gtk4.git
7 years agoframeclockidle: Remove nonexisting functions from header
Benjamin Otte [Wed, 28 Mar 2018 22:04:47 +0000 (00:04 +0200)]
frameclockidle: Remove nonexisting functions from header

7 years agovulkan: Reserve 4 images in the swapchain by default
Benjamin Otte [Thu, 29 Mar 2018 02:56:07 +0000 (04:56 +0200)]
vulkan: Reserve 4 images in the swapchain by default

With the previous approach we would spend most of the time waiting for
the swapchain to be filled again because it seems the compositor takes
care of 2 images at once from time to time.

This is not visible in profiles because waiting for a frame is a
read/poll/whatever operation that does not take CPU. It's only
noticeable because the app becomes less responsive.

7 years agoRename gtkkineticscrolling.h to gtkkineticscrollingprivate.h
Timm Bäder [Wed, 28 Mar 2018 14:47:54 +0000 (16:47 +0200)]
Rename gtkkineticscrolling.h to gtkkineticscrollingprivate.h

To match the naming scheme we use for all private header files.

7 years agoexpander: Update css docs
Timm Bäder [Wed, 28 Mar 2018 14:42:53 +0000 (16:42 +0200)]
expander: Update css docs

7 years agotestgtk: don't set a focus adjustment on a box
Timm Bäder [Wed, 28 Mar 2018 14:24:03 +0000 (16:24 +0200)]
testgtk: don't set a focus adjustment on a box

That will make the focus adjustment code in gtkcontainer.c scroll to the
wrong coordinates.

7 years agoRevert "container: Fix scroll adjustment coordinates"
Timm Bäder [Wed, 28 Mar 2018 11:05:01 +0000 (13:05 +0200)]
Revert "container: Fix scroll adjustment coordinates"

This reverts commit 7351848c75e8d6efee154fe77cbd90056ed993b5.

7 years agoexpander: Remove priv pointer
Timm Bäder [Tue, 27 Mar 2018 18:35:21 +0000 (20:35 +0200)]
expander: Remove priv pointer

7 years agobuild: Defer defining HAVE_PANGOFT and HAVE_HARFBUZZ
Chun-wei Fan [Wed, 28 Mar 2018 03:32:53 +0000 (11:32 +0800)]
build: Defer defining HAVE_PANGOFT and HAVE_HARFBUZZ

We forgot to account for the case where we lookup for HarfBuzz manually
under Visual Studio builds, so only set HAVE_HARFBUZZ (and thus
HAVE_PANGOFT, since PangoFT2 depends on HarfBuzz) after we did the
fallback check for HarfBuzz.

Also, check for hb.h instead of harfbuzz/hb.h to be inline with the
pkg-config case, as the sources also include the HarfBuzz header by
using #include <hb.h>, not #include <harfbuzz/hb.h>

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agogsk/gl: Include cairo.h consistently
Chun-wei Fan [Wed, 28 Mar 2018 03:23:43 +0000 (11:23 +0800)]
gsk/gl: Include cairo.h consistently

Follow the other sources that include cairo.h, by just doing #include
<cairo.h>, not <cairo/cairo.h>.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agogtkimcontextime: fix to compile again
Руслан Ижбулатов [Thu, 8 Mar 2018 00:26:39 +0000 (00:26 +0000)]
gtkimcontextime: fix to compile again

This makes the code compile again, though obviously there have been
some substantial changes in how IM contexts work, so it's possible
that IME IM context doesn't work now.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoGDK W32: remove unused client_message
Руслан Ижбулатов [Sat, 24 Mar 2018 13:03:17 +0000 (13:03 +0000)]
GDK W32: remove unused client_message

This seems to be a leftover from API that was removed in
commit c332ac207ab4cac8f4c32aa85cdf6406157340af back in 2011.
The code, as it is now, does not even make sense.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoGDK W32: drop cursor-related GdkWin32Display functions
Руслан Ижбулатов [Sat, 24 Mar 2018 12:58:05 +0000 (12:58 +0000)]
GDK W32: drop cursor-related GdkWin32Display functions

These functions went away in commit 77bab4e027c06dff7726d96fa4e477bc11b23835

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoGDK W32: remove the use of GDK_SURFACE_STATE
Руслан Ижбулатов [Sat, 24 Mar 2018 12:56:21 +0000 (12:56 +0000)]
GDK W32: remove the use of GDK_SURFACE_STATE

GDK_WINDOW_STATE (or, after rename, GDK_SURFACE_STATE) got
removed from GDK in commit 11a946df39f29b3cc0ce902c0c3b49abede1ead4

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoGDK W32: move GdkWin32MonitorDpiType to a different header
Руслан Ижбулатов [Sat, 24 Mar 2018 12:52:25 +0000 (12:52 +0000)]
GDK W32: move GdkWin32MonitorDpiType to a different header

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoCheck for freetype2 version when PangoFT is used
Руслан Ижбулатов [Thu, 11 Jan 2018 17:49:43 +0000 (17:49 +0000)]
Check for freetype2 version when PangoFT is used

Check for freetype2 version, because pangoft works with any version
(pangoft availability does not indicate that ft2 is new enough), unlike
GTK.

On Windows, since pangoft is optional, we check for the presence of
freetype2 .pc file first after finding that we have pangoft, and then
check for FT_Get_Var_Design_Coordinates() manually by looking for the
freetype headers and .lib first, and then looking for the presence of
that symbol, since freetype2's Visual Studio build system does not
generate a .pc file for us.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoGDK W32: _gdk_surface_invalidate_{for_expose,region}
Руслан Ижбулатов [Sat, 24 Mar 2018 19:33:41 +0000 (19:33 +0000)]
GDK W32: _gdk_surface_invalidate_{for_expose,region}

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoGDK W32: drop the use of gdk_keymap_get_default()
Руслан Ижбулатов [Thu, 11 Jan 2018 17:48:25 +0000 (17:48 +0000)]
GDK W32: drop the use of gdk_keymap_get_default()

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoOnly use gtk_print_backends_init() on UNIX
Руслан Ижбулатов [Sat, 24 Mar 2018 12:59:13 +0000 (12:59 +0000)]
Only use gtk_print_backends_init() on UNIX

It's from gtkprintbackend.c, which is in gtk_unix_print_sources
source list and thus only available on os_unix only.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoAlternative printbackends subdir for non-UNIX OSes
Руслан Ижбулатов [Sat, 3 Mar 2018 21:57:00 +0000 (21:57 +0000)]
Alternative printbackends subdir for non-UNIX OSes

The main buildscript expects 'print_backends' list to be defined.
Since printbackends is os_unix-only, we need to define this list
ourselves for other OSes.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoMake wayland bits in meson.build conditional on wayland use
Руслан Ижбулатов [Sat, 3 Mar 2018 21:55:15 +0000 (21:55 +0000)]
Make wayland bits in meson.build conditional on wayland use

Otherwise the build won't configure due to its inability to find
wayland-scanner program on systems where no such program is availble.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agofont chooser: Make levels flags
Matthias Clasen [Tue, 27 Mar 2018 20:10:19 +0000 (16:10 -0400)]
font chooser: Make levels flags

Turn the GtkFontChooserLevel field into flags, and
add flags for OpenType variations and features. The
motivation for this is to make font-features in the UI
opt-in, since applications need to support them by
applying the pango attribute.

7 years agofont chooser widget: Expose tweak-action as property
Matthias Clasen [Tue, 27 Mar 2018 18:50:11 +0000 (14:50 -0400)]
font chooser widget: Expose tweak-action as property

This is necessary for applications that want to embed
the font chooser widget outside of GtkFontChooserDialog,
in order to access the tweak page.

7 years agofont button: notify when ::font-features changes
Matthias Clasen [Tue, 27 Mar 2018 18:47:36 +0000 (14:47 -0400)]
font button: notify when ::font-features changes

Lack of notification was causing the font-features to not
show up in the inspector.

7 years agospinbutton: spinbuttons aren't entries anymore
Timm Bäder [Tue, 27 Mar 2018 17:09:15 +0000 (19:09 +0200)]
spinbutton: spinbuttons aren't entries anymore

Fix a code sample to reflect that.

7 years agospinbutton: Remove queue_resize call from realize
Timm Bäder [Tue, 27 Mar 2018 17:08:50 +0000 (19:08 +0200)]
spinbutton: Remove queue_resize call from realize

No reason for this to be here.

7 years agospinbutton: Fix a copy/paste bug in the docs
Timm Bäder [Tue, 27 Mar 2018 16:25:03 +0000 (18:25 +0200)]
spinbutton: Fix a copy/paste bug in the docs

7 years agogl renderer: Fix some crashes
Timm Bäder [Tue, 27 Mar 2018 11:24:14 +0000 (13:24 +0200)]
gl renderer: Fix some crashes

When the first op is a modelview or projection change.

7 years agoaccellabel: Stop saying it's a GtkLabel
Timm Bäder [Tue, 27 Mar 2018 09:04:23 +0000 (11:04 +0200)]
accellabel: Stop saying it's a GtkLabel

And clean up the includes accordingly.

7 years agoaccellabel: Remove some unnused class fields
Timm Bäder [Tue, 27 Mar 2018 08:57:36 +0000 (10:57 +0200)]
accellabel: Remove some unnused class fields

7 years agoaccellabel: Remove priv pointer
Timm Bäder [Tue, 27 Mar 2018 08:53:38 +0000 (10:53 +0200)]
accellabel: Remove priv pointer

7 years agoci: Move flatpak building to a script
Emmanuele Bassi [Tue, 27 Mar 2018 11:45:20 +0000 (12:45 +0100)]
ci: Move flatpak building to a script

The YAML is getting hard to modify, so let's use a simple script like we
do for the other CI jobs.

7 years agoDo not test for linker flags
Emmanuele Bassi [Mon, 26 Mar 2018 10:48:30 +0000 (11:48 +0100)]
Do not test for linker flags

Meson warns when doing that, as it's not really portable.

Since we're using platform-specific linker flags on Darwin, we can also
do the same on Linux; the syntax is GCC-specific, so we're going to need
Clang users to test it.

7 years agoMerge branch 'krnowak/leak-fixes' into 'master'
Benjamin Otte [Tue, 27 Mar 2018 11:37:24 +0000 (11:37 +0000)]
Merge branch 'krnowak/leak-fixes' into 'master'

Fix some leaks

See merge request GNOME/gtk!87

7 years agoFix some leaks
Krzesimir Nowak [Tue, 27 Mar 2018 11:14:41 +0000 (13:14 +0200)]
Fix some leaks

7 years agogl renderer: Rename everything called window to surface
Timm Bäder [Mon, 26 Mar 2018 19:13:52 +0000 (21:13 +0200)]
gl renderer: Rename everything called window to surface

7 years agoentry: Remove useless queue_draw call
Timm Bäder [Mon, 26 Mar 2018 17:30:40 +0000 (19:30 +0200)]
entry: Remove useless queue_draw call

This is already taken care of by the add_class/remove_class.

7 years agosnapshot: Generate better names for offset nodes
Benjamin Otte [Sat, 24 Mar 2018 22:27:43 +0000 (23:27 +0100)]
snapshot: Generate better names for offset nodes

7 years agorendernode: Fix serializing
Benjamin Otte [Mon, 26 Mar 2018 17:41:38 +0000 (19:41 +0200)]
rendernode: Fix serializing

Adding the offset node broke serialization in 2 ways:

1. We store the enum value in the node, so make sure to not change it
   for existing values
2. The offset node was missing in the deserialization lookup table

7 years agoiconhelper: Apply scale to paintable
Benjamin Otte [Mon, 26 Mar 2018 16:46:22 +0000 (18:46 +0200)]
iconhelper: Apply scale to paintable

Instead of fiddling around with scale in the iconhelper (and getting it
wrong), create a GtkScaler around the paintable that takes care of the
scaling.

7 years agodnd: Remove gtk_drag_set_icon_surface()
Benjamin Otte [Mon, 26 Mar 2018 16:32:38 +0000 (18:32 +0200)]
dnd: Remove gtk_drag_set_icon_surface()

All users are gone. Use gtk_drag_set_icon_paintable() instead.

7 years agotests: Don't use gtk_drag_set_icon_surface()
Benjamin Otte [Mon, 26 Mar 2018 16:30:34 +0000 (18:30 +0200)]
tests: Don't use gtk_drag_set_icon_surface()

7 years agotextview: Port dnd icon to be a GdkPaintable
Benjamin Otte [Mon, 26 Mar 2018 04:49:55 +0000 (06:49 +0200)]
textview: Port dnd icon to be a GdkPaintable

7 years agocolorswatch: Port drag icon to paintables
Benjamin Otte [Mon, 26 Mar 2018 04:29:06 +0000 (06:29 +0200)]
colorswatch: Port drag icon to paintables

7 years agocolorbutton: Port dnd icon to paintable
Benjamin Otte [Mon, 26 Mar 2018 04:26:34 +0000 (06:26 +0200)]
colorbutton: Port dnd icon to paintable

Patch is untested, because colorbuttons can't do DND at the moment.

7 years agotreeview: Turn drag icon into paintable
Benjamin Otte [Mon, 26 Mar 2018 04:20:56 +0000 (06:20 +0200)]
treeview: Turn drag icon into paintable

7 years agoiconview: Port DND icons to snapshot paintables
Benjamin Otte [Mon, 26 Mar 2018 03:21:18 +0000 (05:21 +0200)]
iconview: Port DND icons to snapshot paintables

7 years agotests: Remove references to long-deleted property
Benjamin Otte [Mon, 26 Mar 2018 03:09:46 +0000 (05:09 +0200)]
tests: Remove references to long-deleted property

7 years agotextutil: Make gtk_text_util_create_drag_icon() return a paintable
Benjamin Otte [Mon, 26 Mar 2018 02:38:54 +0000 (04:38 +0200)]
textutil: Make gtk_text_util_create_drag_icon() return a paintable

7 years agosnapshot: Add gtk_snapshot_to_paintable()
Benjamin Otte [Mon, 26 Mar 2018 02:37:01 +0000 (04:37 +0200)]
snapshot: Add gtk_snapshot_to_paintable()

This is intended for DND icons

7 years agosnapshot: Export gtk_snapshot_append_layout()
Benjamin Otte [Mon, 26 Mar 2018 02:23:11 +0000 (04:23 +0200)]
snapshot: Export gtk_snapshot_append_layout()

This is the equivalent snapshot function to pango_cairo_show_layout().

Not to be confused with gtk_snapshot_render_layout(), which is the
equivalent to gtk_render_layout().

7 years agosnapshot: Make gtk_snapshot_append_node() take care of offset
Benjamin Otte [Tue, 20 Mar 2018 03:33:58 +0000 (04:33 +0100)]
snapshot: Make gtk_snapshot_append_node() take care of offset

Push an offset node when append_node is called. That resets the offset.

7 years agorendernode: Add an offset node
Benjamin Otte [Tue, 20 Mar 2018 01:46:57 +0000 (02:46 +0100)]
rendernode: Add an offset node

This is a special case of the transform node that does a 2D translation.

The implementation in the Vulkan and GL renderers is crude and just does
the same as the transform node.

Nothing uses that node yet.

7 years agoUpdate Polish translation
Piotr Drąg [Sun, 25 Mar 2018 23:42:57 +0000 (01:42 +0200)]
Update Polish translation

7 years agocolorscale: Draw gradient using render nodes
Benjamin Otte [Sun, 25 Mar 2018 22:29:52 +0000 (00:29 +0200)]
colorscale: Draw gradient using render nodes

Don't use Cairo when it's not needed.

7 years agocolorscale: Draw hue background as a texture
Benjamin Otte [Sun, 25 Mar 2018 22:12:20 +0000 (00:12 +0200)]
colorscale: Draw hue background as a texture

No need to use Cairo here.

7 years agocolorplane: Create background image as a texture
Benjamin Otte [Sun, 25 Mar 2018 21:56:09 +0000 (23:56 +0200)]
colorplane: Create background image as a texture

Don't create it using Cairo.

7 years agoMerge branch 'bz773274' into 'master'
Matthias Clasen [Sun, 25 Mar 2018 20:44:44 +0000 (20:44 +0000)]
Merge branch 'bz773274' into 'master'

wayland: Don't paint if the window is unmapped

See merge request GNOME/gtk!32

7 years agocolorscale: Don't free surface data before surface
Benjamin Otte [Sun, 25 Mar 2018 20:37:09 +0000 (22:37 +0200)]
colorscale: Don't free surface data before surface

When drawing onto a recording surface, source surfaces get cached.

But if we g_free() the surface data after we're done, that cache is
gonna point at invalid data...

7 years agoGdkGLContext: Add a GL debug message callback
Timm Bäder [Sun, 25 Mar 2018 13:28:40 +0000 (15:28 +0200)]
GdkGLContext: Add a GL debug message callback

If G_ENABLE_CONSISTENCY_CHECKS is defined (i.e. if our buildtype is
'debug'), add a opengl debug callback that prints all debug messages
with a severity higher than SEVERITY_NOTIFICATION as a warning to the
console.

7 years agogl renderer: Skip all 0-sized nodes
Timm Bäder [Sun, 25 Mar 2018 09:47:04 +0000 (11:47 +0200)]
gl renderer: Skip all 0-sized nodes

These will cause problems later on.

7 years agogl renderer: Remove cairo node special case
Timm Bäder [Sun, 25 Mar 2018 09:46:48 +0000 (11:46 +0200)]
gl renderer: Remove cairo node special case

7 years agoci: Remove --werror
Emmanuele Bassi [Sun, 25 Mar 2018 00:31:36 +0000 (00:31 +0000)]
ci: Remove --werror

Turns out that GCC errors out when building the GLib test suite, as it
now checks for overflows in allocator functions, and we're testing for
those.

This would not be an issue for GTK, but since we're building GLib as a
subproject, we get failures for those as well.

Until we can find out how to disable errors for subprojects, or fix the
GLib test suite not to trip up warnings in GCC, we're going to live
without compiler warnings treated as errors for a while.

7 years agoci: Use the appropriate argument for errors
Emmanuele Bassi [Sun, 25 Mar 2018 00:00:02 +0000 (00:00 +0000)]
ci: Use the appropriate argument for errors

The option to enable fatal warnings in Meson is `--werror`.

7 years agoMenu: Don't leak priv->motion_controller
Daniel Boles [Sat, 24 Mar 2018 21:57:42 +0000 (21:57 +0000)]
Menu: Don't leak priv->motion_controller

7 years agoci: build with -Werror
Benjamin Otte [Sat, 24 Mar 2018 21:27:28 +0000 (22:27 +0100)]
ci: build with -Werror

7 years agoci: Disable ccache
Emmanuele Bassi [Sat, 24 Mar 2018 21:17:34 +0000 (21:17 +0000)]
ci: Disable ccache

Forcibly remove any ccache use, even if it's installed.

7 years agodocs: Style fixes for the contribution guidelines
Emmanuele Bassi [Wed, 21 Mar 2018 01:05:29 +0000 (01:05 +0000)]
docs: Style fixes for the contribution guidelines

7 years agosnapshot: Remove renderer
Benjamin Otte [Sat, 24 Mar 2018 20:41:53 +0000 (21:41 +0100)]
snapshot: Remove renderer

Now that there's no longer a need to keep the renderer around for Cairo
rendering, don't do that then.

7 years agorendernode: Remove renderer argument
Benjamin Otte [Sat, 24 Mar 2018 20:33:08 +0000 (21:33 +0100)]
rendernode: Remove renderer argument

It's not needed to create Cairo nodes anymore.

7 years agorendernode: Create Cairo surfaces as recording surfaces
Benjamin Otte [Sat, 24 Mar 2018 19:42:42 +0000 (20:42 +0100)]
rendernode: Create Cairo surfaces as recording surfaces

This way, we can postpone the actual rendeing of the node until the
renderer. This allows the renderer to choose the right scale to
render at, so it can decide to use 2x scale for hidpi on its own.

Last but not least, it makes all nodes independent of the context they
are created in, because they do not need to know at snapshot time what
they will ultimately be rendered into.

7 years agovulkan: Remove special case for Cairo surfaces
Benjamin Otte [Sat, 24 Mar 2018 15:41:59 +0000 (16:41 +0100)]
vulkan: Remove special case for Cairo surfaces

This is in preparation for the next commits which switch Cairo nodes to
recording surfaces.

7 years agoMerge branch 'lrn/bug-773299' into 'master'
LRN [Sat, 24 Mar 2018 11:08:36 +0000 (11:08 +0000)]
Merge branch 'lrn/bug-773299' into 'master'

bug 773299 2nd

See merge request GNOME/gtk!79

7 years agoGDK W32: Adapt DnD event putting to recent changes
Руслан Ижбулатов [Sat, 24 Mar 2018 10:58:11 +0000 (10:58 +0000)]
GDK W32: Adapt DnD event putting to recent changes

Set the display for each event that we put.
Also reorganize the dnd_event_put() function a bit, giving it a surface
directly instead of setting it by implication.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoMerge branch 'lrn/bug-773299' into 'master'
LRN [Sat, 24 Mar 2018 10:46:59 +0000 (10:46 +0000)]
Merge branch 'lrn/bug-773299' into 'master'

bug 773299

See merge request GNOME/gtk!78

7 years agoGDK W32: Don't check dest_surface for != NULL on button events
Руслан Ижбулатов [Sat, 24 Mar 2018 10:26:51 +0000 (10:26 +0000)]
GDK W32: Don't check dest_surface for != NULL on button events

dest_surface is going to always be NULL for source contexts.
Previously we used to put the root window there to pass this check,
but root windows are gone (and root surfaces never existed to begin
with), so we have to adapt.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoGDK W32: Init display scale to the global Windows scale, not 1
Руслан Ижбулатов [Sat, 24 Mar 2018 10:05:34 +0000 (10:05 +0000)]
GDK W32: Init display scale to the global Windows scale, not 1

This affects gdk_device_query_state() for the virtual device. It has
no window, and is forced to query the display itself, and display
defaults its scale to 1 even for HiDPI desktops. Use the same
"query scale of a NULL monitor" trick that we use in other places
to get the global desktop scale.

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoGDK W32: adapt to the recent changes in GdkEvent
Руслан Ижбулатов [Sat, 24 Mar 2018 10:04:38 +0000 (10:04 +0000)]
GDK W32: adapt to the recent changes in GdkEvent

https://bugzilla.gnome.org/show_bug.cgi?id=773299

7 years agoUpdate zh_CN po and po-properties translations
Dz Chen [Fri, 23 Mar 2018 15:55:05 +0000 (23:55 +0800)]
Update zh_CN po and po-properties translations

7 years agoAdwaita: Regenerate CSS from SASS
Daniel Boles [Thu, 22 Mar 2018 20:58:31 +0000 (20:58 +0000)]
Adwaita: Regenerate CSS from SASS

https://gitlab.gnome.org/GNOME/gtk/merge_requests/66 introduced spurious
red/green fore/background.

Also, the GtkVideo stuff didn't get regenerated in.

7 years agoMerge branch 'toolbar-spacing-fix' into 'master'
Daniel Boles [Thu, 22 Mar 2018 20:54:01 +0000 (20:54 +0000)]
Merge branch 'toolbar-spacing-fix' into 'master'

theme: Be more selective when adding margins to toolbar items

See merge request GNOME/gtk!66

7 years agosurface: Make process_updates no longer be a vfunc
Benjamin Otte [Wed, 21 Mar 2018 20:58:08 +0000 (21:58 +0100)]
surface: Make process_updates no longer be a vfunc

7 years agosurface: Remove some unused vfuncs
Benjamin Otte [Wed, 21 Mar 2018 19:33:20 +0000 (20:33 +0100)]
surface: Remove some unused vfuncs

7 years agogdk: Remove GdkSurfaceImpl::shape_combine_region vfunc
Benjamin Otte [Wed, 21 Mar 2018 19:15:21 +0000 (20:15 +0100)]
gdk: Remove GdkSurfaceImpl::shape_combine_region vfunc

... and all the implementations.

Input shapes still exist.

7 years agogdk: Remove a bunch of unused variables
Benjamin Otte [Wed, 21 Mar 2018 18:58:49 +0000 (19:58 +0100)]
gdk: Remove a bunch of unused variables

7 years agoReduce the header level in the bug template
Emmanuele Bassi [Wed, 21 Mar 2018 11:49:24 +0000 (11:49 +0000)]
Reduce the header level in the bug template

Make version and additional information headers use h2, like everything else.

7 years agoMerge branch 'wip/sync-pre-swap-buffers' into 'master'
Matthias Clasen [Wed, 21 Mar 2018 11:40:55 +0000 (11:40 +0000)]
Merge branch 'wip/sync-pre-swap-buffers' into 'master'

Synchronize wl_surface state before wl_surface_commit

See merge request GNOME/gtk!75

7 years agoContinue renaming window to surface
Alexander Larsson [Wed, 21 Mar 2018 10:49:14 +0000 (11:49 +0100)]
Continue renaming window to surface

This renames a lot of arguments, local variables and functions.

7 years agoGdkSurface: Fix doc comments for remoteve invalidate_children argument
Alexander Larsson [Wed, 21 Mar 2018 08:18:30 +0000 (09:18 +0100)]
GdkSurface: Fix doc comments for remoteve invalidate_children argument

7 years agobroadway: Fix up window -> surface renames
Alexander Larsson [Wed, 21 Mar 2018 08:17:50 +0000 (09:17 +0100)]
broadway: Fix up window -> surface renames

7 years agoFix up indentation after GdkSurface rename
Alexander Larsson [Wed, 21 Mar 2018 08:06:31 +0000 (09:06 +0100)]
Fix up indentation after GdkSurface rename

This fixes up a lot of the indentation issues from the
rename of windows to surfaces, at least the ones in the headers
and many others.

7 years agowayland/vulkancontext: Sync surface state before vkQueuePresentKHR
Jonas Ådahl [Wed, 21 Mar 2018 06:08:13 +0000 (14:08 +0800)]
wayland/vulkancontext: Sync surface state before vkQueuePresentKHR

A side effect of vkQueuePresentKHR is the Vulkan implementation calling
wl_surface_commit() on the corresponding Wayland surface. Thus, before
this, we must synchronize the surface state (e.g. opaque region, window
geometry, etc) that changed since last time. Prior to this commit this
was done after calling vkQueuePresentKHR(), causing the surface state to
always correspond to the previous buffer state. As of this commit this
is now done before calling vkQueuePresentKHR(), thus before
wl_surface_commit().

7 years agowayland/glcontext: Sync surface state before swapping buffers
Jonas Ådahl [Wed, 21 Mar 2018 06:01:58 +0000 (14:01 +0800)]
wayland/glcontext: Sync surface state before swapping buffers

A side effect of eglSwapBuffers* is the EGL implementation calling
wl_surface_commit() on the corresponding Wayland surface. Thus, before
swapping buffers, we must synchronize the surface state (e.g. opaque
region, window geometry, etc) that changed since the last buffer swap.
Prior to this commit, this was done after eglSwapBuffers*, causing the
surface state to always correspond to the previous buffer state. As of
this commit this is now done before swapping the buffers, thus before
wl_surface_commit().

7 years agosurface: Simplify gdk_surface_invalidate_region()
Benjamin Otte [Wed, 21 Mar 2018 03:07:37 +0000 (04:07 +0100)]
surface: Simplify gdk_surface_invalidate_region()

Remove duplicate functions and unused arguments.

Functionality is identical otherwise.

7 years agosurface: Stop tracking clip region
Benjamin Otte [Wed, 21 Mar 2018 02:51:29 +0000 (03:51 +0100)]
surface: Stop tracking clip region

It's not needed because we're not using it for anything anymore now that
client-side surfaces aren't drawn anymore.

7 years agogdk: Remove gdk_surface_shape_combine_region()
Benjamin Otte [Wed, 21 Mar 2018 02:42:34 +0000 (03:42 +0100)]
gdk: Remove gdk_surface_shape_combine_region()

If you want transparent region, you can just render them transparent.
If you want input shaping, use gdk_surface_input_shape_combine_region().

Also remove gtk_widget_shape_combine_region().

7 years agosurface: Remove gdk_surface_get_clip_region()
Benjamin Otte [Wed, 21 Mar 2018 01:57:48 +0000 (02:57 +0100)]
surface: Remove gdk_surface_get_clip_region()

... and gdk_surface_get_visible_region() APIs. They are unused.

7 years agogdk: Remove unused members from surface struct
Benjamin Otte [Wed, 21 Mar 2018 01:50:37 +0000 (02:50 +0100)]
gdk: Remove unused members from surface struct

7 years agosurface: Stop shaping native children of csw surfaces
Benjamin Otte [Wed, 21 Mar 2018 01:45:43 +0000 (02:45 +0100)]
surface: Stop shaping native children of csw surfaces

We don't have them anymore and they are goig away anyway.

7 years agodocs: Drop the 'Contributing' section from the README
Emmanuele Bassi [Wed, 21 Mar 2018 01:02:23 +0000 (01:02 +0000)]
docs: Drop the 'Contributing' section from the README

We have a whole document for that, and it's up to date with regards to
the GitLab workflow.

7 years agoAdd a template for bug reports
Emmanuele Bassi [Wed, 21 Mar 2018 00:59:06 +0000 (00:59 +0000)]
Add a template for bug reports

GitLab issues can use templates to guide users when filing new bug
reports; we should take advantage of this feature.