The model_cancellables have an extra ref when we get them from gtk_tree_model_get().
So, we need to unref them always. On the other hand, the cancellables that get
passed as arguments to the callbacks in question are memory-managed by
GtkFileSystemModel.
Reported by Morten Welinder <mortenw@gnome.org>
https://bugzilla.gnome.org/show_bug.cgi?id=646460
Instead rely on state_flags & GTK_STATE_FLAG_INSENSITIVE to tell us if a
widget is sensitive.
This has the huge benefit that the way the widget is actually rendered
corresponds to the return value of gtk_widget_is_sensitive().
As a side effect, we do not ever allow unsetting the
GTK_STATE_FLAG_INSENSITIVE for a widget the is set to not be sensitive
(via gtk_widget_set_sensitive()). This way we stop propagation of making
stuff sensitive at insensitive widgets.
https://bugzilla.gnome.org/show_bug.cgi?id=642918
This commit makes GTK_ALIGN_START/_END pay attention to
the text direction when used in horizontal context.
This is how similar parameters in GtkMisc and GtkAlignment work,
and is generally expected of GTK+ positioning parameters. And this
is new GTK+ 3 api, so it is basically still unused at this point.
If explicit right/left turn out to be needed at some point, we
can expand the enumeration with new values.
This code has been useless for a while, and now it's breaking things.
Moving the paned in tests/testheightforwidth's wrapping label shows that
bug.
Broken since 9992efdb0e
Use the _NET_SYSTEM_TRAY_ICON_SIZE property set by the tray mananger as
a hint that we should use a specific icon size. This allows the tray to
instruct us that it expects 16x16 icons, for example.
Bug #645232
Oftentimes we want to measure a layout that is as wide or wider than the
current layout's maximal width. In that case we can safely reuse the
current layout.
It's always FALSE.
Note that this patch changes the layout for allocations that are
smaller than 1px, but that's just the default layout that is never
rendered.
Async callbacks are delivered in idles, so we need to make sure
we get the gdk lock before calling any gdk/gtk stuff. This was
missing in a few places.
Dumps the widget path into a string representation. It tries to match the CSS
style as closely as possible (Note that there might be paths that cannot be
represented in CSS).
The main use of this code is for debugging purposes, so that you can
g_print() the path or dump it in a gdb session.
This ensures that widgets that aren't ported and rely on the style-set
signal being emitted work as well as before. They should not rely on
style-set being emitted however.
Note that this function is a no-op if the initial style has been set
already and is very cheap if it has not been set yet. It only becomes
relevant if the resulting style actually gets used.
https://bugzilla.gnome.org/show_bug.cgi?id=639584
The intention of this patch is to make the code clearer, shorter and
most of all to avoid recreating the widget path and setting it path
twice on the style context when the style context was recreated.
Now that we have wesome macros in glib 2.0, we can use them. Woohoo, GTK
2.0 will be so awesome in the future.
(Did anybody touch dnd code this millenium?)
This patch optimizes window resizes by assuming that if a widget
has the same height at a width of 50 as with a width of 150, the
height for width 100 will also be the same.
The patch also further optimizes the cache allocator, now there
are 2 pointer arrays of up to a maximum of 5 requests, the arrays
will only be allocated if a request is ever made in that orientation
and the array will be sparse until each request is made (i.e. if a
label can only wrap to 3 lines, there will only be 3 out of a
possible 5 SizeRequest structures allocated to cache it).
This patch makes contextual height-for-width request caching
optional (the contextual cache is not allocated for widgets that
report GTK_SIZE_REQUEST_CONSTANT_SIZE).
The constant size request mode defines a request mode where
height-for-width geometry is unneeded, thus optimizing GTK+
by reducing the overall amount of requests that need to be
performed and cached while resizing an interface.
gtk_settings_get_default() was used to monitor the
"gtk-application-prefers-dark-theme" setting, rather
than the GtkSettings object associated with the window's
screen.
Fixes Bug 639584 - initial emission of GtkWidget:style-set is
not happening.
GtkWidget was filtering out ::style-updated (and ::style-set)
emissions until the widget was realized in order to avoid often
useless updates during widget construction and placing.
This is now done instead until the widget has a parent/screen,
which ensures the initial emission happen prior to the first
size negociation, while still filtering out all early emissions
during construction.
You can select something, but moving the focus somewhere else, or bringing up
the context menu makes the selection go away. This makes it impossible to
copy-paste the text.
The reason for this regression is that gtk_label_state_changed was sloppy and
assumed that it would only ever be called when a labels goes insensitive, which
is no longer the case.
The patch fixes things by explicitly checking if the widget is insensitive, and
only resetting the selection in that case, which is the same behaviour we have
for e.g. GtkEntry.
It turns out the bug is more complicated than I originally understood.
Not replacing the font description fields while merging here makes it
impossible for application's CSS to override fonts.
This reverts commit 89c1d93b68.
https://bugzilla.gnome.org/show_bug.cgi?id=645458
Use the background color instead.
I think the optimal solution would be delegating the shadow rendering to
the theme completely, and removing this in GtkThemingEngine, but for
now, this simple fix makes the dark variant more usable.
https://bugzilla.gnome.org/show_bug.cgi?id=645405
The metacity theme format allows to use colors from the current
GTK+ theme in window decorations. Since GTK+ now gained support
for dark theme variants, window managers using that theme format
(metacity, mutter, compiz via gtk-window-decorator) should be able
to use colors from the correct variant; so in case a variant is
requested, export it in the _GTK_THEME_VARIANT property on
toplevel windows.
https://bugzilla.gnome.org/show_bug.cgi?id=645354
gtk_entry_set_placeholder_text() makes the entry display the given text
when it is empty and unfocused. Based on previous patch by Alberto
Garcia.
https://bugzilla.gnome.org/show_bug.cgi?id=440963
We were ignoring the focus ring padding, and we were incorrectly setting
handle_x to padding.left, while the draw handler already takes padding
into account while redrawing.
https://bugzilla.gnome.org/show_bug.cgi?id=645134
When tabs are not shown, and the notebook is used as a container,
we should not set the notebook style, otherwise the theme has no
way to know whether tabs are shown or not, and which colour to draw
the background.
https://bugzilla.gnome.org/show_bug.cgi?id=640692
Writing onto the passed-in GtkStateData could cause changes that would
propagate to siblings, as the data was not reset again.
By copying the data structure, this is avoided and the proper values are
passed to sibling widgets.
Commit 4bb3d64414 introduced a limitation
to GtkRange style properties; when stepper-spacing is > 0,
trough-under-steppers is automatically set to FALSE; this means that
setting a spacing between the steppers (e.g. the scrollbar buttons) and
the trough (i.e. the area over which the slider is free to move) would
make the buttons always get the full allocation on the !orientation
direction.
The rationale is without this limitation, you would get an area which
seems clickable, but it's actually not.
While this is true, and undesirable, for big stepper spacings, themes
that use trough-under-steppers (which is TRUE by default anyway),
might want to set smaller spacings to avoid drawing a double line between
the button and the slider borders.
To add confusion, the documentation got it flipped, i.e. it stated
setting a positive stepper-spacing would set trough-under-steppers to
TRUE (which would also make the behavior expected by commit
4bb3d64414 impossible).
I don't think hardcoding either of the two limitations is a good thing.
We should let themes handle this instead, and remove this limitation. If
you want the old behavior, you can manually set trough-under-steppers to
FALSE if you set a positive stepper-spacing in your theme.
https://bugzilla.gnome.org/show_bug.cgi?id=644777
This documentation was in direct conflict with the documentation
of gtk_tree_model_row_deleted() docs, the function docs are TRUE,
the signal docs were FALSE, fixed.
parse_value() could fill in a GError, but yet (incorrectly) return G_TOKEN_NONE,
having the GError leaked as a result.
https://bugzilla.gnome.org/show_bug.cgi?id=642604 , Reported by Felix Riemann.
Use lower level methods that just require changing the state in the
GtkStyleContext instead. Fixes a problem where the insensitive state
was being set to the entry becasuse a parent was insensitive, spotted
by Alex Larsson.
The render order for tabs is now
- left to right until the active tab
- right to left until the active tab
- active tab
This allows themes that use non-straight lines for the tab curvature to
draw them not worrying about flipping one side after the active tab.
Let the tab overlap eat the padding, otherwise having tab-overlap >
tab-curvature without cutting the label is impossible.
This way we give the label widget all the allocation possible before
giving up and cutting it if the values for tab-overlap are too extreme
for the given padding.
Previously, the inactive state caused the spinner's solid line to be
drawn to the right. Now it's drawn to the top instead, which makes the
inactive state look more natural.
Previously, for i == 0, we were drawing the translucent line first.
However, people reading the code would always assume the solid line is
indicated by the current step. So make the code do that.
Note that the spinner now rotates in the wrong direction. This will be
fixed in a future commit.
* gtk/gtkmain.c: (gtk_init): Mention that argc and argv can be 0,
using the new text from the g_applicatoin_run() documentation.
There was already introspection annotation about that.
(gtk_init_check, gtk_init_with_args): Use the same text here.
See bug #643649 comment 2.
This is in preparation for making scale marks styling better.
As a side-effect, TOP and LEFT position are treated the same now,
as well as BOTTOM and RIGHT.
Based on a patch by Bastien Nocera,
https://bugzilla.gnome.org/show_bug.cgi?id=643685
GtkWidget now parses custom attributes like
<style>
<class name="dark-label"/>
<class name="big-heading"/>
</style>
to add style classes to widgets.
https://bugzilla.gnome.org/show_bug.cgi?id=643347
Set the Mac key theme when creating a Quartz settings object instead of
having it the default when quartz is enabled.
This keeps compatibility with the GTK2 behavior that the Mac key theme
is not used for the X11 backend, which could now happen for a
multi-backend build.
Previously we were enabling some settings properties only if the X11
backend was enabled. This worked fine with GTK2 where only one backend
was enabled at a time, but now when multiple backends can be enabled,
this does not make sense.
GtkSwitch was listening for Space/Enter keyreleases itself, instead
of providing an 'activate' action signal, like e.g GtkButton. As a
side-effect, this fixes
https://bugzilla.gnome.org/show_bug.cgi?id=643321
In this patch we adress rows_reordered() and row_deleted() callbacks
(since some icon view subclasses manipulate the connected treemodel
from _init()).
We were breaking prematurely out of the loop that goes through
each file that is pending selection; not only does that loop
disable filtering and show hidden files if needed, but it also
selects the files themselves. So we need to walk the whole
list of files.
Signed-off-by: Federico Mena Quintero <federico@gnome.org>
https://bugzilla.gnome.org/show_bug.cgi?id=643170
Adding these cases here to cater to icon view subclasses that want to
access icon view APIs from the instance structure initializer instead
of properly waiting for the super class to initialize and adding renderers
from the ->constructor() vfunc.
This avoids overflow when transforming a large value to a pango unit.
To reproduce the problem:
seq 200000 > test.txt && tets/print-editor test.txt
Then scroll to around line 140.000 to see it (depends on font size of
course).
gtk_tree_model_filter_row_deleted was emitting ::row-deleted while
the internal state of the model was still in disarray, causing
segfaults e.g. when mapping the file chooser with accessibility
turned on. This is just a bandaid fix, and doesn't try address
any of the deeper problems of the filter model code.
I did take the time to reduce rampant variable shadowing in that
function, though.
If a level 1 key maps to a key value passed to
gdk_test_simulate_key(), raise the GDK_SHIFT_MASK flag so the reqested
key value is generated. Also add a regression test for this fix.
