This one is marked "pass" in the IETC summary grid, but it is actually a failure. If you look closely, you will see little splotches of red between the 4th and 5th "A" and again between the 2nd and 3rd "B". Also the color of all the letters is a brighter green, indicating that the overlaid red letters are missing.
Created attachment 181124 [details]
Comment on attachment 181124 [details]
Looking at the spec, it looks like 0 is also invalid: "Specified values must be greater than 0." http://dev.w3.org/csswg/css3-multicol/#column-width
Can we make the code match this and add a test for 0? Also, how do Firefox and IE handle 0?
Created attachment 181199 [details]
Added test for zero column-width. Other browsers treat zero same like negative value. Thanks for review.
Comment on attachment 181199 [details]
View in context: https://bugs.webkit.org/attachment.cgi?id=181199&action=review
The change looks fine, but the test could be better. I see some red at the edges of the letters due to text antialiasing which is confusing.
I would convert this from a ref test to a dumpAsText test (they run faster and are easier to know if it passes or fails) and use fast/js/resources/js-test-pre.js. The change has to do with CSS parsing, so you could simply have a div with column-width: 0 and read the value back out using getComputedStyle (it should be auto).
> + -moz-column-width: 0em;
0 shouldn't have a unit.
Created attachment 181217 [details]
Thanks for fixing this!
Comment on attachment 181217 [details]
Clearing flags on attachment: 181217
Committed r138746: <http://trac.webkit.org/changeset/138746>
All reviewed patches have been landed. Closing bug.