Re: Ref-test image flakiness when web fonts are used

On 28/03/17 04:04, Philip J├Ągenstedt wrote:
> Ping Geoffrey. Koji and I were just chatting and it seems reasonable to
> have reftests wait for webfonts by default, as long as there is an escape
> hatch so that it's possible to test with slow-loading fonts or what happens
> if a first font load fails and then you try to load another font. Would it
> be weird to treat class="reftest-wait" as that escape hatch, or do we need
> something else?

I think that works OK because presumably in that case you want some 
control over when you decide the load is done anyway.

> Practically speaking, how do we go about proclaiming that reftests now
> should wait for document.fonts.ready? Is it a matter of changing wptrunner
> and filing browser bugs that they probably want to do something similar?

That seems reasonable to me.

Received on Tuesday, 28 March 2017 09:01:45 UTC