Mario Blättermann [Wed, 27 Jan 2016 19:08:25 +0000 (20:08 +0100)]
Updated German translation
Matthias Clasen [Wed, 27 Jan 2016 18:23:58 +0000 (13:23 -0500)]
css node tests: Update expected output
The entry undershoot change needs to be reflected in the
test output.
Matthias Clasen [Wed, 27 Jan 2016 18:07:13 +0000 (13:07 -0500)]
window: Always disconnect signal handlers
We connect to the titlebar widgets change notification regardless
whether it is internally created or not, so don't make the signal
handler disconnection conditional on that either.
Matthias Clasen [Wed, 27 Jan 2016 17:57:58 +0000 (12:57 -0500)]
assistant: Avoid tripping template tests
We need to unset the titlebar manually before chaining up
in destroy, otherwise we trigger the template invariant
checking - GtkWindow would eventually unset it, but too late
for the invariants checking code in gtk_widget_destroy.
Allison Ryan Lortie [Wed, 27 Jan 2016 14:50:42 +0000 (15:50 +0100)]
Tweak startup-notification after the first window
Presently, Gtk will only send a startup notification completion message
for the first window that is shown. This is not good for the case of
GtkApplication, where we are expected to participate in
startup-notification for all windows.
We have avoided this problem by manually emitting the startup complete
message from after_emit in GtkApplication.
Unfortunately, this causes problems for windows that are shown with a
delay. It is also a dirty hack.
The reason for the original behaviour is simple: there is a static
boolean in gtkwindow.c which controls it. We remove this.
Instead, clear the startup notification ID stored in GDK when sending
the completion message. GtkApplication will re-set this the next time
an event comes in which needs startup-notification handling. In the
non-GtkApplication case, newly shown windows will still not send the
message, since the cookie will have been cleared.
Finally, we remove the hack from GtkApplication's after_emit.
This will probably cause some regressions in terms of lingering startup
notification messages. The correct solution here is to always use
gtk_window_present(), including when merely opening a new document (with
a new tab, for example).
https://bugzilla.gnome.org/show_bug.cgi?id=690791
Timm Bäder [Wed, 27 Jan 2016 15:32:39 +0000 (16:32 +0100)]
modelbutton: Use indicator state for button gadget if iconic
Krzesimir Nowak [Wed, 27 Jan 2016 12:25:58 +0000 (13:25 +0100)]
entry: Fix a typo
Benjamin Otte [Wed, 27 Jan 2016 13:02:44 +0000 (14:02 +0100)]
entry: Fix a case of disagreeing coordinate systems
gesture_get_current_point() returned x/y in layout coordiantes, but
callers assumed it in textarea coordinates.
Cosimo Cecchi [Wed, 27 Jan 2016 11:25:49 +0000 (12:25 +0100)]
Adwaita: fix spinbutton styling
Spinbuttons were quite broken after the latest changes. This commit
brings the style back around what it was for 3.18.
Matthias Clasen [Wed, 27 Jan 2016 04:33:06 +0000 (23:33 -0500)]
Adwaita: Add some themeing for entry undershoot
This is just copying what we do for scrolled windows.
It will need some tweaks to look good for entries.
Matthias Clasen [Wed, 27 Jan 2016 04:31:28 +0000 (23:31 -0500)]
entry: give a visual clue that content is scrolled
This has been requested long ago, and we now have the
machinery to implement it easily.
Matthias Clasen [Wed, 27 Jan 2016 03:11:54 +0000 (22:11 -0500)]
box gadget: Improve assertion message
Include the owner of the gadget in the assertion message,
to improve the chances of tracking down whats actually
going wrong here.
Matthias Clasen [Wed, 27 Jan 2016 02:27:12 +0000 (21:27 -0500)]
color button: Allow skipping the palette
Allow skipping the palette in the dialog, and going directly
to the editor. This is useful when the color button is already
part of a palette.
Matthias Clasen [Wed, 27 Jan 2016 01:57:50 +0000 (20:57 -0500)]
Fix a typo
Benjamin Otte [Wed, 27 Jan 2016 01:07:35 +0000 (02:07 +0100)]
entry: Compute selection range correctly
gtk_editable_get_selection_bounds() returns UTF-8 character offsets,
but gdk_pango_layout_get_clip_region() wants byte ranges, so convert
from one to the other.
With English, this is especially visible for passwords, which use ●
as the invisible character.
https://bugzilla.gnome.org/show_bug.cgi?id=761128
Benjamin Otte [Wed, 27 Jan 2016 00:09:51 +0000 (01:09 +0100)]
window: Deprecate gtk_window_parse_geometry()
Geometry handling in GTK is deprecated.
Benjamin Otte [Wed, 27 Jan 2016 00:07:01 +0000 (01:07 +0100)]
window: Deprecate gtk_window_set_default_geometry()
We don't support geometries anymore.
Olav Vitters [Mon, 25 Jan 2016 18:59:05 +0000 (19:59 +0100)]
Add an XSetting for primary button warps slider
This will be used by xsettings-kde to ensure GTK+3.x programs behave
consistently.
https://bugzilla.gnome.org/show_bug.cgi?id=688524
Matthias Clasen [Tue, 26 Jan 2016 17:42:04 +0000 (12:42 -0500)]
gdk: Don't translate signal names
This was a thinko - what we sometimes do for signal names is to
use I_() to intern them (to avoid a strdup), but I_() is not
currently available in gdk, so lets just skip this
microoptimization for now.
Benjamin Otte [Tue, 26 Jan 2016 15:40:33 +0000 (16:40 +0100)]
spinbutton: Don't expand the buttons
When the spinbutton grows larger, distribute horizontal size to the
entry and vertical size to the buttons.
Obviously, horizontal size only matters for horizontal spinbuttons and
vertical for vertical spinbuttons.
Benjamin Otte [Tue, 26 Jan 2016 15:36:13 +0000 (16:36 +0100)]
stack: realize windows in correct position
Benjamin Otte [Tue, 26 Jan 2016 13:13:18 +0000 (14:13 +0100)]
pathbar: Use CSS node ordering
Instead of the custom get_path_for_child() function.
Benjamin Otte [Tue, 26 Jan 2016 13:02:55 +0000 (14:02 +0100)]
stack: Turn into no-window widget
Having priv->view_window be the same as widget->window does not make
sense.
And because we need to clip the content area, we need the view window.
Benjamin Otte [Tue, 26 Jan 2016 12:23:44 +0000 (13:23 +0100)]
entry: Chain up in realize()
... instead of copy/pasting what gtk_widget_real_realize() does.
Emmanuele Bassi [Tue, 26 Jan 2016 13:32:47 +0000 (13:32 +0000)]
build: Missing build flags
The compiler and linker flags variables were renamed in the previous
commit.
Emmanuele Bassi [Tue, 26 Jan 2016 13:29:46 +0000 (13:29 +0000)]
Add more checks for the font features demo
The demo also uses PangoFt2 API.
Emmanuele Bassi [Tue, 26 Jan 2016 11:46:48 +0000 (11:46 +0000)]
demo: Link against Harfbuzz
The font features demo started calling the Harfbuzz API directly
starting from commit
9de3b24c205f1c647292a490f92f21a776b931a4. Harfbuzz
is an implicit dependency of Pango on some platforms, but it's not part
of the public dependencies; this means that we cannot expect to link to
Pango and automatically get Harfbuzz symbols to link against —
especially when things like --as-needed are in play.
This change triggered build failures on non-Unix platforms, fixed by
commit
2a9967731a3b77b102ada633c39b5a35efc80957, as well as build
failures in Continuous, with this error message:
/usr/lib/gcc/x86_64-gnomeostree-linux/4.9.3/../../../../x86_64-gnomeostree-linux/bin/ld:
font_features.o: undefined reference to symbol 'hb_tag_to_string'
//lib/libharfbuzz.so.0: error adding symbols: DSO missing from command
line
collect2: error: ld returned 1 exit status
In order to get the font features demo to build everywhere we should
take an explicit, though optional, check on Harfbuzz, and conditionally
build the font features demo with the right compiler and linker flags.
Chun-wei Fan [Tue, 26 Jan 2016 07:56:07 +0000 (15:56 +0800)]
demos: Don't build font features demo on non-UNIX
The fonts features demo now uses fontconfig APIs via PangoFT2, which makes
the code not build on non-Linux, so only include this demo in the build
on UNIX.
Matthias Clasen [Mon, 25 Jan 2016 05:40:49 +0000 (00:40 -0500)]
gtk-demo: Expand font features demo
Add more features to the list, allow selecting script/language
from the set that is supported by the font, indicate which
features are present in the font for the selected script/language,
and expand the default specimen to cover latin, cyrillic and
greek.
Matthias Clasen [Mon, 25 Jan 2016 20:59:32 +0000 (15:59 -0500)]
Update expected output of a11y tests
This is the outcome of a recent fix to default values
for a GtkPaned property.
Ben Iofel [Mon, 25 Jan 2016 20:34:08 +0000 (15:34 -0500)]
Docs: fix typo
Ray Strode [Mon, 25 Jan 2016 18:55:25 +0000 (13:55 -0500)]
wayland: __NR_memfd_create instead of SYS_memfd_create
It looks like the gnome-continuous headers haven't quite
caught up yet, so try __NR_memfd_create instead.
If that doesn't work, i'll likely just add in a fallback
code path.
Ray Strode [Mon, 25 Jan 2016 16:41:23 +0000 (11:41 -0500)]
wayland: use memfd_create instead of open in tmpdir
The tmpdir is used for a wide assortment of things, and
can easily fill up. If it fills then desktop will start
crashing with SIGBUS errors.
This commit changes the shm pool allocation code, to use
memfd_create, instead, so the shared memory files will
be anonymous and not associated with /tmp
https://bugzilla.gnome.org/show_bug.cgi?id=761095
Matthias Clasen [Mon, 25 Jan 2016 18:01:15 +0000 (13:01 -0500)]
Fix another quoting error in installed tests
Same as the previous one. I should fix my silly mistakes before
copying them around...
Matthias Clasen [Mon, 25 Jan 2016 15:56:19 +0000 (10:56 -0500)]
testsuite: Fix a quoting error
This was breaking continous testing. Oops.
Benjamin Otte [Mon, 25 Jan 2016 15:14:23 +0000 (16:14 +0100)]
spinbutton: Add a base gadget
This really messes up rendering in Adwaita because the code now renders
the spinbutton's entry node when it didn't before.
Benjamin Otte [Mon, 25 Jan 2016 14:47:40 +0000 (15:47 +0100)]
entry: Use cleverer method to get the entry node
Benjamin Otte [Mon, 25 Jan 2016 14:42:18 +0000 (15:42 +0100)]
entry: Move spinbutton size hack
If we want to do special sizing for the text, we need to do it for the
text. Otherwise paddings, borders and entyr icons will screw up
everything.
Benjamin Otte [Mon, 25 Jan 2016 14:04:04 +0000 (15:04 +0100)]
entry: Move code
This movement is necessary for future gadgetization of GtkSpinButton.
Benjamin Otte [Mon, 25 Jan 2016 00:33:28 +0000 (01:33 +0100)]
entry: Redo textarea handling
(1) Keep priv->text_allocation for the area used by the text
(2) Compute all text coordinates with the help of priv->text_allocation
As a side effect the get_text_area_size and get_frame_size vfuncs are
now unused. If we wanted them back, they should get a single use durig
size_allocate() and then their results should be stored for further
processing.
Benjamin Otte [Sun, 24 Jan 2016 23:07:23 +0000 (00:07 +0100)]
entry: Use gtk_render_layout()
It's funny that nobody has yet noticed that we don't draw the text using
the theme.
We had no text-shadow in entries!!!!!
Benjamin Otte [Sun, 24 Jan 2016 23:00:41 +0000 (00:00 +0100)]
entry: Draw selection the same way as GtkLabel
Benjamin Otte [Sun, 24 Jan 2016 22:52:49 +0000 (23:52 +0100)]
entry: Fold function into only caller
Benjamin Otte [Sun, 24 Jan 2016 22:46:12 +0000 (23:46 +0100)]
entry: Don't allow different text over progress
This complicates refactorings, so remove that feature. It's not used
anywhere and doesn't play well with nodes the way it's implemented.
If we want it back, we can add it back later.
Piotr Drąg [Mon, 25 Jan 2016 02:07:00 +0000 (03:07 +0100)]
Updated POTFILES.skip
Matthias Clasen [Mon, 25 Jan 2016 01:10:45 +0000 (20:10 -0500)]
One forgotten rename
Rename app-menu.ui to menus.ui in EXTRA_DIST too.
Matthias Clasen [Mon, 25 Jan 2016 01:01:33 +0000 (20:01 -0500)]
icon browser: Load app menu automagically
No need to manually create a builder and pass the resource to it,
just follow the conventions and have GtkApplication do it for us.
Matthias Clasen [Sun, 24 Jan 2016 21:29:40 +0000 (16:29 -0500)]
path bar: Don't hide arrows
Changing the visibility of child widgets in size-allocate does
not work well with out current allocation and layout machinery.
To avoid the visual fallout, just keep the arrow buttons visible
and only change their sensitivity.
https://bugzilla.gnome.org/show_bug.cgi?id=754868
Matthias Clasen [Sun, 24 Jan 2016 15:58:32 +0000 (10:58 -0500)]
inspector: Show CSS pseudoclass names for node state
This is a better fit for showing CSS node state, since one
commonly looks at this information to devise CSS selectors,
where the CSS names are needed.
Matthias Clasen [Sun, 24 Jan 2016 15:56:23 +0000 (10:56 -0500)]
css: Privately export a function to get pseudoclass names
This function centralizes the mapping for widget states
to pseudoclass names in one place, for easier maintenance.
Timm Bäder [Sun, 24 Jan 2016 09:48:48 +0000 (10:48 +0100)]
gdkdnd: Fix signal docs
Signals all have at least one parameter, the object that received the
signal.
Timm Bäder [Sun, 24 Jan 2016 07:01:36 +0000 (08:01 +0100)]
Inspector: Plug a few leaks
Matthias Clasen [Sun, 24 Jan 2016 05:22:31 +0000 (00:22 -0500)]
paned: Initialize max_position
As pointed out in
https://bugzilla.gnome.org/show_bug.cgi?id=608865
max_position should be set to G_MAXINT.
Matthias Clasen [Sun, 24 Jan 2016 03:30:59 +0000 (22:30 -0500)]
widget-factory: Add a color editor example
Matthias Clasen [Sat, 23 Jan 2016 23:26:09 +0000 (18:26 -0500)]
color editor: Ensure spin buttons are wide enough
If we don't explicitly ask for 2 chars to be visible, we might
get less, and it doesn't look great.
Sebastien Lafargue [Fri, 22 Jan 2016 21:25:29 +0000 (22:25 +0100)]
GtkColorChooser: make set_rgba work in editor mode
When using the color chooser in editor mode,
gtk_color_chooser_set_rgba need to be propaged
to the editor
https://bugzilla.gnome.org/show_bug.cgi?id=761005
Matthias Clasen [Sat, 23 Jan 2016 23:00:57 +0000 (18:00 -0500)]
Add a color editor test
This tests using a color editor embedded in other ui.
Matthias Clasen [Sat, 23 Jan 2016 20:52:09 +0000 (15:52 -0500)]
widget-factory: Ensure spinbuttons are sized right
GtkEntry now respects the padding set by the theme, so we
need to ask the spin buttons to show 2 characters, otherwise
they will be clipped.
Matthias Clasen [Sat, 23 Jan 2016 06:51:48 +0000 (01:51 -0500)]
container: Clarify the forall/foreach documentation
Try to make the difference between these two functions
clearer.
Phil Clayton [Sun, 13 Sep 2015 10:58:37 +0000 (11:58 +0100)]
Add missing (out) annotation to gdk_rgba_parse
https://bugzilla.gnome.org/show_bug.cgi?id=754990
Mario Blättermann [Sat, 23 Jan 2016 17:21:45 +0000 (18:21 +0100)]
Updated German translation
Ignacio Casal Quinteiro [Sat, 23 Jan 2016 22:35:59 +0000 (14:35 -0800)]
win32 geometry: reorganize code to avoid prototypes
Ignacio Casal Quinteiro [Sat, 23 Jan 2016 22:32:32 +0000 (14:32 -0800)]
win32 device manager: reorganize code to avoid prototypes
Ignacio Casal Quinteiro [Sat, 23 Jan 2016 22:27:02 +0000 (14:27 -0800)]
win32 virtual device: reorganize code to avoid prototypes
Ignacio Casal Quinteiro [Sat, 23 Jan 2016 22:26:04 +0000 (14:26 -0800)]
win32 device: reorganize code to avoid prototypes
Ignacio Casal Quinteiro [Sat, 23 Jan 2016 22:24:39 +0000 (14:24 -0800)]
wintab: reorganize code to avoid prototypes
Timm Bäder [Sat, 23 Jan 2016 09:30:53 +0000 (10:30 +0100)]
overlay: Don't access priv pointer before typ check
Matthias Clasen [Sat, 23 Jan 2016 04:30:36 +0000 (23:30 -0500)]
levelbar: Document CSS node ordering
Matthias Clasen [Sat, 23 Jan 2016 04:14:09 +0000 (23:14 -0500)]
notebook: Document CSS node order
Matthias Clasen [Sat, 23 Jan 2016 03:49:01 +0000 (22:49 -0500)]
box gadget: Don't allow shifting too far
We don't want to let baseline adjustment shift the child
out of the original allocation. This is purely a sanity
measure - in practice, the baseline should always be bigger
than the child_baseline.
Matthias Clasen [Sat, 23 Jan 2016 03:47:47 +0000 (22:47 -0500)]
box gadget: Fix baseline calculation
Careful comparison with GtkBox revealed that the box gadget
code was confusing up and down, and thus shifted baselines
vertically.
Matthias Clasen [Sat, 23 Jan 2016 03:24:46 +0000 (22:24 -0500)]
inspector: Show baseline value
The misc info page already shows the allocation and clip areas,
the baseline fits right in.
Matthias Clasen [Sat, 23 Jan 2016 03:08:38 +0000 (22:08 -0500)]
box gadget: Fix baseline allocation for widgets
We were adjusting the allocation to line up baselines before
calling gtk_widget_size_allocate_with_baseline, but that function
is doing this alignment internally anyway and expects to be given
a 'fill' allocation.
Move the allocation adjustment code down into
gtk_box_gadget_allocate_child where it only affects child gadgets,
not child widgets.
Matthias Clasen [Sat, 23 Jan 2016 02:34:54 +0000 (21:34 -0500)]
gadget: Add baseline debugging
This should help to understand whats going wrong with
baselines.
Matthias Clasen [Sat, 23 Jan 2016 00:23:54 +0000 (19:23 -0500)]
Adwaita: Move .ssd class around
Make the theme follow our documentation for the various .csd and
.ssd style classes: They all go on the window node. For now, just
add the new selector; the old one will be removed when mutter has
been updated.
https://bugzilla.gnome.org/show_bug.cgi?id=760714
Matthias Clasen [Fri, 22 Jan 2016 22:10:41 +0000 (17:10 -0500)]
box gadget: Fix height-for-width handling
When measuring children while distributing a given height,
we must measure them for the given width that goes with
the height. Otherwise, things will go wrong if some of the
children do actual width-for-height. This was showing up
as misaligned images in anaconda.
https://bugzilla.gnome.org/show_bug.cgi?id=760967
Carlos Garnacho [Thu, 21 Jan 2016 14:22:42 +0000 (15:22 +0100)]
GtkTreeView: Update prelight row after revalidating rows
It might have changed (eg. after a row being expanded, and the child
rows revalidated), so just update it here based on the last pointer
position.
Based on a patch by Maxim Reznik <reznikmm@gmail.com>
https://bugzilla.gnome.org/show_bug.cgi?id=760891
Carlos Garnacho [Thu, 21 Jan 2016 14:21:16 +0000 (15:21 +0100)]
GtkTreeView: Avoid device query on collapsing rows
The pointer position is queried to properly trigger the prelight
updates on the new row below it. We store the last coordinates
though, and track crossing events to unset these, so it's safe
to just update_prelight() here on these.
Benjamin Otte [Fri, 22 Jan 2016 19:39:14 +0000 (20:39 +0100)]
shortcuts: Don't forall() widgets that aren't children
Benjamin Otte [Thu, 21 Jan 2016 15:53:25 +0000 (16:53 +0100)]
widget: Improve gtk_cairo_should_draw_window()
Check that non-native window are indeed children of the event window and
only then confirm that they should be drawn.
Fixes Glade thinking that it's okay to have the draw function do
different things depending on what window to draw. (This should really
be fixed in Glade.)
Benjamin Otte [Thu, 21 Jan 2016 12:31:14 +0000 (13:31 +0100)]
cssstyleproperty: Don't allow assigning min-width/min-height
Benjamin Otte [Thu, 21 Jan 2016 01:19:04 +0000 (02:19 +0100)]
container: Turn assert into return_if_fail()
There's no reason to insta-crash when something goes wrong. Just don't
do anything stupid.
Also, remove the SPCIAL_CONTAINER() exception. Every case where special
containers needed this, it is wrong and made containers draw children
multiple times.
Benjamin Otte [Thu, 21 Jan 2016 00:44:47 +0000 (01:44 +0100)]
stack: Clip animation drawing
We use cairo_paint() when doing that and we don't want to overdraw the
content area.
Benjamin Otte [Thu, 21 Jan 2016 00:22:41 +0000 (01:22 +0100)]
spinbutton: Make the entry gadget be the entry's gadget
This is a big and somewhat evil hack: We replace the entry's gadget's
node with the spinbutton's entry node.
Lapo Calamandrei [Fri, 22 Jan 2016 12:29:35 +0000 (13:29 +0100)]
Adwaita: fix checks and radios on selected row
Matthias Clasen [Fri, 22 Jan 2016 04:39:16 +0000 (23:39 -0500)]
text view: Don't leak cairo transforms
The text view draw function was leaving its cairo context
with a transformation after drawing to all the border windows,
which lead mis-drawing in gitg. Avoid this by moving the
gtk_cairo_transform_to_window call inside the existing
cairo_save/restore calls.
https://bugzilla.gnome.org/show_bug.cgi?id=760942
Matthias Clasen [Fri, 22 Jan 2016 03:53:06 +0000 (22:53 -0500)]
wayland: Don't hardcode /tmp
As pointed out in https://bugzilla.gnome.org/show_bug.cgi?id=760964,
we should use the GLib facilities for determining the preferred
location for temporary files.
Mario Blättermann [Thu, 21 Jan 2016 16:48:25 +0000 (17:48 +0100)]
Updated German translation
Carlos Garcia Campos [Thu, 21 Jan 2016 13:36:42 +0000 (14:36 +0100)]
cssstyleproperty: Make it possible to query CSS opacity with gtk_style_context_get()
Add a query implementation to opacity property. Also fix the assert in
gtk_css_style_property_register() to allow registering properties with
query but without assign function.
https://bugzilla.gnome.org/show_bug.cgi?id=760933
Lapo Calamandrei [Thu, 21 Jan 2016 02:12:10 +0000 (03:12 +0100)]
Adwaita: first css check/radio implementation
draw checks and radios with css istead of relying on png assets.
Matthias Clasen [Thu, 21 Jan 2016 13:42:55 +0000 (08:42 -0500)]
css: Fix refcounting thinko in -gtk-recolor
gtk_css_style_get_value does not return a reference.
So don't unref the return value without taking one.
Matthias Clasen [Thu, 21 Jan 2016 13:20:06 +0000 (08:20 -0500)]
Handle svg image assets in Adwaita
The build glue for collecting all the assets in Adwaita as
resources was assuming that they are all pngs, and tried to
preprocess them into embedded GdkPixbufs.
Fix it to leave svgs unmolested, so they can be recolored
at runtime.
Matthias Clasen [Thu, 21 Jan 2016 13:18:46 +0000 (08:18 -0500)]
css: Handle image loading errors graciously
If we fail to load the image for a -gtk-recolor() expression,
fall back to using the image-missing icon instead of crashing,
and include more details in the warning message.
Matthias Clasen [Thu, 21 Jan 2016 13:18:28 +0000 (08:18 -0500)]
Forgotten file
Matthias Clasen [Thu, 21 Jan 2016 03:53:47 +0000 (22:53 -0500)]
Adwaita: Set a text color for lists
Symbolic icons in list boxes came out white-on-white in the
inspector. Set an explicit text color to prevent that.
Matthias Clasen [Thu, 21 Jan 2016 03:52:05 +0000 (22:52 -0500)]
inspector: Show full allocation + clip
Previously, we were only showing the size of the allocation
and clip area. But there is no good reason to hide the position
of these rectangles, so add them, in the traditional format
of X geometry strings: wxh+x+y
Matthias Clasen [Thu, 21 Jan 2016 03:51:25 +0000 (22:51 -0500)]
inspector: Refine typography
Use thin spaces in the formatting of monitor sizes.
Matthias Clasen [Thu, 21 Jan 2016 00:33:16 +0000 (19:33 -0500)]
Fix GtkShortcutsWindow crash
GtkShortcutsWindow is among the 'cheating' containers that iterate
over indirect children in forall, and this is now triggering
an assertion in gtk_container_propagate_draw.
For now, just exclude the cheating containers from the assertion.
Eventually, this needs a better solution.
Ray Strode [Wed, 20 Jan 2016 17:35:44 +0000 (12:35 -0500)]
wayland: don't pass in width and height to create_shm_pool
create_shm_pool doesn't need the width or height, it just needs
the total size. By passing it in, we're requiring it to redo
stride calculation unnecessarily.
This commit drops the width and height parameters and makes the
function just take the total size directly.
https://bugzilla.gnome.org/show_bug.cgi?id=760897
Ray Strode [Wed, 20 Jan 2016 16:40:34 +0000 (11:40 -0500)]
wayland: clean up stride calculation when creating shm surface
Right now, we assume the stride for the image surface needs to
be 4 byte aligned. This is, in fact, true, but it's better to
ask cairo for the alignment requirement directly rather than
assume we know the alignment rules.
This commit changes the code to use cairo_format_stride_for_width
to calculate a suitable rowstride for pixman.
https://bugzilla.gnome.org/show_bug.cgi?id=760897