1
0
Fork 0
mutter-performance-source/src/tests
Owen W. Taylor 21f123c69f Don't restack windows while we are unmanaging them
Restacking the frame for a window while unmanaging the window is
harmless, but for undecorated (in particular, client-side-decorated)
windows, this causes problems because the window is typically
destroyed by the client immediately after withredrawing the window.

Skip windows flagged as being unmanaged when assembling the new
stack and when comparing the old order to the new stack.

Add a stacking test for this.
2014-09-16 13:44:42 -04:00
..
stacking Don't restack windows while we are unmanaging them 2014-09-16 13:44:42 -04:00
mutter-all.test.in Add missing file from test framework 2014-09-12 13:40:33 -04:00
README test-runner: Add 'csd' keyword for window creation 2014-09-16 10:59:45 -04:00
test-client.c test-runner: Add 'csd' keyword for window creation 2014-09-16 10:59:45 -04:00
test-runner.c mutter-test-runner: Make criticals and warnings failures 2014-09-16 10:59:45 -04:00

This directory implements a framework for automated tests of Mutter. The basic
idea is that mutter-test-runner acts as the window manager and compositor, and
forks off instances of mutter-test-client to act as clients.

There's a simple scripting language for tests. A very small test would look like:

---
# Start up a new X11 client with the client id 1 (doesn't have to be an integer)
# Windows for this client will be referred to as 1/<window-id>
new_client 1 x11

# Create and show two windows - again the IDs don't have to be integers
create 1/1
show 1/1
create 1/2
show 1/2

# Wait for the commands we've executed in the clients to reach Mutter
wait

# Check that the windows are in the order we expect
assert_stacking 1/1 1/2
---

Running
=======

The tests are installed according to:

https://wiki.gnome.org/Initiatives/GnomeGoals/InstalledTests

if --enable-installed-tests is passed to configure. You can run them
uninstalled with:

 cd src && make run-tests

Command reference
=================

The following commands are supported. Quoting and comments follow shell rules.

new_client <client-id> [wayland|x11]
 Starts a client, connecting by either Wayland or X11. The client
 will subsequently be known with the given client-id (an arbitrary
 string)

quit_client <client-id>
 Destroys all windows for the client, waits for that to be processed,
 then instructs the client to exit.

create <client-id>/<window-id> [override|csd]
 Creates a new window. For the X11 backend, the keyword 'override'
 can be given to create an override-redirect and the keyword 'csd'
 can be given to create a client-side decorated window.

show <client-id>/<window-id>
hide <client-id>/<window-id>
 Ask the client to show (map) or hide (unmap) the given window

activate <client-id>/<window-id>
 Ask the client to raise and focus the given window. This is currently a no-op
 for Wayland, where this capability is not supported in the protocol.

local_activate <client-id>-<window-id>
  The same as 'activate', but the operation is done directly inside Mutter
  and works for both backends

raise <client-id>/<window-id>
lower <client-id>/<window-id>
  Ask the client to raise or lower the given window ID. This is a no-op
  for Wayland clients. (It's also considered discouraged, but supported, for
  non-override-redirect X11 clients.)

destroy <client-id>/<window-id>
  Destroy the given window

wait
  Wait until all requests sent by Mutter to clients have been received by Mutter,
  and then wait until all requests by Mutter have been processed by the X server.

assert_stacking <client-id>/<window-id> <client-id>/<window-id> ...
  Assert that the list of client windows known to Mutter is as given and in
  the given order, bottom to top.

  This function also queries the X server stack and verifies that Mutter's
  expectation of the X server stack matches reality.