We were relying on indirectly getting notify when fontconfig
configuration changes, by GtkSettings translating the timestamp
change into a style-invalidation, which gets fed through the
css invalidation machinery. That machinery has gotten good enough
at optimizing away redundant changes that it no longer emits
::style-updated in this case.
So, instead make the font chooser listen directly to what it
cares about: the fontconfig change notification from GtkSettings.
We can use the GtkSettings:gtk-fontconfig-timestamp property to decide
whether or not we should reload fonts on style and screen changes. This
should avoid doing a lot of work with large font collections when only
the theme has changed.
https://bugzilla.gnome.org/show_bug.cgi?id=748782
Add a new API, gtk_popover_set_default_widget, that can be
used to make a widget act as default while the popover is
shown. This is useful in dialog-like popovers.
http://bugzilla.gnome.org/show_bug.cgi?id=747664
Making a container focusable is rarely the right thing to do.
It typically breaks the focus chain, and makes the container
contents unfocusable, as was the case here.
We have a number of cases where properties should have their
default value overridden in a subclass, but haven't because thats
annoying to do. We also have properties where the absence of
an explicit value has context-dependent meaning.
Add a list of exceptions for these cases.
The previous commit removed default-valued properties,
but apparently that has a negative effect for grid
packing properties, so put the explicit value back.
Just using tracker does not work well if you are searching in
non-indexed locations, such as git checkouts or network mounts.
Ideally, we'd decide the 'best' engine to use for each location.
Since that is not easy to do, just run them in parallel for now,
which is the same strategy that nautilus uses.
Catch an error that indicates the file looks like a template,
and then try again, this time with the template parsing API
of GtkBuilder. This is a little iffy, since we need to create
a 'fake' type and instance to pass in, but it works ok in
simple tests.
Add the class and parent class name to the error message.
gtk-builder-tool will parse the error message and use the
class names for trying again to parse the file as a template.
Requires Vista and newer.
* Create surfaces with cairo_win32_surface_create_with_format
* Provide an rgba visual that can be distinguished from the system visual
* Make rgba visual the best available visual
* Enable alpha-transparency for all windows that we control
* Check for appropriate cairo capabilities at configure time
(W32 - 1.14.3 newer than 2015-04-14; others - 1.14.0)
* Check for composition support before enabling CSDs
* Re-enable transparency on WM_DWMCOMPOSITIONCHANGED
Windows that were created while composition was enabled and that were CSDed
as a result and will look ugly (thick black borders or no borders at all) once
composition is disabled.
If composition is enabled afterwards, they will return back to normal.
This happens, for example, when RDP session is opened to a desktop where a GTK
application is running. For W7/Vista windows will only re-gain transparency after
the RDP session is closed. For W8 transparency will only be gone momentarily.
Windows that were created while composition was disabled will not be CSDed
automatically and will use SSD (WM decorations), while windows that are CSDed
manually will get a thin square border.
If composition is enabled afterwards, these windows will not change.
This is most noticeable for system menus (popup menus are often generated
on the fly, system menus are created once) and some dialogues (About dialogue,
for example).
https://bugzilla.gnome.org/show_bug.cgi?id=727316
Add a convenience function that is like gtk_builder_get_object()
but stashes away a GError if a lookup fails. To make the error
message informative, the function takes a line/column pair.
Doing things this way is necessary because the custom_tag_end,
custom_finished, and parser_finished vfuncs don't take a
GError parameter, despite being called from a place where
we can report a GError back.
We want to clear the style property cache whenever things change in the
tree, not as we previously did only when those changes actually lead to
a different CSS style.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
Instead of issuing g_warning, fill the provided GError.
This lets us test this error handling, and is the right
thing to do. Use the new GtkBuilder helpers and
g_markup_collect_attributes to do so.
cairo_rectangle_int_t was replaced by GdkRectangle in commit
552c29b488, but the type of the pointing-to
property was not changed.
To avoid breaking old code that sets or gets the property with a GValue
of type CAIRO_GOBJECT_TYPE_RECTANGLE_INT, transformation functions between
CAIRO_GOBJECT_TYPE_RECTANGLE_INT and GDK_TYPE_RECTANGLE are registered on
the first call to gdk_rectangle_get_type().
https://bugzilla.gnome.org/show_bug.cgi?id=723394
To calculate the shadow width, we look at the value of priv->fullscreen
and priv->maximized.
Those fields will have the actual value only after GTK receives back a
window state event though, so they will be wrong in _realize(). Look at
priv->fullscreen_initially and priv->maximize_initially too, to avoid
the size changing right after realize, which would make the window
flicker if maximized at startup.
https://bugzilla.gnome.org/show_bug.cgi?id=747808
Libinput will use 0.0f on the "scrolling finished" event, so check for this
instead of rounding (<1 values are sort of frequent on touchpads). This
impedes bug #745315 to resurface after commit d563b943ed.
When the scrolledwindow receives scroll events, it ensures the timeout to
maybe start the "snap back to edges" animation is reset, but it does nothing
about the animation source. It must be reset just the same, to maybe be
started after the timeout fires up.
Added another :not() to fix the first and last swatches in the
first section of colorpicker.
Dodge the artifacts due to stacked anti-aliased rounded corners by
increasing the radius of the base object by 1px.
We were only updating window buttons when the headerbar was
a direct child of the window. That is not the case in more
complicated situations, such as the split headers in gedit
or polari. To fix such cases, make the headerbar itself listen
for state changes on its toplevel.
https://bugzilla.gnome.org/show_bug.cgi?id=747805
This used to do the right thing, practically ignoring those event
if they went all through the event handler, unless this motion event
triggered the cancellation of this gesture from the event handler
of another controllers.
In this case, the controller would be reset, but the motion event
would still go through its handler, setting again
current_button/sequence as the motion event has buttons in its
modifiers. This leaves GtkGestureSingle with inconsistent data
that may interfere with future runs.
https://bugzilla.gnome.org/show_bug.cgi?id=747469
The code managing scrollbars visibility was too pervasively checking for
mouse devices, leaving pen/eraser/cursor devices with no scrollbars at
all. Relax these checks a bit, and actually toggle full-width scrollbars
on pen/eraser devices, so it is an easier target.
https://bugzilla.gnome.org/show_bug.cgi?id=747608
Since windows can be intermediate containers this is somewhat redundant,
plus gtk_window_size_allocate() will do additional things like allocating
the popovers, which was mistakenly skipped by local plugs not chaining up.
https://bugzilla.gnome.org/show_bug.cgi?id=747553
This is a workaround for atspi-atk behaviour.
atspi-atk uses signal emission hooks. So it to already catches
signal emissions on creation of objects, before anyone could even
think of g_signal_connect()ing.
https://bugzilla.gnome.org/show_bug.cgi?id=746706
gtk_application_shutdown clears the impl member, so all
callbacks, signal handlers, etc that might still be triggered
between a shutdown call and the return from the mainloop
better be prepared to deal with impl being NULL.
strcasecmp() is unfortunately not universally available, along with
strings.h. Fix the build by replacing strcasecmp() with
g_ascii_strcasecmp(), and remove the strings.h include.
https://bugzilla.gnome.org/show_bug.cgi?id=747604
When a treeview is destroyed while rubberbanding is going
on, we crash because the rb tree is nuked before we want
to access it to stop the rubberbanding. To avoid this crash
end the rubberbanding early in destroy().
See
https://bugzilla.redhat.com/show_bug.cgi?id=1173904
It can apparently happen that we get focus in events
on windows after gtk_application_shutdown() has been
called. Avoid an unnecessary crash in this case.
See
https://bugzilla.redhat.com/show_bug.cgi?id=1176339
The coordinate translations here were not working properly
for window widgets inside the scrolled window, as can be
seen e.g. for the horizontal scrollbar of the 'Tree View'
example in gtk3-demo.
https://bugzilla.gnome.org/show_bug.cgi?id=747406
When moving over a non-expanded indicator from the outside, we were
not expanding it, due to on_scrollbar being true. This can be seen
e.g. when moving from the content pane over to the sidebar indicator
in gtk3-demo. We must still ensure that the indicator is expanded
when receiving motion events over the indicator.
https://bugzilla.gnome.org/show_bug.cgi?id=747407
The gtk_label_set_text() and gtk_label_set_markup() functions have
various side effects that ought to be documented, especially for
non-C developers using properties directly.
https://bugzilla.gnome.org/show_bug.cgi?id=747121
This path is only intended to be triggered on events directed towards the
child of the scrolledwindow, so make it explicitly so. This avoids scrollbar
"over" state flashing when dragging finishes within the slider.
https://bugzilla.gnome.org/show_bug.cgi?id=746961
The "app_id" of a xdg_surface should be the ID that can potentially be
used to get the DBUS name or the .desktop file.
For GtkApplication programs this is often the ID passed when creating the
GtkApplication object, so when available lets use that.
As fallbacks, first try g_get_prgname as it often corresponds to the
basename part of the .dektop file for non-GtkApplication programs.
Otherwise use gdk_get_program_class, even though that string usually
doesn't conform to the expectations of xdg_surface.set_application_id.
https://bugzilla.gnome.org/show_bug.cgi?id=746435
If the icon happened to come from the drag site icon helper (ie. set through
gtk_drag_source_set_icon*), it would be referenced on the GtkDragInfo,
but not hooked into its context. This results on non visible drag windows,
until set_icon_helper() happened to be called on some path.
GtkRadioButton had a deficient copy of the focus sort code
in GtkContainer, causing focus to jump over the next button
in the list. Just use _gtk_container_focus_sort() here,
which fixes the bug _and_ saves 80 lines of code.
https://bugzilla.gnome.org/show_bug.cgi?id=746817
As of 74405cc, progress bars use a new design with values drawn on top
(or to the left) of the through instead of inside of it. This change
brought a number of regressions: the min-horizontal-bar-height and
min-vertical-bar-width style properties are not respected anymore. For
vertical progress bars, the value was drawn too close to the bar and not
centered vertically.
Fix this by respecting the style properties and drawing the value label
at the correct position.
Also, the xspacing and yspacing properties didn't server any apparent
purpose. Change their semantics to mean "the spacing between the label
and the bar". Hence, they only need to be added to the size request when
showing the label. Since we are changing semantics anyway, reduce their
default values from 7 to 2, to avoid and excessive gap.
https://bugzilla.gnome.org/show_bug.cgi?id=746688
As it is, GtkListBox model binding will work nicely as long as your
create_widget_func returns a floating reference on the newly-created
widget.
If you try to return a full reference (as any higher-level language
would do) then you will leak that reference.
Fix that up by converting any floating references into full references
and then unconditionally releasing the full reference after adding to
the box.
https://bugzilla.gnome.org/show_bug.cgi?id=746893
I managed to stall recent files today while trying to save a GTK
testcase in glade that contained enough spinning spinners that the CPU
was saturated just redrawing things.
I had to navigate the filesystem!
The direction in which the slider moves can be inverted by setting the
inverted property. But the draw method does not check this, instead it
checks if the direction of the widget is set to be right to left.
Call the should_invert function in order to determine if the direction
of the range should be inverted. It too checks the widget's direction,
but also checks the "inverted" property, and allows the range to be
drawn inverted even if it is vertically oriented.
https://bugzilla.gnome.org/show_bug.cgi?id=746712
Put the equality check in front. This allows better detection of when an
insert or delete needs to be emitted.
Also, only emit text-changed:delete if the deleted text is not the empty
string. Only emit text-changed:insert if the inserted text is not the
empty string.
https://bugzilla.gnome.org/show_bug.cgi?id=746706
Compute the new text and its length in advance. This way those
computations will not confuse us when they happen in the middle of the
actual action.
https://bugzilla.gnome.org/show_bug.cgi?id=746706
Compute the proportion of the range that should be filled to match the
fill level, and use it to compute the starting point and length of the
area between the slider and the fill level.
https://bugzilla.gnome.org/show_bug.cgi?id=734741
When recomputing CSS, we need a correct widget path in the fallback mode
where we're still using widget paths.
So we need to invalidate it everytime it actually changes, and not just
when emitting the style-updated signal.
Fixes css-match-regions reftest.
Clang complains that this check can never be true. Since this
is a argument range check which we do to catch bad input,
convince clang to not complain instead of taking it out.
clang complains that the expression involving sqrt() is not
constant, and thus refuses to accept it as a case label. So,
use precomputed values instead.
https://bugzilla.gnome.org/show_bug.cgi?id=746468
Break out a setter that manages the weak pointer, and
use it in finalize. This also fixes a bug where we were
forgetting to disconnect the right signal handler in
some cases.
The other Radio* widgets have this convenience method that removes the
memory management of the opaque GSList used to handle the group from the
API usable from language bindings (especially the ones not based on
introspection).
This commit adds gtk_radio_menu_item_join_group().
https://bugzilla.gnome.org/show_bug.cgi?id=671362
Some arguments, like the group and the label of a RadioMenuItem, can be
NULL: the RadioMenuItem has all the code to deal with them. The argument
validation is too strict, though, for instance doing:
return_if_fail (IS_RADIO_MENU_ITEM (foo))
if (foo != NULL)
set_foo (foo)
Which is obviously incorrect.
This commit also modifies the annotations of the API, to ensure that
language bindings do the right thing.
https://bugzilla.gnome.org/show_bug.cgi?id=671362
If we create a RadioMenuItem without a group, and then set a group, the
menu item will still be set as active, which means an inconsistently
drawn radio menu item - as the RadioMenuItem will set the active flag on
itself, but then it won't reset it when it gets a new group.
https://bugzilla.gnome.org/show_bug.cgi?id=671362
It turned out that using mwm hints to instruct wms to
create border-only decorations is not really working
universally. So, instead of doing this, render a solid
frame without shadow on the client-side to handle this
case.
https://bugzilla.gnome.org/show_bug.cgi?id=746222
Signed-off-by: Olivier Fourdan <fourdan@xfce.org>
If the widget has been destroyed since a DBus message had been sent,
we could be in a condition that the widget pointer exists but it does
not have a window.
This bails as if the widget didn't exist if there is no available
GdkWindow.
We also set the extents to 0 to be defensive since this is a vfunc
implementation.
https://bugzilla.gnome.org/show_bug.cgi?id=746586
Since these part really are the same in all of the x or y direction
and we don't blur in that direction we can just blur one line and
repeat it during drawing.
https://bugzilla.gnome.org/show_bug.cgi?id=746468
There is no need to e.g. blur in the x-direction for the top part
of a box shadow. Also, there is no need to extend the mask in the
non-blurred direction.
https://bugzilla.gnome.org/show_bug.cgi?id=746468
For radius 1px the current implementation rounds down to a 1 px box
filter which is a no-op. Rather than creating useless shadow masks
in this case we bail out blurring early.
Another alternative would be to make radius 1px round up to a 2 px box
filter, but that would change the rendering of Adwaita which is probably
not a great idea this late in the cycle.
https://bugzilla.gnome.org/show_bug.cgi?id=746468
Sometimes path nodes can survive longer than the style context that
created them. Don't crash in those cases.
Fixes startup of mutter.
Testcase included.
https://bugzilla.gnome.org/show_bug.cgi?id=746407
This allows monitoring the CSS tree. For now, moving a child to a
different position relative to its siblings while keeping the same
parent will cause a child-added + child-removed emission.
We need to properly track if a node needs to propagate invalidation
state information to its children. We didn't do this properly before and
that could lead to us forgetting to invalidate nodes in corner cases.
Do not propagate the TIMESTAMP change through the node tree, as that
causes lots of uneeded markings of nodes as invalid.
Instead, walk the node tree and find the nodes that have a non-static
style and only invalidate timestamps on those.
Only invalidate timestamps if the node is marked as invalid. We overload
the meaning of "invalid" as "tracks timestamps".
While I don't like the way this is written, it is an important
optimization because 95+% of nodes don't animate so timestamps don't
matter to them. But timestamps are invalidated 60x per second.
We don't return a NULL style to mean "no changes" anymore, instead
we check new_style == old_style to mean that.
Make sure the code reflects this, otherwise we'll send
GTK_CSS_CHANGE_PARENT_STYLE invalidations everywhere and screw up
performance.
Now that the widget node recomputes styles on update_style() we can just
call it during validate(). That way, we don't need the widget node to
manually compute its style.
If CSS values are queried from a widget, recompute them if necessary. Do
not emit style-updated until the validation phase however.
This way, we don't run into performance traps when style-update causes
invalidations that cause new style-updated to be emitted.
... and pass it to the API that computes new styles.
A special timestamp of 0 means "please don't animate" and is used when
no frame clock is available for a node.
This is mainly an attempt to merge the update_style() and validte()
vfuncs. Code is not there yet, but that's the idea.
Also, gtk_css_node_set_style() should not be public. And this gets
closer to that goal, too.
After measuring it, I realized the optimization never triggers for
Adwaita and rarely ever triggers for simple themes. So it is not
useful to keep it around.
This allows adding more API for it.
It also includes code that tracks modifications and invalidates siblings
and their positions whenever nodes get added or removed.
If we know the parent's get_path_for_child() implementation is safe to
be used with GtkCssNode because it doesn't do anything special, we do
that. Unfortunately that requires whitelisting vfuncs because the vfunc
is public API so anyone can override it.
Instead, use gtk_widget_get_path() which makes GtkWidget cache the path.
This is a temporary solution until we can get rid of widget paths.
This increases memory usage quite noticably.
The node declaration has the same functionality as
gtk_css_node_declaration_add_to_widget_path(). So instead of using that
function on a path, you can use the original path and the declaration in
a matcher.
So far the vfunc is kinda quirky (the path argument is an out argument
for something you have to free when you're done with the matcher), but
I'm about to change that.
We don't want to add the current classes to the widget path - which
might potentially be different after a gtk_style_context_save() - but
the root node's ones. So what better thing to do than actually using the
root node?
Instead of directly requesting the layout phase, register a tick
callback. This is what the docs suggest for animations and it's what we
need for the next commit.
... to a bunch of functions.
This requires a tiny change to the heuristics for the style cache - we
now cache styles when they have the same style provider as their parent
instead of when they have the default provider - but that change doesn't
have any effect in practice.
The parent refs the child, so gtk_css_node_set_parent() adds/removes a
reference.
We should probably refactor this so that we name the function
parent.add(node) instead of node.set_parent(parent) - makes the
refcounting more clear.
Instead of passing the style declaration, take the node's style
declaration. Add arguments to allow overriding the state for the one
case where this didn't work.
The code is confusing the stylecontext=>node transition, so I'll remove
it.
WIth the recent introduction of the parent style cachen, I'm not sure
it's worth it at all. Some crude benchmarking suggests a slight speedup
when removing the cache.
So no guarantees about adding it back later.
- GtkCssWidgetNode
for style contexts owned by a widget
- GtkCssPathNode
for style contexts using a GtkWidgetPath
- GtkCssTransientNode
for nodes created with gtk_style_context_save()/restore()
