New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 670274 link

Starred by 33 users

Issue metadata

Status: Fixed
Owner:
Closed: Jan 2017
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: All
Pri: 2
Type: Bug-Regression



Sign in to add a comment

Position:Sticky + Box-Shadow Breaks Positioning

Reported by azmer...@gmail.com, Dec 1 2016

Issue description

Steps to reproduce the problem:
1. Open: http://codepen.io/Kelderic/full/XNVdZx/
2. Look at sticky nav.

What is the expected behavior?
Should be fully left, touching edge of page.

What went wrong?
It's not fully left. It's been moved to the right, and down.

Did this work before? No 

Does this work in other browsers? Yes

Chrome version: 56.0.2924.10  Channel: dev
OS Version: 5.1
Flash Version: 

Removing box-shadow causes the bug to disappear. Also this tracker doesn't work in Firefox. Filled out the report and got a blank white page. Trying in Chrome now.
 
chrome-56-android.png
91.0 KB View Download

Comment 1 by azmer...@gmail.com, Dec 1 2016

Attached screenshot with only change being removed shadow and added background-color.

http://codepen.io/Kelderic/full/ENoygx/
chrome-56-android-no-boxshadow.png
106 KB View Download

Comment 2 by azmer...@gmail.com, Dec 12 2016

I can now reproduce this issue on Windows 10, with Chrome 56.0.2924.18 dev (64-bit).

Also still broken in Chrome 56 and 57 on Android. Specifically 56.0.2924.18, and 57.0.2946.0.

Comment 3 by azmer...@gmail.com, Dec 13 2016

This is now happening in Chrome Beta, v 56.0.2924.23.

It's a little surprising to me that no one has responded to this. It's a pretty visible bug. Any websites that are using sticky positioning look mildly broken on Chrome 56 and up.
Cc: dknandiraju@chromium.org
Labels: triage-te OS-Windows
@azmercer@gmail.com, unable to reproduce your issue as "http://codepen.io/Kelderic/full/XNVdZx/" is now redirecting to "http://codepen.io/anddo/full/XNVdZx/". Can you provide correct link to verify the issue.

Comment 6 by azmer...@gmail.com, Dec 15 2016

When fully logged out of Codepen, that link is working for me. However, here is a mirror:

https://jsfiddle.net/LjmaLcky/

I've noticed it only appears when the wrapper window is scrollable. My guess is that when it's not scrollable, then the sticky is rendered as static.

Comment 7 by azmer...@gmail.com, Dec 15 2016

Also I'm adding the test case as a self-contained HTML file attachment in case that second link doesn't work either.
sticky-position-broken-test-case.html
1.6 KB View Download

Comment 8 by azmer...@gmail.com, Dec 16 2016

@dknandiraju@chromium.org, does the new link/attachment work for you to verify the issue?
Labels: hasbisect-per-revision
Owner: flackr@chromium.org
Status: Assigned (was: Unconfirmed)
1) Bisect Range: https://chromium.googlesource.com/chromium/src/+log/56.0.2909.0..56.0.2910.0?pretty=fuller&n=10000
2) Bisect script pointed to:  https://chromium.googlesource.com/chromium/src/+/77a8c51a440e0741315ff5e9e77441b86cbbc56d

@flackr, can you please look into this,thanks
Labels: -triage-te
Labels: -Type-Bug M-56 Type-Bug-Regression

Comment 12 Deleted

Comment 13 by mpc...@gmail.com, Dec 21 2016

Same problem on Windows 7 with:

Chrome 56 beta (64-bits)

and

Canary 57.0.2957.0 (64-bits)

Comment 14 by azmer...@gmail.com, Dec 29 2016

This is still broken on 57.0.2950.4 on Windows 10.
Just a bump because we are less than 2 weeks away from this bug being in the public release channel.

Comment 16 by azmer...@gmail.com, Jan 16 2017

3 days until this is broken on Stable.
Components: Internals>Compositing
Labels: -OS-Android -OS-Windows Hotlist-Threaded-Rendering OS-All
Status: Started (was: Assigned)
Apologies for the absence, I was on vacation over the holidays. The issue is only with composited sticky - when sticky position is updated on the main thread (i.e. non retina non-opaque sticky elements) it is updated correctly. This suggests that the main thread offset passed to the compositor needs to be adjusted by the additional compositing offset caused by the box shadow.
Cc: flackr@chromium.org smcgruer@chromium.org meade@chromium.org
 Issue 670953  has been merged into this issue.
Project Member

Comment 20 by bugdroid1@chromium.org, Jan 18 2017

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/653e80aeeaac895c6f9bfb6181655f15f29cb0be

commit 653e80aeeaac895c6f9bfb6181655f15f29cb0be
Author: flackr <flackr@chromium.org>
Date: Wed Jan 18 22:57:23 2017

Add offset of content in composited layer to sticky position offset.

The actual sticky layoutobject may have an offset within its composited layer
which needs to be included in the compositing update of sticky positioning.

BUG= 670274 
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Review-Url: https://codereview.chromium.org/2636233002
Cr-Commit-Position: refs/heads/master@{#444527}

[modify] https://crrev.com/653e80aeeaac895c6f9bfb6181655f15f29cb0be/third_party/WebKit/Source/core/layout/compositing/CompositedLayerMapping.cpp
[modify] https://crrev.com/653e80aeeaac895c6f9bfb6181655f15f29cb0be/third_party/WebKit/Source/core/layout/compositing/CompositedLayerMappingTest.cpp

Comment 21 by azmer...@gmail.com, Jan 19 2017

Awesome! So will this be uplifted up through v56? Or just apply to Canary?
Status: Fixed (was: Started)
Unfortunately, as innocuous as this patch is I think it's too late to merge back, so it will just apply to Chrome 57+.

Comment 23 by azmer...@gmail.com, Jan 19 2017

Well that kinda sucks, but it is what it is. I did just get an update on Canary, on my phone and can confirm that the bug is fixed.
 Issue 678914  has been merged into this issue.

Comment 25 Deleted

 Issue 688022  has been merged into this issue.
 Issue 690381  has been merged into this issue.
 Issue 695085  has been merged into this issue.
 Issue 697186  has been merged into this issue.
 Issue 674356  has been merged into this issue.

Sign in to add a comment