Allow decompression to continue after exceeding max_length

[Dreamsorcerer]

Architecture summary:

• Compression utils now have a .data_available attribute, when True the decompress() call can be repeated with b"" to get more data. The decompression output has been reduced to 256KiB, matching the socket read limit.
• DeflateBuffer/StreamReader.feed_data() now return True if there is more data available by calling the method again with b"".
• (PY) PayloadParser.feed_data() now returns an enum indicating when the payload is complete or whether more input is needed to continue producing output.
• (PY) HttpParser.feed_data() now returns early when the payload parser is paused and has a new attribute to track if more data is available from the payload parser.
• (C) HttpParser.cb_on_body() now processes the decompressed data in chunks and pauses llhttp if the parser is asked to pause.
• (C) HttpParser.feed_data() now call cb_on_body() again when more data is available.
• Both parsers now have a .pause_reading() method to stop the parser from producing more output.
• BaseProtocol.pause_reading() now calls the parser's .pause_reading() method.
• BaseProtocol.resume_reading() now resumes parsing and then checks if it's been paused again before telling the transport to resume.

[codecov]

Codecov Report

❌ Patch coverage is 99.44649% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 98.92%. Comparing base (b5f1aa5) to head (293d7f6).
⚠️ Report is 1 commits behind head on master.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
aiohttp/web_protocol.py	83.33%	1 Missing and 1 partial ⚠️
aiohttp/_http_parser.pyx	98.52%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #11966      +/-   ##
==========================================
+ Coverage   98.89%   98.92%   +0.02%     
==========================================
  Files         133      133              
  Lines       46257    46550     +293     
  Branches     2405     2423      +18     
==========================================
+ Hits        45746    46048     +302     
+ Misses        379      373       -6     
+ Partials      132      129       -3     

Flag	Coverage Δ
CI-GHA	98.98% <99.36%> (+0.02%)	⬆️
OS-Linux	98.72% <99.36%> (+0.01%)	⬆️
OS-Windows	96.98% <98.94%> (+0.01%)	⬆️
OS-macOS	97.88% <99.15%> (+0.41%)	⬆️
Py-3.10.11	97.39% <94.72%> (+1.11%)	⬆️
Py-3.10.20	97.86% <94.72%> (+0.09%)	⬆️
Py-3.11.15	98.11% <98.73%> (+0.67%)	⬆️
Py-3.11.9	97.65% <98.73%> (+1.16%)	⬆️
Py-3.12.10	97.73% <98.52%> (+<0.01%)	⬆️
Py-3.12.13	98.20% <98.52%> (-0.01%)	⬇️
Py-3.13.12	98.45% <98.73%> (+0.68%)	⬆️
Py-3.14.3	98.50% <97.25%> (+1.69%)	⬆️
Py-3.14.4t	97.51% <97.46%> (+<0.01%)	⬆️
Py-pypy3.11.15-7.3.21	97.34% <93.24%> (?)
VM-macos	97.88% <99.15%> (+0.41%)	⬆️
VM-ubuntu	98.72% <99.36%> (+0.01%)	⬆️
VM-windows	96.98% <98.94%> (+0.01%)	⬆️
cython-coverage	38.21% <78.04%> (+0.40%)	⬆️

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.

[codspeed-hq]

Merging this PR will degrade performance by 45.32%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 5 improved benchmarks
❌ 3 (👁 3) regressed benchmarks
✅ 53 untouched benchmarks
⏩ 4 skipped benchmarks¹

Performance Changes

	Benchmark	BASE	HEAD	Efficiency
⚡	test_one_hundred_get_requests_iter_chunks_on_10mb_chunked_payload[pyloop]	1.2 s	1.1 s	+8.5%
👁	test_get_request_with_251308_compressed_chunked_payload[zlib_ng.zlib_ng-pyloop]	207 ms	244.4 ms	-15.32%
👁	test_get_request_with_251308_compressed_chunked_payload[isal.isal_zlib-pyloop]	62.1 ms	113.5 ms	-45.32%
⚡	test_one_thousand_round_trip_websocket_binary_messages[pyloop-large]	110.6 ms	101.7 ms	+8.7%
⚡	test_read_large_binary_websocket_messages[pyloop]	8,940.3 µs	49 µs	×180
⚡	test_ten_streamed_responses_iter_chunks[pyloop]	131.5 ms	119.6 ms	+9.9%
⚡	test_ten_streamed_responses_iter_chunked_1mb[pyloop]	1.2 s	1.1 s	+10.11%
👁	test_ten_streamed_responses_iter_chunked_4096[pyloop]	28.2 ms	35.5 ms	-20.68%

---

_{Comparing Dreamsorcerer-patch-5 (293d7f6) with master (b5f1aa5)}

Footnotes

1. 4 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[patchback]

Backport to 3.14: 💔 cherry-picking failed — conflicts found

❌ Failed to cleanly apply 0c7ce34 on top of patchback/backports/3.14/0c7ce3485dbcda196e125b4ea913a67e6b54f26c/pr-11966

Backporting merged PR #11966 into master

1. Ensure you have a local repo clone of your fork. Unless you cloned it
   from the upstream, this would be your origin remote.
2. Make sure you have an upstream repo added as a remote too. In these
   instructions you'll refer to it by the name upstream. If you don't
   have it, here's how you can add it:

   $ git remote add upstream https://github.com/aio-libs/aiohttp.git

3. Ensure you have the latest copy of upstream and prepare a branch
   that will hold the backported code:

   $ git fetch upstream
   $ git checkout -b patchback/backports/3.14/0c7ce3485dbcda196e125b4ea913a67e6b54f26c/pr-11966 upstream/3.14

4. Now, cherry-pick PR Allow decompression to continue after exceeding max_length #11966 contents into that branch:

   $ git cherry-pick -x 0c7ce3485dbcda196e125b4ea913a67e6b54f26c

   If it'll yell at you with something like fatal: Commit 0c7ce3485dbcda196e125b4ea913a67e6b54f26c is a merge but no -m option was given., add -m 1 as follows instead:

   $ git cherry-pick -m1 -x 0c7ce3485dbcda196e125b4ea913a67e6b54f26c

5. At this point, you'll probably encounter some merge conflicts. You must
   resolve them in to preserve the patch from PR Allow decompression to continue after exceeding max_length #11966 as close to the
   original as possible.
6. Push this branch to your fork on GitHub:

   $ git push origin patchback/backports/3.14/0c7ce3485dbcda196e125b4ea913a67e6b54f26c/pr-11966

7. Create a PR, ensure that the CI is green. If it's not — update it so that
   the tests and any other checks pass. This is it!
   Now relax and wait for the maintainers to process your pull request
   when they have some cycles to do reviews. Don't worry — they'll tell you if
   any improvements are necessary when the time comes!

🤖 @patchback
I'm built with octomachinery and
my source is open — https://github.com/sanitizers/patchback-github-app.