Created attachment 254708 [details] test case @supports (filter:blur()) or (-webkit-filter: blur()) { /* do something */ } is parsed as valid by Webkit but not by Firefox (the rules inside the `@supports` block are applied by Webkit nightly builds). Note the missing parameter for the blur value. Chromium based browsers appear to suffer from the same issue (tested with Opera 30). Note: 1. Webkit handles this correctly for `@supports (transform:scale()) {}` and ignores everything inside the @supports block. 2. -webkit-backdrop-filter: blur() is also affected
Well, after multiple readings of the relevant spec [1], I think I’m wrong – and Webkit (+ Blink) is correct. filter: blur() or filter: grayscale(), etc without a parameter is actually valid per the mentioned spec. The spec mentions - quote (for blur()): “The lacuna value for interpolation is…”. The word ‘lacuna’ apparently refers to ‘missing parameter’. I had to ask two native English speakers for this; the relevant MDN article [2] confirms this reading (“If no parameter is provided, then a value 0 is used”). By extension @supports (filter: blur()) [or other filters] should parse as valid (prefixed atm of course). [1] http://www.w3.org/TR/filter-effects/#FilterProperty [2] https://developer.mozilla.org/en-US/docs/Web/CSS/filter
Per the bug reporter's comment, close this bug as non-issue.