[jdx]

https://entire.io/gh/jdx/mise/trails/637

test_text_request_allow_html_returns_the_page (added in #14253) times out on every run, which fails unit-macos on main (run 38058688044) and reproduces locally on Linux.

Its second half checks that an HTML page is still rejected without allow_html(). For an http:// URL, TextRequest::send retries as https:// first. That retry sends a TLS hello to the plain-HTTP canned server, which drains until \r\n\r\n and so blocks until the 2s client timeout. nextest's CI ceiling is 1s, so both attempts are killed.

The test now uses a small inline listener for that half: it answers the first connection with the HTML and drops the second, so the https retry fails immediately. The test goes from timing out to 0.03s, and all 34 http::tests::retry tests pass. No production code changes.

AI-assisted — Tool: Claude Code; model: anthropic/claude-sonnet-5-5; version: 2.1.293.

🤖 Generated with Claude Code

---

Note

Low Risk
Test-only change in retry.rs; no runtime HTTP behavior is modified.

Overview
Fixes a flaky/timeout failure in test_text_request_allow_html_returns_the_page on the assertion that HTML is rejected without allow_html().

The second request still uses an http:// URL (so the client’s https retry runs), but the test no longer uses spawn_canned_server with two canned responses. It binds a one-off TcpListener that serves HTML on the first connection and leaves the second connection hanging without blocking until the client timeout—matching how a TLS hello on plain HTTP used to stall the canned server under nextest’s short CI deadline.

Comments in the test document that behavior. No production HTTP code changes.

^{Reviewed by Cursor Bugbot for commit a59cf87. Bugbot is set up for automated code reviews on this repo. Configure here.}

Summary by CodeRabbit

• Tests
  • Updated coverage for HTTP-to-HTTPS retry behavior when an HTML response is followed by a dropped connection.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

• Configuration used: Repository YAML (base), Central YAML (inherited), Organization UI (inherited)
• Review profile: CHILL
• Plan: Advanced
• Run ID: c3e3ca0d-1920-43b3-a0c7-72fa8b98f78c

---

📥 Commits

Reviewing files that changed from the base of the PR and between 19e8c8f and a59cf87.

---

📒 Files selected for processing (1)

• crates/mise-util/src/http/tests/retry.rs

---

Included review availability: This review used your included allowance. Your plan provides up to 10 included reviews per hour; 9 remain after this review.

---

---

📝 Walkthrough

Walkthrough

The HTML-rejection test replaces canned responses with a TCP listener. The listener returns HTML on the first connection and accepts, then drops, the second connection to model the documented HTTP-to-HTTPS retry.

Changes

HTTP retry test

Layer / File(s)	Summary
Simulate the retry connection crates/mise-util/src/http/tests/retry.rs	The test returns HTML on the first connection, then accepts and drops the second connection.

Priority: ⬇️ Low

Estimated code review effort: 2 (Simple) | ~5 minutes

Change: Bug fix

Suggested reviewers: jambalaya56562

---

Merge Risk: ⚪ Minimal · up to a59cf

The test now closes the HTTPS retry promptly after serving the expected HTML response, avoiding the timeout pattern. No actionable merge-blocking risk is evident.

Pre-merge checks | 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check		Check skipped - CodeRabbit’s high-level summary is enabled.
Title check		The title clearly and concisely describes the main change: preventing the allow_html test from waiting for the HTTPS retry timeout.
Docstring Coverage		Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 1 functions across 1 files.
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

---

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

---

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

Instruction counts

benchmark	trend	instructions	Δ	wall (min)	Δ
env	█▁▃▄▁▁▂▄▃▄█	27,896,842 → 27,932,607	+0.13%	19.61 → 20.08ms	+2.41%
hook-env	▃▁▁██▇█▇█▇█	30,205,788 → 30,228,904	+0.08%	20.89 → 20.88ms	-0.04%
ls	▇▂▂▁▂▂▁▃▂▄█	32,117,536 → 32,174,333	+0.18%	21.69 → 21.81ms	+0.58%
registry	▇▁▅▁▁▁▁▁▃▄█	40,175,424 → 40,242,053	+0.17%	17.72 → 18.13ms	+2.33%
startup	▆▁▇▅█▇▅█▇▅▆	11,077,051 → 11,079,873	+0.03%	12.14 → 12.42ms	+2.33%

No instruction-count regression above 1%.

_{Only instruction counts gate. Wall clock is shown for context — on identical hardware it moves 4-20% run to run.}

_{Measured by tak — instruction-counted CLI benchmarks, stored in this repository's git notes.}

_{a59cf8729a5f vs 19e8c8f0d03d · measured on the runner, not pushed to the history.}