stream: fix recursive WritableStream abort

[standard-Chan]

Re-check the stream state after signaling the controller's abort signal, as required by steps 3 and 4 of the WritableStreamAbort algorithm.

Refs: https://streams.spec.whatwg.org/#writable-stream-abort

Also, reject close promise caches only when their promises are still pending. A cache lazily materialized after the stream becomes errored may already contain a rejected promise and does not have a callable reject method.

Remove the expected-failure entry for the recursive abort WPT.

Testing

$ ./node test/wpt/test-streams.js writable-streams/aborting.any.js

Files: 1/1 ran, 1 passed, 0 skipped, 0 expected failures, 0 unexpected failures, 0 unexpected passes
Subtests: 65 passed, 0 skipped, 0 expected failures, 0 unexpected failures, 0 unexpected passes

[nodejs-github-bot]

Review requested:

• @nodejs/web-standards

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.15%. Comparing base (f8f8aa2) to head (84727d9).
⚠️ Report is 10 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #64825   +/-   ##
=======================================
  Coverage   90.14%   90.15%           
=======================================
  Files         746      746           
  Lines      242818   242821    +3     
  Branches    45765    45765           
=======================================
+ Hits       218884   218904   +20     
- Misses      15417    15419    +2     
+ Partials     8517     8498   -19     

Files with missing lines	Coverage Δ
lib/internal/webstreams/writablestream.js	99.50% <100.00%> (+<0.01%)	⬆️

... and 22 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[standard-Chan]

The build-tarball job was canceled because exceeding the 15-minute execution limit. Could someone please rerun the failed job?