[Webkit-unassigned] [Bug 191657] New: REGRESSION (r195560?): Writing direction gets stuck on root RenderStyle after matching a rule and turning it off

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed Nov 14 14:25:53 PST 2018


https://bugs.webkit.org/show_bug.cgi?id=191657

            Bug ID: 191657
           Summary: REGRESSION (r195560?): Writing direction gets stuck on
                    root RenderStyle after matching a rule and turning it
                    off
           Product: WebKit
           Version: WebKit Nightly Build
          Hardware: Unspecified
                OS: Unspecified
            Status: NEW
          Severity: Normal
          Priority: P2
         Component: CSS
          Assignee: webkit-unassigned at lists.webkit.org
          Reporter: timothy at apple.com
                CC: koivisto at iki.fi, simon.fraser at apple.com

Created attachment 354853

  --> https://bugs.webkit.org/attachment.cgi?id=354853&action=review

Test Case

If the page toggles a rule on the body or :root that has `direction: rtl`, then later toggles it off, the direction will be stuck as rtl since it inherits from the RenderView's style.

Steps:
1) Load attached test case
2) Note ltr is the default direction.
3) Click Toggle RTL button.
4) Note rtl is applied.
5) Click Toggle RTL button again.

Results: RTL is now stuck unless a rule explicitly applies direction: ltr.

Notes:
RenderBox::styleDidChange does this "Propagate the new writing mode and direction up to the RenderView." This is well and good, but RenderView is never reset when an explicit direction on body or html is no longer present and everything inherits from RenderView's RenderStyle. Style::resolveForDocument() has similar code for the initial document style.

I suspect this broke with the fix for bug 153446 (https://trac.webkit.org/changeset/195560/webkit).

-- 
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/20181114/c631897c/attachment.html>


More information about the webkit-unassigned mailing list