Experimental support for WebSocket extension permessage-gzip

[jakob-keller]

What do these changes do?

Provides experimental support for communicating with WebSocket servers that require the non-standard permessage-gzip extension.

Are there changes in behavior for the user?

None

Is it a substantial burden for the maintainers to support this?

No

Related issue number

Fixes #9933

Checklist

• I think the code is well written
• Unit tests for the changes exist
• Documentation reflects the changes
• If you provide code modification, please add yourself to CONTRIBUTORS.txt
• Add a new news fragment into the CHANGES/ folder

[codspeed-hq]

CodSpeed Performance Report

Merging #9934 will not alter performance

_{Comparing jakob-keller:permessage-gzip (19c847d) with master (73691e4)}

Summary

✅ 36 untouched benchmarks

[codecov]

Codecov Report

Attention: Patch coverage is 81.81818% with 4 lines in your changes missing coverage. Please review.

Project coverage is 98.70%. Comparing base (73691e4) to head (19c847d).

Files with missing lines	Patch %	Lines
aiohttp/_websocket/helpers.py	76.47%	2 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #9934      +/-   ##
==========================================
- Coverage   98.72%   98.70%   -0.02%     
==========================================
  Files         120      120              
  Lines       36570    36579       +9     
  Branches     4361     4363       +2     
==========================================
+ Hits        36102    36107       +5     
- Misses        315      317       +2     
- Partials      153      155       +2     

Flag	Coverage Δ
CI-GHA	98.60% <81.81%> (-0.02%)	⬇️
OS-Linux	98.28% <81.81%> (-0.02%)	⬇️
OS-Windows	96.10% <73.68%> (-0.02%)	⬇️
OS-macOS	97.38% <77.27%> (-0.01%)	⬇️
Py-3.10.11	97.22% <77.27%> (-0.02%)	⬇️
Py-3.10.15	97.82% <81.81%> (+0.04%)	⬆️
Py-3.11.10	97.81% <81.81%> (-0.03%)	⬇️
Py-3.11.9	97.26% <77.27%> (-0.02%)	⬇️
Py-3.12.7	98.35% <81.81%> (-0.02%)	⬇️
Py-3.13.0	98.29% <81.81%> (-0.01%)	⬇️
Py-3.9.13	97.14% <77.27%> (-0.02%)	⬇️
Py-3.9.20	97.74% <81.81%> (+0.03%)	⬆️
Py-pypy7.3.16	97.30% <81.81%> (-0.02%)	⬇️
VM-macos	97.38% <77.27%> (-0.01%)	⬇️
VM-ubuntu	98.28% <81.81%> (-0.02%)	⬇️
VM-windows	96.10% <73.68%> (-0.02%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[jakob-keller]

While technically feasible, the lack of specification for permessage-gzip makes this PR difficult to frame. I will pause further work on this PR until the provider of the WebSocket API clarifies their choice of compression to me.

[Dreamsorcerer]

While technically feasible, the lack of specification for permessage-gzip makes this PR difficult to frame. I will pause further work on this PR until the provider of the WebSocket API clarifies their choice of compression to me.

Well, if it ends up depending on the vendor, then we'd probably need something where the user can register extensions of their choice and define how it works. Then it could be made to work with any vendor's extensions.