On Wed, 16 Oct 2013, Justin Novosad wrote: > > No, I don't mean an implied zero. I meant that at every iteration over > the dash pattern, the pattern is inverted if it has an odd number of > elements (blanks become dashes and vice versa). This would render > exactly the same as what is currently in the spec, which is consistent > with SVG. The only difference is that there is no doubling of the > array. If the array needs to be doubled in order to work with the > underlying graphics API, that should be the UA's business and it should > be hidden under the hood.
Ah, ok. I'm fine with us changing to that if browsers are able to. I don't think there's much of a problem either way; the advantage of the current model is that it leaves no ambiguity in the API about what it means to set an odd-length array. -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'