Workaround: Apply `filter: blur(0)` to the element with `overflow: hidden`

I can't reproduce this on blur.html. It looks to me like it is correctly
clipping the overflow areas before applying the blur. On what page did you
take the screenshot?

The page in the screenshots are not publicly available ATM. However, I'm still getting the same issue in the blur.html.

I created a fiddle with two versions, one where the filter breaks overflow, and one where I applied a workaround. https://jsfiddle.net/Micke/r3gvjgo9/2/

Attaching a screenshot of how it behaves for me. 

(Chrome v 64.0.3282.140)

Deleted: blur-bug.png 40.2 KB

	blur-bug.png 40.2 KB View Download

Thank you for providing more feedback. Adding requester "chrishtr@chromium.org" to the cc list and removing "Needs-Feedback" label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Reproduces and Firefox does what one would expect, I think.

I'll see if this is recent breakage.

Deleted: cr809102.html 451 bytes

cr809102.html 451 bytes View Download

Bisects to

https://chromium.googlesource.com/chromium/src/+/fb48c272877d177f4c93a617aa97323c28fb1552

Use GeometryMapper to map visual rect with filters for paint invalidation

This is an M-63 stable regression.

 Issue 809402  has been merged into this issue.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/d88b38539a1cae36c00ba1ecbaf564dc6386f753

commit d88b38539a1cae36c00ba1ecbaf564dc6386f753
Author: Xianzhu Wang <wangxianzhu@chromium.org>
Date: Thu Feb 08 19:00:41 2018

[PE] GeometryMapper::LocalToAncestorClipRect() should ignore effects

The method is used to get a clip rect in the target state space that
will applied to the contents under local_state. The clip is applied
before any effect are applied so should not consider effects.

Bug:  809102 
Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel;master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
Change-Id: Id97c6766f962904f493037b08ebab09439c9a627
Reviewed-on: https://chromium-review.googlesource.com/907472
Commit-Queue: Xianzhu Wang <wangxianzhu@chromium.org>
Reviewed-by: Chris Harrelson <chrishtr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#535465}
[add] https://crrev.com/d88b38539a1cae36c00ba1ecbaf564dc6386f753/third_party/WebKit/LayoutTests/paint/filters/absolute-under-clip-under-filter-expected.html
[add] https://crrev.com/d88b38539a1cae36c00ba1ecbaf564dc6386f753/third_party/WebKit/LayoutTests/paint/filters/absolute-under-clip-under-filter.html
[modify] https://crrev.com/d88b38539a1cae36c00ba1ecbaf564dc6386f753/third_party/WebKit/LayoutTests/platform/linux/compositing/reflections/nested-reflection-on-overflow-expected.txt
[add] https://crrev.com/d88b38539a1cae36c00ba1ecbaf564dc6386f753/third_party/WebKit/LayoutTests/platform/mac-mac10.12/compositing/reflections/nested-reflection-on-overflow-expected.txt
[modify] https://crrev.com/d88b38539a1cae36c00ba1ecbaf564dc6386f753/third_party/WebKit/LayoutTests/platform/win/compositing/reflections/nested-reflection-on-overflow-expected.txt
[modify] https://crrev.com/d88b38539a1cae36c00ba1ecbaf564dc6386f753/third_party/WebKit/LayoutTests/virtual/spv175/compositing/reflections/nested-reflection-on-overflow-expected.txt
[modify] https://crrev.com/d88b38539a1cae36c00ba1ecbaf564dc6386f753/third_party/WebKit/Source/platform/graphics/paint/GeometryMapper.cpp
[modify] https://crrev.com/d88b38539a1cae36c00ba1ecbaf564dc6386f753/third_party/WebKit/Source/platform/graphics/paint/GeometryMapper.h
[modify] https://crrev.com/d88b38539a1cae36c00ba1ecbaf564dc6386f753/third_party/WebKit/Source/platform/graphics/paint/GeometryMapperTest.cpp

Just found that the #c9 patch just fixed some cases of the issue, but not the reported case.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/4926f8217a1795cedffc9c64113eb7944b0cf504

commit 4926f8217a1795cedffc9c64113eb7944b0cf504
Author: Xianzhu Wang <wangxianzhu@chromium.org>
Date: Fri Feb 16 23:59:11 2018

[PE] Add a test case for  crbug.com/809102 

The test case passes in SPv175 (enabled by default), fails otherwise.

Bug:  809102 
Change-Id: Ic83e3ad785f1b10fb8c359fddb61c5672a6f8e33
Reviewed-on: https://chromium-review.googlesource.com/924622
Reviewed-by: Chris Harrelson <chrishtr@chromium.org>
Commit-Queue: Xianzhu Wang <wangxianzhu@chromium.org>
Cr-Commit-Position: refs/heads/master@{#537467}
[modify] https://crrev.com/4926f8217a1795cedffc9c64113eb7944b0cf504/third_party/WebKit/LayoutTests/TestExpectations
[add] https://crrev.com/4926f8217a1795cedffc9c64113eb7944b0cf504/third_party/WebKit/LayoutTests/paint/filters/clip-absolute-under-clip-under-filter-expected.html
[add] https://crrev.com/4926f8217a1795cedffc9c64113eb7944b0cf504/third_party/WebKit/LayoutTests/paint/filters/clip-absolute-under-clip-under-filter.html

SPv175 has been enabled on ToT and the test is now passing. Removing Target-65 because we'll not ship SPv175 in M65.