Re: [css-backgrounds] border-image with an SVG resource that has no intrinsic size

> On Jun 18, 2015, at 5:35 AM, David Vest <davve@opera.com> wrote:
> 
> I think we're mostly in agreement by now. We probably need more input
> from WebKit or Firefox about their way of thinking. I'd be fine
> leaving the spec as is and try to get implementations to match, even
> on the nonsensical cases.

[referencing the test cases in http://doomdavve.github.io/gists/svg-in-border-image-5.html ]

Yes, I really can’t explain why the nightly webkit renders the way it does, unless it is a work in progress or an optimization or re-use of code that went horribly bad. It is clearly not following the spec in any reasonable way. Shipping Safari (8.0.7) is so, so much closer to the spec and more useful. To me, this is a giant step backwards if they leave it like that. The image should never be just repeated from the upper left corner of each tile. My hope is that this is just some intermediate stage of development, but I haven’t seen it noticeably change in a while. Otherwise, I can’t understand how the SVG can be rendered correctly when viewed by itself, and the window is providing the concrete size, but rendered so badly when the border image box is providing the concrete size. Safari 8.07 was almost exactly right, except that it was stretching the little black boxes in your no-intrinsic-relative-mixed.

Firefox 39 is nothing short of baffling in each case except “intrinsic-size”. It is neither predictable nor useful.

Received on Tuesday, 21 July 2015 03:32:04 UTC