http: remove pause()/resume() from parser bindings

[addaleax]

As far as I can tell, these are effectively no-ops
(http_parser_pause only sets the http_errno flag,
but that isn’t read anywhere), so removing them cleans
things up a tiny bit.

The corresponding regression test remains intact.

/cc @nodejs/http-parser

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• tests and/or benchmarks are included
• commit message follows commit guidelines

[nodejs-github-bot]

@addaleax build started: https://ci.nodejs.org/blue/organizations/jenkins/node-test-pull-request-lite-pipeline/detail/node-test-pull-request-lite-pipeline/399/pipeline

[addaleax]

CI: https://ci.nodejs.org/job/node-test-pull-request/16149/

[indutny]

They're not a no-op. The http_errno is checked on every http_parser_execute() run, and the parsing won't happen.

[addaleax]

@indutny Thanks, that’s why I pinged the team… is there any behaviour that we could test against? It’s not quite ideal that all tests are passing without this, I’d say.

[indutny]

I'm pretty sure that there's a testable effect of having pause/resume, not sure if it is desirable though 😛 It'd be great indeed if someone would write a test for it.

[mcollina]

Why do we have to pause the parser as well? We are already pausing the underlining socket.

[BridgeAR]

Any progress here?

[Trott]

@addaleax Is this land-able after a rebase? Or is there something that has to happen here? (Write a test or something?)

[addaleax]

@Trott Yes, this is more of a situation where we should test that these methods actually do something (and if it turns out that they actually don’t, we can still remove them). It’s on my backlog 😄

[Trott]

FWIW, needs a rebase.

[BridgeAR]

Ping @addaleax