Things like color affect symbolic icons, but not colored icons, while
other css properties like -gtk-icon-effect affect colored icons, but not
symbolic ones.
If the theme has rounded corners for fullscreen, we don't tell the
window manager that we are now fully opaque, which then makes things
less efficient than they should be.
https://bugzilla.gnome.org//show_bug.cgi?id=761571
When we don't get a baseline passed in, we want to basically
center the children inside the allocation. There was an attempt
in the code to do 'internal baseline alignment', but it had the
side effect of moving the contents to the top when we don't get
a baseline passed in. Remove it for now, this needs some more
infrastructure to do properly.
https://bugzilla.gnome.org/show_bug.cgi?id=761363
These were affected by the recent change to level offset handling.
At the same time, make the test files more realistic by updating
the level offsets when we set a custom range.
We had some odd special-casing for the lowest and highest offset
that did not quite work. The new rule is simple: If the value
is between offset n-1 and n, it gets the style for offset n.
https://bugzilla.gnome.org/show_bug.cgi?id=761416
There are a couple of issues with the way that buffers are handled in
wayland in right. These issues mean that:
- buffers can get leaked at a fairly fast clip under the right
conditions. This leads to the OOM killer kicking in and
gnome-shell and gnome-terminal (for instance) showing memory
usage in the high gigabytes range.
- drawing can happen to a shared memory buffer at the same time
the compositor is reading out the pixels. This can lead to
glitching in drawing and other undefined behavior by the compositor.
This changeset reworks how buffer management is done in the code to try
to address both problems.
The first change (commit 2c300081) addresses the leak by dropping code
that has an unchecked cairo_surface_reference call. The code is dropped
rather than fixed, because it has a more serious issue: it's overarching
purpose is to deal with shared memory buffer contention with the
compositor, but it does it in a racy way and so fails at that mission.
The second change (commit 40e91195a) moves what layer of the code buffer
release events are handled. This is an organizational change in the
code, with no functional changes, but it's important for the last change
in the changeset.
The last change (commit c80dd549) adds back code for dealing with shared
member buffer contention in a race free way. The new code is careful to
never reuse a buffer that hasn't been explicitly released by the
compositor.
Right now we use one buffer for both staged changes (freshly painted
changes waiting for the frame clock to send to the compositor) and
committed changes (changes actively being read by the compositor
process). This creates a problem in the event we need to stage updates
at the same time the compositor is processing committed updates: we
can't change what the compositor is actively processing.
The current solution for handling this contention is to allocate a
temporary buffer on the spot at the time the updates are staged, and to
copy that buffer back to the shared buffer later. The problem, though,
is that the copy to the shared buffer currently happens as soon as
the updates are finished being staged, not when the shared buffer is
done being processed by the compositor.
In order to address that problem, this commit changes the code to always
stage changes to a dedicated staging buffer. The staging buffer is
used exclusively by the client until the client is done with it, and then
once that staging buffer is committed, the client never writes to that
buffer again. If the client needs to stage new updates, it allocates a
brand new staging buffer, draws to it, and back fills the undrawn parts
of the buffer from a copy of the contents of the committed buffer.
As an optimization, the compositor has the option of releasing the
committed buffer back to the client. If it does so before the client
needs to stage new updates, then the client will reuse the buffer
for staging future updates. This optimization prevents having to allocate
a new staging buffer and the associated cost of back filling
that new buffer with a readback of the committed buffer.
https://bugzilla.gnome.org/show_bug.cgi?id=761312
Right now we handle buffer releases coming from the
compositor in a central place. We add a listener when
first creating the shared buffers.
This is problematic because a buffer can only have
one listener on it at once so users of the buffer
can't get notified when it's released.
This commit moves the buffer listener code from the
centrally managed display code to the cursor and window
code.
https://bugzilla.gnome.org/show_bug.cgi?id=761312
The client and compositor share access to the window
pixel buffers. After the client hands off (commits)
the buffer to the compositor it's not supposed to write
to it again until it's released by the compositor.
The code tries to deal with this contention by allocating
a temporary buffer and using that in the mean time. This
temporary buffer is allocated by a higher layer of the code
when begin_paint returns TRUE. Unfortunately, that layer of
the code has no idea when the buffer is released, so it ends
up blitting the temporary buffer back to the shared buffer
prematurely.
This commit changes begin_paint to always return FALSE.
A future commit will address the contention problem in
a different way.
https://bugzilla.gnome.org/show_bug.cgi?id=761312
There are a few places where we destroy a cairo surface and
then nullify it. This commit changes those to use
g_clear_pointer instead.
It also drops a cairo_surface_finish call that is unnecessary
https://bugzilla.gnome.org/show_bug.cgi?id=761312
The name surface is really overloaded when dealing
with wayland windows.
To alleviate ambiguity, this commit changes the name
of the "surface" and "subsurface" members to have
a wl_ prefix.
https://bugzilla.gnome.org/show_bug.cgi?id=761312
Don't allow syntax like
at top left circle
but follow the spec about requiring the at <position> right before the
comma.
This is porbably because
circle at 10px 10px
could be interpreted as
circle 10px at 10px
with the now disallowed syntax, too.
Test included.
Don't hardcode 96 for dpi, but instead use the value of the -gtk-dpi
property (that mirrors the GdkScreen's dpi if it wasn't set explicitly).
This makes these values scale when the large font setting in
control-center is enabled.
During the gadget conversion, the drawing of discrete levelbars
was unintentionally changed to draw a wide trough but narrow
blocks, which does not look great. So go back to the previous
way of drawing things.
https://bugzilla.gnome.org/show_bug.cgi?id=761428
Now selecting a widget by class name no longer works.
This is probably most relevant for users outside of GTK that want to
style their own widgets. Those widgets should now either add their own
style classes (if they want to adjust existing CSS) or use
gtk_widget_class_set_css_name() themselves (if they want to get rid of
all "upstream" styling).
And reset the grab_location in the ::released handler of the multipress
gesture.
Previously, when leaving fine-tune mode, the ::released handler of the
multipress gesture would call stop_scrolling, which calls
range_grab_remove and resets the grab_location. The ::drag-end handler
is executed after that, and only unsets priv->in_drag if the
grab_location is MOUSE_OUTSIDE, which it never was, since the ::released
handler already reset it. This lead to priv->in_drag being set even
though no dragging was in progress anymore, which e.g. made shift
pressed after leaving the fine-tune mode entering it again.
https://bugzilla.gnome.org/show_bug.cgi?id=761402
The fallback behaviour of get_work_area () divides the
screen width and height by the window scaling factor, but
those values are already scaled down.
https://bugzilla.gnome.org/show_bug.cgi?id=761474
Calling _gtk_file_consider_as_remote() with a NULL argument
results in warnings being thrown.
Note that query->priv->location being NULL is a state that does
not seem to be invalid by itself.
This could happen if you do search-as-you-type in a filechooser,
which has a filter that does not match anything *and* the current
"place" selected is "Recent".
https://bugzilla.gnome.org/show_bug.cgi?id=761552
