test(cua-driver/windows): robustify slider_drag coords + web CDP-port isolation - #2049
Conversation
… isolation Both surfaced in the overnight harness sweep; both are TEST-side (the driver paths are correct, VM-confirmed). - harness_wpf_slider_drag: the hardcoded drag y=275 landed ~29px ABOVE the slider track (which sits at window-local y≈304, on empty GroupBox space), so the thumb never moved — the SendInput/foreground-drag path itself is fine (the UIA-Invoke companion passes; corrected coords advance the value on the VM). Move the drag to the track row (44,304)→(330,304). (TODO: derive from the sld-value element frame for DPI independence.) - harness_web_test: these fixed-CDP-port (9222/9223) tests must run serially; a prior host still releasing its port made the daemon discover the stale host's page[0] (counter check then fails). Add wait_port_free() before launch to close the teardown-overlap window. Run with --test-threads=1. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
|
The latest updates on your projects. Learn more about Vercel for GitHub. |
|
Caution Review failedThe pull request is closed. ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (2)
📝 WalkthroughWalkthroughThe web test harness now waits for the fixed CDP port to be free before starting a new session, and the WPF slider drag test now uses explicit window-local coordinates with updated comments. ChangesWeb harness port readiness
WPF slider drag coordinates
Sequence Diagram(s)Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Possibly related PRs
Poem
✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
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. Comment |
Linux visual regression artifactsMatrix jobs now run independently. Download visual artifacts from this workflow run.
|
Two Windows harness tests flagged in the overnight harness sweep. Both are test-side — the driver paths are correct (VM-confirmed); no driver bug.
harness_wpf_slider_dragThe hardcoded drag
y=275landed ~29px above the slider track (which sits at window-local y≈304, on empty GroupBox space), so the thumb never moved — but the SendInput/foreground-drag path itself is fine: the UIA-Invoke companion (harness_wpf_slider_increase_large) passes, and corrected coords advance the value. Moved the drag to the track row(44,304)→(330,304).✅ Sent drag via SendInput … (358,306)→(644,306)→thumb tracked→ 1 passed.sld-valueelement frame for DPI/placement independence.harness_web_test(CDP-port isolation)The 5 web tests use fixed CDP ports (9222/9223) + a process-global
CUA_DRIVER_CDP_PORT, so they must run serially (--test-threads=1). A prior host still releasing its port let the daemon discover the stale host'spages[0](the click then lands on the wrong window, counter check fails). Addedwait_port_free()before launch to close the teardown-overlap window.harness_electron_click_element: … counter=1).No driver code touched. (Latent driver hardening noted for later:
cdp_list_pagespickspages[0]on a fixed env port with no correlation to the requested window — could hit the wrong Chromium app in production if two expose CDP.)🤖 Generated with Claude Code
Summary by CodeRabbit