[Webkit-unassigned] [Bug 172596] Crash on display-contents-replaced-001.html

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=172596

--- Comment #5 from Antti Koivisto <koivisto at iki.fi> ---
(In reply to Emilio Cobos Álvarez from comment #4)
> Note that the spec requires contents to still compute to contents, which I
> don't think will be true with this patch.
> 
> i.e., a test like:
> 
> var svg = document.querySelector('svg');
> svg.style.display = "contents";
> getComputedStyle(svg).display; // "none" with your patch, "contents" per
> spec.

Is the spec correct though? We generally return the effective display value from getComputedStyle rather than the original. We already save the original in RenderStyle so if we want we could add a bit that says "return original instead of effective" to fix this.

> I had a WIP patch that did this[1]. Pending thing was to address the
> special-cases you mention in your FIXME.

I'd like to use a blacklist like in my patch instead of virtuals. It is much easier to compare the code to the spec. A bigger problem with your approach is that we have a bunch of sites that explicitly test for display() ==/!= NONE (and some for CONTENTS) and there is no way to know they should behave like 'none' in the effective case.

Does this make sense to you?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20170530/7c499509/attachment.html>