(Possibly we can ignore 'color' for this issue... Only Gecko matches the spec there, and it has a darker border color than the other browsers. I'll follow up on that in a new spec issue.)
Massive rebaselining is required, since each HR will now end up establishing a PaintLayer. I'm not convinced it's worthwhile to make this change. A PaintLayer isn't that cheap, is it? While we're not going to increase the PaintLayer tree depth by more than one here (and they'll typically all be leaf nodes), but still. Memory-wise and performance-wise?
I'm tempted to leave this unresolved as long as specifying overflow:hidden automatically means that we'll establish a PaintLayer. But WDYT?
HR definitely isn't as common as hydrogen, but still widely used, right? Or?
If there are no concerns, I might just go ahead and assign reviewers to that CL.
I don't think it will be a significant performance problem to apply PaintLayers
to <hr>. So if that is the only blocker and the CL is ready, I'd say go ahead.
Comment 1 by sim...@opera.com
, Jun 1 2017