The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/9f25075ba35f2d463af614fa4d88d632c9f29361

commit 9f25075ba35f2d463af614fa4d88d632c9f29361
Author: Lloyd Pique <lpique@chromium.org>
Date: Tue Aug 29 18:56:26 2017

exo: Pointer::OnSurfaceCommit() should set the last hotspot

If a call is made to wl_pointer_set_cursor() to set a new surface and
corresponding hotspot, follow shortly thereafter by a call to
wl_surface_commit() to commit changes to the surface to use as the
pointer shape, the hotspot ended up being set to an old hotspot value.

Both calls make an asynchronous call to take a snapshot of the pointer
surface, which takes the hotspot to set. The second call appears to
cause the first snapshot operation to be aborted before completing,
otherwise it would make sense to have the commit call not even set the
hotspot. Instead the call to set the cursor now store the hotspot that
would have been set, and the commit call will use that value. This does
result in the correct hotspot being set.

Bug:  759856 , b:65096873
Change-Id: Iec69b0f138bb81f2da5ea9ecc2b9ff8739b02785
Reviewed-on: https://chromium-review.googlesource.com/639025
Commit-Queue: Lloyd Pique <lpique@chromium.org>
Reviewed-by: David Reveman <reveman@chromium.org>
Cr-Commit-Position: refs/heads/master@{#498185}
[modify] https://crrev.com/9f25075ba35f2d463af614fa4d88d632c9f29361/components/exo/pointer.cc
[modify] https://crrev.com/9f25075ba35f2d463af614fa4d88d632c9f29361/components/exo/pointer.h

This bug requires manual review: Less than 3 days to go before AppStore submit on M61
Please contact the milestone owner if you have questions.
Owners: amineer@(Android), cmasso@(iOS), ketakid@(ChromeOS), govind@(Desktop)

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

lpique@, please provide the rationale as to why this change needs to be merged into the release branch. How risky is this and how would the users be impacted if we don't merge this? 

This should be a low risk bug fix, needed as part of solving long-standing issues setting the pointer/mouse cursor in ARC++. It goes along with another non-Chromium change to fix the displayed image.

b/37641315 documents the Android side issue.

The changes impact a Wayland interface primarily used by ARC++, though it would affect any other code that does use the same interface. The code here is only built by default for Chrome OS.

This change fixes an issue with the hotspot set for the cursor, which was obviously wrong after applying the other patch when trying to drag the border of an Android window to resize it.

The other patch is necessary to prevent the cursor from sometimes being invisible in Android apps under certain conditions (most easily reproduced after letting it fade out by not using it for a little while), and I believe the same issue is causing the cursor to appear as a black square at other (much harder to reproduce) times.

We do have an option if we cannot land this change of adding a small workaround in the Android code for 61. But it was a small enough bug fix to the Chromium code that it seemed best to fix there and cherry-pick to 61.

And I just realized this bug was tagged as a iOS bug and not a Chrome OS bug. Sorry!

This change only affects ChromeOS devices running Android apps. Safe merge imo.

Approving merge to M61.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/32e58e42435c2abd48e7ae317bfbbfc2a5eed94d

commit 32e58e42435c2abd48e7ae317bfbbfc2a5eed94d
Author: David Reveman <reveman@chromium.org>
Date: Thu Aug 31 19:44:08 2017

exo: Pointer::OnSurfaceCommit() should set the last hotspot

If a call is made to wl_pointer_set_cursor() to set a new surface and
corresponding hotspot, follow shortly thereafter by a call to
wl_surface_commit() to commit changes to the surface to use as the
pointer shape, the hotspot ended up being set to an old hotspot value.

Both calls make an asynchronous call to take a snapshot of the pointer
surface, which takes the hotspot to set. The second call appears to
cause the first snapshot operation to be aborted before completing,
otherwise it would make sense to have the commit call not even set the
hotspot. Instead the call to set the cursor now store the hotspot that
would have been set, and the commit call will use that value. This does
result in the correct hotspot being set.

TBR=lpique@chromium.org

(cherry picked from commit 9f25075ba35f2d463af614fa4d88d632c9f29361)

Bug:  759856 , b:65096873
Change-Id: Iec69b0f138bb81f2da5ea9ecc2b9ff8739b02785
Reviewed-on: https://chromium-review.googlesource.com/639025
Commit-Queue: Lloyd Pique <lpique@chromium.org>
Reviewed-by: David Reveman <reveman@chromium.org>
Cr-Original-Commit-Position: refs/heads/master@{#498185}
Reviewed-on: https://chromium-review.googlesource.com/644139
Cr-Commit-Position: refs/branch-heads/3163@{#1039}
Cr-Branched-From: ff259bab28b35d242e10186cd63af7ed404fae0d-refs/heads/master@{#488528}
[modify] https://crrev.com/32e58e42435c2abd48e7ae317bfbbfc2a5eed94d/components/exo/pointer.cc
[modify] https://crrev.com/32e58e42435c2abd48e7ae317bfbbfc2a5eed94d/components/exo/pointer.h