http: writable set to false

[ronag]

This is another take on #15404 (comment).

Depends on #28681

Refs: #15029

Checklist

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

[ronag]

@mcollina as far as I understand the problem with the original fix (and probably this one) in #14024 causes regression in ecosystem. Are there test for that? How do I run those myself?

[mcollina]

I'm -1 to adding _writableState property to OutgoingMessage.

[mcollina]

@ronag this was it: https://github.com/SimenB/node-stream-repro/blob/master/index.js.

Essentially that writable property underpins both pump and pipeline.

[ronag]

@mcollina I've added a test. Assuming pipeline is pretty much the same as pump? this should work. Basically by adding a _writableState we can "fool" pipeline and eos to treat it differently/correctly.

[mcollina]

I've added a test. Assuming pipeline is pretty much the same as pump? this should work. Basically by adding a _writableState we can "fool" pipeline and eos to treat it differently/correctly.

I'm not convinced with this approach, adding a fake proxy is definitely going to break other things.

[ronag]

I'm not convinced with this approach, adding a fake proxy is definitely going to break other things.

Not much I can do with that. Is there any chance I can take a look at "other things" or is it too ambitious? Otherwise I guess this PR is closable.

[ronag]

This PR needs more work. However, if the idea of implementing _writableState to resolve the original issue is a no go I won't put further effort into this.

[ronag]

I don't remember what this is about anymore. Will get back to it in there future.