commit 1a61614d5ae46af046bf5d2e9e0a0850b51a7679
author Max Ihlenfeldt <max@igalia.com> Mon Feb 19 12:01:12 2024
committer Chromium LUCI CQ <chromium-scoped@luci-project-accounts.iam.gserviceaccount.com> Mon Feb 19 12:01:12 2024
tree 607198600ab9bc6505e15d34a32a3df031fdbde1
parent 9cbd471af90b8dab6b05af0c54e45b74254c27f0

[1/n] Fix TabDragging tests for fallback tab dragging: DragAll

This is part of a series of CLs to make the TabDragging tests pass with
fallback tab dragging, i.e. with the AllowWindowDragUsingSystemDragDrop
feature enabled. The goal is for all tests that pass with the current
tab dragging implementation to pass with fallback tab dragging as well,
so that it can be enabled by default.

This CL adjusts the expectations of
TabDragging/DetachToBrowserTabDragControllerTest.DragAll to match what
is possible on Linux ozone/wayland.

The test expects the browser to be moved down if we drag all tabs, but
that doesn’t work with fallback tab dragging (we can’t both start a DnD
session, which we need for merging with other windows, and an
interactive move; that’s what we need the extended-drag protocol for).
Thus, we change the test to only expect the browser bounds to change if
we can control them.

Note that the test passes without this change using the current tab
dragging implementation (i.e. when fallback tab dragging is disabled),
but it shouldn’t. When you run the test non-headless, you can see that
the window doesn’t really move; the test just thinks it has moved
because `GetBounds()` returns the bounds that were set using
`SetBounds()`, even though that had no actual effect on the window’s
position.

Bug: 1361940
Change-Id: I6e7be5b761a2029a33cade431d2e2839fe9b91e3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/5253657
Reviewed-by: Taylor Bergquist <tbergquist@chromium.org>
Commit-Queue: Max Ihlenfeldt <max@igalia.com>
Cr-Commit-Position: refs/heads/main@{#1262339}