fix: timeoutSignal is lost after being overriden with controller.signal

[sachiniyer]

I believe that the watch timeout abort signal was getting overwritten, and I think 0ae5ddf just forgot to delete line 44.

I wrote a one-off test to make sure this was really happening, but please let me know if I should be adding a unit test for this functionality as part of this PR.

[cjihrig]

I wrote a one-off test to make sure this was really happening, but please let me know if I should be adding a unit test for this functionality as part of this PR.

That would be very helpful!

[sachiniyer]

Hi @cjihrig !

I've added a previously failing test. If you add back the line that I deleted, you will see that the test hangs/times out.

I also made the timeout value a private instance property to make it more easily testable - hopefully that's alright.

[cjihrig]

Mostly LGTM, thanks.

[sachiniyer]

Made both adjustments @cjihrig

Lmk if anything else looks off.

[cjihrig]

The code looks good to me, but in CI it's crashing the Node.js process in C code somehow (impressive):

node: ../deps/uv/src/unix/stream.c:446: uv__stream_destroy: Assertion `!uv__io_active(&stream->io_watcher, POLLIN | POLLOUT)' failed.
✖ /home/runner/work/javascript/javascript/dist/watch_test.js (1764.106097ms)

[sachiniyer]

Hey @cjihrig , I have no idea why this is happening (or how to debug it really).

I bumped the timeout to 10 ms for the test, and the actions are passing on a test pr that I created on my fork of this repo - sachiniyer#1

If there is a way to be doing better due diligence for this change, please lmk.

[cjihrig]

Before merging, do @brendandburns or @mstruebing have any concerns given #2665 (comment)

[mstruebing]

I would run the pipeline a couple of times - I just restarted it a couple of times and it seems stable.
That's probably some bug in NodeJS itself, so I'm fine with merging when it runs kinda stable

[sachiniyer]

Hey @cjihrig and @mstruebing, thanks for reviewing this PR.

Please lmk if there is anything I can do from my side to move this along. If we are worried about the test failing, I can attempt to write it even cleaner (or maybe remove the test and just merge the fix?)

[cjihrig]

/lgtm
/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: cjihrig, sachiniyer

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [cjihrig]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment