[Webkit-unassigned] [Bug 186244] REGRESSION(r232338): [GTK] Broke a few layout tests

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=186244

--- Comment #1 from Michael Catanzaro <mcatanzaro at igalia.com> ---
I found enough additional broken tests that I'm no longer comfortable with keeping the change. E.g. fast/forms/button-generated-content.html, the button text should be blue but is now black after this change:

Before: https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20(Tests)/r232449%20(6864)/retries/fast/forms/button-generated-content-expected.png

After: https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20(Tests)/r232449%20(6864)/retries/fast/forms/button-generated-content-actual.png


Another example is fast/forms/input-disabled-color.html:

Before: https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20(Tests)/r232449%20(6864)/retries/fast/forms/input-disabled-color-expected.png

After: https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20(Tests)/r232449%20(6864)/retries/fast/forms/input-disabled-color-actual.png


I don't think there is any bug in the patch itself, just that it's exposing some existing bug with color computation. There ought to be logic somewhere to not use the theme color if the page has specified its own color.

So I'm going to do a rollout. Sorry, Carlos Eduardo. This is actually really common, when making changes that affect rendering or layout, it often takes a while to get everything completely right....

-- 
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/20180603/e1b4cc25/attachment.html>