This was an oversight in the recent accel label improvement.
When we get an untranslated string back from gettext(), it is
not ok to replace '_' by ' ' in-place. Instead, do it while
appending to the GString.
https://bugzilla.gnome.org/show_bug.cgi?id=641912
This operation is potentially expensive, so the code has been
refurbished so avoid the cases it was needed for. Antialiased
cairo_clip() has been also reduced to transitions with surface
patterns involved.
Gdk-3.0 is already included above via --include-uninstalled, so
don't also specify it in INCLUDES. Doing so breaks if it isn't
installed already, since we try to find the gdk-3.0.pc.
It was unclear in the docs what happened if
gtk_style_context_add_provider() and
gtk_style_context_add_provider_for_screen()
were used with the same priority.
Sensitivity changes were not properly propagated down the
hierarchy. There were two issues here:
a) correctly identifying when a state change request affects
sensitivity
b) not filtering out sensitivity in gtk_widget_propagate_state(),
since gtk_widget_set_sensitivity() uses that to do its work
https://bugzilla.gnome.org/show_bug.cgi?id=641431
Optimized GtkSizeGroup code that is invoked for every queued resize
and every request that is not previously cached by trading qdata on
widgets for 3 extra bitfields on the GtkWidgetPrivate structure.
so pull the SELECTED state flag out of the !sensitive branch. Also,
don't make FOCUSED depend on SELECTED here, it's up to the widget to
decide whether or not that is possible.
It is used to get the default providers, without them
the style context can't do much. A check for NULL screen
is done before any sensitive call to
gtk_style_context_set_screen(), in the hope that any widget
will open the display before doing anything related to
styling. Fixes bug #641429, reported by Bastien Nocera.
Before this patch, unsetting the filter manually before disposing
the recent chooser menu (or unsetting the filter twice), would cause
warnings (or invalid memory accesses).
Having default implementations on abstract classes that do nothing
or print errors make life easier for language bindings implementors,
inside GTK+ this makes little or no difference.
Having default implementations on abstract classes that do nothing
or print errors make life easier for language bindings implementors,
inside GTK+ this makes little or no difference.
constructors which take an object of the same class as its first argument are
mis-detected as method call with "self" argument by the GIR scanner. Using the
new (constructor) annotation from bug 561264, mark some of them as proper
constuctors, so that you can call them with NULL as first argument from
bindings; in particular, this fixes gdk_window_new() and the
gtk_radio_button_new_with*() constructors.
b3f6f67c changed the loop from while() to for() in gtk_fixed_forall(),
but that's wrong since the callback can have side-effects on the list,
in case the current child gets removed. And that's the case when the
widget is destroyed.
Patch by Vincent Untz
https://bugzilla.gnome.org/show_bug.cgi?id=641196
The previous function gdk_drag_get_protocol_for_display() took native
window handles, so it had to be changed. Because it didn't do what it
was named to do (it didn't return a protocol even though it was named
get_protocol) and because it doesn't operate on the display anymore but
on the actual window, it's now called gdk_window_get_drag_protocol().
It turns out that the gtk_grab_remove() can trigger a do_prelight()
call, which may end up changing prelight_node, and then the state
gets messed up. Moving the grab removal until after we're done with
button_pressed_node and prlight_node makes expanders work reliably.
One thing that is still not right is that the expander doesn't get
prelight again after the animation is done, if you manage to release
without any additional motion events.
https://bugzilla.gnome.org/show_bug.cgi?id=641039
I've decided that it is isn't feasible to make cell areas runtime-settable
in the time we have left before 3.0, therefore, I'm going with the
approach to allow init() functions to instantiate the default cell area
and issue a warning if a construct property is ignored.
This is not ideal, but it keeps existing icon view and combo box
subclasses working.
https://bugzilla.gnome.org/show_bug.cgi?id=639139
The forall() loop was buggy as it was skipping items in the list when
the current item gets removed from the groups array as a result of
calling the callback (causing memory leaks).
Make gtk_binding_entry_add_signal_from_string() return the expected
token in case of parsing error, so that we can return a GError
instead of spewing. Also, add a separate scope for binding-set,
since allowing {} in identifiers in SCOPE_VALUE breaks the fact
that the ; after the last assignment in a rule is optional.
This is done in the same place than it was done before the switch
to GtkStyleContext, the GtkBindingEntry has been slightly simplified
since it's now the GtkStyleContext which hands us the list
of binding sets applying to a widget, so no need to use the older
matching mechanisms.
This property takes a list of binding set names to have these
stored as a GPtrArray containing GtkBindingSets. this property
is handled so merging GtkStyleProperties will merge both
GtkBindingSet lists.
these custom rules can be used to define key themes
as in the RC files, a series of bind/unbind rules
may be added then:
@binding foo-bar {
bind "<alt>1" { "move-cursor" (visual-positions, 1, 0) };
unbind "<ctl>d"
}
When loading a theme, make sure we take into account the variant
so we don't use the plain version when the theme changes.
Also make sure to fallback to the plain theme when loading a variant
fails.
https://bugzilla.gnome.org/show_bug.cgi?id=640983
This management is better done per-widget rather than per-screen,
as windows being destroyed won't trigger a leave notify for the
devices on top of it, and this information is too transitive
to keep weak refs and such.
This fixes the critical warning seen in gtk/tests/testing.
Move the private get_cursor_color method belongs to StyleContext. Change
the api so that retrieving both primary and secondary color is possible.
I left the method private for now, though it should probably be public
as all the other getters.
People agreed that it doesn't make sense to make an xembed
implementation like plug/socket does not make sense on anything but X11.
And as GdkNativeWindow is going away for multibackend portability
reasons, we'd either need API fixes in the plug-socket codee or
constrain the widgets to a single backend. We chose the latter.
Instead of checking if klass->get_request_mode is != NULL from
the gtk_widget_get_request_mode() api, this allows classes to
trust that there is a default implementation and chain up (specifically
added this for gtkmm wrapper objects).
Expander animation has been replaced by implicit animations
from the style context.
Column headers are also properly themed, GtkContainer::get_path_for_child()
is implemented and the treeview resets column buttons style on columns
being reordered.
The default CSS has changed to theme treeviews sensibly.
This lets themes override settings values again. We are using
the same priority that was used from the rc file parser, so things
are largely unchanged, relative to other settings sources.
Move key file parsing to gtk_settings_load_from_key_file(), in
preparation for loading per-theme files. Load key files from both
/etc and ~/.config, with the latter overriding the former. Support
parsing enum values.
Also added documentation section for this. Since the GtkRecentFilter
documentation was still living in sgml, as a side-effect I migrated these
docs to the gtkrecentfilter.[ch] sources.
GtkStateType was generally used as an index in GtkStyle color arrays, so
bigger values will cause invalid memory accesses in widgets that are still
doing that. this was seen in focused GtkIconViews for example
together with commit 8903615a34, this finally fixes bug #640282.
Insensitivity is handled separatedly in _gtk_widget_update_state_flags(),
but the insensitive flag is mistakenly unset afterwards if clear is TRUE
in gtk_widget_set_state_flags().
When set to PANGO_ALIGN_CENTER, text was rendered off-center
by the amount of 'free space' that the PangoLayout left (ie
the x returned by pango_layout_get_pixel_extents).
The logic was to try to switch to that nonexistent folder and thus get
an error message presented. However, no such message actually appears,
as the file chooser tries to switch to the closest parent folder that
actually exists, without bringing up an error message --- this is done
to cope with the case of the file chooser being started with a folder
that doesn't exist anymore.
Now, we just bring up an error message directly when we detect that
the user types a subfolder name that doesn't exist.
Signed-off-by: Federico Mena Quintero <federico@gnome.org>
We used to explicitly map and unmap the child GtkFileChooserWidget when
mapping and unmapping the dialog, respectively. Now that GtkWidget actually
unmaps child widgets (instead of avoiding that), we can assume that the
child GtkFileChooserWidget will be unmapped when we wanted it to be.
This fixes a warning that happened with the new GtkWidget invariant checker,
as we were mapping our child widget before calling our parent class' ::map() handler.
Signed-off-by: Federico Mena Quintero <federico@gnome.org>
Buttons may also be activated at any time from gtk_widget_activate()
or related functions. In that case, just do the 'show the button
as pushed for a short amount of time' trick, but don't actually
try to grab the keyboard device.
Signed-off-by: Federico Mena Quintero <federico@gnome.org>
