1436189 - Marking the frame's overflow area for rebuild doesn't include out-of-flow descendants

Status: RESOLVED FIXED

(Core :: Web Painting, enhancement)

Description

[Matt Woodrow (:mattwoodrow)]

If we mark a frame modified for display list building, and that frame has placeholder descendants, then the overflow area we contribute to building won't include the area for the out of flow frames.

If the modified frame has an empty area, then we can end up doing no building, and then discard display items belonging to the out of flow frames.

Comment 1

[Matt Woodrow (:mattwoodrow)]

Attachment: Part 1: Add testcase

This triggers a display item rebuild for the frame with opacity, it has empty overflow, and we don't build any display items.

During the merging phase we discard the nsDisplayOpacity, since it belongs to a modified frame, and the final list no longer has an content.

Comment 2

[Matt Woodrow (:mattwoodrow)]

Attachment: Part 2: Make sure we descend to the modified frame

This makes sure we at least descend past the containing block for these frame (if it's visible!), which will be true if the OOF frames are visible.

At that point we have existing machinery in MarkOutOfFlowFrame for display that will detect that the OOF frames are a content-descendant of the modified frame, and ensure that building reaches them.

Comment 3

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/ee3062043923
Part 1: Add testcase. r=miko
https://hg.mozilla.org/integration/mozilla-inbound/rev/3403b5f669d1
Part 2: Make sure we descend to the modified frame so that we descend through the containing block and mark out of flow frames for display. r=miko

Comment 4

[Tiberius Oros[:tiberius_oros]]

https://hg.mozilla.org/mozilla-central/rev/ee3062043923
https://hg.mozilla.org/mozilla-central/rev/3403b5f669d1

Comment 5

[Pulsebot]

Backout by [e-mail]:
https://hg.mozilla.org/mozilla-central/rev/dad27b0cda4e
Backed out 2 changesets for frequently failing on layout/reftests/table-background/backgr_layers-opacity.html a=backout

Comment 6

[Arthur Iakab [arthur_iakab]]

Backed out 2 changesets (bug 1436189) for frequently failing on layout/reftests/table-background/backgr_layers-opacity.html a=backout

Link to the failing logs: https://treeherder.mozilla.org/logviewer.html#?job_id=161620792&repo=mozilla-inbound&lineNumber=2385
Backout revision: dad27b0cda4e
Failing push: https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=3403b5f669d1395097f26c69c514e52d302a3ec0&filter-searchStr=linux64+opt+Rs8

Comment 7

[Arthur Iakab [arthur_iakab]]

Backed out 2 changesets (bug 1436189) for frequently failing on layout/reftests/table-background/backgr_layers-opacity.html a=backout

Link to the failing logs: https://treeherder.mozilla.org/logviewer.html#?job_id=161620792&repo=mozilla-inbound&lineNumber=2385
Backout revision: dad27b0cda4e
Failing push: https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=3403b5f669d1395097f26c69c514e52d302a3ec0&filter-searchStr=linux64+opt+Rs8

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/mozilla-central/rev/2222c721dfe5
Part 1: Add testcase. r=miko
https://hg.mozilla.org/mozilla-central/rev/3ee38289dac8
Part 2: Make sure we descend to the modified frame so that we descend through the containing block and mark out of flow frames for display. r=miko a=mass-reftest-fix CLOSED TREE

Comment 9

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

https://hg.mozilla.org/mozilla-central/rev/2222c721dfe5
https://hg.mozilla.org/mozilla-central/rev/3ee38289dac8

Comment 10

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Relanded because of mass reftest failures after the latest merge from inbound to central and back to integration trees.