OLED NIGHT / TEST LABCI results ↗

RENDERING · INTERACTION · PERFORMANCE

Catch it here.
Before it ships.

A collection of components and past failures. Explore each state with OLED Night on, then compare it with the original.

— cases2 browser engines— extension

Loading the latest CI receipt…

Load a fixture to begin.

CVS hover · fix in 0.6.38

Late-created nested components, replaced stylesheets and frame-created surfaces are handled before paint. This regression stays separate from the component total. The receipts below show whether the tested version passed; a passing run does not guarantee every live page state.

Loading the separate synthetic regression…

Loading the separate anonymous CVS check…

Check the live site ↗

INTERACTIONS

Component state matrix

Open case ↗

Checking extension…

What a release has to prove

State coverage

Move between siblings, tab through controls, open overlays, update content and disable/re-enable. A resting screenshot is only one state.

Visual evidence

Approved screenshots, contrast and boundary checks, and paint samples during transitions. Baselines are reviewed explicitly.

Frames & resources

Chrome and Firefox compare animation frame throughput, p95 frame gaps and stalls with OLED Night on and off. Chrome also measures script work, style passes and collected heap growth. Both engines check that removed components are released. These checks block deployment when they fail.

This lab cannot prove saved-password filling, CAPTCHA token generation or every third-party site. Those still need targeted live verification.