So, I don't think this is the right approach.

Maintaining the new code is indeed a concern. I am not keen on taking on this extra workload at the moment. I think for now, since this is mostly a re-write of existing code into Cython, I would suggest developing and maintaining it as a standalone GitHub repository and project. We can discuss bringing it under the python-hyper umbrella if it proofs valuable and enough users are interested in it. What do you think? Any ideas how to reduce the overall code duplication and make future development and maintaining with security fixes et al. easier for us and the overall community?

@Vizonex
Copy link
Author

Vizonex commented Sep 9, 2025

Thanks for raising this and the effort in this PR. Unfortunately, as it is right now, it would result in a major API break and change that is backwards incompatible. So, I don't think this is the right approach.

Maintaining the new code is indeed a concern. I am not keen on taking on this extra workload at the moment. I think for now, since this is mostly a re-write of existing code into Cython, I would suggest developing and maintaining it as a standalone GitHub repository and project. We can discuss bringing it under the python-hyper umbrella if it proofs valuable and enough users are interested in it. What do you think? Any ideas how to reduce the overall code duplication and make future development and maintaining with security fixes et al. easier for us and the overall community?

sure, no problem whatsoever I think I can arrange for that and let aiohttp handle the cython related stuff we have planned I'll request to open a repository for it in the aiolibs organization and go from there, sorry I couldn't respond sooner I was at my job when I read the response. :)

@Vizonex Vizonex closed this Sep 9, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Cythonize HPack

2 participants