Fix unfullscreening a maximized window in Metacity

Fix unfullscreening a maximized window in Metacity

BUG=382126
TEST=Manual, see instructions below

Test steps:
1) Run Chrome in the Gnome Classic window manager (this is very important)
2) Maximize a browser window
3) Fullscreen the browser window via F11
4) Press escape to exit fullscreen
5) Unmaximize the window
6) Verify that the window can be resized

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=285658

[pkotwicz, 2014-07-20 23:10:25]

Elliot, can you please take a look?

[Elliot Glaysher, 2014-07-21 17:29:26]

lgtm

[pkotwicz, 2014-07-25 14:43:37]

The CQ bit was checked by pkotwicz@chromium.org

[commit-bot: I haz the power, 2014-07-25 14:45:09]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/pkotwicz@chromium.org/405743004/20001

[commit-bot: I haz the power, 2014-07-25 15:56:28]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  linux_chromium_clang_dbg on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/linux_chromium_clang_...)

[commit-bot: I haz the power, 2014-07-25 16:04:01]

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

[commit-bot: I haz the power, 2014-07-25 16:04:01]

Try jobs failed on following builders:
  linux_chromium_clang_dbg on tryserver.chromium
(http://build.chromium.org/p/tryserver.chromium/builders/linux_chromium_clang_...)

[pkotwicz, 2014-07-25 19:00:36]

The CQ bit was checked by pkotwicz@chromium.org

[commit-bot: I haz the power, 2014-07-25 19:04:00]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/pkotwicz@chromium.org/405743004/40001

[commit-bot: I haz the power, 2014-07-25 20:21:29]

Change committed as 285658

[evanm, 2014-07-25 22:05:43]

For what it's worth, I expect this issue to affect more than metacity.
My memory of this area is pretty old, but I think the issue is this:

Some old apps existed before the fullscreen (NET_WM_STATE_FULLSCREEN or
whatever) flag existed.  They would fullscreen themselves by removing their
window decorations and adjusting their bounds to the size of the screen.

Then the fullscreen flag was added.  So some WMs would identify such old apps
and apply the flag for them.

Then Chrome came along, and properly obeyed the fullscreen flag for its real
fullscreen mode (where the tab strip etc is hidden).  But when Chrome was
maximized (not fullscreen) it wants to occupy the full screen without window
decorations, triggering this WM behavior.

As I recall the workaround in Chrome was to never let the window match the
screen size -- we'd resize to one pixel shorter than the screen, or maybe
taller, but avoid tickling the bug.

The code written here works around metacity specifically but it looks fragile
even for metacity and won't fix other wms that behave like metacity does.

[Daniel Erat, 2014-07-25 22:37:12]

On 2014/07/25 22:05:43, evanm wrote:
> For what it's worth, I expect this issue to affect more than metacity.
> My memory of this area is pretty old, but I think the issue is this:
> 
> Some old apps existed before the fullscreen (NET_WM_STATE_FULLSCREEN or
> whatever) flag existed.  They would fullscreen themselves by removing their
> window decorations and adjusting their bounds to the size of the screen.
> 
> Then the fullscreen flag was added.  So some WMs would identify such old apps
> and apply the flag for them.
> 
> Then Chrome came along, and properly obeyed the fullscreen flag for its real
> fullscreen mode (where the tab strip etc is hidden).  But when Chrome was
> maximized (not fullscreen) it wants to occupy the full screen without window
> decorations, triggering this WM behavior.
> 
> As I recall the workaround in Chrome was to never let the window match the
> screen size -- we'd resize to one pixel shorter than the screen, or maybe
> taller, but avoid tickling the bug.
> 
> The code written here works around metacity specifically but it looks fragile
> even for metacity and won't fix other wms that behave like metacity does.

i agree with evan here. fighting with the window manager is usually a recipe for
trouble (e.g. it does something, so you undo it, so it redoes it, etc.). i would
greatly prefer that the old approach used by the gtk version were ported
forward.

[pkotwicz, 2014-07-26 15:35:27]

I think I implemented the approach that you are discussing here:
https://codereview.chromium.org/407893002/