http: cleanup ClientRequest oncreate

[ronag]

Iterative work towards cleaning up ClientRequest.

[ronag]

@nodejs/http

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/35444/

[nodejs-github-bot]

CI: https://ci.nodejs.org/job/node-test-pull-request/35447/

[Trott]

Looks good to me. Should we run a subset of the http benchmarks before landing? @mscdex

@nodejs/http

[mcollina]

lgtm

[dnlup]

lgtm

[dnlup]

@Trott is benchmark-node-micro-benchmarks the CI job for running benchmarks in this list?

[ronag]

@dnlup Do you want to try and land this?

[dnlup]

@dnlup Do you want to try and land this?

@ronag Yes 👍🏻 . I'll run the benchmark job on the http subsystem first.

[dnlup]

https://ci.nodejs.org/job/benchmark-node-micro-benchmarks/897/

[dnlup]

The job is pretty long. It's still running. I don't see performance regressions so far, inspecting the job console output.

[dnlup]

Landed in 963dec3

[Trott]

@dnlup The http benchmarks take a long time to run, so it's usually a good idea to use the filter option to run a subset of the http benchmarks. In this case, I think using request as the value for FILTER would have limited the benchmarks to just client-request-body.js and create-clientrequest.js.

I'll open a draft revert PR so we can see that benchmark. Assuming no significant performance issues, I'll close the draft revert PR.

[dnlup]

@dnlup The http benchmarks take a long time to run, so it's usually a good idea to use the filter option to run a subset of the http benchmarks. In this case, I think using request as the value for FILTER would have limited the benchmarks to just client-request-body.js and create-clientrequest.js.

I'll open a draft revert PR so we can see that benchmark. Assuming no significant performance issues, I'll close the draft revert PR.

@Trott thank you for the explanation. I apologize for the extra work I caused.

[Trott]

@Trott thank you for the explanation. I apologize for the extra work I caused.

No worries at all. I was only letting you know in case this is an area where you plan on doing other similar work and may need to run another http benchmark.