fix: correctly handle nexttick scheduling in stream reads

[pfrazee]

Description of Change

We discovered some cases where protocol streaming responses would stall before completing. We isolated the issue to the case where a stream's read handler scheduled a push() using nextTick(), which (according to @mafintosh, who I trust to know about this sort of thing) will get scheduled to run prior to control being returned to C++. This means the call into js' read() will return nothing, but the 'readable' event will get missed.

We were able to fix this by flagging when that occurs and automatically retrying the read.

Checklist

• PR description included and stakeholders cc'd
• npm test passes
• tests are changed or added
• PR title follows semantic commit guidelines
• PR release notes describe the change in a way relevant to app developers, and are capitalized, punctuated, and past tense.
• This is NOT A BREAKING CHANGE. Breaking changes may not be merged to master until 11-x-y is branched.

Release Notes

Notes: Fix an issue which would cause streaming protocol responses to stall in some cases

[mafintosh]

Just to add a little context. nextTicks in Node er technically afterTicks as they run in the same event tick as the last thing (legacy naming).

In practice this bug manifest itself almost immediately on any stream returned that’s the last stream in a pipeline as they will have lots of these “read returns null but data is added later in the tick” events

[nornagon]

NodeStreamLoader needs a rewrite. It's extremely subtle and full of incorrect behaviors like this 😞

Preferably, a rewrite would have the C++ side expose a much simpler "raw" API than NodeStreamLoader does, and handle translating between node's stream API and the "raw" API in JavaScript.

[zcbenz]

Node.js streaming is so hard 😞 , thanks for fixing this!

[release-clerk]

Release Notes Persisted

Fix an issue which would cause streaming protocol responses to stall in some cases

[trop]

I was unable to backport this PR to "8-x-y" cleanly;
you will need to perform this backport manually.

[trop]

I have automatically backported this PR to "10-x-y", please check out #24081

[trop]

I have automatically backported this PR to "9-x-y", please check out #24082

[trop]

@zcbenz has manually backported this PR to "8-x-y", please check out #24133