So for context, this discussion comes from this Firefox commit.
Consider this:
z-index: calc(10% / 10%);
line-height: calc(10% / 10%);
opacity: calc(10% / 10%);
It seems to me these should all behave identically, either computing to 1 (z-index: 1 / line-height: 1 / opacity: 1), or being rejected?
That doesn't match browser behavior tho. There's some informal notes in https://drafts.csswg.org/css-values-4/#calc-simplification about how percentages generally might block simplification vs. not:
However, "raw" percentages—ones which do not resolve against another value, such as in opacity—might not block simplification.
However browsers behave differently here, and it's unclear from the spec what the right behavior is.
Similar weird example would be this test, effectively line-height: calc(10% / 1px), expecting it to compute to a number:
<!doctype html>
<style>
body {
font-size: 10px;
}
#target {
line-height: calc(10% / 1px);
}
</style>
<div id=target>What's my line height?</div>
<pre><script>
document.writeln(getComputedStyle(target).lineHeight);
document.writeln(target.computedStyleMap().get("line-height").toString());
</script></pre>I think making it compute to a number and making the percentage disappear (what Chromium does) is wrong. Because percentage line-heights generally inherit differently...
cc @kbhomes
So for context, this discussion comes from this Firefox commit.
Consider this:
It seems to me these should all behave identically, either computing to
1(z-index: 1/line-height: 1/opacity: 1), or being rejected?That doesn't match browser behavior tho. There's some informal notes in https://drafts.csswg.org/css-values-4/#calc-simplification about how percentages generally might block simplification vs. not:
However browsers behave differently here, and it's unclear from the spec what the right behavior is.
Similar weird example would be this test, effectively
line-height: calc(10% / 1px), expecting it to compute to a number:I think making it compute to a number and making the percentage disappear (what Chromium does) is wrong. Because percentage line-heights generally inherit differently...
cc @kbhomes