Loading…
Loading…
Dataset 20260827-2142-2f1b575 · measured 2026-08-27 21:42 UTC
The parser accepts the name, yet our rendering is byte-identical with and without the declaration while Chrome's is not. The declaration is being swallowed.
Each example is one declaration on a small page, drawn by Chrome and by our engine. Drag the handle to wipe between them.
text-decoration-line:overline✕renders differently
Our renderChrome6,825 pixels differ from Chrome, the largest by 204 of 255 on one channel.
text-decoration-line:line-through✕renders differently
Our renderChrome7,024 pixels differ from Chrome, the largest by 187 of 255 on one channel.
text-decoration-line:spelling-error✕parsed, no effect
Our renderChrome2,046 pixels differ from Chrome, the largest by 178 of 255 on one channel.
text-decoration-line:grammar-error✕parsed, no effect
Our renderChrome1,834 pixels differ from Chrome, the largest by 76 of 255 on one channel.
text-decoration-line:underline✓pixel-exact
Our renderChrome1,402 pixels differ from Chrome, the largest by 45 of 255 on one channel.
“Supported” on one example means “supported at the value we tried”. The property’s declared grammar has 7 branches; 5 of them carry a measurement. The publishing ceiling is 12 examples per property — a branch past it is left out by our choice, not by accident.
| Value | Measured |
|---|---|
| underline | ✓pixel-exact |
| overline | ✕renders differently |
| line-through | ✕renders differently |
| blink | not measured |
| spelling-error | ✕parsed, no effect |
| grammar-error | ✕parsed, no effect |
| none | not measured |
Secondary evidence: 11 Web Platform Tests carry a spec anchor pointing at this property. Only 1,254 tests in the whole corpus carry such an anchor, so an empty list here means nothing at all.
Measured against Chrome 148.0.7778.96 on 2026-08-27, with up to 40 candidate values tried per property. All properties · How the measurement works