Fix the bug that absolute positioned element moves unexpectedly

Fix the bug that absolute positioned element moves unexpectedly

For an absolute positioned element, if its containing block is the body
it's not supposed to move as scroll. However, on low dpi devices, it
doesn't get composited if its ancestor is isolated composited therefore
it moves unexpectedly.

BUG=719533
TEST=third_party/WebKit/LayoutTests/compositing/overflow/absolute-element-in-isolated-composited-ancestor.html
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2869813002
Cr-Commit-Position: refs/heads/master@{#471158}
Committed: https://chromium.googlesource.com/chromium/src/+/fca776d5b4979b1f9d86926bfb6d3fb14a0343f3

[yigu, 2017-05-08 15:34:37]

Description was changed from

==========
Fix the bug that absolute positioned element moves unexpectedly

For an absolute positioned element, if its containing block is the body
it's not supposed to move as scroll. However, on low dpi devices, it
doesn't get composited if its ancestor is isolated composited therefore
it moves unexpectedly.

BUG=
TEST=third_party/WebKit/LayoutTests/compositing/overflow/absolute-element-in-isolated-composited-ancestor.html
==========

to

==========
Fix the bug that absolute positioned element moves unexpectedly

For an absolute positioned element, if its containing block is the body
it's not supposed to move as scroll. However, on low dpi devices, it
doesn't get composited if its ancestor is isolated composited therefore
it moves unexpectedly.

BUG=
TEST=third_party/WebKit/LayoutTests/compositing/overflow/absolute-element-in-isolated-composited-ancestor.html
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

[yigu, 2017-05-08 15:36:45]

Description was changed from

==========
Fix the bug that absolute positioned element moves unexpectedly

For an absolute positioned element, if its containing block is the body
it's not supposed to move as scroll. However, on low dpi devices, it
doesn't get composited if its ancestor is isolated composited therefore
it moves unexpectedly.

BUG=
TEST=third_party/WebKit/LayoutTests/compositing/overflow/absolute-element-in-isolated-composited-ancestor.html
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

to

==========
Fix the bug that absolute positioned element moves unexpectedly

For an absolute positioned element, if its containing block is the body
it's not supposed to move as scroll. However, on low dpi devices, it
doesn't get composited if its ancestor is isolated composited therefore
it moves unexpectedly.

BUG=719533
TEST=third_party/WebKit/LayoutTests/compositing/overflow/absolute-element-in-isolated-composited-ancestor.html
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

[yigu, 2017-05-08 15:36:46]

yigu@chromium.org changed reviewers:
+ chrishtr@chromium.org

[yigu, 2017-05-08 15:37:19]

Hi Chris. PTAL. Thanks.

[yigu, 2017-05-08 18:21:02]

Hi Chris. Just pasting your comment from the patch that you brought this bug up
for convenience:

-- comment starts --

This is not just a problem for position:sticky. Changing the compositing trigger
in
non-composited-sticky-element-in-main-thread-scrolled-composited-ancestor.html
to:

.composited {
backface-visibility: hidden;
isolation: isolate;
}

will cause the position:absolute element to appear to scroll, even though it
should not, because its containing block is the body element.

This will only reproduce for position:absolute on low-DPI screens, because
otherwise kCompositingReasonOutOfFlowClipping kicks in. I don't know why exactly
that trigger is limited to high-DPI screens but I think it should not be.

Proposed fix:

1. Get rid of kOverflowScroll trigger in the conditional here
https://cs.chromium.org/chromium/src/third_party/WebKit/Source/core/layout/co...
(fixes bug with position:absolute).

-- comment ends --

In the comments you mentioned that the trigger is limited to high dpi devices. I
think it's because we only set kOverflowScroll on high dpi be default. See
https://cs.chromium.org/chromium/src/third_party/WebKit/Source/core/layout/co....

The proposed fix "failed" one layout test. See
https://storage.googleapis.com/chromium-layout-test-archives/linux_chromium_r....
In the test, we don't create any non-root layer on low dpi devices without the
patch but we do so with patch. That should be the reason the fix "failed".
Should I reset the expected png with the new result?
Thanks!

[chrishtr, 2017-05-09 20:52:35]

lgtm

[yigu, 2017-05-09 21:19:53]

The CQ bit was checked by yigu@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-05-09 21:21:16]

Dry run: CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-09 21:48:40]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-09 21:48:41]

Dry run: Try jobs failed on following builders:
  android_clang_dbg_recipe on master.tryserver.chromium.android (JOB_FAILED,
https://build.chromium.org/p/tryserver.chromium.android/builders/android_clan...)
  cast_shell_linux on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/cast_shell_linu...)

[yigu, 2017-05-10 13:14:10]

The CQ bit was checked by yigu@chromium.org

[yigu, 2017-05-10 13:14:11]

The patchset sent to the CQ was uploaded after l-g-t-m from
chrishtr@chromium.org
Link to the patchset: https://codereview.chromium.org/2869813002/#ps20001
(title: "Rebaseline")

[commit-bot: I haz the power, 2017-05-10 13:14:49]

CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-10 14:33:36]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-10 14:33:37]

Try jobs failed on following builders:
  linux_layout_tests_slimming_paint_v2 on master.tryserver.chromium.linux
(JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_layout_te...)

[yigu, 2017-05-11 02:39:15]

The CQ bit was checked by yigu@chromium.org

[yigu, 2017-05-11 02:39:16]

The patchset sent to the CQ was uploaded after l-g-t-m from
chrishtr@chromium.org
Link to the patchset: https://codereview.chromium.org/2869813002/#ps40001
(title: "Add test expectation")

[commit-bot: I haz the power, 2017-05-11 02:40:14]

CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-11 04:27:35]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-05-11 04:27:36]

Try jobs failed on following builders:
  linux_chromium_asan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[yigu, 2017-05-11 23:09:13]

The CQ bit was checked by yigu@chromium.org

[commit-bot: I haz the power, 2017-05-11 23:10:22]

CQ is trying da patch.

Follow status at:
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-05-12 01:04:16]

CQ is committing da patch.

Bot data: {"patchset_id": 40001, "attempt_start_ts": 1494544153689650,
"parent_rev": "ed59882df38c70c6801b67478454fa00f432d62d", "commit_rev":
"fca776d5b4979b1f9d86926bfb6d3fb14a0343f3"}

[commit-bot: I haz the power, 2017-05-12 01:04:31]

Description was changed from

==========
Fix the bug that absolute positioned element moves unexpectedly

For an absolute positioned element, if its containing block is the body
it's not supposed to move as scroll. However, on low dpi devices, it
doesn't get composited if its ancestor is isolated composited therefore
it moves unexpectedly.

BUG=719533
TEST=third_party/WebKit/LayoutTests/compositing/overflow/absolute-element-in-isolated-composited-ancestor.html
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

to

==========
Fix the bug that absolute positioned element moves unexpectedly

For an absolute positioned element, if its containing block is the body
it's not supposed to move as scroll. However, on low dpi devices, it
doesn't get composited if its ancestor is isolated composited therefore
it moves unexpectedly.

BUG=719533
TEST=third_party/WebKit/LayoutTests/compositing/overflow/absolute-element-in-isolated-composited-ancestor.html
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2869813002
Cr-Commit-Position: refs/heads/master@{#471158}
Committed:
https://chromium.googlesource.com/chromium/src/+/fca776d5b4979b1f9d86926bfb6d...
==========

[commit-bot: I haz the power, 2017-05-12 01:04:32]

Committed patchset #3 (id:40001) as
https://chromium.googlesource.com/chromium/src/+/fca776d5b4979b1f9d86926bfb6d...