GitLab

- Specifically request GL version when creating context. Just specifying core
profile bit results in the requested version defaulting to 1.0 which causes
the core profile bit to be ignored and an arbitrary compatability context to be
returned.

- Fix GL painting by removing GL calls that have been depricated by the 3.2 core
profile.

- Additionally remove glInvalidateFramebuffer() call, it is not supported by 3.2
core.

https://bugzilla.gnome.org/show_bug.cgi?id=742953

The clipping was wrong so it was not drawing grid lines
outside the visible area, which broke due to the pixel cache.

Add helper for getting the main clipboard. This makes
the API usable for bindings (as GdkAtoms aren't usable through
gobject-introspection), and easier to use in C.

https://bugzilla.gnome.org/show_bug.cgi?id=712752

... and _gtk_css_style_property_changes_affect_font().

Replace it with _gtk_css_style_property_get_mask_affecting().

... to avoid recomputing the cached surface all the time.

Computes a bitmask for all properties that affect the given flags.

... and GtkCssStyleProperty::affects-font properties. Code now uses
GtkCssStyleProperty::affects instead.

The property is useless to set (it only allows 'initial', 'inherit' and
'unset' as values), but it is used to track changes to the icon theme.
And as such, it can ensure that widgets can track when they need to
reload icons.

https://bugzilla.gnome.org/show_bug.cgi?id=743341

Gtk_container_cell_widget_set should chain up to its parent's set
function, not its parent's unset function. This was resulting in
accessibles being erroneously marked defunct after being created.

The ICCCM says:

If the specified property is None, the requestor is an obsolete client.
Owners are encouraged to support these clients by using the specified
target atom as the property name to be used for the reply.

Lets do that, instead of crashing.

https://bugzilla.gnome.org/show_bug.cgi?id=740613

The previous fix for this issue in 732af314 was incomplete.

Remove two erroneous quotation marks.

GtkFileChooserWidget: Fix a crash for when the widget was unrefed right after setting a current folder.

Keep a reference to the widget while the async operation is ongoing.

https://bugzilla.gnome.org/show_bug.cgi?id=725164

These fields are never set, so we don't need the code to clean them up.

https://bugzilla.gnome.org/show_bug.cgi?id=725164

We don't support selection notification, so don't claim
we do, only to fail later.

https://bugzilla.gnome.org/show_bug.cgi?id=743229

No need to risk valgrind complaints about initialized
values.

https://bugzilla.gnome.org/show_bug.cgi?id=743422

https://bugzilla.gnome.org/show_bug.cgi?id=743286

Fix transfer annotation for tree_model parameter of gtk_tree_get_row_drag_data() and add nullable, optional

The GtkSelectionData does not own the model set with
gtk_tree_set_row_drag_data so change the out param to "transfer none".
This fixes a crash after multiple DnD actions due to the bindings
stealing a reference each time this function is called.

This also adds nullable and optional annotations for tree_model and path while at it.

https://bugzilla.gnome.org/show_bug.cgi?id=743193

This makes dnd icons work for widgets on a !default screen.

https://bugzilla.gnome.org/show_bug.cgi?id=743393

Make it so that typing switches to search unless we have a better
use for the key events, and stop search if the entry is emptied.

It is the third kind of navigation entry in the file chooser after
the location entry and search, and it interacts badly with both
of the others.

Sadly, GtkSearchBar allows connecting plain entries, so we have
to keep the old code around.

Avoid repetitive, manual event handling in various places
in GtkInspector.

Add ::next-match, ::previous-match and ::stop-search keybinding
signals that are bound to Ctrl-g, Ctrl-Shift-g and Escape. Also
add a gtk_search_entry_handle_event() function to handle key events.

If a side of the box is 0px wide, make the corners owned by the adjacent
sides. This avoids spilling over of unwanted colors from the 0-width
side into the corner.

New test for this case is included.

For widgets with a window, we need to adjust the allocation of their
children.

Previously, we would not include any child widget on the first
allocation, which happens right after realize(), but before map(). No
widget is drawable at that point.

This is now done (as it should) in the compute code.

This fixes the background being shifted by the border-width even though
the border-style was none.

when calling get_padding/border/margin, there is no need to go through
the fallback API.

This is a huge speedup for certain benchmarks.