[GiHoon1123]

Fixes: #66347

When a peer-initiated unidirectional stream sent data and FIN before the consumer performed
its first async iterator pull, the native stream could be cleaned up before the reader was
created. The buffered data was then no longer available to the JavaScript stream.

Create the reader when a readable QuicStream is constructed so the underlying data queue
remains available until the consumer pulls from the iterator. Local unidirectional streams
remain unreadable and do not allocate a reader.

This means every readable stream allocates a reader whether or not it's ever read, instead
of only on first pull. #66347 asked whether lazy allocation was the right call for this;
eager allocation is the answer this PR takes to that question.

Added a regression test covering data received before the first iterator pull.

This also preserves buffered incoming data after local session or stream destruction,
even when reading has not started. Updated the teardown tests to wait for the full payload
and check both the retained data and the truncation error. Added coverage for explicit
errors after reading 128 KiB with both truncatedReads policies, and documented this behavior.

Rebased onto main at 763524d and wrapped the commit message to 72 columns.

Validation on macOS:

• QUIC-enabled Node build
• python3 tools/test.py --mode=release -j6 -t60 -p tap 'parallel/test-quic*':
  269 passed, 2 skipped (ngtcp2_test_client and ngtcp2_test_server binaries not built)
• ESLint for the changed JS files and delayed-reader test
• Markdown lint for doc/api/quic.md

[nodejs-github-bot]

Review requested:

• @nodejs/quic

[github-actions]

Welcome to Node.js, and thank you for your first contribution!

Before review, please take a moment to read:

• the guide for first-time contributors
• the contribution and automation policies
• the pull request guide
• the AI use policy
• the Code of Conduct

Please make sure every commit is signed off. For a first pull request, GitHub Actions require collaborator approval and Jenkins CI must be started by a collaborator or triager, so an initial wait is normal.

[GiHoon1123]

Sorry about that - didn't follow the one-issue-at-a-time guideline. Closing this for now and focusing on #64627 first, will reopen once that's merged.

[MikeMcC399]

The problem with multiple PRs for first-time contributors is that every step need to be manually approved by a team member, which creates a high workload and slows down the progression of the PRs.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.58%. Comparing base (1eb734d) to head (c3cf330).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #66407   +/-   ##
=======================================
  Coverage   90.58%   90.58%           
=======================================
  Files         797      797           
  Lines      278298   278304    +6     
  Branches    53523    53517    -6     
=======================================
+ Hits       252095   252104    +9     
+ Misses      16570    16558   -12     
- Partials     9633     9642    +9     

Files with missing lines	Coverage Δ
lib/internal/quic/quic.js	100.00% <100.00%> (ø)

... and 31 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.

[martenrichter]

@jasnell @pimterry
I think this PR does the trick. I have thought, if there is another way, than always allocating the reader in this case, but I do not think there is.

[GiHoon1123]

Thanks for taking a look. I closed this after learning about the first-time contributor guideline to focus on one PR at a time. Would it be okay to reopen it now?

[martenrichter]

Let's see what the others say, I think they can reopen.

[martenrichter]

I think you should reopen.
From me, you get a:

LGTM

as I just cherry picked it to my webtransport branch.

[GiHoon1123]

Thanks for the review! I've reopened the PR.

[GiHoon1123]

Rebased on main and fixed the commit message line lengths. Thanks!