[vdusek]

test_html_to_text_is_linear_in_node_count (added in #2372) failed intermittently in CI, mostly on macOS, for both parsers and on unrelated PRs. It timed the 20k-node runs and then the 80k-node runs with perf_counter, each size in its own window a few seconds long. When other xdist workers got busy during one of those windows, the wall-clock ratio went past 10 even though the code is linear.

The test now measures the process's own CPU time with time.process_time() and alternates the small and large runs, so both sizes see the same load. The threshold stays at 10.

Locally, with CPU load switching on and off on all cores, the old test failed 3 times out of 24 and the new one passed 100 times out of 100. With the quadratic code from before #2372 restored, it still fails for both parsers at a ratio of about 16.

✍️ Drafted by Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 94.10%. Comparing base (06d331b) to head (70af630).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2384      +/-   ##
==========================================
+ Coverage   94.05%   94.10%   +0.04%     
==========================================
  Files         184      184              
  Lines       13662    13662              
==========================================
+ Hits        12850    12856       +6     
+ Misses        812      806       -6     

Flag	Coverage Δ
unit	94.10% <ø> (+0.04%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 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.

[apify-factory]

Looks good, nothing blocking.

🏭 Apify Factory auto approved ✅ — the juror tried very hard to object and could not

Every gate passed, and neither the reviewer nor a juror found anything a human needs to look at.

_{Approval locked to 70af630f1 · workflow run}