Don't just create a menushell and populate it with random data -- verify that
the resulting menu layout is actually correct.
This is introduced in a separate commit because the old code was failing this
part of the test.
https://bugzilla.gnome.org/show_bug.cgi?id=696468
GtkMenuTracker folds a nested structure of sections in a GMenuModel into
a single linear menu, which it expresses to its user by means of 'insert
item at position' and 'remove item at position' callbacks.
The logic for where to insert separators and how to handle action
namespaces is contained within the tracker, removing the need to have
this logic duplicated in the 3 or 4 places that consume GMenuModel.
In comparison with the previous code, the tracker no longer completely
destroys and rebuilds menus every time a single change occurs. As a
result, the new gtkmenu testcase now runs in approximately 3 seconds
instead of ~60 before.
https://bugzilla.gnome.org/show_bug.cgi?id=696468
It's more logical to have the GtkTextAttributes and GtkTextAppearance
structs together. And it creates a separation between gtk_text_tag and
gtk_text_attributes functions.
https://bugzilla.gnome.org/show_bug.cgi?id=697048
We can get G_IO_HUP and G_IO_IN at the same time, if the compositor writes
data to us and then closes our connection. Make sure that we dispatch events
always if we have G_IO_IN and then error out if we get G_IO_HUP after that.
This way we don't have to reopen all the time for pure updates,
and we can immediately unlink the shm file to avoid "leaking" them
on improper shutdown.
We now only update surface data after we have painted. Before we painted
in an idle, which meant we might send black data some times if we e.g.
resized the window and had not painted yet. Also, it means we're updating
less often to the daemon, saving resources.
We still have to queue a flush in the idle for non-draw operations,
otherwise e.g. resize of a toplevel will never be flushed if the clock
is frozen (e.g. during toplevel resize).
We don't want to update the window size on configure event, only
the position, as the size is client side controlled. We were
updating to an old size during resizes which causes us to send
surfaces of the wrong size to the daemon.
The following CSS would infloop:
@define-color self @self
as it would infinitely lookup the color named "self" and try to resolve
it. This patch adds detection of such cycles to the resolve function by
keeping a list of currently resolving colors in the cycle_list variable.
If a named color references a nonexistant named color, we didn't catch
that error and ended up crashing on a NULL-dereference.
This crashed Boxes, because its CSS referenced values from the theme
that didn't exist in any theme.
The cursor buffer is only non-null when a cursor is created from pixbuf,
so it is not necessary to keep track of whether to free this buffer on
finalize.
