[wpt-sync] Sync PR 42778 - Avoid duplicate block-in-inline break tokens.
Categories
(Core :: Layout, task, P4)
Tracking
()
Tracking | Status | |
---|---|---|
firefox121 | --- | fixed |
People
(Reporter: wpt-sync, Unassigned)
References
()
Details
(Whiteboard: [wptsync downstream])
Sync web-platform-tests PR 42778 into mozilla-central (this bug is closed when the sync is complete).
PR: https://github.com/web-platform-tests/wpt/pull/42778
Details from upstream follow.
Morten Stenshorne <mstensho@chromium.org> wrote:
Avoid duplicate block-in-inline break tokens.
Regardless of whether the break inside a block-in-inline was to be
resumed in a parallel flow or not, we'd also include it as a break token
in the same flow.This was usually harmless, since any next line in the same fragmentainer
would typically not be a block-in-inline, since all consecutive blocks
inside inlines are put inside the same block-in-inline wrapper during
box generation.However, there are ways to end up with adjacent block-in-inline
wrappers: If a block that previously wasn't inflow becomes inflow
(change from float:left to float:none, for instance), a new wrapper is
created around the now-inflow block, even if there's already an adjacent
wrapper there for an adjacent block.\<span>
\<div id="first" style="float:left;">\</div>
\<div id="second">\</div>
\</span>This generates a box tree like this:
LayoutInline SPAN
LayoutNGBlockFlow (floating) DIV #first
LayoutNGBlockFlow (anonymous)
LayoutNGBlockFlow DIV #secondIf we (after initial layout) change #first to float:none, we end up with
this:LayoutInline SPAN
LayoutNGBlockFlow (anonymous)
LayoutNGBlockFlow DIV #first
LayoutNGBlockFlow (anonymous)
LayoutNGBlockFlow DIV #secondThe fix is to make sure that we only create one break token for a
fragmented block wrapper. Move the code that examines the outgoing
block-in-inline break token a bit further up, to eliminate the need for
LineInfo::BlockInInlineBreakToken(), and to keep things a bit more
tightly together. Add a DCHECK to LineBreaker::HandleBlockInInline(),
which would have failed without this fix.Bug: 1495515
Change-Id: I6d019b8c2e6c03046e122fecce76e044f3aeabb3
Reviewed-on: https://chromium-review.googlesource.com/4973967
WPT-Export-Revision: 4d006c47809fb0a9f377912b41deb3c77ae66b20
Assignee | ||
Updated•2 years ago
|
Assignee | ||
Updated•1 years ago
|
Assignee | ||
Updated•1 years ago
|
Assignee | ||
Comment 1•1 years ago
|
||
Assignee | ||
Comment 2•1 years ago
|
||
CI Results
Ran 11 Firefox configurations based on mozilla-central, and Firefox, and Safari on GitHub CI
Total 2 tests
Status Summary
Firefox
PASS
: 1[Gecko-linux2204-64-wayland-debug, Gecko-linux2204-64-wayland-opt] 2[Gecko-android-em-7.0-x86_64-lite-qr-opt-geckoview, Gecko-android-em-7.0-x86_64-qr-debug-geckoview, Gecko-android-em-7.0-x86_64-qr-opt-geckoview, Gecko-linux1804-64-qr-debug, Gecko-linux1804-64-qr-opt, Gecko-windows11-32-2009-qr-debug, Gecko-windows11-32-2009-qr-opt, Gecko-windows11-64-2009-qr-debug, Gecko-windows11-64-2009-qr-opt, GitHub]
Safari
PASS
: 1
FAIL
: 1
Links
Comment 5•1 years ago
|
||
bugherder |
Description
•