I'm testing within the fork. It almost works, but fails to import the compiled parser for some reason. Will get back when it's fixed.

@ShogunPanda OK, that's ready to go and can be merged in now.

The 100_continue test failures are #249. The other test failure is just due to the wording of an error message being changed. I'll update that final test when we have a new release that fixes the other 2 tests.

@Dreamsorcerer This looks go to me. Once aiohttp is all green with llhttp I'll merge this.

Well, if you merge it first, then you should see the tests pass once #249 is fixed. We need the regression resolved before we can push a new release.

Once aiohttp is all green with llhttp I'll merge this.

@ShogunPanda The issue is on llhttp's side. If you do not agree, express your opinion accordingly.

@benz0li @Dreamsorcerer I get a failure on llhttp CI. Can you please verify it?

@benz0li @Dreamsorcerer I get a failure on llhttp CI. Can you please verify it?

@ShogunPanda If you are talking about https://github.com/nodejs/llhttp/actions/runs/6225384656/job/16918217042?pr=250:

The 100_continue test failures are #249.

This needs to be fixed on llhttp's side.

@Dreamsorcerer @ShogunPanda Apparently you both disagree about which side is causing the problem.

Please clarify this as soon as possible. Thank you.

Yes, as mentioned above, the 100_continue failures:
https://github.com/nodejs/llhttp/actions/runs/6225384656/job/16918217042?pr=250#step:9:701

They are caused by #249, so if you merge this in now, you can use the CI to verify any fixes to that issue (i.e. you'll see those 2 failures disappear in the PR that fixes it)..

@indutny As the initiator of llhttp, I would greatly appreciate your objective view on this.

Released v9.1.3.

@ShogunPanda Thanks a lot. That's merged now, so if you rerun the last test run on master, then it should go green now:
https://github.com/nodejs/llhttp/actions/runs/6394836947/job/17357138399