We haven't had any scalable directories in this list.
Add some. Since we seem to have settled on including
just actions and status as subdirectories for each
size, add scalable/actions and scalable/status.
Fixes: #4960
Use a separate queue to dispatch the token object exclusively, just like we
do on the GdkSurface activation paths.
(cherry-picked from commit fb68600d88)
This brings back a subset of what quit-mnemonic.ui tested for, but
trying a lot harder to trigger the label overdrawing its allocation,
which will cause the text to be cut off when clipping is happening.
It should not be an issue at all with GTK4, but keeping that test around
is a good idea.
I can't quite figure out what this test was meant
to test, and how to make it do so in a way that
does not fall afoul of rendering issues in the GL
renderer and rounding differences in pango.
Can't win with reftests.
When loading cursors at scale, we expect the
cursor images to have a size of scale * size.
If we don't find such images, load them at their
unscaled size and scale them up ourselves.
Without this, cursors will appear in unexpected
sizes depending on scales and themes.
Related: #4746
If a context is not realized, calling gtk_at_spi_context_to_ref() will
return a null ref, because its path has not been initialized yet. This
was already done for all other cases in get_parent_context_ref(), but
was missing for the GtkStackPage case.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4944
It appears that we mess up accounting for blinking
cursors sometimes, and can hit blink_cb when there
is a nonempty selection.
Instead of asserting, warn and stop blinking.
Related: #4767
On Wayland it is a protocol violation to upload buffers with
dimensions that are not an integer multiple of the buffer scale.
Until recently, Mutter did not enforce this. When it started
doing so, some users started seeing crashes in GTK apps because the
cursor theme ended up with e.g. a 15x16 pixel image at scale of 2.
Add a small sanity check for this case.
Not updating shadow size unconditionally would lead to shadow size not
being set on map, which would lead mutter to think that we are a Window
without extents and then become confused when we suddenly set some.
Make sure that doesn't happen by always having shadows set on map, just
like GTK3.
Fixes#4136
(cherry picked from commit 213490099b)
Now that we use event controllers we can forward keybindings from the
external entry to the filechooserwidget at the bubble phase.
Fixes#4905
References:
* commit 1fb075dbca
* commit 686116ba61
This can happen if the group can be resolved even when doing the initial
registration of an action as observer will not yet be in the GSList of
watchers (and therefore has no weak references).
Fixes a warning like the following:
g_object_weak_unref: couldn't find weak ref
These were getting created with possible non-zero values and then inserted
into a hashtable where the readers may not know the state of the group.
Ensure those values are set to zero until we assign them below.
Previously, there was an issue with glitching after showing/hiding a
popover that was not also destroyed. This was due to the popover having
an update_freeze_count of zero after hiding the surface.
That resulted in it's toplevel continuously dropping frames such as during
high-frame-rate scrolling in textviews. This problem is much more visible
on high-frame-rate displays such as 120hz/144hz.
With this commit, we freeze the frame clock of the popup until it is
mapped again.
When a non-existing file is selected in the file chooser
for print-to-file, we weren't updating the button label
to show the new filename. Fix that.
Also, use newer file chooser api.
If we get consecutive preedit string updates that announce a NULL
string, we still do end up issuing ::preedit-changed with those.
Ignore changes from NULL to NULL, it is the other combinations which
must issue this signal.
(cherry-picked from commit d6fe6f495a)
This is a hot path when scrolling a ColumnView, and
g_param_spec_pool_lookup () was taking a measurable part in this hot
path. Instead, notify using pspecs to avoid the name lookup.
Related: https://gitlab.gnome.org/GNOME/gtk/-/issues/3334
freeze/thaw_notify () showed up on the perf trace for rapid ColumnView
scrolling. Track the three properties manually to make it a little
faster.
Related: https://gitlab.gnome.org/GNOME/gtk/-/issues/3334
This looks like a leftover excess invalidation from when the surrounding
code was refactored to not just be called on parent changes but also
when repositioning inside the same parent in commit
507016cafc
Ivan Molodetskikh found this problem in
https://gitlab.gnome.org/GNOME/gtk/-/issues/3334#note_1445873 which
contains a longer analysis of this problem and the performance
reductions it causes.
Related: #3334
When changing folders, we were making the select
button insensitive when there is no folder selected.
However, the select button should be usable to
select the current folder.
Fixes#4020
if the loop for determining max width grows too big, print an error and
abort assuming that a satisfactory value was reached.
This will cause wrong layout and might cause widgets to overlap, but it
will not infloop.
It actually works around and doesn't really fix the primary cause of the
following bugs, but good enough to close them:
Fixes: #4252Fixes: #4517
(cherry picked from commit 515b1f5292)
Direct access of the fields of the union trips compiler warnings with
GCC 12, such as:
../gtk/gtkimagedefinition.c:135:13: error: array subscript
‘GtkImageDefinition {aka union _GtkImageDefinition}[0]’ is partly
outside array bounds of ‘GtkImageDefinitionEmpty[1]’ {aka
‘struct _GtkImageDefinitionEmpty[1]’} [-Werror=array-bounds]
This allows the user to navigate via tab the links in a label and exits
the widget after the last link, when moving forward, and first link,
when moving backward.
This also ensures that ellipsised links arn't focused.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4681
The `has-tooltip` property gets set to `false` for label with links if no
link is selected. This makes sure to only change the property to `true`
but never to `false`.
Instead of populating the properties right away (when the widget might
not have been allocated yet, and hence cannot know the right values),
the widget should queue an allocation, where it will populate the
values.
For reasons that only apply to the old serial handling, asking for
the surrounding after IM changes resulted in lazy handling of
commit() afterwards.
With the recent interpretation of serials, this problem became more
apparent, since it is in fact very likely that the last interaction
step after an IM change is notifying of the changed surrounding
text after the IM change was applied.
Make handling of surrounding text similar to caret position changes,
always commit() after the state change, but skip through non-changes.
This makes the compositor state fully up-to-date after an IM change.
The gesture as connected currently on the child GtkText is easily overridden
by the parent editables (and gently done so in the attempt to consume all
clicks).
Connect this gesture to the parent editable widget in these cases, so the
gesture can cohabit with the click-consuming one. It's not part of the same
group, but it won't be abruptly cancelled.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4795
This fixes two issues with the offscreen rendering code for nodes with
bounds not aligned with the pixel grid:
1.) When drawing to an offscreen buffer the size of the offscreen buffer
was rounded up, but then later when used as texture the vertices
correspond to the original bounds with the unrounded size. This could
then result in the offscreen texture being drawn onscreen at a slightly
smaller size, which then lead to it being visually shifted and blurry.
This is fixed by adjusting the u/v coordinates to ignore the padding
region in the offscreen texture that got added by the size increase from
rounding.
2.) The viewport used when rendering to the offscreen buffer was not
aligned with the pixel grid for nodes at coordinates not aligned with
the pixel grid. Then because the content of the offscreen buffer is not
aligned with the pixel grid and later when used as textures sampling
from it will result in interpolated values for an onscreen pixel. This
could also result in shifting and blurriness, especially for nested
offscreen rendering at different offsets.
This is fixed by adding similar padding at the beginning of the
texture and also adjusting the u/v coordinates to ignore this region.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3833
When given an invalid atom, gdk_x11_get_xatom_name_for_display can
return NULL and trigger a seg in gdk_x11_clipboard_formats_from_atoms.
Check for NULL.
Why I'm seeing a bad atom there is probably a separate question.
https://bugzilla.redhat.com/show_bug.cgi?id=2037786
We may well be using an EGL context that does not support Desktop (W)GL on
Windows, such as in the case of using libANGLE. So, check whether WGL is
supported for this running instance before trying to query WGL extensions.
This will get rid of warning messages from libepoxy.
Otherwise a stray scroll controller may prevent others from getting hold
events, even if it always propagates scroll events and does absolutely
nothing.
We only need a C compiler and not the whole toolchain,
and gst-plugins-bad was split into libraries and plugins.
pkg-config -> pkgconf.
This should speed the CI setup up a bit.
WebKit's GTK 4 port can give us textures with an internal format of
GL_RGBA with GL_UNSIGNED_NORMALIZED and a bit-depth of 8. This fixes
warnings for every GdkGLTexture created/delivered to the GskGLRenderer.
The format is essentially the same as GL_RGBA8 since it is normalized
between 0.0..1.0 for 8-bit components.
Fixes#4783
When surface depth switches from non-high-depth to high-depth (or vice
versa) the current surface has to be destroyed before a new one can be
created for this window. eglDestroySurface however was getting passed a
GdkDisplay, rather than the EGLDisplay it expects. As a result the old
surface did not get destroyed and the new surface could not be created
causing rendering to freeze.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4773
If using the opacity CSS property the renderer cannot optimize these
handles without the use of offscreens due to the use of both a border
and rgb render node.
Instead, we can apply the alpha to the color values and get the same
effect in a way that the GL renderer can optimize without the use of
offscreen textures for a sizeable reduction in runtime overhead.
Currently the GtkIMMultiContext may stick to a delegate GtkIMContext
that no longer applies after the multicontext is dissociated from
any widget.
Handle set_client_widget() so that it can handle changes between
widgets from 2 different display, but also so the delegate is made
NULL whenever the context has a NULL widget.
Doing so, any new client widget results in a new delegate IM context
lookup from the right GdkDisplay and GtkSettings, which avoids any
mix up.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4805
(cherry picked from commit b67da38916)
The root accessible object is registered asynchronously, as it needs to
call a method on the AT-SPI registry daemon. This means we need to defer
registering the GtkAtSpiContext on the accessibility bus and in the
cache until after the registration is complete.
Fixes: #4825
When changing folders, we were making the select
button insensitive when there's no files around.
That doesn't make sense in save mode when we don't
want to select a file but create one.
Fixes: #4851
It is very irritating when the entry completion popup
appears not in response to user input in the entry.
In particular, when that happens right as the dialog
is shown.
To prevent that, temporarily disable completion
when setting the entry text programmatically.
macos: Fix problem that window cannot be minimized by user operation
Closes#4811
See merge request GNOME/gtk!4613
(cherry picked from commit d75147db0a)
d3cf7088 macos: Skip running `showAndMakeKey` when a window is minimized by user action
macos: prohibit fullscreen transition if in transtion
This prevents performing additional fullscreen transitions while
a transition is already in progress.
Closes#4808
See merge request GNOME/gtk!4612
(cherry picked from commit 15b7a4572b)
69a0a5cf macos: Prohibit changing the full-screen state of a window during a full-screen transition
229e4b58 macos: Corrected code style
The call to gdk_win32_clipboard_request_contentformats() can return NULL even
without an error condition being hit (such as when the system clipboard is
empty), so check whether the returned GdkContentFormat pointer is not NULL
before calling gdk_clipboard_claim_remote(), which expects it to be not NULL,
otherwise we face a warning from that funtion and the subsequent
g_object_unref().
This at least partially fixes issue #4796.
As documented:
> Overlay children whose alignments cause them to be positioned
> at an edge get the style classes “.left”, “.right”, “.top”,
> and/or “.bottom” according to their position.
Likely accidental regression in b7ee2cbc28
Fixes https://gitlab.gnome.org/GNOME/nautilus/-/issues/2099
(cherry picked from commit 351ffef704)
Ensure that resolution of the subproject occurs via the dependency
interface, not the "poke at subprojects manually" interface, and make
that actually work via --wrap-mode=forcefallback.
There's no need to mark it as not-required and then manually invoke
subproject(), since fallback should work correctly and it is always
needed.
However, if fallback was performed (or forced) it would error out since
get_variable() was instructed to only use pkg-config while the relevant
variable was exported by the subproject as an internal fallback
dependency.
We had code to do it and it never actually got used correctly. This ensures
that the popup services are attached to the parents so that they get proper
stacking orders when displayed. Additionally, it fixes popups from being
shown as their own windows in Exposé.
If we are clicking through the shadow of a window, we need to take special
care to not raise the old window on mouseUp. This is normally done by the
display server for us, so we need to use the proper API that is public to
handle this (rather than CGSSetWindowTags()). Doing so requires us to
dispatch the event to the NSView and then cancel the activcation from
the mouseDown: event there.
If we closed a key window in response to events, we need to denote another
window as the new key window. This is easiest to do from an idle so that
we don't clobber notification pairs of "did resign"/"did become" key
window.
We have a sorted set of surfaces by display server stacking, so we can
take the first one we come across that is already mapped and re-show it
to become key/main.
If we have server-side decorations we might need to request a layout in
response to the resize notification. We don't need to do this in other
cases because we already handle that in the process of doing the resize
(and that code is that way because of delayed delivery of NSNotification).
If we are using NSWindow titled windows, we don't end up waking up the
frame clock when the window is resized on the display server. This ensures
that we do that after getting a notification of resize.
There are cases we might want to consume a NSEvent without creating a
GdkEvent or passing it along to the NSApplication for processing. This
creates a new value we can use and check against to propagate that without
having to do out parameters at the slightly odd invalid pointer value for
a GdkEvent (similar to how MMAP_FAILED is done).
This can get in the way of how we track changes while events are actively
processing. Instead, we may want to delay this until the next main loop
idle and then check to see if we have a main window as the NSNotification
may have come in right after this.
We want to ensure that we recalculate the sort order of windows before
processing the motion. Generally this would be done in response from the
display server in GdkMacosWindow, but I've seen it possible to race there.
We need to handle the case where we might be racing against an incoming
configure event due to how notifications are queued from the display
server. Rather than calling configure (and possibly causing other things
to move around) this just queries the display server directly for the
coordinates that we care about.
Additionally, we can display:NO as we are in control of all the display
process now using CALayer.
We failed to handle the toplevel with transient-for case here which could
cause our X/Y calculations to be off in other areas such as best monitor
detection.
We do actually need the parent frame clock here because it is the way we
ensure that we get layout called for our popup surfaces at the same time
as the parent surface.
This doesn't appear to happen much, but if it does it is nice to setup
the window placement initially. Generally, transient-for is set after
the creation of the toplevel rather than here.
The GdkMacosBuffer object already has storage for tracking the damage
region as it is used in GdkMacosCairoContext to manually copy regions from
the front buffer to the back buffer. This makes the GdkMacosGLContext also
use that field so that we can easily drop old damage regions when the
buffer is lost. This happens during resizes, monitor changes, etc.
This helper is useful to ensure we are consistent with how we keep a
window clamped to the workarea of a monitor when placing windows on
screen. (This does not affect snap-to-edges).
If _gdk_macos_surface_move_resize() was called with various -1 parameters
we really want to avoid changing anything even if we think we know what
the value might be. Otherwise, we risk messing up in-flight operations that
we have not yet been notified of yet.
This improves the chances we place windows in an appropriate location as
they don't et screwed up before window-manager placement.
We need to bring the application to the foreground in multiple ways, and
this call to [NSApp activateIgnoringOtherApps:YES] ensures that we become
foreground before the first window is opened. Otherwise we end up starting
applications in the background.
Fixes#4736
If we are double buffering surfaces with IOSurface then we need to copy
the area that was damaged in the previous frame to the back buffer. This
can be done with IOSurface but we need to hold the read-only lock so that
we don't cause the underlying IOSurface contents to be invalidated.
Additionally, since this is only used in the context of rendering to a
GdkMacosSurface, we know the life-time of the cairo_surface_t and can
simply lock/unlock the IOSurface buffer from begin_frame/end_frame to have
the buffer flushing semantics we want.
To ensure that we don't over damage, we store the damage in begin_frame
(and copy it) and then subtract it from the next frames damage to determine
the smallest amount we need to copy (taking scale factor into account).
We don't care to modify the damage region to swapBuffers because they
already have the right contents and could potentially fall into another
tile anyway and we'd like to avoid damaging that.
Fixes#4735
This can be used to lock a surface for reading to avoid causing the
surface contents to be invalidated. This is needed when reading back from
a front-buffer to the back-buffer as is needed when using Cairo surfaces
to implement something similar to BufferAge.
Previously, a single CVDisplayLink was used to drive updates for all
surfaces across all monitors. It used a 'best guess' rate which would
allow for updates across monitors of mixed rates. This is undesirable for
situations where you might have a 144hz monitor as it does not allow for
reaching up to that frame rate.
Instead, we want to use a per-monitor CVDisplayLink which will fire at the
rate of the monitor down to the level of updates we require. This commit
does just that.
When a surface crosses onto a new monitor, that monitor is used to drive
the GdkFrameClock.
Fixes#4732
Using the mode allows better detection of refresh rate and refresh
interval for the CVDisplayLink bridge to GdkFrameClock. Using it can help
ensure that our 144hz displays can actually reach that rather than falling
back to just 60hz.
This will also need future commits to rework the displaylink source to be
per-monitor.
When the fingers are placed on the touchpad, we get a scroll event with
the phase NSEventPhaseMayBegin. We can use this to synthesize an is_stop
event. This results in the scrolledwindow stopping scroll with stop
gestures.
This can cause another warning as well, however, which should be addressed
from #4730.
Fixes#4733
Windows can end up on different monitors despite having a parent or
transient-for ancestor. We want them to be driven by the CVDisplayLink
for the best-monitor, and so this needs to be unshared.
Currently we're using a display link that is for all active displays which
is just the display server trying to find some timings that try to overlap
as many as possible.
That was fine for a prototype, but we really need to do better for
situations with mixed frame rate (such as 60hz and 120hz promotion
displays). Additionally, the 144hz external monitor I have will never
reach 144hz using the current design.
This is just the first step in changing this, but the goal is to have
one of these attached to each GdkMacosMonitor which we can then use to
thaw surfaces specific to that monitor.
We will eventually be needing additional feedback from the display server
which would be nice to keep away from the rest of GdkMacosDisplay for
cleanliness sake. Particularly for feedback from mission control and other
environment factors that requires private API for proper integration.
When using server-side-decorations, we need to avoid potential cycles with
compute-size as it may not have the new sizing information yet. We can
just short circuit during "live resize" to get that effect.
Fixes poor window resizing from top-left on titled windows.
This doesn't give us appropriate results if we use the window delegate.
Instead, we need to adjust the frame at the same time we change the
style mask so that we end up in the same location.
Previously we had issues on macos where the overshoot would keep showing.
To fix this we need to actually use discrete events instead of the
generated deltas from macOS in the scroll wheel case. Additionally, we need
to drop the kinetic momentum events from macOS and rely on the gtk kinetic
events which are already happening anyway. We also need to submit the
is_stop event for the GDK_SCROLL_SMOOTH case when we detect it.
To keep the discrete scroll events correct, we need to alter the hack in
gtkscrolledwindow.c to use the same path as other platforms except for
when a smooth scroll event is in place. In the future, I would imagine that
this falls into the boundary of high-precision scrolling and would share
the same code paths as other platforms.
With all of these in place, kinetic scrolling with overshoot appears the
same on macOS as other platforms.
When creating new windows, it is better if we create them with a slight
offset to where they were created before so that they are visible to the
user separately from what they might be overshadowing.
We were missing the surface offset (e.g. shadows) at the time of expressing
the text caret location in surface coordinates. Add this offset so the
coordinates are as expected by the compositor.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4668
Tools like gtk4-launch can't set surface on the activation token so
don't require it. If the compositor requires it we can't do anything
about it anyway. This avoids a critical:
(gtk4-launch:23497): Gdk-CRITICAL **: 17:07:24.704: gdk_wayland_surface_get_wl_surface: assertion 'GDK_IS_WAYLAND_SURFACE (surface)' failed
Fixes: be4216e051 ("gdk/wayland: Support the xdg-activation wayland protocol")
Signed-off-by: Guido Günther <agx@sigxcpu.org>
Currently, we have all the plumbing in place so that GTK consumes the
startup notification ID when focusing a window through the xdg-activation
protocol.
This however misses the case that a window might be requested to be
focused with no startup ID (i.e. via interaction with the application,
not through GApplication or other application launching logic).
In this case, we let the application create a token that will be
consumed by itself. The serial used is that from the last
interaction, so the compositor will still be able to do focus prevention
logic if it applies.
Since we already do have a last serial at hand, prefer xdg-activation
all the way over the now stale gtk-shell focusing support. The timestamp
argument becomes unused, but that is a weak argument to prefer the
private protocol over the standard one. The gtk-shell protocol support
is so far left for interaction with older Mutter.
The simplify and validate commands can function
without a display connection, only preview absolutely
needs one. Allow this, by using gtk_init_check().
The trickery we do with objcopy and ld to speed up
resource inclusion does not seem to work right on
32bit Arm, so just skip it there.
Fixes: #4757, #4748, #4752
This one can be used for both premultiplied and non-premultiplied alpha
formats, since alpha is always 255. It is useful for opaque PNG upload
on both cairo and GL renderers.
That way, all permutations are possible. Previously it was only useful
in the cairo renderer, which required rgba8 → premultiplied bgra8, while
the GL renderer required rgba8 → premultiplied rgba8. Now both are
available.
This was only useful when building for AArch32 without -mfpu=neon, on
AArch64 or with -mfpu=neon gcc is smart enough to do the auto-
vectorisation, leading to code almost as good as what I wrote in
1fdf5b7cf8.
It appears that NVIDIA does not implement EGL_EXT_swap_buffers_with_damage
on their EGL implementation, but does implement the KHR variant of it.
This checks for a suitable implementation and stores a pointer to the
compatible implementation within the GdkGLContextPrivate struct.
We were looking at GtkWidget:has-focus from
event controller signal handlers here, but
the widget property is only changed after
the event controllers.
Update the :has-focus property of the focus
widget when the active status of the window
changes.
We change the property after generating the
GDK_CROSSING_ACTIVE crossing events.
This may come from different sources at around the same time, e.g.
a hold gesture while on overshoot. Avoid doing that if an
animation is already set.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4730
When large viewports are passed to gsk_renderer_render_texture(), don't
fail (or even return NULL).
Instead, draw multiple tiles and assemble them into a memory texture.
Tests added to the testsuite for this.
CI currently fails with "fatal error LNK1318: Unexpected PDB error; OK (0) ''"
Google tells me it might be related to hitting a memory limit. Let's try
disabling debug for now.
Instead of performing keyboard layout substitution whenever we find a matching
entry in the registry, first try to load the original layout and only attempt
substitution when that fails.
See #4724
This broke with the previous fixes for initial window positioning. We need
the initial positioning so that tails will be displayed correctly when the
popover surface is displayed.
If the size changes, we need to relayout the tiles. Otherwise we can keep
using what we had before. Generally, that shouldn't happen, but the
previous check was failing in a number of ways.
It looks like, particularly on the M1, we might need to double buffer the
contents of the IOSurface<->OpenGL texture bindings. This doesn't appear
to show up on the Intel macbooks I've tried, but I've seen it in the wild
on an M1.
If we have a 2x scale laptop with a 1x scale external display, we would
need to create a new IOSurface for the external display once it crosses
a boundary, otherwise we won't have something capable of displaying
correctly on the second monitor.
This provides a major shift in how we draw both when accelerated OpenGL
as well as software rendering with Cairo. In short, it uses tiles of Core
Animation's CALayer to display contents from an OpenGL or Cairo rendering
so that the window can provide partial damage updates. Partial damage is
not generally available when using OpenGL as the whole buffer is flipped
even if you only submitted a small change using a scissor rect.
Thankfully, this speeds up Cairo rendering a bit too by using IOSurface to
upload contents to the display server. We use the tiling system we do for
OpenGL which reduces overall complexity and differences between them.
A New Buffer
============
GdkMacosBuffer is a wrapper around an IOSurfaceRef. The term buffer was
used because 1) surface is already used and 2) it loosely maps to a
front/back buffer semantic.
However, it appears that IOSurfaceRef contents are being retained in
some fashion (likely in the compositor result) so we can update the same
IOSurfaceRef without flipping as long as we're fast. This appears to be
what Chromium does as well, but Firefox uses two IOSurfaceRef and flips
between them. We would like to avoid two surfaces because it doubles the
GPU VRAM requirements of the application.
Changes to Windows
==================
Previously, the NSWindow would dynamically change between different
types of NSView based on the renderer being used. This is no longer
necessary as we just have a single NSView type, GdkMacosView, which
inherits from GdkMacosBaseView just to keep the tedius stuff separate
from the machinery of GdkMacosView. We can merge those someday if we
are okay with that.
Changes to Views
================
GdkMacosCairoView, GdkMacosCairoSubView, GdkMacosGLView have all been
removed and replaced with GdkMacosView. This new view has a single
CALayer (GdkMacosLayer) attached to it which itself has sublayers.
The contents of the CALayer is populated with an IOSurfaceRef which
we allocated with the GdkMacosSurface. The surface is replaced when
the NSWindow resizes.
Changes to Layers
=================
We now have a dedicated GdkMacosLayer which contains sublayers of
GdkMacosTile. The tile has a maximum size of 128x128 pixels in device
units.
The GdkMacosTile is partitioned by splitting both the transparent
region (window bounds minus opaque area) and then by splitting the
opaque area.
A tile has either translucent contents (and therefore is not opaque) or
has opaque contents (and therefore is opaque). An opaque tile never
contains transparent contents. As such, the opaque tiles contain a black
background so that Core Animation will consider the tile's bounds as
opaque. This can be verified with "Quartz Debug -> Show opaque regions".
Changes to Cairo
================
GTK 4 cannot currently use cairo-quartz because of how CSS borders are
rendered. It simply causes errors in the cairo_quartz_surface_t backend.
Since we are restricted to using cairo_image_surface_t (which happens to
be faster anyway) we can use the IOSurfaceBaseAddress() to obtain a
mapping of the IOSurfaceRef in user-space. It always uses BGRA 32-bit
with alpha channel even if we will discard the alpha channel as that is
necessary to hit the fast paths in other parts of the platform. Note
that while Cairo says CAIRO_FORMAT_ARGB32, it is really 32-bit BGRA on
little-endian as we expect.
OpenGL will render flipped (Quartz Native Co-ordinates) while Cairo
renders with 0,O in the top-left. We could use cairo_translate() and
cairo_scale() to reverse this, but it looks like some cairo things may
not look quite as right if we do so. To reduce the chances of one-off
bugs this continues to draw as Cairo would normally, but instead uses
an CGAffineTransform in the tiles and some CGRect translation when
swapping buffers to get the same effect.
Changes to OpenGL
=================
To simplify things, removal of all NSOpenGL* related components have
been removed and we strictly use the Core GL (CGL*) API. This probably
should have been done long ago anyay.
Most examples found in the browsers to use IOSurfaceRef with OpenGL are
using Legacy GL and there is still work underway to make this fit in
with the rest of how the GSK GL renderer works.
Since IOSurfaceRef bound to a texture/framebuffer will not have a
default framebuffer ID of 0, we needed to add a default framebuffer id
to the GdkGLContext. GskGLRenderer can use this to setup the command
queue in such a way that our IOSurface destination has been
glBindFramebuffer() as if it were the default drawable.
This stuff is pretty slight-of-hand, so where things are and what needs
flushing when and where has been a bit of an experiment to see what
actually works to get synchronization across subsystems.
Efficient Damages
=================
After we draw with Cairo, we unlock the IOSurfaceRef and the contents
are uploaded to the GPU. To make the contents visible to the app,
we must clear the tiles contents with `layer.contents=nil;` and then
re-apply the IOSurfaceRef. Since the buffer has likely not changed, we
only do this if the tile overlaps the damage region.
This gives the effect of having more tightly controlled damage regions
even though updating the layer would damage be the whole window (as it
is with OpenGL/Metal today with the exception of scissor-rect).
This too can be verified usign "Quartz Debug -> Flash screen udpates".
Frame Synchronized Resize
=========================
In GTK 4, we have the ability to perform sizing changes from compute-size
during the layout phase. Since the macOS backend already tracks window
resizes manually, we can avoid doing the setFrame: immediately and instead
do it within the frame clock's layout phase.
Doing so gives us vastly better resize experience as we're more likely to
get the size-change and updated-contents in the same frame on screen. It
makes things feel "connected" in a way they weren't before.
Some additional effort to tweak gravity during the process is also
necessary but we were already doing that in the GTK 4 backend.
Backporting
===========
The design here has made an attempt to make it possible to backport by
keeping GdkMacosBuffer, GdkMacosLayer, and GdkMacosTile fairly
independent. There may be an opportunity to integrate this into GTK 3's
quartz backend with a fair bit of work. Doing so could improve the
situation for applications which are damage-rich such as The GIMP.
There are situations where our "default framebuffer" is not actually
zero, yet we still want to apply a scissor rect.
Generally, 0 is the default framebuffer. But on platforms where we need
to bind a platform-specific feature to a GL_FRAMEBUFFER, we might have a
default that is not 0. For example, on macOS we bind an IOSurfaceRef to
a GL_TEXTURE_RECTANGLE which then is assigned as the backing store for a
framebuffer. This is different than using gsk_gl_renderer_render_texture()
in that we don't want to incur an extra copy to the destination surface
nor do we even have a way to pass a texture_id into render_texture().
The GtkFileCHooserNativeQuartz injects a NSComboBox into the NSSavePanel
(which is displayed in a remote process since 10.15 whether or not you are
a sandboxed application). The style has changed and we need more space
here to not clip part of the combobox out of view.
I tried every size from 22 to 30 and this seemed to look the most natural
without skewing the location of the text within the combobox.
Previously, the popover would cause the window to go into the :backdrop
state which is not what we want for consistency with other platforms. This
fixes that by walking up the surface chain when we get notified of
loosing or acquiring "key" input from the display server.
If the rendering operation is over an opaque region, we can potentially
avoid clearing a large section of the framebuffer destination. Some cases
you do want to clear, such as when clearing the whole contents as some
drivers have fast paths for that to avoid bringing data back into the
framebuffer.
One may be using IJG libjpeg or libjpeg-turbo to build GTK, and their
build files may or may not generate pkg-config files for us. To make
things easier, we can make use of CMake's built-in support for finding
IJG libjpeg or libjpeg-turbo.
The CMake build files for libtiff may or may not generate pkg-config
files for us, so we can use Meson's CMake support to help us find
libtiff, as CMake has built-in support for finding libtiff.
Add a variable in meson.build that covers Visual Studio-like compilers,
so that we can use it to help us find depedencies using CMake rather
than via pkg-config, where applicable.
Change the existing use case for finding libpng accordingly.
We might have panels with controls in them where the window is running in
another process. The control could have a wrapper window which we would
see from this process. This can happen with the GtkFileChooserNative, but
any NSSavePanel in macOS 10.15+ is out of process (not just sandboxed
applications).
This significantly cleans up how we handle various move-resize, compute-
size, and configure (notification of changes) in the macOS GDK backend.
Originally when prototyping this backend, there were some bits that came
over from the quartz backend and some bits which did not. It got confusing
and so this makes an attempt to knock down all that technical debt.
It is much simpler now in that the GdkMacosSurface makes requests of the
GdkMacosWindow, and the GdkMacosWindow notifies the GdkMacosSurface of
changes that happen.
User resizes are delayed until the next compute-size so that we are much
closer to the layout phase, reducing chances for in-between frames.
This also improves the situation of leaving maximized state so that a
grab and drag feels like you'd expect on other platforms.
I removed the opacity hack we had in before, because that is all coming
out anyway and it's a bit obnoxious to maintain through the async flows
here.
This fixes GTK's NSWindow for toplevels so that they are allowed to enter
fullscreen. We were already handlign the state transitions from the
setStyleMask: halper, but we didn't previously tell the window we are
allowed to transition into that.
There is a bit of a mismatch here in that GTK doesn't have any such flag
that determines if a window is "allowed" by policy to enter fullscreen
since window managers on Linux are free to do that at will.
This makes it easier to figure out those values (which are mentioned in
the GtkApplication documentation) rather than working that out from the
way they're generated (or documented as being generated).
If we have GStreamer on macOS we likely have support for CGL to get an
OpenGL context we can use. This provides the missing pieces to get
accelerated video playback in gtk4-widget-factory working.
This more than halves the total runtime of this function since the
previous commit, from 8.36% to 4.02%, and is most likely memory
bandwidth limited on this specific board now.
I tried to do a SSE2 version as well, but couldn’t find any equivalent
of the LD4/ST4 ARM instruction.
On x86 on a Kaby Lake CPU, this makes it go from 6.63% of the total
execution time (loading some PNGs using the cairo backend) down to
3.20%.
On ARM on a Cortex-A7, on the same workload, this makes it go from 57%
to 8.36%.
We want our tracking area to be limited to the input region so that we
don't pass along events outside of them for the window. This improves the
chances we click-out of a popover with a large shadow.
This still doesn't let us pass-through clicks for large shadows on top-
level windows though.
We only should be asserting in static functions. Furthermore, this function
did not need to have GDK_BEGIN_MACOS_ALLOC_POOL as nothing is being
allocated there which would cause pooling to get used.
This needs to handle the boundary case where the value is exactly equal
to the edge of a rectangle (which gdk_rectangle_contains_point() does not
consider to be containing). However, if there is a monitor in the list
that is a better match, we still want to prefer it.
When using an external mouse on MacOS, the scrolling behavior is
reversed from the user's scrolling preference. Additionally, it is
noticeably sluggish.
This commit fixes both issues by negating the deltas and multiplying
them by 32 before constructing a new scroll event. 32 seems to be the
"traditional" scaling factor according to [Druid], but I'm not sure
where that value actually comes from. Regardless, scaling the deltas by
this amount makes scrolling feel a lot more responsive in the GTK demos.
Scrolling with a trackpad is not affected by either issue because it
triggers a different code path that uses more precise deltas, and
already negates them.
[Druid]: https://linebender.gitbook.io/linebender-graphics-wiki/mouse-wheel#external-mouse-wheel-vs-trackpad
We currently list everything as a dependencies, regardless of whether
it actually is; this is a source of confusion for users that read the
GTK documentation.
Gi-docgen has a new "related" key in the project configuration which
allows us to list libraries that are merely related to the namespace
we are documenting; the "dependencies" key is used to document the
actual namespace dependencies, now.
This was causing us to draw the same background content twice which is a
significant amount of bits to flip in the GPU for maximized windows,
especially twice.
This updates GtkPopover to use the new GtkNative abstraction for
reporting opaque regions of the window, in hopes that it can speed
up compositors for things like animated lists, menu transitions,
and more.
Fixes#4689
This switches to using the new GtkNative machinery for updating the
opaque region. Some amount of local calculation is still required for
determining when we should apply shadows, and this inherits what was
done previous for that.
Related #4689
This abstracts the machinery to update the opaque region for a GtkWindow
into GtkNative so it may be used from other native impelementations such
as GtkPopover.
Related #4689
Some Windows keymaps have bogus mappings for the Ctrl modifier. !4423 attempted
to fix this by ignoring the Ctrl layer, but that was not enough. We also need to
ignore combinations of Ctrl with other modifiers, i.e. Ctrl + Shift. For example,
Ctrl + Shift + 6 is mapped to the character 0x1E on a US keyboard (but it should
be treated as Ctrl + ^). Basically, always ignore Ctrl unless it is used in
conjunction with Alt, i.e. as part of AltGr.
Related issue: #4667
If any of the APIs that assumes that the entry is set already is used
before having one already set, things break pretty badly.
Fixes a downstream issue reported at https://github.com/gtk-rs/gtk4-rs/issues/873
`free` is defined in `stdlib.h`, see for example
<https://pubs.opengroup.org/onlinepubs/009604499/functions/free.html>. Without
this include compilation can fail with the following error:
```
../gdk/loaders/gdkjpeg.c: In function ‘gdk_save_jpeg’:
../gdk/loaders/gdkjpeg.c:264:7: warning: implicit declaration of function ‘free’ [-Wimplicit-function-declaration]
free (data);
^
../gdk/loaders/gdkjpeg.c:264:7: warning: incompatible implicit declaration of built-in function ‘free’
../gdk/loaders/gdkjpeg.c:264:7: note: include ‘<stdlib.h>’ or provide a declaration of ‘free’
../gdk/loaders/gdkjpeg.c:302:67: error: ‘free’ undeclared (first use in this function)
return g_bytes_new_with_free_func (data, size, (GDestroyNotify) free, NULL);
^
../gdk/loaders/gdkjpeg.c:302:67: note: each undeclared identifier is reported only once for each function it appears in
../gdk/loaders/gdkjpeg.c:303:1: warning: control reaches end of non-void function [-Wreturn-type]
}
^
```
We don't want to risk having something really weird come out if we have a
WCG colorspace, so instead only do the performance hack on systems where
the output is likely reasonable.
We will want to eventually just be drawing in the appropriate colorspace,
but that is not available yet.
When using software rendering w/ cairo, assume we're drawing in
the best-monitor's colorspace rather than RGB to avoid colorspace
conversions on every frame.
Instead of relying on cairo_t to perform drawing from our backing
image surface to the Core Graphics context, we can convert the
cairo_image_surface_t into a CGImageRef without having to copy
data if we are certain of the alignment of the image up front.
Without this, there are many situations, based on the size of the
window that could cause cairo to take a slow path and malloc/copy
the data to ensure that alignment.
The previous commit titled "macos: align image surface rowstride to
16-bytes" ensures that this invariant is true so that our drawing
code can assume we can reference the framebuffer from the
cairo_image_surface_t using a CGDataProvider.
Since GdkMacosCairoContext and GdkMacosCairoSubview are coordinating,
we can also setup the transformation/scale early when drawing the
cairo_image_surface_t instead of when copying it to Core Graphics.
Furthermore, the CGImageRef is created with an RGB colorspace so
that we are not performing colorspace conversion to the output
device. We don't get color matching between displays, but we don't
expect that anyway, particularly with the software renderer.
When creating a cairo_image_surface_t we want both the framebuffer pointer
and each row to be aligned to 16-bytes so that Core Graphics will use more
optimal paths.
However, cairo_image_surface_create() will not guarantee that the rowstride
is aligned to 16-bytes so we must do that ourselves.
We need to avoid conflating the managing of frame callbacks from
the freeze/thaw mechanics and ensure we don't perform extra thaw
requests at the wrong time.
Some keymaps on Windows contain bogus mappings for Ctrl+key for certain
keys, e.g. Ctrl+Backspace = Delete, or Ctrl+[ = 0x1B. These are never
used on Windows, so we should ignore them.
Fixes#4667
GTK's old key symbol list is missing a few symbols like the per mille
sign that is included in some keyboard layouts. This commit updates
gdkkeyuni.c to match libxkbcommon's current key symbol list.
From the GCC manpage:
> Wimplicit-fallthrough=5 doesn't recognize any comments as
> fallthrough comments, only attributes disable
> the warning.
So, check for the =5 version after checking for the simple version. This
way we get -Wfallhrough with clang and -Wfallthrough -Wfallthrough=5
with GCC, which works.
Hold gestures are used to bring existing gestures on touchpad
semantically closer to touchscreen gestures.
Touchpad gestures observe hold gestures with a matching amount of
fingers and emit their begin and end signals when fingers are detected
or removed on/from the touchpad.
When a hold cancel event is detected, it is required to wait a few
milliseconds until the next event(s) are received to avoid emitting
multiple begin signals.
Part-of: <!3454>
Handle hold events:
- GDK_TOUCHPAD_GESTURE_PHASE_BEGIN: scroll-begin is emitted.
- GDK_TOUCHPAD_GESTURE_PHASE_END: A hold gesture ends only when all
fingers are lifted from the touchpad without movement, so
scroll-end is emitted right away.
- GDK_TOUCHPAD_GESTURE_PHASE_CANCEL: A hold gesture is cancelled when
some fingers are lifted/put down or movement is detected. In this
case, scroll-end is emitted after a small timeout only if
GDK_SCROLL wasn't detected.
Part-of: <!3454>
Since the addition of GdkEventSequence in touchpad events, these
are now stored in the gesture using that sequence. This bit of touchpad
gesture handling was however missing to be updated, still looking up
the special NULL sequence.
Use the last sequence here, which will be the one coming from touchpad
gesture events.
Despite touchpad gestures having a sequence, these must use the logical
pointer focus. Avoid using the sequence for GtkPointerFocus lookups with
those events, in order to ensure those events make it all the way to the
intended target.
This is fallout from adding GdkEventSequence information to touchpad
gestures.
This change is done for 2 reasons:
- The logic to request this phase when compressing scroll events is
slightly broken. If there are multiple scroll events that are
coalesced into one, the surface frame clock will not get this request.
The worst case is having >= 2 scroll events on every frame, as the
compressed event will be left in the queue, and be further compressed
on future events.
- Even scroll events aside, this phase is requested in oddly specific
places that are not enough to cover all events, others do rely on
unrelated GdkFrameClock activity that happens to flush the events
as well.
Unify this phase request so it explicitly happens on the arrival of any
event. This ensures that events (compressed or not) will be handled
promptly after arrival.
Currently when the widget is realized after the focus in event the input
method isn't activated as enable is never sent. The call trace is
gtk_text_focus_changed ->
gtk_im_context_focus_in ->
gtk_im_context_wayland_focus_in
which returns early as self->widget is NULL since it's set up in
gtk_text_realize() via gtk_im_context_set_client_widget(). Handle that
case by invoking gtk_im_context_focus_in() from gtk_text_realize() too.
A case where the above happens is a GtkSearchEntry in a GtkSearchBar.
E.g. in gtk4-demo when starting the demo and then hitting the search
button right away.
Remember the current context on focus-in even though the text-input
isn't set up yet. This helps in the case where the text-input is not yet
created but a widget already got focused. Without that the enable()
invocation in text_input_enter() woulnd't be invoked leaving the input
method disabled.
This fixes
gtk4-demo --run=search_entry
which would initially not show the on-screen keyboard when e.g using
phoc/sway as compositor.
As per Benjamin's suggestions, cleanup the previous implementation on
initializing the GLES context on Windows, so that we use more items that are
already in GDK proper and integrate two functions into one.
Instead of first trying to explicitly ask for a WGL 4.1 context, ask for
the WGL context version that matches what is reported via
epoxy_gl_version(), so that we get the maximum WGL version that is
supported by the graphics drivers, and make sure any WGL contexts that
are shared with this (initial) WGL context are created likewise.
We can try to do a default-bog-standard 3.2 core WGL context creation
if the need arises, but let's leave that alone for now.
The EGL context that we are actually creating must have matching OpenGL/ES
versions and allowed GL API set with the previously-created EGL context
that will be shared with it so that they can interoperate together, if
applicable.
This will fix the situation by making sure that we request for the
OpenGL/ES version and OpenGL API set that match with what we have in our shared
EGL context. Otherwise, the newly-created EGL context assumed a OpenGL/ES 2.0
context that supported desktop OpenGL, which may not be what we wanted, such as
in the case of libANGLE.
We are now able to create EGL contexts properly on Windows, but not GLES. This
tries to fix things by doing the following:
* Record the GL context type in a more proper fashion, using an Enum. This
makes things a bit cleaner.
* Force GLES-3.0+ contexts, since libANGLE requires this to properly work with
the shaders-its 2.0 contexts don't work well with our shaders.
Since now we have the shaders working on Windows under GLES with libANGLE using
a 3.0+ context, drop the check to fall back to the Cairo renderer when GLES is
being used.
We only save the size when we transition from floating to fixed, so that
we can restore the size to the one prior to being fixed.
However, we should not restore to this size whenever we see a 0x0 size
from xdg_toplevel, as it can do that any time it doesn't care about the
size, e.g. when the surface is floating and just changing state.
Fix this by only using the saved size when transitioning from fixed to
floating, not when staying floating while previously floating.
Closes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4634
Calling functions inside a g_assert() means those functions will be
compiled out when building with G_DISABLE_ASSERT.
This fixes the release job in the CI pipeline.
Make GtkIMContextIME ignore ASCII control characters just like other
IMContext implementations (e.G. GtkIMContextSimple). Fixes bogus
characters appearing in text input fields (old bug 676077).
The actual code that does the IM context code handling on Windows now uses the
native Windows APIs to handle keystrokes, so this patch is no longer needed, as
it was found that it instead caused issues.
Pointed out in issue #2865.
This reverts commit fd6ce9975e.
gdk_wayland_toplevel_inhibit_idle() contained a contradictory assert
that always fail. More specifically, in the branch that is supposed to
create the idle inhibitor, there is an assertion that it must already
exist and that the refcount must be greater than zero. This causes a
crash on WMs/DEs that use the ZWP idle inhibit manager protocol such as
KDE Plasma and Sway. Fix this by just asserting that the refcount is
zero instead.
GtkBuilder uses GMarkup, which defines a boolean attribute value as:
- yes/no
- true/false
- 1/0
The ITS file for GtkBuilder UI definitions is only using the first pair,
likely because Glade only ever used those values. GTK's own tools, though,
will typically simplify the full yes/no and true/false strings to 1 and 0,
to minimise the parsing time.
Fixes: #4596
Various transforms are normalized with their next transform, and if they
end up being the identity transform will return NULL.
For example a translation by (x,y,z) and followed by (-x,-y,-z) will
result in NULL.
This makes the hotspot of DND surfaces work when using the Vulkan and
OpenGL renderers.
This bumps the CI image used to the newly built image. This is needed to
install a new enough libwayland-client.so needed for wl_surface.offset.
This is done by adding wayland as a meson subproject, building it
on-demand if the version in the system is not new enough. As
libwayland-client.so is pulled in implicitly when linking to gtk4, the
compile step needs LD_LIBRARY_PATH set to make ld find the right library
to link to.
For some users, GetKeyboardLayoutNameA() returns an alias instead of the
fully resolved keyboard layout identifier. In that case, we have to
query the registry to resolve the alias before we can look up the DLL
path.
See comments under https://gitlab.gnome.org/GNOME/gtk/-/issues/4610
Contrary to what you can read on the internet, SGCAPS keys don't work
by having capslock toggle the KBDCTRL bit, they actually have two
consecutive table entries, the first of which is for the normal
version and the second of which is for the capslocked version.
Background: SGCAPS is short for Swiss German caps because Swiss German
was the first layout to use this feature. For keys with the SGCAPS flag,
capslock has a different effect than pressing shift. For example:
Shift + ü = è, CapsLock + ü = Ü, CapsLock + Shift + ü = È
DLL loading failures should not happen under normal circumstances, but
we should at least try not to crash and and print better diagnostic
messages if they do happen.
See https://gitlab.gnome.org/GNOME/gtk/-/issues/4610
Previously, we treated CapsLock and KanaLock as part of the global
keyboard state, much like NumLock and ScrollLock, rather than using
the supplied modifier mask. This was because GDK does not have a
modifier mask for KanaLock, only for CapsLock, so it would not have been
possible to properly support it.
However, this approach ended up causing problems, with certain keyboard
shortcuts not registering when capslock was active. This was first
observed in Inkscape [0] and appears to affect shortcuts consisting of a
single key (like 'a') with no additional modifiers (wheareas shortcuts
like 'ctrl+a' work).
So now we are using the supplied GDK_LOCK_MASK instead, and dropped
support for KanaLock, which we probably don't need anyway (since regular
text input should be handled by the IME input module -- the keymap is
mainly for shortcuts and keybindings, where you don't really want
KanaLock).
[0] https://gitlab.com/inkscape/inkscape/-/issues/3082
The old code used repeated calls to `ToUnicodeEx` to populate
the translation table, which is slow and buggy. The new code
directly loads the layout driver DLLs from Windows.
See https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/4338
Unref private spool_io of GtkPrintJob before setting it to a new one
in gtk_print_job_set_source_file() and gtk_print_job_set_source_fd()
to prevent a leak.
Fixes: #4627
We are no longer using PangoFT2 APIs in this demo, so make sure that we build
it on all builds since we already depend on a HarfBuzz/Pango version that
provide everything that we need here.
Drop the unnecessary pangofc-font.h include as a result.
Like what was done on gtk/language-names.c, acquire the language names via the
native Windows NLS APIs, eliminating a run-time dependency on iso-codes on
Windows.
Instead of relying on the iso-codes package, use the native Windows NLS APIs to
acquire the localized (translated) language names so that we do not need to
incur an extra runtime dependency on Windows. It's not coverering 100% of the
languages that we would like to support through this, but should cover much of
the things that are required.
We aren't really using PangoFT2 for [language|script]-names.c, and are
always using items from them, so make sure they are being built.
Also always include the pangoft2 dependency in gtk_dep if it is found.
We were handling events in the wrong order,
by doing async calls for some of them, but not
for all of them.
And we were not taking into account that GFileMonitors
RENAMED events may or may not move a file on top
of an existing file.
Fixes: #3784
Document the return value and more importantly, specify that a call to
`gsk_renderer_realize()` needs to be matched with a call
`gsk_renderer_unrealize()`.
Prevents issues like https://gitlab.gnome.org/GNOME/gtk/-/issues/4625
Without building the demos, nothing gets installed into $prefix/share/icons/hicolor. Which makes running
`gtk4-update-icon-cache` on the machine causes an error. This is easily reproducible on a Windows machine with MSVC where
there is nothing pre-installed on hicolor icon theme and that makes building gtk without the demos fails with "No such file or directory".
libpng wants to receive samples in either RGB or RGBA order, whether
each sample is big-endian or not. This resolves test failures in
testsuite/gdk/memorytexture.c (and a lot of reftests) on s390x, and
probably the PowerPC family too.
Modifying the test to show the color in use and write out the PNG bytes
to a file, and running the memorytexture test on s390x, produces a PNG
that loads with the correct color values in GIMP (on an x86_64 machine),
which seems like evidence that this is the correct change and not just
compensating errors.
Resolves: https://gitlab.gnome.org/GNOME/gtk/-/issues/4616
Signed-off-by: Simon McVittie <smcv@debian.org>
If we ended up on no output at all, keep the HiDPI scale as is, as it
likely means we were on a workspace that was switched away from. By
keeping the same scale, we avoid unnecessary scale changes that would
otherwise take place if the scale when on monitors would end up being
more than 1.
clang-tidy says:
gtklabel.c:1188:15: warning: Although the value stored to 'mid' is used in the enclosing expression, the value is never actually read from 'mid'
min = mid = text_width;
^ ~~~~~~~~~~
Which seems right since mid will be assigned to at the beginning of the
next loop iteration anyway.
If we have a GDK_ACTION_MOVE, we need to delete the selection. However,
previously this only worked when the drop target and drag source were
different applications, as the selection would get messed up along the
way.
Instead, we stash marks for the duration of the operation so that we can
delete the appropriate selection when completing the move.
That way we ensure that the GL context(s) get disposed, which they
previously weren't due to them still being the current context.
This also implicitly adds testing of gLContext destruction, which
previously wasn't ever done by any test.
The change to use ld and objcopy for resources
had some side-effects: it leaked a few symbols
and made our stack executable. We don't want that.
Use -z nonexecstack and --strip-all to avoid this.
Fixes: #4598
Everything that makes use of gtk_printer_settings_get should be nullable
Because the hashtable might not contain the key and there's no default value provided
Pango may not do this for us, so don't rely on it.
We only show one face with a given name, and we
prefer a variable face over a non-variable one.
The check for variable faces requires new Pango
API that will be in Pango 1.52.
Allowing to tweak the axes of named instances does
not do any harm. If we don't, we have to worry that
we need at least one non-named-instance in the face
list, and make it more obvious how to pick it out.
If the application window is measured with for_size -1 horizontally,
this code clearly passes something lower to the parent class measure()
implementation. Only subtract the menubar_height if we're passed a
for_size > -1.
This has lots of issues:
* It randomly crashes when data is loading while the dnd goes away.
* The data gets randomly reset at the wrong time
* Can't scroll the window on Wayland
* ...
But it's better than nothing, so better get it committed.
After performing an action such as undo/redo, we need to actually scroll
to the position where the operation occurred.
I do note that the scroll here seems to often get invalidated if it is
pages away, and we never make the full scroll. But I've seen this all over
the place elsewhere too and that needs to be handled, most likely, as a
more comprehensive fix for scrolling during line validation.
Related #4575
It's cheap to store the selection position, so always set it even if we
are in a user section. Otherwise, we risk not having the right position
when starting a delete action within a begin_user_action(),
end_user_action() pair.
Related #4575
This adds a test to expose the failure of #4575 which results in the
selection being incorrect when performing a delete as we are likely
already in a begin_user_action()/end_user_action() pair.
Related #4575
We don't need to apply these here, as it will clear the selection which is
needed for the undo. Otherwise we won't be able to test that we end up at
the right selection afterwards.
Instead of just passing major/minor, pass them twice, once for GL and
once for GLES. This way, we don't need to check for GL and GLES
separately.
If something is supported unconditionally, passing 0/0 works fine.
That said, I'd like to group the arguments somehow, because otherwise
it's just a confusing list of numbers - but I have no idea how to do
that.
We want critical GL debug messages to be critical, so that the testsuite
sudokus itself when they appear.
This is relevant in particular for GLES warnings in the GLES runner,
because its warnings can cause crashes on GL drivers less forgiving than
Mesa.
Related: #4571
At last as long as widgets like GtkFlowBox and
GtkGrid still trigger this, it is not a great
idea to have this warning in a stable release.
So remove it for 4.6
When destroying the EGLSurface or GLXDrawable of a GdkSurface, make sure
the current context is not still bound to it.
If it is, clear the current context.
Fixes#4554
We now have a boolean setting that determines whether the high-contrast
theme should be used. Support it by automatically setting the existing
`gtk-theme-name` and `gtk-icon-theme-name` properties when enabled.
With that, it is no longer necessary to change the regular theme settings
for high-contrast, so toggling between high-contrast and a non-default
theme finally works reliably.
Limit the diff region to 30 rectangles (randomly chosen because it
looked big enough to not trigger by accident and small enough to not
cause performance issues).
If the diff region gets more complicated, we abort to the parent node
and use its bounds as the diff region instead and then continue diffing
the rest of the node tree.
Fixes: #4560Fixes: #2396
Functional package managers such as GNU Guix rely on environment
variables such as GI_TYPELIB_PATH to discover the system libraries and
resources; extend rather than override them.
* testsuite/introspection/meson.build (env): New variable that extends
rather than override the GI_TYPELIB_PATH and LD_PRELOAD environment
variables.
(api): Use the above as the value of the 'env' keyword argument.
The introspection tests depend on the pygobject module, but we currently
are not checking if it's available at configuration time, which means we
can get build failures like:
> ModuleNotFoundError: No module named 'gi'
when running the test suite.
Generates a graph visualizing calls to gtk_widget_measure().
Generation of the graph can be slow - like when it forces Pango to wrap
a huge label 1000s of times.
You can dnd the graph to look at it closer or to impress people in
gitlab issues.
It makes sense to connect the begin/update/end events
for touchpad swipes and pinches in a sequence. This
commit adds the plumbing for it, but not backends
are setting sequences yet.
We now require a Pango version that requires Visual Studio 2015 or later to
build, and non-UCRT-based (VS2013) binaries may not bode well with
UCRT-based binaries (VS2015+). Drop the support for VS2013 as a result.
This reverts commit e208e0e07886248d4d86118aa5591c9882f0ed5c.
We run into trouble on X11 if the widgets
in the drag icon have drop targets attached.
Prevent this by suppressing event delivery
to drag icons outright.
We finish the write to the output stream long after the stream has been
closed, so we want to keep the event handler around to do just that.
Instead, remove the handler on finalize.
The OutputStream needs to write a 0 byte end of stream Property. We need
to track if that has been written, and we do that with that new
property.
We also use that property to always request flushes when the stream is
being closed, so that we don't wait for another flush() call.
We need to be very careful when writing data, because if we aren't, sync
functions will be called on the output stream and X11 does not like that
at all.
We were sometimes ending printer enumeration prematurely,
and the code was confused about the meaning of found_printer.
The new setup follows these rules:
- We *only* end the search prematurely if found_printer
is set, which indicates that we found the right printer
- We *always* call find_printer_idle exactly once, and
make it return less than perfect matches like the
default printer, or the first printer we found
Fixes: #4439
If we've already done the tracking into the parent muxer, there is no need
to do it again. This can save a great deal of recursive work when adding
items to the muxer.
This makes showing the context menu in gnome-text-editor repeatedly fast
even as spelling corrections are changed.
It is likely that this could fix#4422 as well.
Fixes#4519
Fixes https://gitlab.gnome.org/GNOME/gnome-text-editor/-/issues/220
When handling action-added callbacks the code was previously using a
freeze_notify/thaw_notify in all cases. This turns out to allocate a
significant amount of memory when called a lot.
That said, it shouldn't be getting called this much but given the current
state of affairs elsewhere in GtkActionMuxer, this brought temporary
allocations down from 9MiB to 9KiB in gnome-text-editor after showing
the context menu a few times.
Related #4422
I saw this coming across through a ffi boundary in Sysprof, and we wanted
to keep most things within GDK using native marshalling to improve
profiler results when frame pointers are not used.
Instead of allocation width for height for width for height or whatever
that code was doing, actually allocate the size we were given or the
requested size, whatever is larger.
Don't just pass on measure() calls, but actually behave in the way we
behave during size allocate.
This should improve cases where GtkScrolledWindow is used with GTK_POLICY_NEVER.
When loading .mp3 files the duration is initially unknown. Before this
change it was reported as a large integer (since GST_CLOCK_TIME_NONE is
-1). Now it's correctly reported as 0.
We want this to take precedence in the wayland platform to other
modules that might be loaded via the IO extension point. None of
those is going to bode well in this platform.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4443
This adds a new row to the Global/Information section which displays the
GTK im-module that is likely to be in use unless changed by an application.
It responds to updates of GtkSettings:gtk-im-module unless the
GTK_IM_MODULE environment variable is set.
Fixes#4512
When returning surrounding context to input methods,
include at least 2 words before and after the insertion
point.
Update the affected input method tests.
For libANGLE to work with our shaders, we must use "300 es" for
the #version directive in our shaders, as well as using the non-legacy/
non-GLES codepath in the shaders. In order to check whether we are
using the GLSL 300 es shaders, we check whether we are using a GLES 3.0+
context. As a result, make ->glsl_version a const char* and make sure
the existing shader version macros are defined apprpriately, and add a
new macro for the "300 es" shader version string.
This will allow the gtk4 programs to run under Windows using EGL via
libANGLE. Some of the GL demos won't work for now, but at least this
makes things a lot better for using GL-accelerated graphics under Windows
for those that want to or need to use libANGLE (such as those with
graphics drivers that aren't capable of our Desktop (W)GL requirements in
GTK.
.. when creating the surface (with the HWND associated with the
newly-created surface) as well as destroying the surface (with NULL,
since the HWND is going to be destroyed), so that we can tie the EGL
calls to the HWND that we want to do the EGL stuff.
If we set the placeholder text before setting a buffer, we end up with
both the placeholder *and* the buffer's contents visible at the same
time.
Fixes: #4376
We use gtk_gesture_get_last_event() underneath at places that need to
work during ::proximity emission. Since GtkGesture only tracks events
while there are button/touch presses involved, this is not going to
bring the right result there.
Use gtk_event_controller_get_current_event() consistently inside,
which always pokes at the event being handled (which is the correct
intent here).
In some circumstances (e.g. activating with a stylus something that
closes a window), we can receive zwp_tablet_tool.proximity_out without
receiving a zwp_tablet_tool.up beforehand.
In those cases, we are not expecting neither .up nor .button, so
reset the stylus device button modifiers on proximity_out.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4103
We are looking up the seat logical pointer modifiers (i.e. the wl_pointer),
not the ones for the tablet tool device. This breaks accounting further
along in GTK leaving stuck implicit grabs.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4102
The idea of within-margin is to scroll as little
as possible to bring the mark within the margins
defined by the factor. The code was achieving
that when scrolling down, but not when scrolling
up. This change makes things symmetrical.
Fixes: #4325
Update the functions that were updated in the previous commit to have all
GdkSurface variables named as 'surface' instead of the GTK-3.x-era window, to
make things more consistent across the board. Also fix formatting a bit.
Make the toplevel surface respond to size computations unless it is just being
created, or maximized, made fullscreen or underwent an AeroSnap operation.
This will ensure that the surface size is properly computed in time, so that
surfaces can be resized as needed.
This will fix issues 3728 and 3799.
On Windows with nVidia drivers at least, when we create a legacy context
via wglCreateContext(), we may still get a (W)GL 4.x context. Allow
such contexts to also use GLSL version 130 instead of 110, so that
things do continue to work.
It turns out we can't just use the size returned
by the memory stream as-is, since it may contain
unfilled garbage at the end, which utf8 validation
will choke on. So, cut it off at the first '\0'
we find.
When the iter is at the end of the buffer,
gtk_text_view_get_iter_location returns a
rectangle with width 0, which in turn makes
gdk_rectangle_intersect return FALSE.
Avoid that by always giving the rectangle
non-empty dimensions.
Fixes: #4503
As far as I can tell, the code here is redundant and probably ended up
this way for historical reasons. A drag surface without
`->is_drag_surface` would be created if `gdk_display_create_surface`
were called with `GDK_SURFACE_TEMP`, but drag surfaces never seem to be
created that way.
In `gtk4-demos`, drag and drop and popovers seem to be working normally
with this.
Setting variations to their default value causes
them to show up in the serialization of the font
description - a font description has no idea about
the default values, so can't filter them out.
Avoid that.
Try to compute a min size that matches the current aspect ratio.
This means that when interactively resizing, we adapt the min size to
the current window area dynamically.
And that means that we always have a min size that is large enough, but
users can interactively cause it to be small-width x large-height,
large-width x small-width or anything inbetween.
Printing the affected widget leads people to assume that it is to blame
for the error. However, the widget is the object the function is being
called on, not the caller. And the caller is doing it wrong.
Usually the caller is the parent widget, so we could print that one, but
it's only usually, it can be an issue propagating from a grandparent and
it doesn't tell you from where the function is called (allocation or
measuring), so you need a debugger anyway.
So don't put anything there instead.
When the stack is homogeneous in only one direction, the other direction
may produce min sizes to small for all children. Make sure to query at
least the min size for those.
If halign=fill, force adjustment to height-for-width.
If valign=fill, force adjustment to width-for-height.
Otherwise look at request mode.
This way we don't try to adapt the filled dimension and only adjust
the one that is not set to fill.
It's not expensive to check it because we'll cache the dfault size
request anyway, and people do it wrong a lot.
As a bonus, don't do any return_if_fail(), just use the min size
instead.
Ensure that we take the DPI scaling into account so that surfaces will
be placed at their correct positions upon an AeroSnap operation on HiDPI
displays.
Also, use the X coordinate of the surface as-is during snap up so that
we do not inadvertently move the surface to the very left. Also fix the
AeroSnap indicator drawing for snap up so that it is drawn at the
correct places.
Since we are updating these functions, make the old GdkWindow-era
variable names to match better the names we use nowadays.
g_log_writer_standard_streams just puts all the logs
out onto stderr and stdout if we don't stop it. Pango
recently grew a bunch of g_debug calls, and those were
now showing up, making all the reftests fail.
When the compose file is a symbolic link, take the link itself's
modification time into account (in addition to its target's) in
determining whether to invalidate the compose cache.
This is useful e.g. on NixOS systems where the compose file might point
to a store path with an irrelevant modification time, and we want the
cache to expire when the symlink itself changes.
This grab-induced crossing event may come from outer means while there are
buttons pressed (e.g. WM window drags/resizes in X11), the implicit active
state should be undone in that situation.
Also, separate the handling of GDK_LEAVE_NOTIFY, as it's fundamentally
different from GDK_TOUCH_END/CANCEL handling.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4416
The sequence should be cancelled from the gesture despite its current state.
Also, there was a piece of pointer emulation that was not dropped here,
maybe breaking things further for the pointer emulated touchpoint.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4387
By adding the `docs_url` key in the project configuration file,
gi-docgen will generate an OpenSearch XML file, which allows to add
docs.gtk.org/<reference> as a "search engine" in web browsers.
I forgot to remove the '-Werror=' part from all the extra warnings, so
the warning/error flags we generated were '-Werror=-Werror=warning-flag'
or 'W-Werror=warning-flag' - but because our compiler flag checking
infrastructure works so nicely, it just ignored these obviously wrong
flags.
Fixes commit 362e91c40b
We only want to determine the size pixel-exact, not pango-unit-exact, so
don't spend lots of time wondering if text is half a pixel or a quarter
pixel wider.
Do kep them for debug and debugoptimized builds though.
Keeping -Werror flags in release builds causes issues with forward
compatibility, when new compiler releases or different toolchains
suddenly cause those warnings to be emitted during compilation.
While we certainly want those issues to be investigated and fixed, they
should not prevent anyone from building GTK until they are.
Resolves#4388
We have a tight coupling with pango, whenever new
pango API appears, our build usually breaks. So
just make our flatpak manifests build pango from git.
Assume a vbox with 2 wrapping labels saying
Hello World
Hi Ho
being measured for their minimum width for 3 rows of text.
This should be layouted like
Hello
World
Hi Ho
and measured accordingly.
However, previously this was layouted as
Hello World
Hi Ho
with 1.5 lines being assigned to both labels.
That will obviously not compute the above wrapping which clearly
results in a smaller min width.
A reftest testing exactly this was included.
... when they are wrong.
Instead, remove them.
Or in other words: GTK4 does not have a fill child property anymore, so
we don't need to run the measuring loop above to determine the size.
This reverts commit cf7fa931d3.
We store the baseline in the cache and we do not know if baselines might
be queried in the future. So always store them.
No reftest because I don't know how to write one.
premature optimization == √😈
Having a short text and a large max-width-chars should request the
natural width of the text, not the limit from max-width-chars.
This caused huge message dialogs.
Reftests added.
Instead of translating font-variant-caps directly
to OpenType features, translate them to a PangoVariant,
now that that enumeration reflects all the css values.
This allows pango to emulate Small Caps for fonts that
don't support the OpenType feature.
When scrolling embedded widgets out of view,
they sometimes get left behind because we don't
reallocated them. To avoid that, move _all_ children
out of view in size_allocate, and let the current
child allocation plumbing move the visible ones
back in place.
Non-root widgets should unrealize their ATContext, if they have one,
when they are unrooted, as they don't have a connection to a top level
any more.
Fixes: #4421
Use the debug envvar 'GDK_DEBUG=gl-egl' to determine whether we want to try to
initialize EGL first before trying WGL, as a means for people to more easily
enable EGL support on Windows to test EGL there (such as to debug the shaders,
for instance)
This will clean up the EGL code in GDK-Win32, as well as fixing crashes caused
by using an invalid EGL context in gdk_gl_context_make_current() as we did not
store up the EGL context in the correct place (lost during the transition to
the common EGL initialization code).
On the Windows/libANGLE side, the initialization of EGL has now fully moved to
the common code in GDK, but we will still default on WGL for now. Help is
really appreciated for fixing the shaders on libANGLE!
We need to ensure that gdk_display_get_egl_display() is available even if EGL
is not enabled in the build, so that things will continue to link and work.
For builds without EGL, just return NULL.
This will port the EGL code in GDK-Win32 to use the common GDK code to
initialize EGL. However, at the current state, although EGL is
correctly initialized, this code is disabled for now since
gdk_gl_context_make_current() fails as the shaders do not work for EGL
via libANGLE on Windows.
We can now clean things up in gdkglcontext-win32-egl.c as a result.
Use a label that is long enough to require wrapping and force it into a
hardcoded width. Use a sentence where all the words have the same size
to not get unwanted wrapping behavior.
Also append a 2nd row to check that the first row gets the proper height
allocated.
Found by Marco Melorio.
The width of a logical rect after line breaking is sometimes not
wide enough to cause line breaking to break at the exact same points.
(Is that by design or a bug in Pango? I don't know.)
So don't use the width, and only relyon values we actually set to
pango_layout_set_width().
Don't just use the natural size as the max size, the natural size
is the ideal size, not necessarily the maximum size.
Also check the nat size for opposite min size.
For size -1 in the opposite orientation, GtkBoxLayout used to measure
the children based on their min size in the box's orientation instead of
-1. That wasn't really intended, but was a side effect of how the sizing
code did (not) distribute extra size above the minimum size.
This is clearly not what we want.
What we want is measuring the orientation as is for size -1. Then we
want to just take the maximum of all children and use that.
A reftest is incldued that ensures a vbox wraps a label just like an
hbox does.
The old code couldn't properly do height-for-width because it only
computed the widest and smallest layout instead of looking at the actual
passed in for-size.
The label-sizing reftest has been adapted as the label code is now smart
enough to always display the whole text and no longer requests a too
small width-for-single-row when wrapping.
Ping/pong serials are not meant to be interpreted as user input serials
(e.g. those given back later to the compositor on grabs). As a matter
of fact, Mutter uses a different count (i.e. timestamps) in these, so
using these serials may confuse the compositor into denying certain
operations like DnD.
This reverts commit ba44e7a228.
The change was meant to revert to old GTK3 behavior but it actually
broke new GTK4 behavior that is in use where max-width-chars is used to
determine an ideal size, but where we don't want to limit the width to
that size.
So what happens is the reintroduction of GTK3-style lots of whitepsace
bugs, and we really don't want those.
We also don't want to break backwards compat if we can avoid it.
So let's revert this.
The reftest that was made for this purpose has been adapted.
Fixes#4399
Instead of using GL_BACK, use GL_BACK_LEFT, because the spec demands
this (many drivers don't).
Also move the call from the GDK backends into the GLContext code, as
this is a generic EGL issue (nvidia being the main driver in need of
this call, see 9c4c4eaaa1 for a longer
discussion).
Fixes#4402
The larger check works well in the headerbar, but not inline in various UI elements. This reverts the larger check since the latter is more common. For selection mode, a separate larger icon (selection-mode-symbolic) has been added to adwaita-icon-theme.
Editors that support configuration through the editorconfig spec:
https://editorconfig.org
should be able to have a subset of the GTK coding style and options
immediately available to them.
Plus, it's better than using relics from the Dark Ages, like modelines.
If a URL can't be loaded, we might end up with a NULL file. Handle that
case properly by creating an invalid image instead and don't crash or
complain to stderr when files are NULL.
This was broken since 0886ade182
A new reftest has been included. We need a reftest instead of a
CSS parser test, because the error only becomes visible when
compute()ing the actual image.
Fixes#4373
gdk_display_create_gl_context only returns NULL when there is
an error set or asserts/aborts. So nullalbe annotation isn't needed.
Similar to 53312cf696
Make it work with the property list as well, handle spinbuttons, adjust
paddings so that buttons don't touch each other, don't override horizontal
padding unnecessarily.
Have square images in the following sizes:
* 20
* 100
* 150
* 200
* 300
and place them in a can-shrink Picture allocated at the sizes:
* 200x100
* 100x200
and set align to center/center.
That's 10 combinations and they should all do the right thing.
This fixes fallout from 3742fabae0 where
we would no longer allocate widgets to their natural size when
align flags where used.
GtkPicture wants to be allocated at 100% in that case, so a picture with
a 100x100 image inside a 200x200 window should be allocated 100x100.
The new adjustment code now does the following (for width-for-height
instead of height-for-width, swap width and height in the following):
1. query the minimum width for the allocated height
2. query the natural width
3. compute the maximum of (1) and (2)
4. set the widget width to the minimum of (3) and the allocated
width.
5. compute the natural height for (4)
6. set the widget height to the minimum of (5) and the allocated height.
But don't call it too early, we only want to call it once we have
prepared the target.
This way, we guarantee that a GL context is always available and that it
is bound to the correct target.
This CSS:
calc(5px+3px)
is wrong because it gets broken to:
calc( 5px +3px )
which is 2 numbers inside the calc, and what you want is:
calc( 5px + 3px )
but you need to add a space to get this, like so:
calc(5px + 3px)
which is the recommended way to write calc() statements.
So whenever we encounter an error, check if the next token is a signed
number and if so, include it in the error message.
This is an alternative to gdk_surface_create_gl_context() when the
context is meant to only draw to textures.
This is useful in the testsuite or in GStreamer or with GLArea,
basically whenever we want to do GL stuff but don't need to actually
draw anything on screen.
A bunch of code will need to be updated to deal with context->surface
being NULL.
in order to make builds reproducible.
See https://reproducible-builds.org/ for why this is good
This was suggested by Matthias Clasen as an alternative to MR !3929
When adjusting allocations, don't query height for the current width,
but query it for the adjusted width.
And adjust width not to the width-for-any-height, but to
width-for-allocated-height.
Even when we have tons of width available, still do the wrapping at
max-width-chars.
This is what happened in GTK3, too, but it happened automatically
because GTK3 did for_size = MIN (for_size, nat_size) and GTK4 does not.
So we do this manually in the label now.
Fixes the label-sizing.ui reftest.
Ideally this would be using box layout, but it overrides measure() so it's
not possible - so reimplement it instead. Fix an accidentally int division
along the way.
Make it use gdk_memory_texture_from_texture().
Also make gdk_memory_format_alpha() privately available so that we can
detect if an image contains an alpha channel.
This is a port of the fix in the quartz backend to the new macOS backend.
From the original commit:
In macOS-12.sdk CGContextConverSizeToDeviceSpace returns a negative
height and passing that to CGContextScaleCTM in turn causes the cairo
surface to draw outside the window where it can't be seen. Passing the
absolute values of the scale factors fixes the display on macOS 12 without
affecting earlier macOS versions.
Don't pass texture + rect, but instead have
gdk_memory_texture_new_subtexture()
and use it to generate subtextures and pass them.
This has the advantage of downloading the a too large texture only once
instead of N times.
Close widget-factory and observe:
Thread 1:
* acquire main loop
* handle close button
* close window
* dispose video and media stream
* stop GstPlayer
WAIT on pipeline stopping
Thread 2:
* prepare next image in pipeline
* hand image to GtkGstSink
* create GdkTexture from image
* gdk_gl_texture_new() determines format
WAIT on acquiring main loop
Sounds like a deadlock?
Indeed, so don't do that.
It does not belong in GdkGLContext, it's a renderer thing.
It's also the only user of that API.
Introduce gdk_gl_context_check_version() private API to make version
checks simpler.
It turns out glReadPixels() cannot convert pixels and you are only
allowed to pass a single value into the function arguments. You need to
know which ones or things will explode.
GL is great.
The resource compiler in the Windows 11 SDK does not allow one to include
winuser.h directly in resource scripts (.rc) with a rather cryptic error
message, so fix generating the .rc file to embed the UAC manifest by including
windows.h with WIN32_LEAN_AND_MEAN instead.
The rc.exe that comes with the Windows 11 SDK does not allow one to include
winuser.h directly in the .rc scripts, so make sure that it is not included
by gtk-win32.rc.body.in, but instead include windows.h with WIN32_LEAN_AND_MEAN
defined.
... if the current locale has a different starting day than Sunday.
This needed 2 fixes:
* We need to take into account `calendar->week_start` when
creating/adding the appropriate `day_name_labels` field
* we were only calculating `calendar->week_start` _after_ attaching the
`day_name_labels`, so it was still set to 0 (the default value).
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/4338
Pass a format do GdkTextureClass::download(). That way we can download
data in any format.
Also replace gdk_texture_download_texture() with
gdk_memory_texture_from_texture() which again takes a format.
The old functionality is still there for code that wants it: Just pass
gdk_texture_get_format (texture) as the format argument.
Broadway is the only GTK+ backend that throws an error on stderr for a
"display server" connection failure.
This causes problems when gtk_init_check() is used and unexpected error
output is generated such as with hotdoc, which fails when generating a
GTK plugin's documentation instead of overlooking the issue.
"Unable to init server: Could not connect: Connection refused"
Broadway is the only GTK+ backend that throws an error on stderr when
failing to initialise, which causes problems when gtk_init_check() is
used and unexpected error output is generated.
This causes hotdoc to fail when generating a GTK plugin's documentation
instead of failing quietly.
"Unable to init server: Could not connect: Connection refused"
Otherwise if we hide and show a window we recreate a new surface,
breaking the compositor's association, but potentially not resend this
data for the new surface.
This matches what we do for input_region.
This is supposed to test the most fallback GL stuff, so we might want to
set even more env vars here.
Also enable the run for the Fedora builder in CI.
Add gdk_gl_context_is_api_allowed() for backends and make them use it.
Finally, have them return the final API as the return value (or 0 on
error).
And then use that api instead of a use_es boolean flag.
Fixes#4221
The only type we have with this prefix is the
deprecated duplicate of gsk_gl_renderer_get_type(),
and including it causes some tests to break.
So skip it.
Before c4a2234a28
menu models could use markup for items and the markup would
be parsed, but this was not intended behavior.
This commit adds official support for using markup
for menu items via the `use-markup` property.
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/4306
Make a deep texture, if the render nodes have
high depth content.
For now, we use 32F here for the deep format,
since using 16F causes small rounding errors
that break the memorytexture roundtrip tests.
Look at the framebuffer and the rendernode to
determine what format to use for intermediate
textures.
Our preference here is to use fp16, if we have it
and it makes sense for the framebuffer we're given.
Add private api to find out if the content
of a render node should be considered 'deep'.
The information is collected at creation time,
so there is no tree-walking involved when we
are using this information in the renderer.
Currently, this comes down to whether there are
any texture nodes with high depth textures in the subtree.
In the future, we may want to allow marking gradient
nodes in this way as well.
For MemoryTexture, this is a simple change.
For GLTexture, we need to query the format at texture creation. This
sounds like a bad idea and extra work until one realizes that we'd
need to do that anyway when using the texure the first time - either
when downloading, or when trying to use it in a rendernode, where we
will soon need that information to determine if the texture prefers high
depth.
The term "hdr" is so overloaded, we shouldn't use them anywhere, except
from maybe describing all of this work in blog posts and other marketing
materials.
So do renames:
* hdr => high_depth
* request_hdr => prefers_high_depth
This more accurately describes what is going on.
Also, now make gdk_memory_convert() the only conversion functions
and allow conversions between any 2 formats by going via a float[4].
This could be optimized via fast-paths, but so far it isn't.
If EGL supports:
* no-config contexts
* >8bits pixel formats
* (optionally) floating point pixel formats
Then select such a profile as the HDR format and use it when HDR is
requested.
Forces request_hdr = TRUE for all requests.
Backends should also use this when choosing whether to honor HDR
requests for low quality compositors - as long as the compositor
pretends to support HDR, shovel HDR at it.
Unify the X11 and Wayland EGL contexts.
This is a bit ugly to implement, because I don't want to create an
interface and I can't make them inherit from the same object, because
one needs to inherit from X11GLContext and the other from
WaylandGLContext.
So we have to put the code in GdkGLContext and make sure non-EGL
contexts can't accidentally run it. This is rather easy because we can
just check for priv->egl_context != NULL.
Quietly export this function mainly for the benefit
of libadwaita, which can can use this to install its
implementation of the gtk-inspector-page extension
point.
We have a global GdkGLBackendType now, just set it.
This way, using the variable forces the backend type, and we don't need
special code handling the env vars in the backends.
It also means setting the env var will now "work" on GDK backends that
don't even support that GL backend and simualte another GDK backend
having registered that GL backend already. So you can run
GDK_DEBUG=gl-wgl gtk4-demo
on test what Wayland will do when WGL is in use.
It is necessary to signal the search engine that we are finished and
that we found something for it to reliably show the results. It would
sometimes work anyway since it is sufficient if any backend signals
completion. However if GtkSearchEngineModel was the only backend
returning results then things would break.
The recent change to faster resource generation
lost the depfiles to ensure that we regenerate
resources when any of the contents change.
Bring it back.
We never put large icons into the icon cache,
so all its items are always atlased, but we do
put large glyphs in to the glyph cache, and we
were never freeing those items, even when they
go unused. Fix that.
Print the extensions one per line, and sort them
alphabetically, so it is actually possible to find
something in the list.
Also print a short description of the chosen config.
Print the extensions one per line, and sort them
alphabetically, so it is actually possible to find
something in the list.
Also print a short description of the chosen config.
Avoid serializing the gresource blob into a C string
and running gcc over it. Instead, use ld to put it
directly into an .o file and add it to the build.
The build system machinations here were copied from
gobject/tests/meson.build, and should ideally be part
of the meson gnome module.
Avoid serializing the gresource blob into a C string
and running gcc over it. Instead, use ld to put it
directly into an .o file and add it to the build.
The build system machinations here were copied from
gobject/tests/meson.build, and should ideally be part
of the meson gnome module.
Avoid serializing the gresource blob into a C string
and running gcc over it. Instead, use ld to put it
directly into an .o file and add it to the build.
The build system machinations here were copied from
gobject/tests/meson.build, and should ideally be part
of the meson gnome module.
We still have links to old gtk-doc references, as well as links to
developer.gnome.org locations that don't exist any more. On the other
hand, we are missing a bunch of links to existing types and symbols.
On Visual Studio-style builds, it is likely that we do not have pkg-config
files for libpng, so improve the search for libpng by using CMake's built-in
mechanisms for looking for libpng. This, however, means that we need to use
'png' rather than 'libpng' for the package name to search for.
Include the appropriate headers as some function prototypes were moved lately.
Also, re-order the include order of the gdk/*private.h headers alphabetically
in the files that were updated.
We don't really need a bus-address property
that gets copied for every single object.
We keep the address in object data on the
display anyway. Just use it from there.
This gets rid of a nice amount of strdups
at startup.
We were only applying <binding> elements when the
object is constructed, which can be triggered by
various things (e.g. a <style> element). Defer
this until we reach </object>, so we can be sure
that we pick up all the bindings.
Testcase included.
Fixes: #4147
The GtkBuilder parser constructs the object e.g.
when handling a <binding> element. There may be
more <property> elements after it, which we were
just not applying. Fix that by always applying
property when we see </object>. To do that, we
need to track the applied status per property.
Test included.
Fixes: #4208
Calling gtk_widget_class_bind_template_child does
*not* give you a reference that you need to unref.
It manages the reference for you. So calling
g_clear_object on such a member is wrong.
Creative people managed to create an X11 display and a Wayland display
at once, thereby getting EGL and GLX involved in a fight to the death
over the ownership of the glFoo() symbolspace.
A way to force such a fight with available tools here is (on Wayland)
running something like:
GTK_INSPECTOR_DISPLAY=:1 GTK_DEBUG=interactive gtk4-demo
Related: xdg-desktop-portal-gnome#5
We want to group in more than one undo group when removing a selection
and replacing it with a new character or characters, unless we're
replacing a single character. In that case, the natural thing is to treat
it as an atomic change.
We don't want to allow new items to be grouped into a previous action
group after the end_user_action() is called. This ensures that we add a
barrier action in those conditions.
Fixes#4276
On Windows, GLES is not that widely available unless one installs wrapper
libraries such as libANGLE, so GLES/EGL support on Windows is used more like
a fallback mode if Desktop OpenGL (WGL) support is inadequate on the system.
Hence, unless one forces WGL or EGL, we will first try to initialize WGL, and
then try to initialize GLES if enabled and if WGL initialization failed, and
then just return whatever the last result we can obtain from these
initialization attempts, since unlike X11 EGL contexts, we do not have
separate modes for WGL except for legacy and non-legacy contexts.
We were setting the WGL pixel format in GdkWin32Display too early, so the code
does not bail out correctly when we retry establishing the WGL context.
Fix this by pushing back setting the WGL pixel format only after it passes the
shader availability check.
Should fix issue #4257.
When pressing the keyboard arrows to move around when the insertion point is
hidden, it causes an assertion error in blink_cb.
Insertion point blinks should only be scheduled when blinking is enabled and the
insertion point is visible.
Closes#4275
This change removes the assertions limiting replacement strings in the compose table to be less than 20 characters.
The limit seems arbitrary, is not required, will break some users' setups, and problems with it result in applications not launching.
Fixes#4273
The gtk_window_set_buildable_property implementation
was only used to set the unused builder_visible flag.
Remove both the flag and the vfunc.
This means we no longer have any set_buildable_property
implementations and could eventually drop that vfunc and
the support for it in GtkBuilder.
Add a private GdkPaintable implementation that
loads a texture in a thread, and does not show
anything until the texture is loaded. This avoid
blocking on image loading in the main thread.
Silly optimization to get rid of
gtk_main_do_event
gtk_inspector_handle_event
gtk_inspector_window_get_for_display
g_object_get_data
showing up in profiles even though it's useless since we've never even
created any inspector window in the first place.
gtk_file_chooser_widget_get_choice() is supposed to return the option
id of the choice, but it currently is returning the option label.
Return the option id instead.
When choices are added to the file chooser widget, the options of
that choice are stored object data under the "options" key. However,
gtk_file_chooser_widget_set_choice() was checking for "choices".
Retrieve the options from the "options" key stored data object data.
This reverts commit 87af45403a.
I've found that this change is needed to ensure that the
bounding boxes of text nodes encompass all the glyphd drawing.
Without it, we overdraw the widget boundaries and cut off
glyphs.
We are rendering the glyphs on a larger surface,
and we should avoid introducing unnecessary
rounding errors here. Also, I've found that
we always need to enlarge the surface by one
pixels in each direction to avoid cutting off
the tops of large glyphs.
We can't have other test pop up windows, and possibly
stealing focus and preventing us from getting data
offers. So, run the clipboard test in isolation.
For 2D transforms, we can read the scale
factors more directly off the matrix.
This should eventually be moved out into a
function to decompose a 2D transform into
scale + rotation + skew + translation.
Since we report width and height as integers, the
default implementation of this introduces rounding
errors. This shows up in the node-editor, as having
uneven scale factors like sx=1.0 and sy=1.0035.
Text nodes don't handle uneven scales like that well
and overdraw.
1. Change INSUFFICIENT_MEMORY to TOO_LARGE
GTK crashes on insufficient memory, we don't emit GErrors.
2. Split UNSUPPORTED into UNSUPPORTED_CONTENT and UNSUPPORTED_FORMAT
So we know if you need to find an RPM with a loader or curse and
the weird file.
3. Translate error messages, they are meant for end users.
We were going via GLoadablieIcon/GInputStream for everything previously
and we have no API for that with GdkTexture.
With this commit, gdk-pixbuf isn't used anymore when starting
widget-factory for anything but SVG.
When loading, convert all >8-bit data to
GDK_MEMORY_R16G16B16A16_PREMULTIPLIED.
When saving, save all 8-bit formats as 8-bit RGBA,
and save all >8-bt formats as 16-bit RGBA.
Use our own loader to (de)serialiaze textures
to and from png and tiff.
We still fall back to gdk-pixbuf for handling all
the other image formats, and for pixbufs.
This is a companion to gdk_texture_save_to_png, using
the tiff format, which will let us avoid lossy conversion
of HDR data, since we can store floating point data.
Add support for the tiff format, which is flexible
enough to handle all our memory texture formats
without loss.
As a consequence, we are now linking against libtiff.
Using libpng instead of the lowest-common-denominator
gdk-pixbuf loader. This will allow us to load >8bit data,
and apply gamma and color correction in the future.
For now, this still just provides RGBA8 data.
As a consequence, we are now linking against libpng.
Color values must be divisible by 15 to be convertible into U8 and U16
values with the same result. 0x80 is not one of these values, so switch
it to 0x99.
We avoid an offscreen if we know the child node
can 'handle' the transform. Shadow nodes can if their
child node does - either the child node is a text node
in which case the shortcuts we take for shadow nodes
will work fine with the transform (we just render the
text node offset), or the child is not a text node,
in which case we render the shadow to an offscreen
anyway.
This change makes the label-shadows reftest pass with
the GL renderer, not by fixing the issue but by avoiding
it.
For shadow nodes, we try pretty hard to avoid
rendering shadows, and and we have a shortcut
that just renders text offset, but we can try
harder to do nothing - if the text is offset
by zero, we don't need to draw it at all.
Tests that overdrawing of content inside an opacity node happens before
the opacity is applied.
This is broken in the GL renderer and causes the opacity.ui reftest to
fail.
We need to use an offscreen whenever there is overlapping
children somewhere in the tree below, just checking the
direct child of the opacity node is not enough.
Fixes: #4261
This also switches the rendering code from using gsk_render_node_draw()
to gsk_renderer_render_texture().
Some tests are broken with the GL renderer, so this patch forces the
Cairo renderer until they get fixed.
The test used to test that GtkBox ordered it's children left-to-right in
CSS, no matter the text direction or pack-type.
But there is neither a pack-type anymore nor does GTK4 do that.
So that test has been broken for yers, it just didn't render anything
wrong.
GLES only allows downloading float if the texture matches specific
criteria and I'm too lazy to determine them, so always fall back.
And the custom stride fallback code isn't necessary, because falling
back does exactly that step already.
Basically, I was building some packages on Guix. I figured out that
wayland-protocols was listed among propagated-inputs for gtk+ package
(gtk-3-24). propagated-inputs holds a list of runtime dependencies,
that should be available to any other package that depends on gtk+.
While discussing we clarified that wayland-protocols is not runtime
dependency. So I moved it to native-inputs of gtk+ package, which
means that, this dependency will be available only to gtk+ package and
only at build time. Once moved, building of other applications that
depening on gtk+ started to fail.
Investigation showed that, all .pc (pkg-config) files prepared by gtk+
package, was including:
Requires.private: ... wayland-protocols ...
Since it becomes requirement, other applications was failing with
missing dependency wayland-protocols of dependency gtk+, for instance:
-- Checking for module 'gtk+-3.0'
-- Package 'wayland-protocols', required by 'gdk-3.0', not found
While actually wayland-protocols is not even a build time dependency
of application that depends on gtk+. Advertisement of such
requirement, is a bit misleading, because one does not need it at
runtime, especially applications based on gtk.
Remove the mention of GNU (since that has not been case
for a long time, effectively), state that GTK is hosted
by the GNOME project, and point to GNOME as a place
for donations.
Up until now, as the focus was moved to the inner button, it was not possible for
assistive technologies to determine the correct labels and descriptions
because developers could set them only for the parent widget.
Now, the proper relations are added so the labels should be picked up properly.
Fixes#4254
This makes sure that the `GListModel` returned by
`gtk_stack_get_pages()` actually has the items removed before
`items-changed` is emitted.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4255
This happens in the real world when using the inspector to look at a
node recording of a GStreamer video while the video is still playing.
GStreamer will use the GL context in a different thread while we are
busy trying to download it.
A test is included.
Use a GL renderer to upload textures (and then optionally download them
via release() again). This way, we can test that the GL renderer
properly uploads textures to the right formats (not losing information
for HDR for example) and downloads them again.
1. The download via gdk_cairo_draw_from_gl() was broken sometimes
2. We get easy conversion on fallback by chaining up and using
download_texture().
3. One more place where Cairo is no longer necessary.
1. It avoids Cairo, and in particular conversion to Cairo.
2. Keeping a texture allows easy chaining in the vfuncs.
3. Using a texture means releasing will work for HDR formats
too, once we add them.
A private vfunc that downloads a texture as a GdkMemoryTexture in
whatever format the texture deems best.
There are multiple reasons for this:
* GLES cannot download the Cairo format. But it can download some
format and then just delegate to the GdkMemoryTexture implementation.
* All the other download vfuncs (including the ones still coming) can
be implemented via download_texture() and delegation, making the
interface easier.
* We want to implement image loading and saving support. By using
download_texture(), we can save in the actual format of the texture.
* A potential GdkCompressedTexture could be implemented by just
providing this one vfunc as a compress() step.
We need to invalidate the Pango contexts when
font settings change. Use the new helper
gtk_widget_update_pango_context to make it less
likely that we forget to update some things.
The cairo_t that we create to render glyphs for
the glyph cache needs to match the font options
that are supposedly governing how glyphs are
drawn.
Since we allow font options to be different per
widget in gtk, we need to have them at least at
the level of individual render nodes. Adding them
to the lookup key for the glyph cache has the
side effect of solving another problem: We are
not flushing the cache when font options change.
Since font options affect how the glyphs get rendered,
we need to pass the font options down from the gtk level
to where the glyph cache is populated.
Add a new gsk_text_node_new_full api that takes a
cairo_font_options_t in addition to the other parameters.
If the alpha channel is zero, it doesn't matter what the values of the
red, green and blue channels are: the pixel is still fully transparent.
On most architectures, fully transparent pixels end up all-zeroes
(fully transparent black), matching what's in the reference PNG file;
but on mips*el the blend-difference and blend-normal tests get all-ones
(fully transparent white) and a test failure.
Resolves: https://gitlab.gnome.org/GNOME/gtk/-/issues/4227
Signed-off-by: Simon McVittie <smcv@debian.org>
This lets people switch back to font rendering that is closer
to what GTK 3 does. It is not perfect - subpixel antialiasing
is not going to work. But it give us an Escape hatch while
we shake out the bugs in our linear layout.
Related: #3787
Make it clear that your class must have all the editable properties
already before you call the (confusingly named) function
gtk_editable_install_properties.
This adds support for sequences like <Compose>,G,u -> capital G with
breve. Previously, only a capital U was accepted for E, G, I and O
(but a lower-case u was accepted for A and U for some reason).
Signed-off-by: Simon McVittie <smcv@debian.org>
The GtkComposeTable cache is always in big-endian format and is
byteswapped on load for the more common little-endian CPUs, but
init_builtin_table() in GtkIMContextSimple can't byteswap the built-in
data without copying it, which is undesirable. Pregenerate both big-
and little-endian compose data, and compile the correct flavour into
each build of GTK. This fixes failure of the composetable test when
building for a big-endian architecture such as s390x and (traditional,
big-endian) powerpc.
Resolves: https://gitlab.gnome.org/GNOME/gtk/-/issues/4217
Signed-off-by: Simon McVittie <smcv@debian.org>
Dragging will just drag the render node.
Dropping will replace the current contents of the textview with the
dropped node.
Neat side effect: You can drag the node onto itself to do a
deserialize/serialize of the current text.
This is needed as GskRenderNode is its own fundamental type and has its
own GValue infrastructure. And I want to put render nodes into the
clipboard which uses GValues.
It seems these are sent with `xwindow` set to the root window, so this
was failing to find a surface and get the screen from that.
I'm not sure if there's a reason not to get the screen this way
elsewhere in the function, but it seems this should be correct.
This fixes the behavior of `gdk_x11_display_get_monitors()`, which
wasn't correctly changing when monitors were added or removed. For
instance, this python code was always showing the same number of
monitors when one was turned off and on, but updates correctly with this
change applied:
```python
import gi
gi.require_version("GLib", "2.0")
gi.require_version("Gdk", "4.0")
gi.require_version("Gtk", "4.0")
from gi.repository import GLib, Gdk, Gtk
def f():
print(len(Gdk.Display.get_default().get_monitors()))
return True
GLib.timeout_add_seconds(1, f)
GLib.MainLoop().run()
```
The clang build fails due to -Werror=implicit-fallthrough being
on by default and some fallthrough cases not being marked as such.
Use G_GNUC_FALLTHROUGH or duplicate the code in those cases.
Claim the pressed, released and canceled gestures
meant for the expander-icon of the TreeExpander.
This avoids selecting the row when expanding or collapsing it.
Closes#4199
The installed ITS rule filename is "gtk4builder.its". The .loc file
is wrongly pointing to old "gtkbuilder.its" which makes gettext fail
on systems without GTK3 installed.
_gdk_macos_event_source_new() calls g_source_set_static_name(), which
for GLib versions before 2.69.1 is a macro defined in gdk-private.h.
Fixes#4195
modified: gdk/macos/gdkmacoseventsource.c
Goals:
1. Provide as much information as possible in the error message, so
users can try to fix their system themselves.
2. Try to formulate the error message in a way that explains that this
is not something GTK can fix, but a lower layer problem.
Related: #4193
We can use the new binding helpers to make this
a little less bothersome. That way, it will need
tweaks less often (only when new fundamental types
are introduced).
The child of a GtkExpander is owned directly by the expander whenever
the "expanded" flag is unset.
We are adding an additional reference to the child of an expander when
expander is not expanded.
Additionally, if a GtkExpander is disposed while not expanded, we need
to explicitly release the reference on the child widget that we own.
This reference leak was masked in GTK3 by GtkContainer removing each
child from the parent container by recursively calling
gtk_widget_destroy().
This gesture is set on the whole widget surface, since there's
multiple input targets inside an entry (icons, the GtkText itself)
it makes sense to consider the full entry an area handling clicks.
Ensure these events don't propagate further up, and result in other
actions.
The default theme changed from Adwaita to Default and this tripped up
the logic to detect if the tarball builds contain pre-built css files or
not. Fix this by looking at pre-compiled css files in themes/Default/
instead of themes/Adwaita/.
This gesture is set on the whole widget surface, since there's
multiple input targets inside an entry (icons, the GtkText itself)
it makes sense to consider the full entry an area handling clicks.
Ensure these events don't propagate further up, and result in other
actions.
This gesture is set on the whole widget surface, since there's
multiple input targets inside an entry (icons, the GtkText itself)
it makes sense to consider the full entry an area handling clicks.
Ensure these events don't propagate further up, and result in other
actions.
All possible ramifications after button1 press (move cursor,
begin drag, begin dnd, select word/line, ...) result in user
actions. The right thing after that is consuming the events,
set the gesture state for that.
The font sizes demo had the space between the font-size spans,
causing us to have a run with just a default sized space between
the words, which in turn leads to wobbly cursor sizes. Avoid that
by including the space in the preceding span.
Also, make it bigger.
Update all the places where we switch over
PangoAttrType to handle PANGO_ATTR_TEXT_TRANSFORM,
and do nothing for now - text-transform support
will land in 4.6.
The old code was just pasting local clipboard data that we put there
ourselves and was causing criticals on remote clipboard data. Now the
code does the proper async paste.
When we initialize OpenGL, check whether we have OpenGL 2.0 or later; if not,
check whether we have the 'GL_ARB_shader_objects' extension, since we must be
able to support shaders if using OpenGL for GTK.
If we don't support shaders, as some Windows graphics drivers do not support
OpenGL adequately, notably older Intel drivers, reject and destroy the GL
context that we created, and so fallback to the Cairo GSK renderer, so that
things continue to run, albeit with an expected warning message that the GL
context cannot be realized.
Also, when we could not make the created dummy WGL context current during
initialization, make sure that we destroy the dummy WGL context as well.
Fixes issue #4165.
With gtkmm, when using `Application()`, the display is initialized
before we know the application name and therefore, the program class
associated to the display is NULL.
Instead of providing a default value, we set it equal to program name
when NULL. Moreover, we give up on capitalizing the class name to keep
the code super simple. Also, not using a capitalized name is
consistent with `gdk_x11_display_open()`. If someone has a good reason
to use a capitalized name, here is how to do it.
```c
class_hint = XAllocClassHint ();
class_hint->res_name = (char *) g_get_prgname ();
if (display_x11->program_class)
{
class_hint->res_class = (char *) g_strdup (display_x11->program_class);
}
else if (class_hint->res_name && class_hint->res_name[0])
{
class_hint->res_class = (char *) g_strdup (class_hint->res_name);
class_hint->res_class[0] = g_ascii_toupper (class_hint->res_class[0]);
}
XSetClassHint (xdisplay, impl->xid, class_hint);
g_free (class_hint->res_class);
XFree (class_hint);
```
Fix eff53c023a ("x11: set a default value for program_class")
It is basically not used by default and is pretty much broken at this point, so
it's about time to drop it.
Let's focus on fixing the OLE2 DnD protocol.
Same thing as the previous popovermenu commit, except for the base popover
because the popovermenu needs special behaviour with e.g. sides arrow so
we need to have the "cycle around" for regular popovers here too.
Currently when moving the focus with (Shift+)Tab, it also traverses the window's
widgets, although it would be expected that the focus stays within the popover,
as it's (almost) like it's a separate window. This would be consistent with
the behaviour of the Up/down arrows, which do cycle around the focus once it
reaches the end.
So this commit makes the popovermenu cycle around focus in any direction, apart
from left/right because they are used to open and close submenus and it wouldn't
make sense anyway to cycle horizontally as there's usually only one widget per
line.
Long time ago, Cairo shadows in both GTK3 and 4 were drawn at a size about
twice their radius. Eventually this was fixed but the shadow extents are
still calculated for the previous size and appear unreasonably large: for
example, 141px for a 50px radius shadow. This can get very noticeable in
places such as invisible window frame which gets included into screenshots.
https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/3419 just divides the
radius by 2 when drawing a shadow with Cairo, do the same when calculating
extents.
See https://gitlab.gnome.org/GNOME/gtk/-/issues/3841
Our compose table format is still limited to 16bit
values for keysyms, but what we see in key events
can be 32bit values, and we treat them as such now.
Fixes: #4149
It broke keyboard focusing any widget added through the custom widget
menu feature. So for example if you put e.g. a custom check box widget
in a menu, you won't be able to focus it.
This is because the gizmo is mostly used to custom drawing with e.g.
CSS for small visual elements like scale markers. That's probably why
gizmo's default focus overrides block the focus from going through
the children. So this commit fixes it by overriding those and passing
the focus through the children.
Slide animations cause changes in the size requests due to the
behavior of GtkRevealer. We can avoid those by using cross-fades, which
don't have that problem.
Besides, cross-fades look better anyway.
harfbuzz has all the information we need, so we
can avoid poking directly at freetype apis. Also
drop the caching of color glyph information until
it turns out to be a problem.
Instead of havoing a label for the video frame that clashes with the
background of the video, add a frame around the text styles box and add
a label for them. As a side benefit, it also makes it more obvious that
it is scrollable.
Note: Most of this patch is just reindenting.
Currently we update the :active property on both the previous and
new focus button. That "visually activate" the button and will
emit ::toggled, but if the button is associated with an action,
the action state won't change.
Fix that by activating the new focus instead of explicitly fiddling
with the :active property.
Remove the limitation on the number of dead keys
that we match, and allow the result be be multiple
characters.
Regenerate the builtin sequences, since this changes
what dead key sequences we can reproduce algorithmically.
Update tests to match.
Fixes: #10
Make gtk_check_algorithmically take a GString
for the result. This is in preparation for allowing
multi-character results here, in the future.
Update all callers.
Delegating the action to the compositor not only improves consistency
with server-side decorations, but also allows for actions that aren't
available client-side (like lower-in-middle-click).
https://gitlab.gnome.org/GNOME/mutter/-/issues/602
Look who changed his mind since commit 8e2ffb3b46 :-)
The "call" scope means that the callback is only used during the
function call itself (here: gtk_widget_class_install_action()).
That's clearly wrong here, as the callback is invoked every time
the action is activated.
Arguably the "notified" scope is a better match here, where the
lack of a GDestroyNotify parameter suggests that the callback may
be used forever (which is the case here).
Related: #3498
This way, it can be set in GtkBuilder.
Also make sure to only ever look at the GTypes set in the formats, as
GtkDropTarget cannot deal with mime types.
Now, we just print a whitespace-separated list of GTypes and mime types.
This makes this neat for 2 things:
1. Parsing it (see next commit)
2. Using it in GtkBuilder (see commits after that)
In particular, the common case of supporting a single GType (or mime
type) looks like just printing the GType (or mime type), which in
GtkBuilder looks like
<property name="formats">GdkTexture</property>
Usually the "dnd-finished" signal will be used to unref the GdkDrag. In
those cases, we would lose the object, so that when we do the final
drag_drop_done() afterwards, we wouldn't have a remaining reference.
With the reference guard, this now works.
Since UCKeyTranslate() converts these keys to Space key unexpectedly,
applications can't distinguish these keys by keysyms.
To solve it, this fix translates these keys by the same way with
function keys & keypad keys.
This patch is equivalent to the patch proposed in:
https://bugzilla.gnome.org/show_bug.cgi?id=702841Closes#4117
It is good practice for (floating) window managers to respect explicit
position hints from clients (as long as the window wouldn't end up
off-screen etc.).
Before commit 13d3afa56e, GTK had a flag for setting the PPosition hint,
but now does so unconditionally. However the real intention is to *not*
request a fixed position, so don't do that.
Currently we use layout coordinates and widget height when determining
where a click or drag has happened. If the widget has top padding (which it
does inside a GtkEntry, for example), the area where it's possible to select
text is shifted down, so the part of GtkText above the layout is not counted
as the draggable area and instead the equal area below the widget is counted.
Since GtkText is always single-line, there's no need to do any of that and
we can use widget coordinates. Then the draggable area matches the widget
and the problems goes away.
The dummy Win32 window that we use to capture display change events and
to create dummy WGL contexts was created with CS_OWNDC, so we really do
not need to (and should not) call ReleaseDC() on the HDC that we
obtained from it, so drop these calls.
Since the shaders need to be updated for using with GLES (libANGLE at
least), default to WGL for now. Unfortunately it is not that common for
Windows to have GLES support, in which the easiest way to obtain such
support is via Google's libANGLE.
It turns out that the problem of the WGL window not drawing was due to
the fact that I messed up where I placed SwapBuffers() during the
conversion... doh:|
At the same time, stop storing the HDC in the GdkWin32GLContextWGL, but
instead always create it along the surface we created, so that it is ready
for use for operating with WGL when we are not dealing with "surfaceless"
contexts. If we are dealing with "surfaceless" contexts, just use the
HDC of the dummy window that we created when we created the
Gdk(Win32)Display.
WGL contexts should now be in working order at this point.
This commit attempts to split GdkWin32GLContext into two parts, one for
WGL and the other for EGL (ANGLE), and attempts to simplify things a
bit, by:
* We are already creating a Win32 window to capture display changes,
so we can just use that to act as our dummy window that we use to
find out the pixel format that the system supports for WGL. We also
use it to obtain the dummy legacy WGL context that we will always
require to create our more advanced Core WGL contexts.
* Like what is done in X11, store up the WGL pixel format or the
EGLConfig in our GdkWin32Display.
* Ensure we do not create the dummy WGL context unnecessarily.
In this way, we can successfully create the WGL/EGL contexts, however
there are some issues at this point:
* For WGL, the code successfully initializes and realizes the WGL
Contexts, but for some reason things became invisible. When running
gtk4-demo, this can be verified by seeing the mouse cursor changing
when moved to spots where one can resize the window, although they
were invisible.
* For EGL, the code initializes EGL but could not realize the EGL
context as shaders failed to compile. It seems like the shader issue
is definitely outside the scope of this MR.
nvidia sets the default draw buffer to GL_NONE if EGL contexts are
initially bound to EGL_NO_SURFACE which is exactly what we are doing. So
bind them to GL_BACK when drawing, as they should be.
See https://phabricator.services.mozilla.com/D118743 for a discussion
about EGL_NO_CONTEXT and draw buffers.
This way, one can force using WGL on Windows even if EGL support was
enabled. Also update the help text for gl-egl as it will apply for
Windows, albeit a bit later.
This has the benefit that we can refactor it and make sure we deal with
GdkDisplay::init_gl() not being called at all because
GDK_DEBUG=gl-disable had been specified.
It's not used there, but both backends have independent
immplementationgs for it.
I want to get rid of GdkGLContextX11 and moving code from it is the
first step.
Now that we have the display's context to hook into, we can use it to
construct other GL contexts and don't need a GdkSurface vfunc anymore.
This has the added benefit that backends can have different GdkGLContext
classes on the display and get new GLContexts generated from them, so
we get multiple GL backend support per GDK backend for free.
I originally wanted to make this a vfunc on GdkGLContextClass, but
it turns out all the abckends would just call g_object_new() anyway.
Instead of
Display::make_gl_context_current()
we now have
GLContext::clear_current()
GLContext::make_current()
This fits better with the backends (we can actually implement
clearCurrent on macOS now) and makes it easier to implement different GL
backends for backends (like EGL/GLX on X11).
We also pass a surfaceless boolean to make_current() so the calling code
can decide if a surface needs to be bound or not, because the backends
were all doing whatever, which was very counterproductive.
The code to create and manage a fake egl surface to bind to is
complex and completely untested because everyone seems to support this
extension.
nvidia and Mesa do support it and according to Mesa devs, adding support
in a new driver is rather simple and Mesa drivers gain that feature
automatically, so all future drivers shoould have it.
... or more exactly: Only use paint contexts with
gdk_cairo_draw_from_gl().
Instead of paint contexts being the only contexts who call swapBuffer(),
any context can be used for this, when it's used with
begin_frame()/end_frame().
This removes 2 features:
1. We no longer need a big sharing hierarchy. All contexts are now
shared with gdk_display_get_gl_context().
2. There is no longer a difference between attached and non-attached
contexts. All contexts work the same way.
Do not treat the context as already current when the value
of context::in-frame changes.
This is so we can bind to EGL_NO_SURFACE if context::in-frame == false
and to context::surface if context::in-frame == true.
This allows getting rid of the attached property in future commits.
The vfunc is called to initialize GL and it returns a "base" context
that GDK then uses as the context all others are shared with. So the GL
context share tree now looks like:
+ context from init_gl
- context1
- context2
...
So this is a flat tree now, the complexity is gone.
The only caveat is that backends now need to create a GL context when
initializing GL so some refactoring was needed.
Two new functions have been added:
* gdk_display_prepare_gl()
This is public API and can be used to ensure that GL has been
initialized or if not, retrieve an error to display (or debug-print).
* gdk_display_get_gl_context()
This is a private function to retrieve the base context from
init_gl(). It replaces gdk_surface_get_shared_data_context().
Create it during init and then reuse it for all contexts.
While doing that, also improve error reporting - that's not used yet but
will in later commits.
This is not used yet, but it allows surfaceless GL contexts.
For that purpose, we need to make the display a construct-only property,
so that it can be set when the surface isn't.
This adds a bunch of very picky checks in the constructor so nothing bad
can happen.
... and move some members from the GdkDisplay struct.
We've always wanted to add one to isolate the display from the backends
a bit more, but so far it's never happened.
Now that I'm about to add more data to GdkDisplay, it's a good excuse to
start.
We try EGL first, but are very picky about what we accept.
If that fails, we try to go with GLX instead.
And if that also fails, we try EGL again, but this time accept anything.
The idea here is that EGL is the preferred method going forward, but GLX is
the tried and tested method that we know works. So if we detect issues with
EGL, we want to avoid using it in favor of GLX.
Also add a GDK_DEBUG=gl-egl option to force EGL at all costs and not try
GLX.
That way, we can give a useful error message when things break down for
users.
These error messages could still be improved in places (like looking at
the actual EGL error codes), but that seemed overkill.
Query the EGL_VISUAL_ID from the egl Config and select a config with the
matching Visual.
This is currently broken on Mesa because it does not expose any RGBA
X Visuals in any EGL config, so we always end up with opaque Windows.
https://gitlab.freedesktop.org/mesa/mesa/-/issues/149
This reverts commit c35a6725b9.
This approach doesn't work because if NVIDIA doesn't work for EGL, the
EGL implementation won't be provided by NVIDIA, so checking the vendor
doesn't work.
Instead, use the display's "leader surface" when no surface is required,
because we have it lying around.
Really, we want to use EGL_NO_SURFACE, but if that's not supported...
Instead of going via GdkVisual, doing a preselection and letting the GL
initialization improve it, let the GL initialization pick an X Visual
directly using X Visual code directly.
The code should select the same visuals as before as it tries to apply
the same logic, but it's a rewrite, so I expect I messed something up.
1. We're using EGL most of the time anyway, so if we wanted to cache
things, we'd need to port it there.
2. Our GL handling is massively configurable, so determining when to use
the cache and when not is a challenge.
3. It makes startup nondeterministic and depend on whether a GTK4 app
has previously been started on this display and nobody thinks about
that when debugging.
4. The only benefit of the caching is delaying GL initialization - which
made sense in GTK3 where almost no app used GL but doesn't make sense
in GTK4 where almost every app uses GL.
So unless I find a big benefit to reintroducing it, this cache will be
gone for good.
Avoids having to use private data, though the benefit is somewhat
limited as we still have to put the destructor in the egl code and can't
just put it in gdk_surface_x11_finalize().
We only have one config, because we use the same Visual everywhere.
Store this config in the GdkDisplayX11 struct for easy access.
Also do this on initialize, because if creating the config fails, we
want to switch to GLX instead of failing to do GL at all.
This also simplifies a lot of code as we can share Visual, Colormap, etc
across surfaces.
There's no need to use g_object_set_data() for it.
We can also stop caching it elsewhere because we know the display has
it.
And finally, we can remove the display->have_egl boolean and use
display->egl_display != NULL instead. We initialize the display at
startup, so that variable is the perfect indicator.
We need to initialize GL to select the Visual we are going to use for
all our Windows.
As the Visual needs to be known before we know if we are even gonna use
GL later, we can't avoid initializing it.
Note that this previously happened, too. It was just hidden behind the
GdkScreen initialization.
We don't want to bind ourselves to GTK3 - both because we don't want to
accidentally cause bugs in a different codebase and because we want to
deviate from it.
While doing so, also store visuals as visuals and not as integers.
And only store one Visual because GTK4 only uses one visual.
And then remove the code that is leftover for dealing with the
compatibility Visual for GTK3.
PS: I'm kinda proud of my STRINGIFY_WITHOUT_BRACKETS hack.
The old code was ordering visuals by depth, but considering that these
days we either use the default visual or a 32bit RGBA visual, that
reordering does not have an effect anymore.
In theory, the only effect is that the GLX Visual selection might select
a different replacement Visual when it checks for improved GL Visuals, but
even there I can't come up with a case where that matters, because
again, the visuals are only reordered by depth and we want to keep the
depth.
In any case, make this a separate commit so bisecting can find this
problem if it ever shows up.
Instead of the display telling the screen to tell the visuals to tell
the display to initialize itself, just init the display directly.
What a concept.
That's a sneaky trick so my edit/compile/test cycle goes faster:
I usually use demos for testing so the tools don't have to be linked for
me to start testing.
If the pointer capability is added, pointer swipe and pinch gestures
will be created. However, if the pointer capability is removed, the
gesture objects won't be destroyed.
If the pointer capability is removed and added several times in a row,
for example due to plugging and unplugging physical mouse, this can lead
to leaking the old gesture objects.
In order to prevent that, this change makes the seat destroy swipe and
pinch gestures when the pointer capability is withdrawn.
It's only used during DND to allow use of the root window's cow window
as a DND target, because apparently gnome-shell used to think that was a
great idea to DND to the overview.
Somebody complain to gnome-shell devs about it not being a good idea if
they want it fixed.
Potentially using Wayland is a better idea though.
This reverts 85ae875dcb
Related: https://bugzilla.gnome.org/show_bug.cgi?id=601731
It's not 2011 anymore, and we shouldn't randomly build one of 10.000
different combinations of X11 backends (I counted the possibilities) but
exactly the one we expect people to use.
Instead, ensure that sassc is made madatory on git builds (because
it is, we don't ship CSS files anymore) and not even looked for in
release builds (because do ship CSS files there).
We don't want people to build Vulkan support when they just want to get
GTK built.
This is in particular true for GTK as a CI subproject or for people
using jhbuild.
Worse, just having Vulkan support compiled in tends to cause crashes
in the Inspector, even if you are not using it.
GTK supports webm playback, which means a backend should always be
compiled.
The ffmpeg backend however is incomplete (no audio) and as such, we
don't want people to end up with it accidentally.
Since we don't want to drag an entire gstreamer build into our ci
on MacOs or msvc, explicitly disable the gstreame media backend there.
Set all settings to their default values, so we
are less dependent on the environment to be set
up just right. In particular, this fixes animations
being disabled when we happen to run in a vm.
Make _gdk_win32_display_get_monitor_scale_factor() less complex, by:
* Drop the preceding underscore.
* Dropping an unused parameter.
* Using a GdkSurface instead of a HWND, as the HWND that we pass into
this function might have been taken from a GdkSurface, which are now
always created with CS_OWNDC. This means if a GdkSurface was passed
in, we ensure that we only acquire the DC from the HWND once, and do
not attempt to call ReleaseDC() on it.
* Store the HDC that we acquire from the GdkSurface's HWND into the
surface, and use that as the HDC we need for our GdkGLContext.
* Drop the gl_hwnd from GdkWin32Display, as that is really should be
stored in the GdkSurface.
* For functions that were updated, name GdkWin32Display variables as
display_win32 and GdkSurface variables as surface, to unify things.
* Stop calling ReleaseDC() on the HDC that we use for OpenGL, since
they were acquired from HWND's created with CS_OWNDC.
Scale factors can be negative, but we were not
looking out for that, triggering an assertion when
trying to create a render target with negative
width of height. Avoid that.
Fixes: #4096
Eliding totally transparent content from the node tree is
not 100% correct, since filters can make things visible, so
we need to at least preserve the bounds. We can do that by
creating a transparent color node.
Apply heuristics to avoid breaking users existing configurations
with the change to not always add the default sequences.
If we find a cache that was generated before 4.4, and the Compose
file does not have an include, and doesn't contain so many sequences
that it is probably a copy of the system one, we take steps to keep
things working, and thell the user about it.
All tables use the compact format now, and we generate
caches in that format too. Bump the cache version to 3
for this.
Replace the python script for generating the builtin table
by a small C program using the same code to generate the data
for the builtin table. This drops the restriction on only
generating a single character in the builtin sequences.
When we find a Compose file, replace the builtin
sequences with the table we found. This matches the
semantics described in Compose(5), and makes it possible
to drop unwanted sequences from the builtin table.
It is slight change of behavior for users with existing
Compose files. To match the previous behavior, you have
to add
include "%L"
to your Compose file, to keep the builtin sequences in
addition to your own.
This lets us naturally replace matching sequences
while parsing. That means that the semantics are now
"last one wins" if the parser sees multiple entries
for the same sequence.
Add a testcase that checks the new replacement semantics.
Keep the list of composetables private to GtkIMContextSimple,
and just have an api that creates new GtkComposeTables, either
from a file or from data.
Update tests to use the new api.
This shows how to use a layout manager in a widget,
implemented in javascript. The example sets up the
environment for running from the toplevel dir, assuming
that the build dir is called 'build'.
It apparently worked by chance in the past, but now causes e.g.
alphanumeric characters to be interpreted as half-width katakana
when using the Japanese IME.
We're using the tag contents array to count the number of invisible tags
set but we always increase it without being initialized.
This may lead to unexpected behavior when traversing them and it's
causing a reliable failure in the textiter tests under s390x.
So, memset that area content to 0 once allocated. It's not needed to do
the same for the tags themselves as we always assign them.
Conditional jump or move depends on uninitialised value(s)
at 0x4CFAA00: _gtk_text_btree_char_is_invisible (gtktextbtree.c:2569)
by 0x4B8A1BB: find_visible_by_log_attrs (gtktextiter.c:3244)
by 0x10E93D: check_backward_visible_word_start (textiter.c:484)
by 0x10E93D: test_visible_word_boundaries (textiter.c:523)
by 0x533288F: g_test_run_suite (in /usr/lib/s390x-linux-gnu/libglib-2.0.so.0)
by 0x53328E7: g_test_run (in /usr/lib/s390x-linux-gnu/libglib-2.0.so.0)
by 0x109CC1: main (textiter.c:807)
In many cases, we have an "extra-menu" property that is used to allow
applications to join menus into the native menu for the widget. Previously,
this was done by nesting that menu in a section.
Doing so increases the complexity of the rules for GtkMenuTracker as you
may want different handling from inside of the section vs toplevel
sections.
If instead we synthetically glue the menus together, we have a much more
natural joining of menus as the application developer would expect for
their menu.
This also ports GtkLabel, GtkText, GtkPasswordEntry, and GtkTextView to
use the joined menu helper.
The joined menu helper comes originally from GNOME Builder and has had
extensive use there.
Fixes#4094
We hardcoded the typelib directory for only an arch (and a distro),
while we can just get it from gobject-introspection pkg config if tests
are enabled.
This shows how to use a layout manager in a widget,
implemented in javascript. The example sets up the
environment for running from the toplevel dir, assuming
that the build dir is called 'build'.
We need to update the visibility of the placeholder
label when we create it, otherwise we can end up
with placeholder text on top of entry content.
Fixes: #4066
This was showing up when calling some filechooser api on
a native filechooser, where we use a filechooser dialog
behind the scenes, but it does not have focus.
Fixes: #4077
This shows how to do custom drawing in a widget,
implemented in python. The example sets up the
environment for running from the toplevel dir,
assuming that the build dir is called 'build'.
In many cases across GtkTextBuffer, we end up with operations performed
inside of a begin/end user action. Those can be coalesced into a single
sub-action within the group, and hoisted out of the group. Doing so
increases the chances that we chain similar actions together for words.
Additionally, this fixes an issue introduced in
6179886b14 for #3977 where GNOME Text Editor
started to group all possible actions into a single group.
FixesGNOME/gnome-text-editor#97
Make sure the button still has the .image-button style class with an icon,
also add it to the initial state with only an arrow. Add a new
.arrow-button style class for the icon+arrow state so it's possible to
style it.
Remove spacing from the label+arrow variant to match, re-add it from the
stylesheet for both.
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3501
In the refactoring from GdkWindow to GdkSurface, GtkWidget no longer
corresponds to a GdkSurface. We have to calculate the relative position
from GtkWidget to the GdkSurface.
Closes#4063.
The Trash is a special location: files cannot be copied or moved, there,
and the file selection dialog is not able to restore files from the
Trash.
Fixes: #674
The Trash is a special location: files cannot be copied or moved, there,
and the file selection dialog is not able to restore files from the
Trash.
Fixes: #674
This allows developers to modify the pango context that is used when
rendering text within the text view.
Such access can be useful to alter how rounding occurs with API such as
pango_context_set_round_glyph_positions() and is needed by GtkSourceView
for proper placement of glyphs within the overview map.
We must call gdk_drag_drop_done() when the drag ends,
successfully or not. Without this, we get an unwarranted
emission of ::cancel after a successful drop.
Since only the first call to gdk_drag_drop_done() is taking
effect, it is safe to call as a fallback, after emitting
::dnd-finished. If the application connects to that signal
and calls gdk_drag_drop_done() itself, its call will take
precedence.
This matches what the X11 implementation does.
We are pretty good at batching commands now, and we can easily
produce batches that exceed the maximum number of elements per
draw call that the hw can handle. Query that number, and respect
it when merging batches.
This fixes the rendering of the overview map in GtkSourceView.
Determine the root_x and root_y coordinates of the drag surface by
relying on the coordinates of the surface where the drag is being
carried out, plus the coordinates that we receive from the drag event,
which is in-line with what the X11 backend does.
This will prevent the drag surface from being initially drawn at the
correct position, but jumping towards the top-left corner of the screen
shortly afterwards.
The DnD support will still need some more updates to function correctly
on Windows, but at least this is a small improvement.
Fixes issue #3798.
This gets the basic mechanics of the drop portion of DnD working on the
macOS backend. You can drag, for example, from TextEdit into GNOME
Text Editor when using the macOS backend.
Other content formats are supported, and match what is currently
supported by the clipboard backend as the implementation to read
from the pasteboard is shared.
Currently, we look up the GdkDrag for the new GdkDrop. However,
nothing is stashing the drag away for further lookup. More work is
needed on GdkMacosDrag for that to be doable.
We will want to be able to reuse the pasteboard reading code from
the macOS DnD drop backend. This just removes the pasteboard
bits from the implementation and allows that to be passed in as in
both clipboard and DnD cases we'll have a specific NSPasteboard
to read from.
We don't want to overdraw when dragging a narrow column
around, and we also need the clipping to avoid picking
the wrong column, when a later column button overlaps
an earlier one.
Fixes: #4045
The normal way to associate accels with actions is
to attach a shortcut controller to the widget. The shorcut
controller will inject the accel into the action muxer
tree, so that it can get displayed in widgets that activate
the action (say, in menus.
This approach does not works for generated menus, since the
widgets are not in the hands of the app developer, so attaching
shortcut controllers to them is impractical.
Instead, GtkModelButton has an accel property that gets
bound to the accel coming from the action muxer tree (most
likely put there via gtk_application_set_accel_for_action),
and creates a shortcut controller itself.
The change in this commit is to prevent the shortcut controller
from injecting the accel into the action muxer tree in this case.
Otherwise, the accels get 'stuck' and we won't update them if the
global accels are later changed.
This is a hack, and needs a better solution.
We were pretty aggressive about not registering
observers further than necessary, stopping at the
first muxer that provides an action.
But even though action changes further up in the tree
won't be relevant in that case, we need to listen to
accel changes, since they may come from higher up
in the tree (e.g. when using
gtk_application_set_accels_for_action with an action
that is defined on a widget.
So, register all the way to the top, and stop propagating
action changes when we hit a muxer that provides the action.
In some cases (such as when getting a new parent), the
action muxer doesn't know exactly which detailed actions
have changed accels, so we call primary_accel_changed with
just an action name.
Make the menu tracker item handle that case by matching
either against the detailed name or the the action name.
Take the size from -gtk-icon-size.
Note that min-width/height still works, as those properties are handled
by the generic widget sizing machinery in GTK4.
If we are undergoing a surface move, just apply the next_layout anyways,
even if we are not moving a toplevel surface.
Update the way how we obtain the x and y coordinates of a surface, if it
is a toplevel, apply the x and y coordinates from the results from we
obtained the underlying Win32 HWND, as we did before. But if it is a
popup, use gdk_win32_surface_get_geometry() to obtain the correct x and
y coordinates to place our popup surface.
Also correct how we compute the shadow dimensions, and the final popup
rectangle as we attempt to layout the popup surface, since GDK-Win32
keeps track of the shadow dimensions in system (unscaled) units, not GDK
units.
Fixes issue #3793.
Also rename gtk_media_stream_ended to
gtk_media_stream_set_ended, to avoid naming
collision with GtkMediaStream:ended.
The existing entry points still exist, deprecated
and marked as non-introspectable.
Update all internal uses.
Fixes: #4023
Rename the GtkDropTraget:drop property to :current-drop,
to avoid naming collision with the signal of the same
name.
We leave the old property and getter in place, deprecated
and marked as non-introspectable.
Fixes: #4028
We were not handling the case right in which we
want to use underlines, but not use markup. Since
we are now using pango_parse_markup for this case,
we need to escape the xml markup.
Fixes: #4041
This reverts commit 95747b1a40.
This was wrong - it turns out that while GtkProgressBar
allows you to change its orientation, its box layout is
always vertical.
Fixes: #4037
Some bindings can't handle the coexistence of
GtkMediaStream:prepared and gtk_media_stream_prepared.
Help them out by renaming the function to
gtk_media_stream_set_prepared, and rename
gtk_media_stream_unprepared as well, to match.
The existing entry points still exist, deprecated.
Update all internal uses.
Fixes: #4023
Claim the gesture when we are activating a list item.
Otherwise we end up with double activations in
columnviews: first GtkColumnViewCell handles
the event, and then GtkListItemWidget handles
it again.
Fixes: #4015
If we have a GAction as model, we just have to let
the action helper handle the state updates. GtkButton
already calls gtk_action_helper_activate() for us.
v32 of the image includes wayland-protocols 1.21, and
other updates from Fedora 34.
While we are at it, drop gtk-doc from the image, and drop
the separate DOCS_IMAGE - no longer used.
The swipe gesture forces values in the spin button that are
"impossible" according to the adjustment. This can break things
in creative ways.
Ensure the steps provided are always multiples of the adjustment
step value, and keep the remainder for further interaction.
In the GTK3 days, the spin button was an entry, with buttons on
top, and the swipe gesture affected the input on the entry bits.
Now the spin button is a container, so this gesture in the capture
phase applies to all contained children (incl. buttons).
Attach this gesture to the entry itself, so the buttons are left
outside this business. The gesture is still in the capture phase
in order to prevent text selection/edition/etc to happen.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/4008
gtk_tree_view_row_activated and the ::row-activated signal
can (and do) receive NULL for the column occasionally.
This is an introspection api change.
Fixes: #3828
__popcnt() is not supported for ARM nor ARM64 Visual Studio builds, so we must
use the fallback implementation as intrinsics are not supported for this
purpose on ARM/ARM64 Visual Studio builds.
GtkApplicationWindows size_allocate does not chain
up if the menubar is visible; don't forget to allocate
the tooltip window in that case.
Fixes: #3997
The releasing of grabs while a button is pressed (e.g. after starting dnd, or
dragging the window, or going to overview with a pressed button, etc...) was
generalized here in https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/1879.
However we shouldn't break all grabs here. In the case of grabbing popups,
compositors will still emit crossing events between client surfaces (e.g.
popping up and selecting a menu item via press-drag-release), breaking all
grabs here means inconsistent client state, that was
https://gitlab.gnome.org/GNOME/gtk/-/issues/2746.
That was fixed in mutter, by essentially making implicit grabs
owner_events=FALSE, however that breaks the mentioned use pattern entirely.
Mutter is changing this behavior back, so GTK should handle these crossing
events.
The grab that we are interested in breaking here is the implicit pointer
one. Popups will be dismissed via other means if the compositor says their
active grab needs breaking. This still leaves dnd/move/resize drags in
one place, while not allowing #2746 to happen with popups.
We already skip them on the edited cell widget, but it's also a problem for
header buttons. Overall, there's no real reason to let it propagate here.
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3996
In GTK 3, GtkStackSwitcher implemented GtkOrientable via the parent GtkBox
type. In GTK 4, that was changed to inherit from GtkWidget and lost this
interface implementation.
This adds that back, along with a note in the documentation that the
interface was added in GTK 4.4.
Fixes#3988
As app menu has been deprecated, the function of app menu should be
merged into the menubar, which behaves like the original API of NSApp.
This also brings back the default app menu, which looks like native
macOS apps.
For future refactoring, please note that the menubar must contain at
least one menu before being set to NSApp to avoid weird menubar
rendering.
Fixes#3967.
This was an oversight when porting the GtkTextHistory into GTK. We simply
need to bind the GtkTextBuffer action into the text history for grouping
to work correctly.
Fixes#3977
The gtk_print_run_page_setup_dialog() function, and its asynchronous
variant, are declared in the common gtkprintoperation.h header, but
implemented in different source files depending on the platform.
Make the "gl-context" property of the GstGLSink readable as well so that
we can query whether the GstGLContext sharing really succeeded. If it
did, then we proceed to playback our video using the glimagesink as we
did before. If it didn't, throw out the GtkGstSink we were creating, and
re-create the GtkGstSink without the "gl-context" property, meaning that
we won't be using the glimagesink in this case.
Add support to look for and use the EGL context in Windows if it was activated
instead of desktop OpenGL.
GstGL may have been built with or without EGL/libANGLE support, so if it were,
check in GstGL whether we have gst_gl_display_new_with_type() to create a
GstGLDisplay that is of the GST_GL_WINDOW_WIN32 type when we are using
Desktop OpenGL (WGL), otherwise we show messages indicating that envvars
need to be set to initialize GstGL properly.
Due to a bug in GstGL, the GstGLContext can only be set up successfully
if one of the following is true:
* An OpenGL 3.x or later emulator, such as Mesa is used (for WGL)
* The latest GstGL master is being used, at the time of writing (for
WGL)
* GTK, libepoxy and GstGL are all built only with WGL support (for WGL)
* EGL is being used in GTK at runtime
Special thanks to Matthew Waters for the help during the process.
Add support to share the WGL context in GDK with the WGL context in GStreamer,
so that we can also use OpenGL in the gstreamer media backend to playback
videos. For now OpenGL/ES is not supported for this under Windows.
The process of setting this up in Windows is a little bit more involved, as:
* The OpenGL support in GstGL requires a GL 4.1 Core context, but we may just
get the GL version from wglCreateContextAttribsARB() that we pass into the
attributes, which is 3.2 by default. So, try to ask for a 4.1 Core context
first if we are asking for anything less.
* There is only one GstDisplay available for Windows, so we just use
gst_gl_display_new().
* We must explicitly tell libepoxy that we are using wglMakeCurrent() outside
of libepoxy that is being used in GdkGL, otherwise we would end up crashing
as the GL/WGL function pointers would become invalid.
* We must also deactivate temporarily the underlying WGL context that was made
current by gdk_gl_context_make_current() so that when
gst_gl_display_create_context() calls wglShareLists(), we won't get bitten
by error 0xaa (resource busy), as some drivers don't handle this well when
the GL context is current in another thread.
For the last two points we make use of macros defined by the platforms that the
build is done for to help us carry out the necessary tasks as needed.
Thanks to Matthew Waters for the info on integrating GstGL and windowing
toolkits on Windows.
Visual Studio 2013's linker does not suport `/WHOLEARCHIVE:`, so just
explicitly extract the objects from the static libraries that will
form the final GTK DLL.
We want to find focusable children, so we need to look at
the focusable property, not at can-focus. This is a change
from GTK 3, where can-focus was the correct property to
look at.
The intention of the code is to find a focusable ancestor,
so it needs to look at the focusable property, not at
can-focus. This is a change from GTK 3, where can-focus
was the correct property to look at.
Fixes: #3965
Check that we are indeed running inside an Xorg server before enabling
the workaround.
XWayland or other nested X servers deadlock when that workaround is
applied.
Remove a boatload of "or %NULL" from nullable parameters
and return values. gi-docgen generates suitable text from
the annotation that we don't need to duplicate.
This adds a few missing nullable annotations too.
The script assumes that the git program is in the PATH which may not be always
the case, so if git was not found, fallback to the former behavior of making
VCS_TAG an empty string if the profile option was not set to 'devel'.
Also fix the call to the script so that we do indeed call GIT on the source GIT
repo so that we can get the right short SHA1 sum.
Re-organize the code a bit in the process, so that things seem cleaner.
We iterate here from the target widget up the toplevel checking
for the previous and new grab, there's however 2 bugs here:
- The check for is_shadowed was different to the check for was_shadowed
- The loop started with the assumption that the widgets did not hold
a grab, just to change it if the grab widget was found. (or maybe
it's the other way around? it's unclear with the differing checks
for past/present state).
Make these checks consistent, and ensure we start with the right
assumption for the past/present grabbing state, and accounting that
new/old grab widgets may or may not be part of the pick stack.
With gtk_grab_notify_foreach() just taking care of emitting crossing
notifications due to the GTK grab change, rename it to a more apt
gtk_synthesize_grab_crossing().
The _gtk_widget_grab_notify() function just (maybe) did a) reset
controllers and b) hide toplevels. The second part was a testing
remnant introduced in commit 024d832d94, not part of the original
fix.
Do the former more concisely, called from the place where we figure
out whether a widget's ability to receive events changed due to
GTK grabs. It's across those changes that we are interested in
resetting the controllers.
With the gestures being reset both ways, GtkWindowHandle (and
probably other) gestures are now able to reset after a GTK grab
takes input away (e.g. GtkMenuButton). This could be seen as
a sudden jump the next time they'd be dragged with the mouse,
as the gesture would "resume" the previous interaction.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3942
When a new sequence is added to a GtkGesture, its state is looked
in other gestures in the same group, and made to match in this
gesture. This however happened a bit too early, before the
gesture touchpoint was fully set up. As this may result in signal
emission and whatnot, it's a good idea to make it happen with a
fully set up touchpoint.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3946
The CI pipeline does not do a full clone, so we cannot use `git
rev-parse` to get the short SHA checksum for the current commit.
Of course, since nothing's ever easy, we cannot use vcs_tag() with a
custom command, so we need to generate our out `demo_conf.h` header
straight from a script.
Match GtkListBox, so it's possible to use the same styles for them.
Update GtkListView and GtkGridView docs to reflect that, fix a few gtk-doc
formatting leftovers along the way.
Make gsk_ngl_texture_library_pack always return
the position including the padding. And compute
texture coordinates accurately in all cases (we
were fudging the padding for standalone textures.
The indentation of new lines inside documentation blurbs must be smaller
than 4 spaces, otherwise the Markdown parser will consider the line to
be part of a pre-formatted code block.
Fixes: #3945
When configuring the inspector display, preserve
debug flags that affect which GL variant we pick.
Otherwise, we may end up with a GLX context on the
default display, and an EGL context on the inspector
one. This hopelessly confuses libepoxy, and things
don't go well when that happens.
If we want to add an EGL implementation for the X11 backend, we are
going to need to move the GLX bits into their own class. The first step
is to declare GdkX11GLContext as an abstract type, and then subclass it
into a GdkX11GLContextGLX type, which includes the whole GLX
implementation.
When loading the emoji data we just try to get the data for a language
while there may be territory specializations and emojibase provides
them.
So, split the loading function and try to load the data for the fully
defined language string (i.e. `it-ch`) before loading the generic one
for the language (i.e. `it`) and eventually falling back to the generic
english.
The change in 875a92b95f made labels strip
out underlines earlier, but overlooked that this made
mnemonics not work before the timeout to show them
has passed. That was unintentional. Make mnemonics
work regardless of their visibility, again.
We can't use this flag for any code that may get run
outside the __builtin_cpu_supports() check, and meson
doesn't allow per-file cflags. So we have to split this
code off into its own static library.
Width and height of a GdkMonitor are derived via wl_output which
talks about physical dimensions of a device and compositors usually
implement this as the untransformed values (e.g. weston, wlroots).
Since the GTK client has no way to figure out if a monitor was rotated,
transform the physical dimensions according to the applied wayland
transform to have the physical dimensions match the logical ones.
Mutter flips the physical dimensions itself but doesn't announce the
transform so this shouldn't break anything there.
gcc says: error: iteration 2147483649 invokes undefined behavior
[-Werror=aggressive-loop-optimizations]
which of course, never happens because ncols is always >= 2.
Add some assertions to tell the compiler.
This was comparing the included image-missing icon
with the one in the current icontheme on the test
system. Works fine as long as we don't change
the icons (which we just did). To avoid this, set
the icontheme to hicolor for this test, which does
not have the image-missing icon, so we end up getting
the builtin icon for both ui files.
When a drop causes the event controller to be finalized
(directly or indirectly), we end up segfaulting while
trying to wrap up the drag operation. So, keep a reference
on the GtkDragSource from when the drag begins to when
it is done.
This fixes a crash in gnome-todo when dragging tasks.
Avoid a nested listbox, show the connector,
don't show information we don't have. Also,
disconnect all signal handlers from the display
when the inspector is going away.
Rewrite this in a way that doesn't depend on kernel
header defines at the time the wayland scanner was run.
This was causing the build to break on Centos 8, where
a bunch of fourcc formats are missing.
If the session bus address is unset, GLib will
helpfully try to autolaunch a bus, which will
fail and timeout. If we set an empty address,
it gives up early.
Some GL drivers such as Mesa-D3D12 do not allow one to call SetPixelFormat() on
a given HDC if one pixel format has been already set for it, so first check the
HDC with GetPixelFormat() to see whether a pixel format has already been set
with the HDC, and only attempt to acquire the pixel format if one has not been
set.
This will fix running with GL/NGL on Windows using the Mesa drivers.
When building for homebrew/linuxbrew on Ubuntu 16.04, memfd_create() is
not available and causes the build to fail.
This adds a proper check for the function.
This fixes nautilus crash and perhaps other callers issues.
Nautilus (and sometimes glib) crashes with malformed URI inside of the
bookmarks file .config/gtk-3.0/bookmarks when it has no LABEL.
This is result from the closed glib MR #2065 analysis and agreement.
https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2065#note_1091979
Signed-off-by: Geyslan G. Bem <geyslan@gmail.com>
When reading text, we need to check we terminate the G_TYPE_STRING
string with a null byte, because the clipboard does not guarantee one.
So just append a \0 to the stream.
Fixes#3899
Just from reading the code, it seems that we
should unset .csd and .solid-csd at the same
time, since the are mutually exclusive and
we unset them here so realize() can set one
of them again.
The code in gtkwindow.c for dealing with the various
combinations of client-side decorations and client-side
shadows is entirely too complicated.
This commit does not really clean it up, but simplifies
one of the shadow conditions far enough to make some
sense.
With this change, I get the expected decorations in
all the cases I can easily reproduce locally.
Deriving the resize border size from the shadows
carries the risk that we might end up with uneven
resize borders (or none at all, on some sides).
So, justs enforce that we have a big enough shadow
width on all sides.
Previously it was impossible to compose characters on higher levels of
some keyboard layouts as pressing the level selection key would just
exit compose mode.
Examples for affected keyboard layouts include the Latvian
apostrophe-variant "lv(apostrophe)" (latched third level), the extended
German keyboard layout "de(e1)" (latched fifth level) as well as the
multilingual Canadian keyboard layout "ca(multix)" and the German
neo-layout "de(neo)" and its descendants (shifted fifth level).
To reproduce, set a compose key and select the Latvian apostrophe layout.
Notice that you now can input [ by pressing first the ' and then the 8-key.
Then pressing <compose>'8'8 should produce ⟦, but prior to this patch it
did not.
The invisible resize borders have been wider than they
should, for a while. Go back to a size close to what
we have in GTK3.
To summarize: resize borders will be at most 12 pixels
on each size, but never wider than the windows shadow.
The resize corners have 'legs' of 24 pixels where you
still get a corner resize cursor.
Fixes: #3856
Only send selection-changed events when we either
had a non-empty selection before, or have one now.
This should help orca speak the right things, and
not the wrong things.
Related: #3549
Orca relies on these to keep track of the focus location,
ignoring the focused state. With this change, orca can
once again speak text in entries as I type.
We are starting with a pretty empty a11y object tree,
and we want orca to bring more of it into existence
by navigating the tree. But that only happens when we
send it events. Primarily focus events, which come in
from GTK via the platform_change mechanism. So realize
the context when we are sending platform_changes,
otherwise, orca never gets the mesage.
With most context realization happening inside
GtkAtspiContext in response to D-Bus calls, the
code in gtk_widget_realize_at_context that sets
the role is not executed for most accessibles,
causing them to be stuck with the 'filler' role
that makes orca ignore them.
To fix this, split gtk_widget_realize_at_context
into the actual context realization (getting on
the bus) and the setting of widget-specific
properties, and do the latter part when the
widget is rooted.
This makes accerciser report proper roles for
entries and buttons. Orca still has an issue
with getting the hierarchy populated.
Orca ignores events unless the object is inside an object
with role window and states ACTIVE and SHOWING. To arrange
for this, introduce a new ACTIVE platform state, and set it
for windows when they are active.
This gets orca to be a lot more talkative.
Linux 3.4 added support for the MADV_DONTDUMP option to madvise(), which
requests that the covered memory not be included in coredumps. It makes
sense to use this to prevent cases where application crashes could
result in secrets being persisted to disk or included in dumps that are
uploaded to remote servers for analysis. I've avoided making this fatal
since there's a chance this code could be built on systems that have
MADV_DONTDUMP but run on systems that don't.
This cleans up and moves the code to load files
while taking the scale into account. Along the way,
we drop the last uses of GdkPixbufAnimation, and
consolidate the pixbuf using code in one place.
In commit 4a76abffd4, we deferred unsetting focus
and default until after the next draw, overlooking the
case where the focus is set to another widget before we
ever get to the unsetting.
Fixes: #3413
Avoid passing through random key press or release
events while we are showing preedit. That prevents
'accidents' like typing Ctrl-. bringing up the
Emoji chooser during preedit, or hitting Ctrl-a
after the Compose key moving the 'dot' around in
vim in terminals.
Not for symbolic icons.
Don't apply `-gtk-icon-filter: opacity(0.5);` to the symbolic icons as
they already have the "gray" colors indicating the disabled state.
Symbolic icons can be styled using the `color` property.
Also remove the obsolete comment.
The condition we check for to catch X servers going away
may not be accurate anymore, and the warning shows up in
logs, causing customers to be concerned. So, be quiet by
default, unless the user explicitly asked for a message.
Some locations have to be mounted, but their mounts are not user-visible
(e.g. smb-browse). Though this is maybe a bit weird, it is how it works
for years. The problem is that the commit 267ea755, which tries to get the
default location for opening, caused regression as it doesn't expect such
possibility. Before this commit, such locations were opened without any
issue, but nothing happens currently after clicking to "Connect" except of
clearing the "Connect to Server" entry. Let's fallback to the original
location if the mount was not found to fix this regression.
Fixes: https://gitlab.gnome.org/GNOME/nautilus/-/issues/1811
When we clean up the uniform allocations after a frame,
it can happen that our space requirements actually increase,
due to padding that depends on the order of allocations.
Instead of asserting that it doesn't happen, just make
it work by growing our allocation.
Fixes: #3853
The gtk_debug_cflags variable is an array, with each argument stored in
a separate string; if we add multiple arguments inside the same string
we break the escaping rules.
Fixes: #3864
It is also possible to enable F16C instructions on 32-bit Visual Studio builds,
so also check for the compiler's ability to build F16C code on Visual Studio
32-bit instead of just bailing out.
We need to use __cpuid() to check for the presence of F16C instructions on
Visual Studio builds, and call the half_to_float4() or float_to_half4()
implementation accordingly, as the __builtin_cpu...() functions are strictly
for GCC or CLang only.
Also, since __m128i_u is not a standard intrisics type across the board, just
use __m128i on Visual Studio as it is safe to do so there for use for
_mm_loadl_epi64().
Like running on Darwin, we cannot use the alias __attribute__ as __attribute__
is also for GCC and CLang only.
The __builtin_cpu...() intrisics are strictly for GCC/CLang-based compilers, so
don't use them in the checking code on Visual Studio. The test code will still
compile without this change, but will certainly fail if we want it to link.
Propagate the focus-on-click setting to the button
inside, so that setting menubuttons as !focus-on-click
works as expected. This helps for menubuttons in
header bars, where dragging on the button will otherwise
steal focus from the content.
When 9-slicing shadows, omit the center tile when it is
entirely contained in the outline (that is not always
the case, depending on corners and offsets).
gsk_rounded_rect_contains_rect was calling
gsk_rounded_rect_contains_point, which potentially
checks all four corners, for a total of up to 16
corner/point checks. But there is no need to do
more than 4 such checks to answer the question.
Commit 3dbf5038fa tried to defer focus changes
until after rendering is done. But it failed to do so, since
the toplevel ::render handler is still before rendering of
popups that are attached to that toplevel. To do this
properly, we need to do it in the AFTER_PAINT frame clock
phase.
Fixes: #3725
We used to override cursor to use all-scroll while the
content is being scrolled. Unfortunately, there is several
problems with this:
- It is really only expected certain devices, and we don't
have the device information on Wayland
- With the way cursor setting works in GTK4, non-NULL cursors
of the content (eg the text views ibeam) win, making the
scroll cursor not show up
- Under X11, we seem to miss scroll end events and then
the scroll cursor gets stuck
Therefore, just remove this feature.
Opportunistically use the coloring program for
drawing underlines instead of the color program.
This avoids program changes in the middle of
text.
For the Emoji text scrolling benchmark, this reduces
the program changes per frame from > 1000 to around 100.
Use an IFUNC resolver to determine whether we can use
intrinsics for FP16 conversion. This requires the functions
to be no longer inline.
Sadly, it turns out that __builtin_cpu_supports ("f16c")
doesn't compile on the systems where we want it to prevent
us from getting a SIGILL at runtime.
We only have one shader that uses the color2 attribute,
and it doesn't use the uv attribute, so save vertex
memory by putting those in the same space.
This reduce the per vertex space from 32 to 24 bytes.
This reduces the size of our Vertex struct from
48 to 32 bytes. It would be nicer if we could store
the colors in fp16 format in the rendernodes, and
avoid conversion here. But this is still good.
At times (most often when closing subsurfaces that are scheduling
relayouts) the PHASE_PAINT handling gets broken with the following
sequence:
1. Surface receives wl_callback.done for the previous frame.
Surface is thawed.
2. A new update on the surface is scheduled. PHASE_PAINT is
requested directly on the frame clock. priv->pending_phase is
left unset in the surface.
3. Surface gets frozen
4. Frame clock processes the update scheduled at 2. The surface
is frozen, so paint is prevented. PHASE_PAINT is considered
handled.
5. Compositor emits wl_callback.done again. Surface is thawed.
6. At this point the machinery is off
- The surface didn't paint but has pending update regions
- priv->draw_needed is set in the toplevel and other portions
of the widget tree
- So queueing redraws is ineffective at eventually calling
gdk_surface_schedule_update() again on the toplevel surface.
- We don't paint anymore, so this broken state is not flushed
until other subsurface changes manage to schedule the missing
update.
To fix this, always set PHASE_PAINT in priv->pending_phase when
doing gdk_surface_schedule_update(). If the frame clock turns
around before the surface is thawed, it will still be waiting to
be processed the next iteration.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3750
When we don't get stettings from the portal, the current
fallback is 'awful fonts'. There is no need for that. Instead,
set the fallback values to grayscale antialiasing with slight
hinting.
Move the resources of each renderer to its subdirectory.
We've previously done that for the ngl renderer, but it
is better to be consistent and do it for all the renderers.
We need to invalidate the style when font-size changes,
because we propagate this value through the initial
value of the CSS font-size property, and it will not
be recomputed otherwise.
Forgetting to do so was causing the Wayland im context
to leave behind a dead event controller. This was showing
up as a crash when closing the inspector after changing
the im-module property of a GtkText widget. The crash
was delayed until closing the inspector because the
inspector keeps a ref on the event controllers of the
currently shown widget.
Arrange things so that non-child parameters
are always printed before the children. This
greatly helps with readability, which really
suffers when there's hundreds of lines of indented
children between the node start and its parameters.
Update all affected tests.
This is a pet peeve of mine: When we call
g_test_init() before handling --generate,
the random seed spew pollutes the output.
Highly annoying. I've fixes many test binaries
over the years, but more keep popping up.
Instead of rendering the unclipped child to a texture
(and risking blowing the texture size limit, and bad
downscaling), just render the clipped region, and live
with the fact that we can't cache the rendered texture.
This avoid bad artifacts when scrolling long textviews
in rounded clips.
The use of the keyboard-activating CSS class for buttons was added
in [1], but the style did not apply to buttons with has-frame=FALSE.
[1] 00923615f4 ("button: Add back visual feedback for keynav", 2021-04-01)
The change in 740559a54f to populate the list incrementally
broke initial font selection. Fix that, by trying to select
until the incremental filling is done.
Fixes: #3687
Since we are likely going to see theme names like
Adwaita and HighContrast, make fallback work as follows:
Adwaita -> Default
Adwaita:dark -> Default:dark
HighContrast -> Default:hc
HighContrast:dark -> Default:hc-dark
HighContrastInverse -> Default:hc-dark
Other themes will fall back to Default, as before.
We lost the visual feedback for activating a button
via Space or Enter when the :active pseudo-state became
managed. Bring it back with a style class.
Fixes: #3813
This was breaking muscle memory of people with
the us intl keyboard layout, for important keys
such as '. The unfortunate side-effect is that
our handling of <dead_acute> is a bit hampered
by sequences that don't fit the pattern. But
such is life.
Fixes: #3807
Rename the included theme to Default, with 4 variants:
light, dark, hc, hc-dark. This replaces Adwaita,
Adwaita:dark, HighContrast and HighContrastInverse.
We still make the themes available under these names,
and we still set up Adwaita-dark and HighContrastInverse
as the dark variants of Adwaita and HighContrast.
The unification of the theme variants under Default
is not quite perfect; it would be nice to merge
the assets/ and assets-hc/ subdirectories and render
all assets from a single svg file.
If we scroll down in a list that's still being filled, we hit the edge and
initiate overshoot, and then the adjustment's upper value increases. This
leads to an unwanted bounce back.
Additionally, if in a similar situation the upper value decreases, the
overscroll glow gets stuck.
Update kinetic scrolling upper and lower value on changes, and immediately
cancel it if dimensions on that side change.
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3752
Instead of getting current display before calling settings signal removal,
do it inside remove function and only if there is a signal connection to remove.
Compare clipped repeat nodes. Must skip cairo here
since it blurred the child by scaling after rendering.
Also skip the gl renderer, since it hasn't been fixed
for this yet. ngl passes this test.
There was confusion here about the handling of the
modelview transform. The modelview transform we are
getting is already set up for rendering the node
we are given, so keep it - except for possible adding
an extra scale on top when the texture would otherwise
be too big.
It seems to make assumptions about text positioning that
are not holding with subpixel positioning. I'm not 100%
sure how that leads to exactly the artifacts that are seen
here, but I am just disabling the test until that is fully
understood.
It makes assumptions about text positioning that are
not holding with subpixel positioning. There is no
guarantee that the next word in a multi-word text
starts on an even pixel boundary, as it does when
you break the text into multiple, separately rendered
blocks.
... until all globals have been received.
The dependency tracking introduced in 4e9be39518 only allows to
specify required globals and processes the closures as soon as
the requirements have been met. There are, however, also optional
dependencies - most notably the primary_selection protocol.
Currently we rely on the fact that compositors like Mutter announce
it before `wl_seat`, even though the order is not specified in
the spec.
Process globals closures only after all globals have been announced,
so optional dependencies can be accommodated.
Closes https://gitlab.gnome.org/GNOME/gtk/-/issues/3791
Move some work out of the loop in visit_text_node.
This takes advantage of the fact that the yoffset
of most glyphs is zero, so yphase generally does
not change in a line of text.
Arrange for the contents to be in a single transform
node that is updated as we scroll. This makes the job
of the render node differ a lot easier, since it does
not have to compare to big containers one-by-one.
Allow comparing container nodes to any other
node, by pretending the other node is a single
child container (if it isn't one already).
This fixes a glitch where we redraw the full
entry text when the blinking cursor goes to
opacity 0, since GskSnapshot then optimizes
away first the opacity node, and then the
single-child container.
Commit 8b82993dde added a noisy warning
to gtk_distribute_natural_allocation to quiet a
compiler warning. It turn out that the file chooser
managed to trigger this warning, so make it a quiet
return.
Reshuffle things to allow for a limited amount of
dead key 'chaining'. We keep up to 2 dead keys in
the preedit, so you can type
<dead_acute> <dead_cedilla> <c>
to produce ḉ, while still getting ```c with
<dead_grave> <dead_grave> <dead_grave> <c>.
Add a utility function to check whether the icontheme
will produce something better than missing-image for
a GIcon. Obviously, we can only answer this question
if the GIcon is a themed icon the begin with.
These were showing up as missing icons when opening
the Inspector with the hicolor icontheme:
system-search-symbolic
go-previous-symbolic
go-next-symbolic
display-brightness-symbolic
Stash away the device timestamp when obscuring
the pointer, and compare it when we decice whether
to unobscure it. This fixes a problem where synthetic
motion events would make the cursor reappear
prematurely.
Fixes: #3792
Stash away the device timestamp when obscuring
the pointer, and compare it when we decice whether
to unobscure it. This fixes a problem where synthetic
motion events would make the cursor reappear
prematurely.
After iterating all the providers, all of them returning unsupported
error, gdk_content_provider_union_get_value() returns FALSE without
filing the given GError. Then gdk_clipboard_read_value_internal()
assumes there's a GError when FALSE is returned and
g_task_return_error() fails. We can just chain up to parent
implementation to ensure the GError is filled with unsupported error.
When there is no visible child, gtk_selection_model_is_selected()
was returning TRUE for any invalid position.
So check if the page is non-NULL and match
Previously, we translated the uniform key (an enum) into a location within
the shader program in GskNglProgram. A number of performance improvements
were focused around having low nubers for the uniform locations. Generally
this is the case, but some drivers such as old Intel drivers on Windows
may use rather large numbers for those.
To combat this, we can push the translation of uniform keys into locations
at the GskNglUniformState level so that we work with unranslated keys
through the process until applying them.
Fixes#3780
Not sure we can do much with a ui file that has
<placeholder/> elements littered throughout. But
at least we can avoid crashing while trying to
convert it.
Fixes: #3781
The DnD code for X11 adds the composite overlay window (aka COW) to the
cache.
Yet the X11 requests to get and release the COW may trigger XErrors that
we ought to ignore otherwise the client will abort.
Fixes: #3715
Update our compose sequences based on the current
update xorg Compose.pre file. Beyond that, remove
some deadkey sequences that we are now handling
(better) in code.
Make this script parse gtk-compose-remove.txt for
sequences to remove from the xorg Compose file.
This will be used for removing some deadkey combinations
that we can handle better in code.
Also, make this script explicitly python2. I tried
porting it to python3, but gave up in the end.
For sequences like ``, we want to commit the first
deadkey and then continue preedit with the second.
The alternative is to do chained deadkeys, where
entering ~~a yields ̃̀̃̃a. But we don't do that, and
I think that would be more controversial.
This lets the NGL backend be selected instead of the Cairo backend on
devices which expose both GL and GLES, but have better support of GLES.
Tested on a PinePhone.
We only need these names when serializing a11y information
for tests. And copying these strings is entirely unnecessary.
So, just pass a callback instead.
The effectiveness of the front cache is limited by
subpixel positioning making it very likely that we
will meet the same glyph in different x phases inside
a single line of text.
Factoring the xphase into the front cache key makes things
better. For the string eeeeeeeeeeeeeeeeeee
before: 0% front cache hits
after: >90% front cache hits
Right now, we land inside a 404 if we go to:
https://gnome.pages.gitlab.gnome.org/gtk/
as we have all our API references in a sub-level. We should have a
landing page for the root, similar to developer.gnome.org/references/.
We don't want to be responsible for duplicating the effort of the hash
table, we just want to speed up subsequent lookups. Otherwise, we risk
not marking glyph usage when tracking usage for compaction.
This required finishing up the begin_frame/end_frame semantics for
GskNglTextureLibraryw which was apparently overlooked.
The driver was changed to provide more information to the library when
beginning frames. We do not need to use end_frame so that was removed.
The frame age is the same as GL (60) but I do wonder if that is based
on seconds if we should be using something longer for situations where
we have higher frame rates.
Fixes#3771
Respect the debug settings for disabling Vulkan or GL,
and do not try to initialize those contexts. This can
be necessary to work around crashes.
Fixes: #3748
As long as we can create a GL context, pass one to
gstreamer. This at least gets us GL textures with
the ngl renderer, the previous code was arbitrarily
refusing that.
Setting up check or toggle button group relationships
in a cycle will lead to lockups. Add a warning about
this, and catch the simplest case with a precondition
check.
Fixes: #3763
Apparently, by comparing with the other backends, we should not call
_gdk_win32_append_event() after calling gdk_scroll_event_new() but we should
call it after calling gdk_scroll_event_new_discrete(), which was why we didn't
restore the cursor after we scroll using the mouse wheel and didn't manage to
remove the shade that appears after we scrolled to the very top or very bottom.
Also, as suggested by the reporter, use IDC_SIZEALL for the system cursor that
we fall back to if no cursor theme is installed, as with other Windows
programs.
This should really fix issue #3581.
If cairo is a subproject, it's not necessarily installed when gtk
is built. In the build tree, libcairo-script-interpreter is not stored
in the same directory as other cairo libraries.
If cairo is a subproject, it's not necessarily installed when gtk
is built. In the source tree, cairo's headers are not stored in
a directory called 'cairo'.
GTK traditionally lets you activate keyboard shortcuts
even if they are for a non-active layout. But it is meant
to only activate with a keysym from a non-active layout
when that symbol is not present in the current layout.
That last condition was lost when key event handling
was redone for GTK4. Bring it back.
This makes sure that we don't have cursors disappearing on Windows upon
scrolling because we can't find a cursor that exists on the system during
a scroll, and unlike GTK-3.x, we do not default to the arrow pointer on GTK4.
Just mimic what we have on X11 and Wayland: the trusty standard arrow pointer.
Fixes issue #3581.
We were calling _gdk_surface_update_size() every frame, even if the
window size didn't change. This would cause us to discard all cached
buffers and redraw the whole screen.
This was BAD.
Recognize a common pattern: A rounded clip with
a color node, followed by a border node, with the
same outline. This is what CSS backgrounds frequently
produce, and we can render it more efficiently with
a combined shader.
Now that colors aren't uniforms anymore, we don't
win much by using the inset_shadow shader. The fragment
shaders of inset_shadow and border are identical. And
the regular border setup does nine-slicing.
Colors are not state that we carry across draw ops,
so setting the color on the render job doesn't make
much sense. Instead, pass the color to the various
draw calls. Add a few new ones for that purpose.
Also, shorten the names of some by going from
'load_vertices_from_offscreen' to 'draw_offscreen'.
Don't place the insertion cursor render nodes in the
middle of the text nodes for all the text. This helps
the renderer batching the text draw calls together.
This reduces how many changes we make when recording uniform state, which
increases the chances that the data offset will be the same when applying
uniforms.
Since we make full snapshots when recording uniform state of batches, we
need to perform some deduplication to avoid so many repeated uniform calls.
This uses a closed hashtable to determine if we are likely changing the
value to something new.
This does not currently compare values, it instead only compares that we
are going to point at a new offset into the uniform buffer. We could go
further if we compare upon updating values (we did that early on in the
prototype) so that offsets are less likely to be changed.
When the color passed is transparent black, use
the color from the texture as source, instead of
as mask. This lets use use the coloring program
both for regular and color glyphs, avoiding
program changes in text with Emoji.
Instead of using uniforms for color used in multiple
programs, pass it as vertex attributes. This will let
us batch more draw calls, since we don't have to change
uniforms so often. In particular, for syntax-highlighted
text.
We may want to change the interface between the
shaders and the renderer for ngl, and therefore,
sharing the shaders between gl and ngl will not
be practical, going forward.
We ought to get the coordinates of where the window menu should be
displayed using gdk_win32_surface_get_root_coords(), instead of rounding
the position that we obtained with gdk_event_get_position().
Also rename items a bit in the same function, and call
gdk_event_get_event_type() for consistency with the other backends.
Fixes issue #3704.
This is not strictly needed from an introspection perspective, but:
- GTK strictly depends on PangoCairo internally
- we want to integrate the GDK docs with PangoCairo's
So even though we don't have an explicit dependency on PangoCairo types
in our ABI, we do assume that people will be able to use the PangoCairo
API with GTK.
The introspection scanner does not handle `static inline` functions:
they are not in the shared library, so cannot be dlsym() out of it; and
the `static` keyword tells g-ir-scanner to skip the function declaration
entirely.
We can trick the scanner into thinking the gtk_ordering_from_cmpfunc()
symbol is a real, public one, by declaring and defining a regular
function under the `__GI_SCANNER__` guard; the symbol does not appear
when actually building GTK, or any code using GTK, so we don't risk
collisions.
Stop using gtk-doc, and switch to gi-docgen.
The gi-docgen tool generates API references through the introspection
data, which has various benefits:
- it does not parse C code
- it does not generate and run C code to introspect types at build time
- it does not present a different API from the one we're exporting
Additionally, gi-docgen:
- does not generate DocBook XML in order to generate HTML
- does not go through xsltproc
- parses proper Markdown
Which makes it markedly faster than gtk-doc has ever been.
Ever since we added surface-to-widget transforms,
the autofading of the fps overlay did not work anymore,
since it was given the transient transform node, most
of the time.
Fix this by passing both the root of the node tree and
the node for the toplevel widget to prepare_render, and
hande the widget node to the overlays. Update the
affected overlays that need to have their positioning
adjusted.
We were leaking buffers. This wasn't caught by valgrind and friends
because it was shared memory (with the compositor), but top(1) would
instantly see memory consumption of the app and the shell go through the
roof.
We were calling _gdk_surface_update_size() every frame, even if the
window size didn't change. This would cause us to discard all cached
buffers and redraw the whole screen.
This was BAD.
Fixes two warnings when opening the inspector on my PinePhone, these two
files were attempted to be loaded as PNG (as per the extension) instead
of SVG.
Signed-off-by: Emmanuel Gil Peyrot <linkmauve@linkmauve.fr>
Hook up the "Show fallback rendering" switch for Vulkan.
This brings home the sobering truth that the Vulkan renderer
is doing *all* fallback, since we switched from offset nodes
to transform nodes.
- style sidebar rows as regular items like buttons wrt focus
- define transition as part of the focus-ring(),
only animate outline related properties rather than `all`
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3708
The Wayland text protocol expects to get both ends
of the selection with its surrounding call, so make
that available via GtkIMContext. We add this as new
api and vfuncs, and fall back, so that existing
implementations keep working.
We were not handling mnemonics vs markup right
in all cases. Rewrite the _-stripping code to
do it during the link parsing, instead of as
a separate function. This avoids the issue of
stripping _ from attribute names in markup.
Add tests.
Fixes: 3706
guint32 is used as part of the protocol in broadway backend.
Memory size declared with it was mistakenly replaced with size_t type
which does not guarantee being 32bit on all platforms, leading to a crash.
These events don't make sense on physical devices (for starters, they
are relative to the logical pointer position). Use this device for
those events, also happens to be what the upper parts expect of them.
It turns out that we we were sometimes emitting
preedit-end multiple times, and sometimes not at
all. Same for preedit-start. To fix this up, introduce
a in_compose_sequence flag, maintain it, and use it
in the right places.
After these changes, both
C-S-u 1 2 3 Enter
Compose a e
generate the right signals:
preedit-start, preedit-changed,..., preedit-end, commit
Commit 97b5fad131 was a forward port from a gtk3 patch, but the hunk
was applied on the wrong bits of code.
Ensure the initialization paths also do mark settings read from the
portal as valid, so the checks for optional/newer settings actually have
the expected result. It is also desirable to mark settings as valid
after configuration changes (as that patch did effectively do), but not
enough to fix all situations.
The primary goal here was to cleanup the current GL renderer to make
maintenance easier going forward. Furthermore, it tracks state to allow
us to implement more advanced renderer features going forward.
Reordering
This renderer will reorder batches by render target to reduce the number
of times render targets are changed.
In the future, we could also reorder by program within the render target
if we can determine that vertices do not overlap.
Uniform Snapshots
To allow for reordering of batches all uniforms need to be tracked for
the programs. This allows us to create the full uniform state when the
batch has been moved into a new position.
Some care was taken as it can be performance sensitive.
Attachment Snapshots
Similar to uniform snapshots, we need to know all of the texture
attachments so that we can rebind them when necessary.
Render Jobs
To help isolate the process of creating GL commands from the renderer
abstraction a render job abstraction was added. This could be extended
in the future if we decided to do tiling.
Command Queue
Render jobs create batches using the command queue. The command queue
will snapshot uniform and attachment state so that it can reorder
batches right before executing them.
Currently, the only reordering done is to ensure that we only visit
each render target once. We could extend this by tracking vertices,
attachments, and others.
This code currently uses an inline array helper to reduce overhead
from GArray which was showing up on profiles. It could be changed to
use GdkArray without too much work, but had roughly double the
instructions. Cycle counts have not yet been determined.
GLSL Programs
This was simplified to use XMACROS so that we can just extend one file
(gskglprograms.defs) instead of multiple places. The programs are added
as fields in the driver for easy access.
Driver
The driver manages textures, render targets, access to atlases,
programs, and more. There is one driver per display, by using the
shared GL context.
Some work could be done here to batch uploads so that we make fewer
calls to upload when sending icon theme data to the GPU. We'd need
to keep a copy of the atlas data for such purposes.
Don't warn about Compose file constructs we don't
support. We haven't supported these for a long time,
and nobody has every complained. No need to wake
up sleeping dogs.
If our opaque region is the entire surface, then we can make the OpenGL
context opaque like we do for decorated windows. This improves performance
as the compositor does not need to blend the surface with the contents
behind the window.
Tweak the preedit text we get from IBus (via the compositor) to
match what GtkIMContextSimple produces for Compose sequences now.
This provides a unified experience.
Tweak the preedit display for Compose sequences to
be not so distracting. We only show the Compose key
when it occurs in the middle of the sequence or is
the only key so far, and use · instead of ⎄ for it.
Also, make sure to display dead keys more adequately.
Use the infrastructure already available to look up keys, instead.
This does the right thing and looks up the setting across all
sources.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3680
In some cases, we were inadvertedly merging the
preedit attributes into priv->attrs, instead of
keeping them separate. This was causing the underlines
to grow beyond the preedit and never go away. One
place where this was showing up is the fontchooser
preview.
Fixes: #3679
Instead of bending GtkGizmo to the breaking point,
split off a GtkPanedHandle class that just does
what is needed here. Its simpler, and lets us keep
GtkGizmo simpler too. Everybody wins.
Instead of bending GtkGizmo to the breaking point,
split off a GtkPopoverContent class that just does
what is needed here. Its simpler, and lets us keep
GtkGizmo simpler too.
- window resizing control area is implemented as where the shadows are drawn
- create a 'fake' shadow for the tiled case to allow easier resizing of
the tiled window ratio even if it's offset to the side of the border
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3670
The debug spew for printing out supported buffer
formats was missing a bunch, among them the fp16
formats that are interesting for HDR support.
Add them.
Note that we still only support ARGB8888. But
at least we can print out what don't support.
css_changed is a vfunc that *must* chain up.
Failure to do so broke the resizing of overlay scrollbars,
amongst other things.
This bug was introduced in 39f72b3834.
This is the only non-whitespace difference between the copies in
testsuite/reftests/ and testsuite/gsk/.
Signed-off-by: Simon McVittie <smcv@debian.org>
X11 allows keysyms to be specified in addition to strings.
We only support the strings. In the past, we ignored everything
after the string. Go back to doing that, but issue a warning
that we've ignored the keysym.
Use gdk_popup_layout_set_shadow_width to take shadows into
account when positioning popovers, and set the input region
to exclude the shadow, since we aren't doing interactive
resizing and the like.
When the popover has a beak, we make the surface size be
content size + shadow + tail, and then position the content
according to the final position inside this slightly too large
surface. The surface being too large doesn't matter, since we
set up an input region.
Similar to e89fcbfad6
but the tests we use mallinfo() in really aren't important to not worth
the effort to differentiate between mallinfo and mallinfo2. Only check
for mallinfo2().
The places view has questionable code for handling
'destroy', which seems to trigger use-after-free in
some cases. Clean this up a bit by acknowledging the
cancellation fully, rather than relying on the
destroyed flag.
When we are rendering a texture node to an offscreen,
and we have a clip, we must force the offscreen rendering.
Otherwise, the code will notice: Hey, it already is a texture
node, so no need to render it to a texture again. But when
clipping is involved, that is exactly what we want to do.
Testcase included.
Fixes: #3651
- while the case of a single half tiled window casting shadow is valid,
the main use case for half tiled windows is when they have a neighbor.
The :backdrop should be enough of a focus indicator and we avoid shadow
casting on neighbors that are on the same Z level.
These shadows cause a significant draw performance drop for maximized
windows. Disabling them increases the chances we can have faster scroll
performance of text.
There is some risk here for systems where they have a dock and you expect
the shadow to draw beneath that dock for transparency reasons.
We do not need to translate this on the CPU when we can instead push it
to the GPU in the same format and allow it to swizzle.
This fixes a huge number of memory allocations found while uploading the
GTK animation in widget-factory.
There isn't any state to modify in the type so we can use const here.
Doing so allows some of the renderer code to use const across a
number of functions so that repeated calls are elided if inlined.
These do not do modify anything so they can be marked as pure to
potentially ellide calls. Since they do dereference, I do not believe
we can make them const although that is unclear since we could technically
just return a pointer + offset. Therefore it *might* be possible to also
make these G_GNUC_CONST.
This also removes the return if fail macros from these as a good portion
of them didn't have them anyway. I think it's fair to say that access to
these incorrectly is a programmer error.
It significantly reduces the amount of code generated into generally a
movss,ret.
Just compiling these out means you have to write code slightly differently
so that you don't end up with "if ();" afterwards.
This adds a "do {} while (0)" so that we're still semantically a statement
but will also compile out.
My reading of the code is that gdk_drop_new() is not
consuming the content formats it is given, so the caller
must not pass a new reference.
Needs testing on Windows.
Whenever we communicate targets, we need to the union, otherwise
we don't tell the other side about our serialization. This makes
drops of images from gtk4-icon-browser to gimp and libreoffice
succeed in transferring data.
Fixes: #3654
When creating the output stream for a drop, we must
pass the mimetypes we support, otherwise the picking
of the right handler does not work.
Fixes: #3652
Shadow values created by gtk_css_shadow_value_new_filter or
gtk_css_shadow_value_parse_filter interpret their radius value
as standard deviation. Add a flag for this mode, and use it
where necessary.
This is the counterpart ot gtk_css_shadow_value_push_snapshot.
To make this easy, move the determination whether we need a
shadow out of the push function and save it.
GModule requires the .so file extension on macOS for historic reasons.
However Meson defaults to .dylib for modules, so we need to override
it to get the correct extension.
Fixes#3645.
This commit unsubscribes CUPS backend from a DBus
signal in idle when listening for new items on Avahi.
Since GDBus emits gathered signals in idle while
checking whether the signal has been unsubscribed
it could happen that a signal was not processed
because it was removed from hash table of
subscribed signals.
This caused the situation where printers advertised
on Avahi were not listed in CUPS backend sometimes.
We need those signals since this happens when switching
from a general subscription which listens to signals
for all Avahi services to a specific one which listens
to just _ipp._tcp and _ipps._tcp (chicken and egg problem).
This change extends set of Avahi advertised printers which
works with Gtk's CUPS print backend.
It creates a temporary queue (local printer) for each
Avahi printer in CUPS instead of accessing them directly
(via CUPS library).
This makes some printers work which did not work before and
also gives users more options to change in the print dialog.
This also changes naming of printers to be in accordance with CUPS.
It uses '_' instead of '-' and has hostname appended for CUPS remote
printers.
This function was not doing the right thing.
Once we are doing the right thing and not compare
shadows as unequal, some reftests that inhibit
snapshots for a few frames now hang forever, since
we are no more redrawing unnecessarily. Fix that
with an explicit queue_draw.
This commit modifies the gtk_list_box_row_grab_focus() implementation to
correctly return FALSE in case the row couldn't be focused. This behavior will
also be honored by gtk_list_box_row_focus() to properly handle focus in that
case. To achieve that, the method has been restructured slightly.
Closes#3633
As the program executable name has 'update' in its filename,
gtk4-update-icon-cache.exe is considered to be an installer program on 32-bit
Windows [1], which will cause the program to fail to run unless it is running
with elevated privileges (i.e. UAC).
Avoid this situation by embedding a manifest file into the final executable
that tells Windows that this is not a program that requires elevation.
Fixes issue #3632.
[1]: https://docs.microsoft.com/en-us/previous-versions/windows/it-pro/windows-vista/cc709628(v=ws.10)?redirectedfrom=MSDN,
under section "Installer Detection Technology"
Keep string values in the table, and return them
from the check function. This commit temporarily
disables the table caching, since the cache format
does not handle string values yet.
Fixes: #186
Make it possible for gtk_compose_table_check to return
a string instead of just a single Unicode character.
Currently, we only ever return strings holding a single
character, still.
This attempts to fix the counter-intuitive resizing of surfaces in GTK4 where
the surface grows or shrinks at the right and/or bottom edge when the window
resized from the top and/or left edge(s).
This is not yet perfect as the window stutters upon resizing from the top
and/or left edges, but at least makes resizing more intuitive.
Remove the 'resized' member from the GdkWin32Surface structure, as we already
have a structure with a member that keeps track of whether a surface is being
resized, so we can just use that and avoid some confusion in the process
Drop GTK_MAX_COMPOSE_LEN from docs. It is no longer
used by GTK at all. We leave the define in place
for now, to avoid breaking 3rd party code that might
use it.
Sometimes the size will exceed the minimum bounds. For example crazy
applications like the widget factory that contains the world, or when a
user interactively resizes a window to be larger than the monitor the
window is on is.
The former is questionable, but the latter is not, and from here we
can't really see the difference, so just stop complaining.
Closes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3035
A bunch of keysyms for dead keys have been added since this
code was last touched. Update the check to cover the full
range from dead_grave to dead_greek.
Make _gtk_window_unset_focus_and_default queue the changes
for after the next draw. This achieves two things: first,
it avoids invalidating css at the wrong time (e.g. when
setting child-visible during size-allocation), and second,
it defers the focus change until after the widget is
hidden, so that moving the focus has the desired effect
of picking a different, visible widget.
Fixes: #3623
gstreamers gl support included headers from this
package without depending on it. Work around this
packaging error by explicitly adding the needed
dependencies.
Visual Studio does not come with unistd.h, but Windows do have write() and
close() in io.h, so include io.h instead of unistd.h on Windows.
For MinGW, unistd.h in turn includes io.h.
Now that we have gtk_drag_check_threshold_double(), be consistent with
other draggable widgets and make sure we don't take over a drag before a
child does.
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3513
If multiple nested widgets have drag sources on them, both using bubble
phase, we need to reliably pick the inner one. Both of them will try to
start dragging, and we need to make sure there are no situations where the
outer widget starts drag earlier and cancels the inner one.
Currently, this can easily happen via integer rounding: start and current
coordinates passed into gtk_drag_check_threshold() are initially doubles
(other than in GtkNotebook and GtkIconView), and are casted to ints. Then
those rounded values are used to calculate deltas to compare to the drag
threshold, losing quite a lot of precision along the way, and often
resulting in the outer widget getting larger deltas.
To avoid it, just don't round it. Introduce a variant of the function that
operates on doubles: gtk_drag_check_threshold_double() and use it instead
of the original everywhere.
This was needed to work around widgets claiming event sequences on press,
by ignoring them and starting the drag anyway unless they have certain
event controllers on them.
The most visible offender was GtkButton, but since the last commit it
doesn't claim the sequence anymore and we can remove the hack.
Currently GtkButton claims the sequence both on press and on release. Stop
claiming it on press and only do it on release, allowing drags to start
from it.
This will allow to remove a hack from GtkWindowHandle.
Before turning off pangos rounding of glyph positions,
we must check if the cairo we are using is new enough
to have working subpixel positioning (the relevant
cairo commit is 52a7c79fd4ff96bb5fac175f0199819b0f8c18fc).
This adds a "release" destructor for the gtk_surface1 interface which
signals to the server that a surface has been destroyed on the client
side, which the current "destroy" does not do.
Ideally the protocol would have specified a destroy request marked as
destructor to handle this automatically, however this is no longer
possible due to the destroy method being implicitly generated in the
absence of an explicit request in the protocol. Adding a destroy request
marked as destructor now would generate a new destroy method that
unconditionally would send the request to the server, which would break
clients running on servers not supporting that request.
Setting can-focus to FALSE on a widget is supposed
to prevent focus from entering the entire subtree.
So when we grab focus directly to a widget, we need
to check the can-focus flag not just of the widget
itself, but all its ancestors.
Fixes: #3610
This property needs to be treated as flags, not as
enum, since it should be possible to specify more
than one value, e.g.
text-decoration-line: underline overline;
Tests included.
Fixes: #3621
When the natural size of the popover exceeds
the monitor width, clamp it to fit on screen.
This gives the Emoji chooser a chance to
work on phone screens.
Fixes: #3307
Use a flowbox instead of a horizontal box for
the section buttons at the bottom, so they can
wrap and allow for narrow layout.
We also need to stop giving the scrolledwindow
a content height, so it can shrink to make room
for more rows of section buttons.
The number of 7 columns was hardcoded in a few places
related to keynav across sections. The flowbox does
not have an api for it, but we can find out anyway
how many columns there are.
Implement most of transform-origin. We only
handle the xy components currently, which lets
us reuse the position value implementation that
is used for background-position.
Rearrange our build to create a libgtk.a for the the contents of gtk/,
and build libgtk-4.so from our per-subdirectory static libraries.
This will let us link tests statically for testing internal apis.
Every program starts in __libc_start_main, so this would suppress
everything. This doesn't happen in practice because asan by default
is unable to unwind the stack. Setting
ASAN_OPTIONS=fast_unwind_on_malloc=0,malloc_context_size=15 can unwind
the stack, and if malloc_context_size is large enough it will indeed
suppress everything.
Unify the many copies of diff_with_file in one source
file, and patch it to detect diff at runtime and fall
back to a simple strcmp if we don't have it. Make all
tests use this new testutils.c, and stop requiring
diff for building the tests.
This should let us allow to build on Windows with the
default value for -Dbuild-tests.
We only realize the ATContext on the top level, which will create an
GtkAtSpiRoot object and the corresponding GtkAtSpiCache object. Whenever
an AT connects to the accessibility bus, and asks for the various
objects, all the ATContext will be realized on demand.
Instead of hardcoding icon names in the widget, use
arrow.none, arrow.up, arrow.down, arrow.left, arrow.right
styles and set the icon to use with -gtk-icon-source. This
lets themes change the icons that are used here, without
forcing all uses of pan-up/down/start/end-symbolic to be
treated the same.
Document this in the menubutton CSS docs.
Instead of hardcoding icon names in the widget, use
sort-indicator.ascending and sort-indicator.descending styles
and set the icon to use with -gtk-icon-source. This lets themes
change the icon that is used here, without forcing all uses of
pan-up/down-symbolic to be treated the same.
Document this in the treeview CSS docs.
Fixes: #3577
Single click activation should only be done on release and not on press,
otherwise it breaks touch scrolling. Double-click activation still can be
done on press.
This matches the GtkListBox behavior as well.
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3345
Commit cbbbf44dd4 tried to replace
gtk_style_context_save_named with an explicit
css node to save to, but it failed, and the failure
was showing up as crashes in the inspector.
Fix by Benjamin Otte.
Fixes: #3599
In GTK4, we are now defaulting to the OpenGL renderer with the Cairo renderer
only used as a fallback, so there is no point keeping the code paths that use
layered windows as layered windows do not work well with OpenGL nor Vulkan.
We don't need to override create_layout_child here,
and setting the child type has the advantage that
the layout properties are showing up in the inspector.
Make HighContrastInverse available as the dark variant of HighContrast,
in addition to being a standalone theme. This regularizes our theme
variants, and doesn't hurt.
Commit e6209de962 added some checks on TranslationEntry.valid in
order to figure out whether using the new font settings or the
old g-s-d ones. However that's only set in the non-sandboxed case.
This makes sandboxed applications fallback to the old (and also
non-existing with modern g-s-d) settings, possibly resulting in
ugly defaults being picked.
Fix this by also marking TranslationEntry elements as valid when
using the settings portal, precisely those entries that we are able
to read and match with our own table.
When I added the sassc option, I failed to ensure that
the theme css is included in the tarballs. This is required
to make tarball builds with -Dsassc=disabled work.
When falling back to g_get_prgname(), we need to take into account that
the program name may be the full argv[0] path, which will end up messing
the DBus object path.
When the icon size changes, we want to reload themed
icons.
This was showing up as the GtkImage in about dialogs
getting a big size with the same small icon upon
changing -gtk-icon-size in the inspector.
Have an implementation of ->request_layout() and ->compute_size() for the Win32
surface backend so that we can properly display and move and resize the
windows, as we request from the Win32 APIs.
Hxndling Aerosnap properly is mostly done except for snap_up(), which needs to
to be looked at later.
In line with what is done with the Wayland backend, enable the mapped state
independently as needed from the toplevel surface presentation, and also enable
the mapped state if necessary when presenting the popup surface.
The fact that we are using gdk-pixbuf for loading files currrently does not mean we will use it going forward.
Also, "anything gdk-pixbuf can load" does not mean anything, because what gdk-pixbuf can load is a compile-time option.
As new_from_resource() will assert() if it cannot load a resource, we must be very sure that people do not use anything but PNG and JPEG for resources and the docs were not clear on that.
When destroying a wl_surface (e.g. when a window or menu is closed), the
surface may continue to exist in the compositor slightly longer than on
the client side. In that case, the surface can still receive input
events, which need to be ignored gracefully.
In particular, this prevents segfaulting on wl_surface_get_user_data()
in that situation.
Reported in
https://gitlab.gnome.org/GNOME/gtk/-/issues/3296
The same issue for pointers/keyboards was reported in
https://bugzilla.gnome.org/show_bug.cgi?id=693338
and fixed with in
bfd7137ffb3625f17857a8fc099a72
In pointer_surface_update_scale(), only rescale the cursor surface when
the scale has actually changed and the cursor is on at least one output.
fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3350
Right now, this issue is not completely understood, so it might also
involve some questionable handling of cursor surface by sway/wlroots.
However, irrespective of that issue, this patch avoids unnecessary calls to the
compositor, and there should be no drawback: Whenever the pointer enters
a new output, pointer_surface_update_scale() will be called again, such
that correct scaling of the cursor is still ensured.
There is a slight difference: When the cursor leaves the last output,
previously the image was reset to scale factor 1. Now, it keeps whatever
was last. That might be more sensible than the previous behaviour,
assuming that it's likely that when the cursor enter an output again, it
has the same scaling. Alternatively, if one cares about resource usage
at this level, it might make more sense to destroy the surface than
rescaling to 1.
Visual Studio 2013 is just shy of being sufficiently C99-compliant to
build GTK master, as it did not support snprintf() in its CRT
implementation.
Use g_snprintf() to cover for this.
This rule is almost always wrong, and we're having some trouble agreeing
on replacement text. Let's remove the whole paragraph for now, to avoid
confusion.
In GTK 3, it was pretty much always better to use
gtk_builder_new_from_whatever() rather than
gtk_builder_add_from_whatever(). But in GTK 4, this is no longer
generally true. Let's add some context on when you want to use one or
the other.
To support Sierra, we need to have access to pasteboard types as a
NSString. Constants are provided in later versions of macOS, but we
can emulate that with an array which is initialized on first access.
On older systems, the availability of some methods seem to be incorrect
based on Apple documentation. This works around the issue by using
the rect conversion on older systems.
We only every parse one of those at a time, so keep a GString around for
it. This way we don't have to create a new GString every time we parse
an identifier. Doesn't spare us the strdup though.
Using a variadic function is just awkward there. Use different init
functions and use them as appropriate; we already know which one to call
everywhere.
...for at least WGL, since we are not allowed to use wglMakeCurrent(),
which is eventually called by gdk_gl_context_make_current() to share WGL
contexts across different threads, which will cause a crash.
This means, we cannot enable WGL in the gstreamer media backend on
Windows.
The demo is called 'Video', so we should actually
use a GtkVideo widget. Doing things this way has
the advantage that we get GL support, which wasn't
working with GtkPicture.
This makes it so that we obtain a GL context before
creating the GstSink, so we at least advertise our
ability to handle GL buffer.
Gstreamer still won't send us any.
Most of this is cribbed from glib.
We don't use COMMON_MESON_FLAGS here, since disabling
-Werror one subproject at a time is too painful for
all the promoted sub-sub-projects. This needs a better
solution in meson.
Since this is a full rebuild of the entire stack, limit
this to master and merge requests.
On x11 toplevel layout is not created before toplevel
is presented, but GTK tries to update it on idle
which leads to a crash due to accessing property
of undefined object. Treat soon to be created layout
as a layout with default values upon creation (resizable).
Instead of walking the pspecs twice and checking for the owner_type,
just allocate n_pspecs entries. They are the same value right now
anyway, but even if they aren't it doesn't hurt to have a few extra ones
allocated.
This does not change anything in the case where the .init file does not
exist. We still hit the disk once. Otherwise, we now only hit it once
instead of twice and avoid a potential race-condition.
These functions were not implemented when the sizing changes
landed before GTK 4 was released. This fixes an issue with non-
resizeable windows not reacting to layout changes.
Fixes#3532
We were not taking spacing into account when adjusting
the size of an expanding center child, causing it to slip
under the end child at times.
Fixes: #3506
If there is no secondary text, then the primary text is just a message,
not a title, and should not use title style.
This partially reverts 1e3ec7c1f9. The
message dialog nown looks like it used to in GTK 3. However, it's still
styled only using a style class rather than using pango markup, as in
GTK 3.
Fixes#3509
We were not aligning model button labels with nesting
submenus with indicators in their containing menu.
This was visible in the application demo of gtk4-demo.
According to OpenGL spec, a shader object will only be flagged
for deletion unless it has been detached; when a program object
is deleted, those shader objects attached to it will be detached
but not deleted unless they have already been flagged for deletion.
So we shall detach a shader object before it is deleted, and delete
it before the program object is deleted best.
Some GTK based applications such as Qemu UI create and manage
EGLSurfaces associated with the relevant GdkSurfaces. In order to create
an EGLSurface, there needs to be a way to pass the native window
object to eglCreateWindowSurface(). While running in an X environment,
the native window object can be obtained by calling
gdk_x11_surface_get_xid(). Likewise, the native window object can be
obtained by calling gdk_wayland_surface_get_wl_egl_window() while
running in a Wayland environment. Therefore, this API needs to be
exposed to apps.
Signed-off-by: Vivek Kasireddy <vivek.kasireddy@intel.com>
Using the module will prevent trivial bugs caused by incorrect string concatenation.
This is just a literal translation. In the future we might rely on Meson auto-detecting fields from library declaration.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3517
The memory pointed to by GstVideoFrame::plane_data becomes invalid after
unmapping causing the GBytes to point at some random memory if the
unmapping is not deferred until its destroy notify.
When the GStreamer buffer is backed by normal system memory this is not
a problem but if it is backed by e.g. an OpenGL texture, dmabuf or some
other hardware-specific memory this will otherwise cause interesting
problems.
Here is a command to reproduce this testcase:
GDK_DEBUG=gl-gles gtk4-demo --run gears
Without this patch, Mesa throws this compile error:
0:130(13): error: no matching function for call to `mod(error, float)'; candidates are:
This is caused by `u_rotation - 90` being of type error since
`u_rotation` is a float and it’s illegal to subtract it with an integer.
Motion history coordinates are based on GdkSurface coordinates. Transform
the coordinates from the GtkNative they are received on, accounting for the
possible transforms (e.g. due to window borders and shadows).
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3491
If set to TRUE, does not call the free func for the removed items.
This can be used to move items between arrays without having to do the
refcounting dance.
1. That's a bad idea UI wise as you can't see the search after you
scrolled.
2. That's a bad idea code-wise because Listviews need to be put into
a ScrolledWindow or they won't scroll.
The special case for ASCII glyphs is unfortunately not
working very well, because of an oversight in pango:
When I added subpixel positioning, I made pango_shape()
default to not rounding, and make PangoLayout call
pango_shape_with_flags() and pass the rounding information
down. The upshot is that we need to use the _with_flags
variant here and tell it to round position, so it matches
what the text node contains.
We only need this to render shapes and trapezoids, i.e. only when
falling back to cairo. Remove code to measure the layout and convert the
ink_rect to a graphene_rect_t from gtk_snapshot_append_layout() and do
it when drawing shapes and trapezoids instead.
Rename a few local variables to be clearer about whether they belong to
the widget or to the layout.
Remove a workaround for an old bug that is no longer valid. We don't
underallocate the layout anymore. Aways center vertically, with respect
to the yalign.
We have a region for the border here, but the color is defined as
transparent black because we don't want to render anything for it.
This way, the generated .node file for the listbox demo in gtk4-demo
with enabled layout borders is 3.5MB instead of 3.8MB.
This way we can render the first frame of tests/testoutsetshadowdrawing
in 153 ops instead of 183.
And the first frame of gtk4-demo in 260 instead of 300.
We cannot bubble up ::insert-text and ::delete-text signals from the
delegate to the wrapper editable without causing an infinite recursion,
due to how the signals are designed; ideally, we would have decoupled
signal emission from virtual functions *before* releasing GTK4, but now
that we have, we can't break the contract.
The GtkEntryBufferClass.deleted_text() behaviour changed between GTK3
and GTK4, which means any subclass of GtkEntryBuffer is now responsible
for emitting the "notify" signal for the "text" and "length" properties.
Without this, the GtkText delegate widget used by GtkPasswordEntry would
not be able to communicate changes in the contents of its buffer.
Fixes: #3484
- normally lighter (on bg_color)
- darker for headerbar
- undecorated checked buttons were never style properly (even gtk3)
see page2 volume buttons
Finetunes https://gitlab.gnome.org/GNOME/gtk/-/issues/3427
Float properties can not be read from keyfiles. This was causing changes
to this property in settings.ini to be ignored. Fix this by changing it
to a double.
Technically this change could be considered an API break, but there are
no users of this property as float yet and it is early enough to expect
there to never be any.
Also document this change in the 3 to 4 migration guide since in gtk 3
this setting will be kept a float.
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3441
../gdk/gdktoplevellayout.c:217: Warning: Gdk: gdk_toplevel_layout_get_maximized:
unknown parameter 'maximized' in documentation comment, should be 'maximize'
100% symbol docs coverage.
833 symbols documented.
0 symbols incomplete.
0 not documented.
What's left are just type system macros and windowing system opaque
structures.
Depending on the input driver, we will get XI_Motion based scroll
events for regular mouse wheels. These are intended to be handled
as discrete scroll, so detect smooth scroll events that move by
exactly 1.0 in either direction.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3459
When being fullscreen, and wanting to unfullscreen but not caring about
whether to go unmaximized or maximized (as this information is lost), if
the GdkToplevelLayout represents the full intended state, we won't be
able to do the right thing.
To avoid this issue, make the GdkToplevelLayout API intend based, where
if one e.g. doesn't call gdk_toplevel_set_maximized() with anything, the
backend will not attempt to change the maximized state.
This means we can also remove the old 'initially_maximized' and
'initially_fullscreen' fields from the private GtkWindow struct, as we
only deal with intents now.
When deciding whether to call gdk_toplevel_present(), check the
mappedness of GtkWidget instead of the mapped-ness of GdkSurface, as the
latter is mapped asynchronously, while the former is direct state of the
GtkWindow itself.
This fixes an issue where calling e.g. gtk_window_maximize() shortly
after showing it didn't take effect, as hadn't been mapped yet. While
maximizing after showing is racy, and will likely glitch from time to
time, the correct thing to do is still to maximize and handle whatever
state the compositor configured the window to be in.
Closes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3466
- Add borders around the main scrollable content, like most other
places.
- Set padding and spacing on .emoji-searchbar and .emoji-toolbar in a
more proper way.
- Remove unused button.emoji-section label styles.
This is a more reliable calling point than ::resume-events, and a
good one to schedule things so they happen on a frame clock in no
special phase (Thus still fixing the original issue at 80d4a08e30)
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3461
These ended up square in the various refactors. Make them again L-shaped
by extending these along the edge being checked. This means we have to
check for corner positions in all edges, though.
In order to do this, leverage smooth scroll handling into the capture
phase scroll controller, controlled by ::scroll-begin in the propagation
phase controller.
There's 2 cases here:
- A child widget handles scroll. The scrolled window does not get
::scroll-begin, the child widget handles the full scroll sequence.
- No child handles scroll, the scrolled window gets ::scroll-begin,
and transfers control of scrolling to the capture phase controller.
As scrolling is performed, the pointer may fall into scrollable children,
but the scrolled window will be capturing the scroll events, so these
won't be seen by the child widgets.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/593
We let smooth scroll events that don't trigger a ::scroll signal through.
This is unintended, these are handled, even if just accumulated.
This fixes cases like GtkSpinButton inside GtkScrolledWindow, where both
would handle events, until the GtkSpinButton eventually shifts away from
underneath the pointer.
Brought up at https://gitlab.gnome.org/GNOME/gtk/-/issues/593
This code was here in gtk3 to cater for the completion window being
positioned. That was only to meant once as long as the completion window
was shown.
This doesn't work as well for gtk4, ::size-allocate gets propagated from
the toplevel, so happens much more often for the completion window, this
ends up with the completion position being reset to the first row
frequently.
Do this simply once when popping up the completion, instead.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3083
The coordinates are already widget-local here, not transformed by the
adjustment positions. Using the adjustment value here ends up pushing
the entry far from the left border.
The correct minimum value here is 0, which matches the treeview left
border.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3009
Let model buttons handle unpaired releases, these may happen indirectly
e.g. due to other child menus being opened at the time. Clicking would
dismiss the menu, but the menu item beneath the pointer would not get
activated.
We can handle that button release though via ::unpaired-release, so
there's no second click required.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3463
Tracking it through the GtkGestureClick becomes a bit cumbersome for
handling of simultaneously pressed buttons. We can track ::stopped,
but that also emits for a number of situations where we want drag to
continue.
However, the GtkGestureDrag is grouped with the click gesture, and
knows better when to finish the drag gesture (not just because of a
button release), so hook drag and zoom mode finalization there.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3426
We released all the 3.9x development snapshots using:
libgtk-4.0.so.0.xxxx.y
which means that the 4.0.0 release of GTK will have a lower SONAME than
the snapshots. To avoid that, bump the SONAME to 1.xxxx.y.
Fixes: #3392
We don't want to increase the interface age manually, because we're
going to end up forgetting about it.
Instead, we should tie it to the rest of the version:
- in stable (even minor) cycles, we don't add new API; the interface
age is the same as the micro version
- in unstable (odd minor) cycles, every new release might have new API,
or updates to newly added API; keep the interface age to 0
This removes one more manual thing to change during release, and keeps
us honest with our promise not to add symbols during stable cycles.
Having a property with the same name of a method makes some languages
which put properties and method names in the same flat namespace not
really happy about our choices. To avoid collisions, let's put the the
"fullscreen" property into the past, alongside the "maximized" property.
Drop the "don't write code that crashes" folksy bit; don't start
sentences with conjunctions; remove copy-pasted references to
"iconified" states; point to the corresponding property notification in
a consistent manner.
Currently, the implicit grab is broken on the first button release,
in the case of pressing multiple buttons simultaneously. This means
that we emit crossing events early, and the next button releases
are sent to the pointer focus widget instead.
Consider the implicit grab effective until all buttons are released,
and only unset the pointer implicit grab (and emit crossing events)
after there are no further buttons pressed. We do this by checking
event modifiers, given button release events do contain the modifiers
in effect at the time the event was generated, we have to look for
exactly one active button modifier.
Fixes weird pointer states after pressing multiple buttons on a
widget.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3426
Those requests are received while dealing with the ::layout frame
clock phase, this has the unintended side effect of making the
frame clock "rewind" to handle ::flush-events again during this
frame, which delays everything and practically halves the frame
rate.
We do intend to make the motion events dispatches on the next frame,
so do this in an idle at a slightly lower priority than layout/draw,
so the ::flush-events phase is actually requested for the next frame.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3264
It tested that a widget shown after the window was mapped deals with
window resizing appropriately when doing the follow-up allocation. Now,
doing this only allocates both at the same time, as allocation happens
during the frame dispatch.
To do the equivalent now, one would have to write code that shows the
label after the first frame was drawn, and that's not possible via a
reftest, so lets remove it.
We don't need to go through the NSOpenGLContext for these.
We can just use the C API directly. It's also clearer what is using
CGLEnable() vs CGLSetParameter().
It was used by all surfaces to track 'is-mapped', but still part of the
GdkToplevelState, and is now replaced with a separate boolean in the
GdkSurface structure.
It also caused issues when a widget was unmapped, and due to that
unmapped a popover which hid its corresponding surface. When this
surface was hidden, it emitted a state change event, which would then go
back into GTK and queue a resize on popover widget, which would travel
back down to the widget that was originally unmapped, causing confusino
when doing future allocations.
To summarize, one should not hide widgets during allocation, and to
avoid this, make this new is-mapped boolean asynchronous when hiding a
surface, meaning the notification event for the changed mapped state
will be emitted in an idle callback. This avoids the above described
reentry issue.
We only called xdg_toplevel.(un)set_maximize() if the toplevel layout
changed, but this misses the case when the compositor had changed the
maximized state. Change it to call the xdg_toplevel request if either
the local layout changed, or if the layout differs from the current
state.
This fixes an issue where one couldn't unmaximize a window by double
clicking the titlebar that, had previously been maximized e.g. using a
keyboard binding.
Do the same for fullscreen.
This makes it possible to set 'maximized' to true in .ui files, and the
window will show up maximized.
gtk_window_is_maximized() will return the intended maximized state until
actually mapped, it will then show the actual maximized state. The same
applies to reading the property.
This commit changes the behavior of window size computation and the
default size properties to:
* The default-width and default-height properties are updated to the
current window size unless the size is fixed by e.g. being maxmized,
tiled etc.
* The compute-size semantics are to just pick the default size, or if
not adequate, use the measured size, and consequently update the
default size, unless unresizable.
* gtk_window_get_size() is removed, what's more likely relevant is the
gtk_window_get_default_size() which will now contain more sensible
values.
Various places that used gtk_window_get_size() were updated to use
gtk_window_get_default_size() to remember and restore previous sizes.
This also changes the default value of 'default-width' and
'default-height' from -1 to 0. The gtk builder simplify tool is taught
how to omit when the default size is set to both -1 and 0.
This will sometimes mean a frame is skipped if a resize was requested
during the update phase of the frame dispatch. Not doing so can cause
trying to allocate a window smaller than the minimum size of the widget.
If compute_size() returns TRUE, the layout will not be propagated to
GTK. This will be used by the X11 backend to queue asynchronous resizes
that shouldn't yet allocate in GTK.
Not doing this means the next time the same surface is shown, if the
shadow size wasn't changed, it wouldn't be sent to the compositor, which
then would result in compositor deriving its own window geometry which
would include the shadow margin.
This fixes an issue where the file chooser dialog would grow each time
it opened.
The allocation of popups are part dependent of the allocation of the
root, which means the root must still be allocated when updates are
frozen, otherwise we'll try to allocate non-laid out popups.
This fixes an issue where the focus of the window continuously received
fake motion events even when a popover was open, making input events end
up behind the popover.
It also adds a comment describing why motion events are requested. Note
that popovers won't work with this, and it's possible both in the past
and now that sporadic missplaced motion events will appear, e.g. when a
window changes allocation but a popover is open.
This removes the GDK_CONFIGURE event and all related functions and data
types; it includes untested changes to the MacOSX, Win32 and Broadway
backends.
While the workaround hides majority of the issue there are still two big downsides:
- shadow does eat from the widget dimensions so alignment is broken
- situations like popover going upwards on screen edge break completely
The appropriate action is to revert these theme duct tape solutions to make room
for a proper fix.
This reverts commit b3dba1dca6.
Issue https://gitlab.gnome.org/GNOME/gtk/-/issues/1987
This removes the gdk_surface_set_shadow_width() function and related
vfuncs. The point here is that the shadow width and surface size can now
be communicated to GDK atomically, meaning it's possible to avoid
intermediate stages where the surface size includes the shadow, but
without the shadow width set, or the other way around.
This changes allocation of the widget trees to happen as a side effect
to the GdkSurface::layout signal, which first passes the GtkNative
instance where it is then forwarded to the implementations of the
GtkNative interface.
The implementations of GtkNative are the ones doing the actual
gtk_widget_allocate(), and they do so in their GtkNativeClass::layout
function.
The size should correspond what gtk_widget_measure() does, and it
measures what's within the window excluding the shadow; so make this
helper function correspond to this.
GTK4 doesn't support arbitrary constraints when resizing a window (e.g.
steps, or aspect ratio), so we don't need to care about the result from
compute-size when doing interactive resizing.
This follows the trail of the Wayland backend in that GdkSurface changes
happen during the layout phase, and that a GDK_CONFIGURE no longer being
used to communicate the size changes of a surface; this now also uses
the layout signal on the GdkSurface.
If a surface scheduled a relayout, got frozen, and a layout phase
happened, then got unfrozen, it wouldn't see it's layout being
requested; avoid this race by remembering the pending phases until they
actually happened.
Reading the comment, it seems to be related being a window manager
decoration utility; this is not something GTK4 aims to handle, just drop
support for this.
Showing before the child would result in bogus
gdk_drag_surface_present() with an "empty" (1x1) size. This can easily
be avoided by postponing showing until there is anything to show.
By moving popup layout emission to the layout phase, the current
GdkPopup::poup-layout-changed signal has no value on its own as it'd be
ignored by GtkPopover.
Make the Wayland backend communicate the popup layout changes via the
common signal; but leave the rest intact until other backends catch up.
Put them in a anonymous struct, and separate the toplevel specific ones
into another anonymous struct inside the first one. Later popup related
fields will be added.
Don't have GtkRoot listen directly to the layout signal on the frame
clock, but let it pass through GdkSurface. This will allow GdkSurface to
be more involved in the layout phase.
Scheduling an update when frozen would reschedule when unfrozen; change
this to a generic pending phase enum, and use this for resrcheduling
paint and compute-size.
GdkSurface's are initialized to have the size 1x1, as otherwise we'd
receive an X11 error, would a corresponding X11 window be created.
This confuses the "saved size" mechanisms in the Wayland backend, as
treats 0 as uninitialized, and not 1.
Fix this simply not saving size that if it's smaller or equal than 1.
This will be handled between 'update' (which may trigger animation
ticks, CSS update, etc) and 'layout' which will allocate the widget
tree. It's meant to perform surface size computation, and is done
between these two phases in order to have an up to date state, and
letting the layout phase have an up to date size to layout in.
Concentrate state application to the start of a frame; this is to avoid
having GTK going back and forth between different state if so would
happen between two frames.
Queue it, and then wait for it to actually take effect, i.e. be
confirmed via a configure event from the compositor, before setting the
actual GdkSurface::state value.
The plan is to concencrate size computations as part of the frame clock
dispatch, meaning we shouldn't do it synchronously in the present()
function.
Still, in Wayland, and maybe elsewhere, it is done in the present()
function, e.g. when no state change was made, but this will eventually
be changed.
Mapping a surface under Wayland is an asynchronous process, where one
creates a surface and commits an initial state without having drawn
anything, then waiting for a configuration, which then is acknowledged
and content is painted and committed. Not until having received this
configuration is a surface actually mapped, so wait with setting the
mappedness until this.
We don't need the OpenGL view to be transparent if the window itself
is not transparent. This has the potential to speed up the compositing
of the GL view onto the NSWindow.
We need to tell people what signal will be emitted when calling
gtk_widget_activate(), and that the shortcuts API might be more
appropriate to what they are looking for.
Setting a field on a class structure is not always an easy task from
languages other than C. While bindings can provide access to the class
pointer, twiddling the fields in the class structure can be awkward.
Additionally, signal ids are not always readily available.
We can paper over the direct access to the class structure, as well as
the "signal name to id" mapping with a simple couple of setter
functions.
These positions are not guaranteed to be in a specific order when linked
into the final GPU program. They need to be specified so that our code
in gskglrenderer.c can use known positions for them to match up with
our GskQuadVertex.
This fixes the GL renderer on macOS's OpenGL shader compiler.
Fixes#3420
This broke when we started using GDK_PROFILER_CURRENT_TIME for
timekeeping - that gets defined to 0 when we're building without
sysprof, so we can use it to make such determinations. Go back
to using g_get_monotonic_time().
Fixes: #3438
This fixes an issue where we would ignore events with Y delta
and no X delta while scrolling due to a typo when checking for
any delta.
This fixes deceleration of kinetic scrolling on the macOS backend.
Fixes#3418
On the macOS OpenGL implementation, the use of noise2 as a
function within the glsl shader collides with the builtin noise2 of a
different signature.
This changes the name to something similar (noize2) so that we
do not risk colliding names when linking.
With this commit, the shadertoy alienplanet demo works on mac
OpenGL (albeit still with the Cairo renderer).
Catch the error when it happens, so that we can emit a specific and more
helpful error message.
Also verify that all branches in the code now do indeed set a proper
GError when they fail, so that the final catch-all is no longer needed.
Instead, assert that the error is set so that we catch future code
additions early that do not set the GError.
This ensures that we don't leak window references inside the action muxer.
Otherwise, we risk not disposing the windows upon gtk_window_destroy()
and blocking the main loop from quitting.
Fixes#3419
We need to keep this consistent so that we can look things up
faster in other places. Therefore, just take the hit here and clear
the entire list ensuring prev/next poniters are cleared.
This comes complete with animation support. For a good time, try:
@keyframes conic {
100% { background-image: conic-gradient(from 1turn, red, lime, blue, yellow, red); }
}
window {
background-image: conic-gradient(red, lime, blue, yellow, red);
animation: conic infinite linear 5s;
}
The keycode and modifier (state) parameters are in the wrong order
for gdk_key_event_new() in the gdk win32 backend, which causes
key up/down events to be populated incorrectly.
This isn't done automatically for us, so we need to synthesize it in
our hide helper.
With this commit, we properly re-focus the new main/key window after
we have closed a transient-for window.
First of all, we must list *all* ignored headers. Since we have public
headers in the x11 and wayland directories, we must explicitly declare
all headers that we consider private under those directories.
The "quartz" subdirectory was renamed "macos", with the new macOS
backend. The "mir" directory was removed, so there's no need to ignore
it.
We are also missing a bunch of ignored headers in the top-level gdk
directory.
Finally, pass the list of ignored files to gtkdoc-mkdb, so we won't get
missing declaration warnings.
We need to re-attach to the transient-for window whenever we present or
we risk getting placed behind the window by the display server. Apparently
that setting does not persist across a hide of the NSWindow.
We only needed the widget to get its action muxer. And this way we don't
have to call gtk_widget_get_action_muxer() dozens of times, just once in
set_widget().
The get_size() vfuncs is deprecated in GTK3, and only used as a
fallback path for cell renderers that do not implement preferred
size virtual functions.
since http RFC state that the header names should be processed case in-sensitive, broadway should not rely on the actual case. E.g. the go-language libraries tend to rewrite the header, which cause problems with e.g. Caddy
Fixes#3406
We have code in place to handle a NULL node in the state when dealing
with blend nodes, but we don't always check for NULL, which leads to
warnings in the CSS Blend modes demo.
Use the same initial check for the accessible object type that we use
when connecting the signal, in case we try to disconnect signals on
different widgets. Additionally, check before accessing data that might
have already been removed.
Fixes: #3403
In gdk/win32/gdkmonitor-win32.c in function
populate_monitor_devices_from_display_config() refresh->Numerator * 1000
overflows for refresh->Numerator > 4294976.
Cast the factor 1000 to UINT64 to prevent the overflow.
Fixes#3394
The accessibility bus might not be available, and if it isn't the case,
it means something has failed at a level where the user can't do much
about it. There's no need to emit a critical warning.
We depend on GObject, which already depends on GLib; on older versions
of Meson, this ends up trying to depend on the system copy of GLib,
instead of the sub-project copy, in case the version of GLib we have
installed is too old.
To avoid bleeding edge deprecations we use GLIB_VERSION_MIN_REQUIRED and
GLIB_VERSION_MAX_ALLOWED. Since we depend on GLib 2.66, we cannot use
API introduced in 2.67, even when conditionally compiled.
- Reorder declaration blocks for code legibility.
- Don't lighten text color on :hover and :selected for better contrast.
- Add missing :focus-visible:focus-within styling to non-:selected row.
- Don't set the sidebar padding twice.
- Add padding to the assistant sidebar directly, as it doesn't have
.navigation-sidebar internally.
- Set missing margin to the sidebar separator.
See https://gitlab.gnome.org/GNOME/gtk/-/issues/3382
If an active grab gets undone on button press (e.g. closing a menu), we
will receive a button release on the new target even though it didn't handle
the button press, and disable ::active state.
This causes warnings when handling the button release, as it tries to undo
::active state that is not really there.
In order to fix this, check that the pointer focus actually had an implicit
grab at the time of receiving the button release, before trying to unset
the ::active state.
If the accessible object is hidden, we can skip the emission of the
AddAccessible and RemoveAccessible signals on the cache, as those
objects won't be visible in the accessibility tree.
Simplify getting the accessible role when checking if an accessible
implementation should present itself; this avoids going through
GtkAccessible twice to get the same data.
The GtkAtSpiContext is responsible for removing itself from the root,
which will remove itself from the cache. Any code path that leads to the
GtkAtSpiContext instance being collected passes through the
unrealization phase, which will also unregister the context from the
accessibility bus and from the cache.
- use the old Industrial style menuitems to make the design feel purposeful
- adjust corners to have compatible border-radii of child elements
- include sidebars and navigation sidebars to look consistent
(plus the assistant sidebar)
Implements https://gitlab.gnome.org/GNOME/gtk/-/issues/3366
The use of volatile was incorrect in GLib and has been that way for
a long time. Recently however that has changed, and this makes GTK
follow suit to avoid using volatile in the type registration.
See also: https://gitlab.gnome.org/GNOME/glib/-/merge_requests/1719
Combined with the above merge request for GLib, this fixes a large
number of compilation warnings when using Clang.
Use the set_minimized method of the xdg_toplevel
interface to implement minimization as well as possible.
It is not possible, since there is no corresponding
state that we could use to update our surface state,
but in practice, it works well enough.
Fixes: #2688
Since GTK4, the functions that used the GtkPackType enumeration,
`gtk_box_pack_{start,end}`, were replaced by `gtk_box_{append,prepend}`,
hence this enumeration isn't used anymore by any function within the
GtkBox type, and the child packing properties were also removed for GTK4.
So this commit adapts the documentation accordingly.
For readonly properties, we show the serialized value
in a label. If we don't take precautions, this can cause
our window to grow extremely wide, and break things.
So, ellipsize things at a reasonable size.
Fixes: #3278
GtkTreeView.get_tooltip_context() takes an inout X and Y coordinates,
but the "out" side is a side effect: the conversion from widget-relative
to bin window-relative coordinates is not documented, and can be done
using public API, if needed.
GtkIconView.get_tooltip_context() follows the same pattern, and takes
two inout arguments for the coordinates, but it does not change them any
more, after GtkIconView's bin window was dropped in commit 8dc5e13e.
There's really no point in having these `inout` arguments, and while
GtkTreeView and GtkIconView are certainly de-emphasised in GTK4, and we
nudge developers to move to the new list views, we should take advantage
of the API break to remove warts.
This commit fix the warning:
../gdk/wayland/gdkdisplay-wayland.c:1079: Warning: GdkWayland: gdk_wayland_display_set_cursor_theme: unknown parameter 'name' in documentation comment, should be 'theme'
Using GtkCssSection in public headers here may be
ok from the C perspective, since it all ends up in
the same library anyway. But it causes circular
dependency problems for our gir files that are still
split by namespace.
To avoid this problem, copy the GtkCssLocation struct
struct as GskParseLocation, and pass take two of them
instead of a GtkCssSection in the error callback.
Update all users.
Fixes: #2454
In commit 024d832d94, we introduced a
cascade-popdown property that makes closing a submenu
propagate up and close its parent menus. This is the
behavior we want when a menuitem in the submen is
activated.
What we overlooked is that we still need to be able to
close a submenu during navigation, before opening another
one. And in this case, propagating the closing is breaking
things. Fix this by adding a private close_submenu api
to GtkPopoverMenu that avoids the propagation.
Fixes: #3301
When falling back to using the program name to create a unique base path
for the objects on the accessibility bus we need to ensure that the name
is a valid DBus object path.
glFrameBufferTexture maps to all faces of a cube and that is not needed
here. Additionally, texture_id is not deleted after we use the additional
flipped texture, but should be.
The AT-SPI cache interface is used to quickly populate the accessible
objects tree.
The tricky bit is ensuring that we emit change notifications on the
cache only when the cache is available, which means waiting until the
root is asynchronously registered.
The root path is shared by all AtSpiContext instances, so we should
compute it once, instead of every time we instantiate a new context.
This allows us to defer the path creation at realization time and ensure
that we have a registered application.
The expected behavior when activating check or radio
menuitems via keynav is that Space toggles the item
but keeps the menu open, while Return toggles the
item and closes the menu.
Ever since 5b5d2665d3, tooltips don't work properly,
since we get more motion events now, and those keep the
tooltips from appearing. This can be seen e.g. on the icon
in the password entry in widget-factory, or on the section
switcher buttons in the Emoji chooser.
Ignore synthetic motion events for the purposes of
determining whether the user moved the mouse.
Currently GTK can be built with G_ENABLE_DEBUG which enables various debug code and parsing
of those env vars, or without, which instead of parsing them prints a warning if they are set.
While building with G_ENABLE_DEBUG isn't strictly needed it's the only way to make GTK_DEBUG=interactive work,
which is a nice thing to have always.
This enables parsing of those env vars in any case and allows specific values being marked as also
available when not built with G_ENABLE_DEBUG (interactive for example). If not built with G_ENABLE_DEBUG
then all unavailable values will be marked as such in the help output and a note is added that
GTK needs to be built with G_ENABLE_DEBUG to use them, which should help discoverability.
An implicit grab means some button is down, reset the active state
only in that case when we get a grab broken event.
Avoids active state accounting warnings when we do get active grabs
broken (e.g. after selecting a menu option).
Since the $button_fill variables are no longer used outside of the
button() mixin (and it was for an ugly hack), there is no need to keep
that anymore.
Fixes https://gitlab.gnome.org/GNOME/gtk/-/issues/3354
We may have the situation of multiple touchpoints in the same
widget, or combinations with other devices. Stack those ::active
states are preserved on widgets on all but the last pointer/touch
going away.
Use the GROUP role, and make sure to note that the child is always
revealed. Unlike GtkExpander, GtkRevealer can only be programmatically
toggled, so we cannot turn it into a "button" object.
Document the role of the GtkTreeExpander, and the behaviour of the
expander button.
Additionally, improve the label of the expander button, by adding a
fixed "Expand" label, and setting the "labelled-by" relation to the
child of the GtkTreeExpander.
The search bar widget has a "SEARCH" landmark role, which is described
as:
A landmark region that contains a collection of items and
objects that, as a whole, combine to create a search facility.
Commit fa3d1940bf added separate grab handling for GtkSearchEntry, but
didn't consider whether the bar was revealed or concealed. The expected
behavior for the latter is that the entry is cleared rather than focused,
fix the condition accordingly.
With the last commit, pressing the same button with multiple fingers
doesn't cause extra emissions, so we can remove exclusive and allow
pressing multiple buttons at once on touch.
Add a possible replacement for GtkEntryCompletion
as a demo.
Move the Dropdowns demo to Lists/Selections, and make
it show both GtkDropDown and the suggestion entry, with
some variations.
Now that GtkATContext is explicitly realized and unrealized, we should
always create an instance at widget initialization time, and drop it
during the widget finalization. This should make it easier to set up the
initial accessible state of a widget during the instance initialization,
as well as reduce the chances of accidental creation of GtkATContext
instances during the destruction sequence.
Since GtkATContexts are now lazily realized, we need to go through the
GtkAccessible's implementation to access the :accessible-role property,
in case there are fallbacks.
There's no need to do a lot of work on construction, if we're delaying
all remote work after the GtkATContext is realized.
The GtkAtSpiContext should also keep a reference on the root, and drop
it at unrealize time.
By unrealizing the context we avoid additional work during the dispose
phase, in case widget code updates the accessible state. We use
GtkAccessible's API, to ensure we unrealize the right ATContext, instead
of the one we store inside GtkWidgetPrivate.
We drop the ATContext instance inside GtkWidget during finalization, to
mop up eventual vivifications there.
Making the list row child css depend on the position
is very expensive, and does not acutally work correctly
(since we don't have widgets for all children, so the
position of the child widget does not reflect the actual
model item position).
To make this more palatable, use the bottom border
instead of the top border, since most lists have a natural
border at the top (with headers), and may end up with
empty space at the bottom.
The overarching goal here is to not queue a resize
unless something has actually changed. In columnview
scenarios, we often deal with hundreds of labels.
Labels are cattle, not pets.
This is leftover code from when widgets were hidden
by default, and was setting them back to their initial
state.
This is getting in the way now, as hiding the widget
updates the HIDDEN accessible state, which ends up
re-creating the at context that we've already disposed
of, leading to memory leaks.
In gdkdevice-win32.c we are interested in knowing which window
receives mouse input at a specific location.
Only WindowFromPoint is the right API for the task, other API's
(such as (Real)ChildWindowFromPoint(Ex)) have shortcomings because
they are really designed for other purposes. For example, only
WindowFromPoint is able to look through transparent layered windows.
So even if we want to find a direct child we have to use
WindowFromPoint and then walk up the hierarchy.
Fixes: #370, #417
See: !2800
For builds that do not have PangoFT2, the demo fails to link because we weren't
building listview_ucd.c. Fix the build by building listview_ucd.c with
script-names.c and unicode-names.c for all builds, as we now require a Pango
version that already always depends on HarfBuzz and those sources do not use
anything from PangoFT2.
gtk-builder-tool simplify --3to4 now produces a
<requires lib="gtk" version="4.0"/>.
Adapt for that, and also add some tests that check that we
actually verify the requires, and accept 3.99 and 4.0, but
not 5.x.
The hypothetical widget that needs to clone ATContext instances
because it needs to control the accessible role post-construction is
really GtkModelButton.
Fixes: #3342
Some widgets might want to override GtkAccessible and create their own
context in order to control the accessible role post-construction time.
To avoid explicitly copying the existing state over from the original
ATContext to the new one, we need a way to clone the context's state
from inside the ATContext itself.
Using GList is a bit lame, and makes the API more complicated to use
than necessary in the common case.
The only real use case for a GList is gtk_widget_add_mnemonic_label(),
and for that we can use the GValue-based API instead.
Fixes: #3343
The accessible-role property in GtkWidget has three possible targets:
- the :accessible-role of a GtkATContext, if realized
- the accessible_role field of GtkWidgetPrivate
- the accessible_role field of GtkWidgetClassPrivate
When we set the accessible role of a widget using the GObject property
mechanism, we want to either set the GtkWidgetPrivate.accessible_role
field, if there's no ATContext *or* if the ATContext is not realized.
Conversely, when we get the accessible-role property we want to have a
series of fallbacks in place:
- if GtkAccessible.get_at_context() returns an ATContext, and that
ATContext is realized, return the :accessible-role of the context
- if GtkWidgetPrivate.accessible_role is not WIDGET, return the
stored accessible role
- return GtkWidgetClassPrivate.accessible_role
This should help catch the case of getting the accessible role of
widgets that override GtkAccessible.get_at_context(), like
GtkModelButton.
See: #3342
Use a single environment variable for everything:
- select the ATContext implementation
- select the test ATContext
- disable ATContext entirely
We use the same pattern as GSK_RENDERER, GTK_DEBUG, etc.
The documentation needs to be updated to include the environment
variable.
The GtkPlacesOpenFlags enum is only used in private
API, so move it to private headers. Since we still need
a GType for it, add gtkplacessidebarprivate.h to the
headers we use for generating private enum types. In turn,
this registers the other private enums in that header, so
take the opportunity to fix their naming, and use the
generated types for the corresponding sidebarrow properties.
Fixes: #3337
On desktop GL, GL 1.5 or GL_ARB_occlusion_query is required to get the
glGenQueries() etc. symbols. This isn’t the case on GLES, where they
are provided by GL_EXT_occlusion_query_boolean, and more importantly
have never been made core.
This patch allows gtk4-demo to start when GDK_DEBUG=gl-gles is set, on
my Mali 400 MP running the Lima driver from Mesa.
Check the text handle role, instead of looking for the other handle
visibility. The other handle may be invisible during selection mode
(e.g. pointing to offscreen contents). This fixes both this code
switching to cursor mode out of the blue, and possible crashes later
on as this handle might be hidden in the process, while its own event
controller is handling events on the parent surface.
The gtk_text_view_set_handle_position() function called some lines above
takes care of handle visibility already, also accounting for other
conditions (e.g. whether the handle points to contents onscreen).
Forcibly showing handles here misbehaves if the handle should stay hidden,
and somewhat expensively as it involves creating and throwing a native
surface every time.
With the scrolledwindow drag gesture not claiming the sequence immediately,
we end up placing the cursor (and undoing the previous selection) each time
we scroll.
There is already handling too short drags in ::drag-end, so let this code
handle touchscreens as well.
If the gesture becomes captured (e.g. from a parent scrolledwindow), we
leave some things in the air. Clean these up properly. This is recurrent
with touch scroll.
In destruction paths of a focused entry, the IM context might first
end up detached from its widget, then destroyed. This currently prevents
the IM context from detaching itself from the GtkIMContextWaylandGlobal.
Make it do so when unsetting the client widget, so the IM context gets
properly unfocused before destruction.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3176
In wayland, popup positioning and event handling are doubly async.
This makes it unreliable to figure out parent surface coordinates
out of the popup position and the events received. This results in
jumpy text handles there.
The best way to deal with parent surface coordinates is to handle
the events there. Make the handles transparent to input, and make
the drag gesture be set up on the parent widget's native.
The gesture is set up in the capture phase, setting it on the native
(as opposed to the parent widget) achieves a feeling similar to it
being a distinct surface, as it should take precedence over other
gestures in the emission chain (e.g. scrolledwindows).
As everything is in parent widget's native's coordinates, the drag
handles become smooth again.
The intended use case of the "presentation" role is
| […] when an element is used to change the look of the page but does not
| have all the functional, interactive, or structural relevance implied
| by the element type, or may be used to provide for an accessible
| fallback in older browsers that do not support WAI-ARIA.
One of the examples is, literally:
| An element whose content is completely presentational (like a spacer
| image, decorative graphic, or clearing element);
Which fits perfectly for the GtkTreeExpander's "indent" builtin icon
widget.
Mark gtk_drag_icon_create_widget_for_value with
GDK_AVAILABLE_IN_ALL, since it is meant to be public.
Also, clarify the documentation a bit.
Fixes: #3325
Once we figure out what is going on with textures, changes are we'll be
able to let this stay a zero rect. But that is still a bit up in the air right now.
When we close grabbing popups due to an outside
click, check at each level if the click is still
outside. This makes closing the nested popover
menu in the popover on page 3 of widget-factory
work as expected, when you click the menubutton
again.
When an event happens on a non-grabbing popup that hangs off
a grabbing popup, don't trigger the autohide. This makes touch
text handles work inside the popover on page 3 of widget-factory.
We are doing too much work during the construction phase of the
AT-SPI backend for GtkATContext. Instead of having the AtSpiContext
register itself at construction time, let's add explicit realize
and unrealize operations, and connect the ATContext realization to the
rooting operation of a GtkWidget.
We can just call gdk_surface_invalidate_rect here
like we do elsewhere in gdk, and I'm a bit uncertain
about the current code that adds an empty update
region.
We don't want the popover to stay open when we open
an about dialog or shortcuts window. Since cascade-popdown
would also affect e.g. the context menu of the text widgets,
do this explicitly.
The magnifier in entries is magnifying the GtkText
widget, which does not draw any background itself.
Therefore, we give the magnifier a background, to
make things look as expected.
Fixes: #3318
Install a log writer function that is a bit simpler than
the default GLib log writer, and does not put PIDs and
timestamps in the output, so we can compare easily in
our testsuite.
It doesn't make much sense to set this in per-directory
meson files, since that makes use use different logging
facilities in different parts of the project. Set it
globally.
We are setting the value to TRUE initially, but
the property had a declared default of FALSE.
This is messing up the simplification of .ui files
with gtk4-builder-tool, since it thinks it can
omit can-target properties when it really can't.
We were not checking the passed-in type in the right
way. An interface type can still pass the
g_type_is_a (..., G_TYPE_OBJECT) check, if G_TYPE_OBJECT
is one of its prerequisites. What we need to check is
whether the fundamental type is G_TYPE_OBJECT.
The tests in subdirectories of testsuite/css were
using an environment defined in the parent directory,
causing them to not find their test data. Give each
test directory its own environment.
When using GtkBuilder to create constraints and guides,
the layout manager is already rooted when the guides get
created, and we were forgetting to create the guides
constraints in this case. Fix it by adding a call to
gtk_constraint_guide_update().
This was showing up in the new builder-based constraints
demo as the guide not having the expected effect.
This isn't strictly required for correctness, but
it makes the GtkBuilder codepath do the same that
the vfl codepath does: call gtk_layout_manager_layout_changed
only once, after all the constraints have been added.
The Cairo implementation for the Macos backend uses a toplevel
window with full transparency and a series of NSView to create opaque
regions. This improves compositor performance because it allows the
display server to avoid costly blends.
However, we want to ensure we clip better when exposing the
transparent region so that we only expose the shadows/corners as
necessary.
This typedef was not used in any public APIs, and is
only used in the MacOS backend. It is not worth preserving
as public API, move it to the only user.
This implements the basics for a GdkGLContext on macOS. Currently, rendering
only is fully working for the GskCairoRenderer case where we read back pixels
into a cairo surface for rendering. More work on synchronization is required for
the GL on GskGLRenderer case.
When we attempt to render a surface itself with GL, the context will ensure that
the new GdkMacosGLView is placed within the NSWindow. In other cases, we
use a dummy NSView and NSWindow for backing the NSOpenGLContext to
ensure that we can get accelerated drawing.
This gets GtkGLArea working when running with GSK_RENDERER=cairo.
This helps a situation where the window contents has not changed
in time for a drawing. Setting the texture gravity helps that side or
corner to be less jittery while moving.
Ideally, we can get to a point where we are synchronized and keeping
up with drawing fast enough to not need this. That may require some
work to drive frame clocks from drawRect: though.
This follows the mockups that these demos are
copied from. Unfortunately, it has to be implemented
for every item factory, so we repeat it here.
Fixes: #3291
Make the default factory add a checkmark to the
currently selected item (not the hovered item)
in the popup. This will unfortunately have to be
done in non-default factories too.
Related: #3291
Add a way to add children at certain places in
the generated menu for both GtkPopoverMenu and
GtkPopoverMenuBar.
New apis:
gtk_popover_menu_add_child
gtk_popover_menu_remove_child
gtk_popover_menu_bar_add_child
gtk_popover_menu_bar_remove_child
Fixes: #3260
This makes it so that you have to explicitly specify
"true" and "false" for the checked state, but it
matches how this enumeration is meant to be used.
It turns out gtk-doc can't even parse its own custom
links if you don't arrange for the <link> element to
on a single line, since it is just a glorified sed
script :(
This ensures that keybindings for small-step changes
work despite draw-value being FALSE now. This was
fallout from 8ca612c966 that showed up
as arrow keys not working anymore for the color scales
in the color chooser.
Instead of monitoring the list of toplevels, rely
on GtkWindow updating the HIDDEN state before windows
get removed. This is better, since we still have the
object available when it happens, so we can pass it
to the ATs.
We want to use the HIDDEN state to control when
things get added and removed from the accessible
tree, so ensure that we a) set HIDDEN to true
initially for windows, and b) we update HIDDEN
when a window is shown or hidden.
The second part is handled by gtk_widget_hide
for other widgets, but hiding a window does not
always go through that code path.
When a toplevel window gets hidden (and not destroyed),
the frontend code set the HIDDEN state, and we need to
emit child notification when that happens.
We need to use gtk_accessible_should_present() whenever we
calculate accessible tree positions, to avoid inconsistencies.
While we are at it, make these helpers usable for finding
the position of accesibles that are now ignored, by not
looking at should_present for the object itself. This will
be relevant when we calculate the position of objects whose
HIDDEN state changes.
We need to translate the list model position into
an accessible tree position, since hidden toplevels
will be skipped.
While we are at it, add an api for this notification
that will be used in the next commit.
Hidden elements are not presented in the accessible
tree, so when then HIDDEN state changes, we should
emit child-added or -removed signals.
This commit does not yet handle all cases (HIDDEN
toplevels or hidden stack pages are not handled),
but it should cover the common case.
The ARIA spec is clear on this: when an element has the
HIDDEN state, it should not be presented in the accessible
tree.
This change is incomplete, we also need to emit child-added/
removed signals when the state changes, but that needs to
wait for the child added infrastructure to land.
gtk_widget_reposition_after is called both to add new children,
and to reposition existing children. We only want to emit
accessible changes in the former case (since AT-SPI doesn't
have events for reordering).
The current code was marking queued events as flushed,
but left them in the queue. That doesn't make sense to
me - we should deliver all events we have before we
reach the paint phase of the frame cycle.
Unparenting the stack finalizes the entry and label,
but at least the label is available via the
gtk_editable_get_delegate API, and the a11y
implementation uses that at dispose time. So, clear
the pointers to prevent them from being dangling.
This was showing up as a segfault of the doc-shooter.
When the compositor unmaximized the window, we get a
state-changed signal, and we update the maximized field.
But then we go and recompute our layout based on the
maximize_initially field, and that is still TRUE, when
we were maximized initially. So we need to update both
fields.
This fixes a problem where using the window menu to
unmaximize an initially maximized window would not
work.
Fixes: #3226
These tests are a bit fragile because the initial focus
that is taken depends on what ends up in the center of
the window, horizontally or vertically, which depends
on things like fonts, or theme spacing.
This commit makes some tweaks to push things in the
widget-factory example around far enough to make the
tests work again.
We should figure out a way to make this more robust.
We are not using a box layout here since we want
to ignore the icons for measuring. But we still
want the layout to respect border spacing that
comes from the theme.
If the operation mode is OPERATION_MODE_RECENT and we end up in the
'goto file_entry' case, we don't set info.result. Then later after
calling check_save_entry, is_empty is TRUE which causes a goto out and
here we then try to use info.result, which is uninitialized.
Initialize info.result before doing all this.
Found by scan-build
If error is NULL, replacing the file contents worked and we don't
show the error dialog. The previous version pretty explicitly
dereferenced a NULL pointer by first ensuring that error is NULL and
then accessing error->message.
Found by scan-build
Unparent the child widget before tearing down its
stack page. This is necessary so a11y can still access
the stack page accessible to emit change notification
when it learns that the child is removed.
Since we mention abstract roles in the documentation
for GtkAccessibleRole, we should say what roles are
abstract. Doing this shows that we actually use two
abstract roles heavily, ourselves: WIDGET and WINDOW.
GtkModelButton is no longer derived from GtkButton,
but can still treat it like a button for the purposes
of having a click action. This lets ATs activate
menu items again.
When the button role changes, we want to update the
accessible role to match. Since accessible roles are
unchangeable post-creation of the AT context, we have
to cheat a bit and recreate the whole context.
Set the accessible role to GTK_ACCESSIBLE_ROLE_MENU_ITEM.
This is incomplete, we need to recreate the context when
the buttons role changes, and there are other things that
need to be set.
Some people read the "Getting Started" section as a series of
incremental lessons, and having the examples go from GtkApplication to
the old style "init / spin the main loop" confuses them.
We should be using GtkApplication everywhere in our examples.
According to section 7.1 of WAI-ARIA, the progressbar role
has the "Children presentational" characteristic, which
indicates that children should not be represented in
the a11y tree.
According to section 7.1 of WAI-ARIA, the meter role
has the "Children presentational" characteristic, which
indicates that children should not be represented in
the a11y tree.
According to section 7.1 of WAI-ARIA, the switch role
has the "Children presentational" characteristic, which
indicates that children should not be represented in
the a11y tree.
According to section 7.1 of WAI-ARIA, the slider role
has the "Children presentational" characteristic, which
indicates that children should not be represented in
the a11y tree, which makes sense, since these are all
just internal gizmos.
Recompute the layout when the css style change
affects text attributes. This matches what we do
in GtkLabel, and without this, changing the
font-features-setting css property in the Inspector
does not have immediate effect.
The last event, matching lifting the finger/releasing the mouse button,
is important when there's a large delay between it and the previous events,
as in when performing a movement, stopping, then releasing fingers as
opposed to doing a swipe.
If this event is skipped, doing this will result in kinetic deceleration
matching the previous finger movement, while the expected behavior would
be no deceleration.
See also 5dc6194b98 for a similar fix in
GtkEventControllerScroll.
GtkAtSpiRoot is not a context, which means it needs to emit
ChildrenChanged events by itself whenever a toplevel is added to, or
removed from, the list of toplevels.
A bit hacky: we skip parsing values that have a reference or
reference-list type, but we do not error out. Instead, we return a NULL
value, which we catch in the GtkBuildable interface implementation to
get the actual object, and construct a reference list value.
There's still some ickyness around the value type that can only be
solved by having an attribute and role taxonomy.
Accessible attributes are not GObject properties. This means that we
need a custom parser for setting attributes in our UI description files.
The new section is defined as a sub-tree with the `<accessibility>`
element at its root, and elements for each type of accessible
attributes, i.e. properties, relations, and states:
```xml
<object class="..." id="...">
<accessibility>
<property name="label">The accessible label</property>
<state name="pressed">false</state>
<relation name="labelled-by">label1</relation>
</accessibility>
</object>
```
The name of the attribute is the enumeration value; the value is defined
by the WAI-ARIA specification.
The nameless, faceless gizmos inside a range do not
contribute to the accessible experience at all, lets
not add them to the tree. All the accessible functionality
is on the main widget (either a scale or a scrollbar).
Show the object path of the object on the a11y bus,
this is can be useful information. While we are here,
make sure that the Inspector does not throw criticals
when used with GTK_NO_A11Y=1.
There were several places where we were confusing
GList and GSList and list->data and list->next, causing
a crash in the accessible name computation for buttons
with mnemonic labels.
Anybody who keeps their own CSS nodes around or wants to order CSS nodes
different from widgets will from now on have to do it manually all the
time.
This is outdated behavior, nobody should be doing either of those two
things.
Also, the correct case is much more common, and not doing it
automatically was causing bugs.
Fixes#3280
The stack page objects were not properly integrated
in the accessible tree - they were appearing as parent
of the pages when navigating up, but not as children
of the stack when navigating down.
Instead of falling back to the role nick for both,
fall back to the class name for the name, and to
the empty string for the description. This makes
labels show up in Accerciser the same way they
did in GTK 3, and seems more useful to me than
the alternative.
Set the expected accessible role on the tagged entry
widget in the demo of the same name. Accessible tools
may decide to ignore widgets that have the wrong role,
so setting an appropriate role is important.
The ARIA spec determines the name and description of accessible elements
in a more complex way that simply mapping to a single property; instead,
it will chain up multiple definitions (if it finds them). For instance,
let's assume we have a button that saves a file selected from a file
selection widget; the widgets have the following attributes:
- the file selection widget has a "label" attribute set to the
selected file, e.g. "Final paper.pdf"
- the "download" button has a "label" attribute set to the
"Download" string
- the "download" button has a "labelled-by" attribute set to
reference the file selection widget
The ARIA spec says that the accessible name of the "Download" button
should be computed as "Download Final paper.pdf".
The algorithm defined in section 4.3 of the WAI-ARIA specification
applies to both accessible names (using the "label" and "labelled-by"
attributes), and to accessible descriptions (using the "description" and
"described-by" attributes).
Our Text implementation requires that we have
a GtkEditable with a delegate that is a GtkText
widget.
This change make the Text implementation work for
the custom widget in the tagged entry demo.
Implement the non-questionable parts of the Component interface
for accessibles which are widgets.
This does not include:
- global coordinates
- setters
- scrolling
- alpha, layers, zorder, and the like
Make right-aligned content work in resized columns.
There is currently no way to make a title right-aligned,
but we can still make it work correctly. This is a follow
up to 7eb0ae39c5.
Fixes: #3276
When resizing columns, we clip a shrunk column
on the right, so the separator disappears in that
case unless we put it on the left side of the other
column.
GLDEBUGPROC callback is defined with APIENTRY which is a windows
specific calling convention. That macro expands to nothing when building
on other platforms.
Fixes: #3268.
Ensure that the column resize cursor stays in place
for the duration of the resize drag. This is a bit
annoying, since the implicit grab can end up on the
header of a different column from the one we are
resizing, so just set the cursor on all column headers.
Make it so that for overlapping resize rectangles (with
very narrow columns), we prefer the narrow column, so you
can regrow a column after shrinking it all the way.
Related: #3274
Ensure that we place the resize rectangle at the visible
right edge of the column, not where the allocation ends
(we clip the header drawing, after all).
Related: #3274
This gracefully disable ffmpeg, gstreamer, cups and cloudprint optional
dependencies when they are not available, while still giving full
control to distributors using -Dauto_features=enabled.
It is nicer for contributors to have a build that works by default.
Distributors should be using -Dauto_features=enabled to get ride of
automagic options.
There is no reason to fallback to find_library for cairo and still rely
on pkg-config for all other dependencies, and just when using MSVC. When
building and pkg-config is not working it is preferable to fallback to a
subproject just like for all other dependencies.
Also add cairo.wrap now that meson support has been merged upstream.
Call SetCapture() explcitly for the (new) modal window so that we make the
modal window respond to mouse input, and also call SetCapture() to the parent
of the transient window that we are destroying so that mouse input capture is
returned to the parent window.
This attempts to fix the following:
* Upon creating a new modal window, the new modal window does not receive
pointer input unless one switches to another program and back
* Upon closing a transient window, the parent window that activated the
transient window does not receive pointer input unless one switches to
another and back
This reverts commit fc2008f2.
Turns out, we *don't* have code to maintain Z-order. Restacking
code is not doint that, it just enforces a few weird Z-order-related
behaviours.
Make sure that we get the state of the modal window properly, and send out the
corresponding notification signals.
This will ensure that we do not try to activate windows that should have become
inactivated due to it opening modal windows and render the program unresponsive
because we are not activating the correct window that is due to receive user
input.
We only want to show relevant, local actions for
widgets, but _gtk_widget_get_action_muxer() will
return the muxer of a parent widget (all the way
up to the toplevel), if the widget does not have
any actions of its own. To detect this situation,
compare what _gtk_widget_get_action_muxer() returns
for the parent widget, and act accordingly.
The buttons here are not really buttons (the action
is not tied to the "clicked" signal), so triggering
the buttons via a11y does not have the expected effect.
And we expose the Value interface that ATs can use
to set the value.
The nested window was not modal, causing it to be
inoperable. And the nested views within were all
shrunk down to nothingness. Give them some width.
Fixes: #3257
The Selection interface is defined in terms of child
positions, so we need to always translate from that
to model positions if we want to use the selection
model apis.
Add a paragraph to the migration guide that explains
how to properly render symbolic GtkIconPaintables.
Also mention this in the GtkIconPaintable docs.
There's a reason you can't spell 'paintable' without
'pain'...
Using GtkIconPaintable+GtkPicture is not good for symbolic
icons - they don't get properly colored that way. So change
things around to use the icon name if we have it.
Fixes: #3252
The 'has_uncommitted_ack_configure' state was added to make sure we're
responding to 'xdg_surface.configure' events with
'xdg_surface.ack_configure' requests, as is necessary according to spec.
What we didn't do was to clear this state when hiding, meaning that if
we hid the surface after a configure event, but before the frame
finished and we processed the 'has_uncommitted_ack_configure', we'd try
to acknowledge the surface configuration after having destroyed the
surface.
Closes: #3262
This implementation works for both GtkListView and
GtkGridView, and by extension, also for GtkColumnView
and GtkDropDown, since those just use a list view
internally.
epoxy_dep cannot be used in a configure time check when it comes from a
subproject. Use variables set in pc file instead.
This requires https://github.com/anholt/libepoxy/pull/231.
Use the TREE_GRID, ROW, COLUMN_HEADER and GRID_CELL roles
for the various widgets involved in a GtkColumnView. To
enable this, we subclass GtkListView for the internal
list in the column view.
This is a bit different from the way things were done
in GTK 3 - we follow what was done for GtkStackSwitcher,
and make the tab bar carry the GTK_ACCESSIBLE_TAB_LIST
role, and implement Selection there.
Set up the necessary roles, relations and properties
for the tab patterns. This parallels what we have done
for GtkStackSwitcher, and implements the Tabs pattern
as described in the ARIA authoring guidelines.
Set the tablist role on the stackswitcher itself, and
the tab role on the buttons. This is another step towards
implementing the tabs pattern for GtkStack.
platform change is called from gtk_widget_set_focusable
which is likely to be called early on in init(). We don't
want to create an AT context that early if we can help
it, e.g. since it makes it impossible to override the
accessible-role with a construct property.
This requires some cleanup to remove assumptions
about accessibles being widgets in the backend,
and some code to navigate the tree with these
extra objects in between widgets.
The accessibles for stack pages have the role
GTK_ACCESSIBLE_ROLE_TAB_PANEL. This is the first
step towards implementing the tabs patterns
as described in the aria authoring guidelines
for GtkStack.
This was incorrectly reporting the toplevel surface instead of the
popup surface that was placed above it. This fixes event delivery
to popups for selecting menu items and more.
Either we or clang needs to get its shit together about this warning.
But using it during development with clang just makes actually usable
warnings get lost in a flood of -Wcast-align warnings.
This fix is correct and fixes:
1) GL textures being upside down in the inspector. They are getting
downloaded because they've been created in a different GL context
2) GL textures being upside down in the cairo renderer (same reason)
However, it breaks the testsuite. We do the flipping via the projection
matrix, but most of the shaders don't care about that.
Like we do for GdkX11. We can't use all of the public C API, but we can
expose enough type information to allow non-C developers to actually
check if they are running the Wayland GDK backend or not—plus some
additional Wayland-specific API.
The GdkWayland API takes generic GDK types and performs a run time
check, which means we need to properly annotate the actual expected
type in order to have methods recognised as such.
For the various uses of GDK_WINDOWING_QUARTZ, we need to use
alternatives from GDK_WINDOWING_MACOS.
Some minor loss of functionality is here, such as icons sent with
application menus. That can certainly be added back at a future
point.
No point in showing an apologetic tab for a blob of binary data.
gtk4-demo shows the resources under /DEMONAME/ for each demo,
so move the data to /DEMONAME_data/.
We are not propagating focus change events, and that is the only
place where we are listening for focus change events. If GtkWindow
does not see focus-in events for its popovers, we end up with
inadvertendly inactive windows.
Fixes: #3240
When querying a device, we need to ensure we are providing coordinates
in the coordinate system of the surface. Further, we need to actually
provide the button and keyboard state.
This fixes some issues related to dragging scrollbars and selecting list
box rows more reliably.
Set the SELECTED state to reflect whether the selected
is selected, unselected, or unselectable. This is
enough to make selection changes appear in Accerciser.
While we are at it, also set the multi-selectable
property for the flowbox itself.
Non need to announce the same things for every context
we create, and the path is not really that interesting.
without knowing what it belongs to. I would suggest to
make it visible in the inspector instead, so you can
look it up for the widgets you are interested in.
Set the SELECTED state to reflect whether the row
is selected, unselected, or unselectable. This is
enough to make selection changes appear in Accerciser.
While we are at it, also set the multi-selectable
property for the listbox itself.
And honor it in gtk_popover_popdown(). By default, a GtkPopover
pops down automatically if a child popover was closed, if this
property is FALSE, the popover will remain opened.
We use to set the the 'password text' role for entries with
visibility = FALSE. Nowadays, we have a separate class for
password entries, so fix up the role mapping based on that.
Make text change notification work for editables, by connecting
to the ::insert-text and ::delete-text signals on the wrapped
GtkText widget, and for GtkTextView by connecting to the
corresponding GtkTextBuffer signals.
This code is more or less directly copied from GtkTextViewAccessible
and GtkEntryAccessible in GTK 3.
Since the big editable reorg, GtkText was not emitting
::insert-text and ::delete-text, as is expected of
editables. We want to use those signals for a11y
change notification, so make them work again.
This copies what was done for GtkEntry: get
the focused state from the GtkText within.
We also add a private getter for the text widget,
which was missing here.
As a companion to go with the platform_change api,
add a gtk_accessible_get_platform_state() function
that can be used by backends to get the platform
state.
This is in preparation for making entries inherit
their focus states from the text widget within.
Similar to gtk_widget_should_layout(), add a
gtk_accessible_should_present() function that backends can
use to determine whether an accessible should be presented
or not.
Ways to make a widget not presented in a11y:
- hide the widget
- set its role to NONE
- make it have a NULL AT context
We will use this in future to hide the GtkText inside
an entry, since the Text implementation will be done
by the wrapper.
We are determining editable state based on the
accessible role (although we could make it platform
state now), so cover all the roles that we use for
entry wrappers.
Add an enum for 'platform changes' to the at context
change notification mechanism. This will let us pass
along things that ARIA considers 'platform state' such
as focus or editability. The difference between the
platform state and other ARIA states is that we don't
keep the platform state separately in the at context
- backends are expected to just query the widgets.
This is just about avoiding notify listeners for
change notification.
We can use the read-only property, together with the
accessible role, to determine whether to set editable
and read-only states for at-spi. This lets us avoid
directly poking at the widgets.
ATs look at not just the implemented interfaces, but
also the states to decide what to do. It turns out that
the EditableText interface is only used by accerciser
if the editable state is set. So set it.
It is error prone to keep the same conditions in sync
in two places. Instead, just assemble the list of interfaces
as we register objects, and use when GetInterfaces is called.
Apply the Value implementation to the widgets where
we had one in GTK 3: GtkLevelBar, GtkRange, GtkScaleButton,
GtkSpinButton, GtkPaned, GtkProgressBar. To make these
work, the widgets need to set the accessible value properties.
There is some open question here whether the interface
should be implemented on the outer or the inner widget
of the entry-text pairs. For now, our hand is forced,
since only GtkText provides access to the layout that
we need for implementing many of the interface methods.
This is a not-quite-complete implementation of the
Text interface for GtkLabel. The missing parts are
anything around extents and positions, as well as
the ScrollSubstring apis.
This translates relations as far as the match.
I'm not sure yet what we can do about the fact that
atspi expects relations to be bidirectional (ie have
label-for *and* labelled-by) while aria has only one
direction.
It turns out that accerciser depends on this undocumented
method that is not in the xml at all, otherwise interface
sections in the accerciser ui never get enabled.
The ARIA spec defines the mechanism for determining the name of an
accessible element—see §4.3 of the WAI-ARIA spec.
We follow the specification as much as it makes sense for GTK to do
so:
1. if the element is hidden, return an empty string
1. if the element has a labelled-by relation set, retrieve the
label of the related element
2. if the element has a label property set, use the value of
the property
3. if neither labelled-by nor label attributes are set, we use
the role to compute the name:
- for a `range` role, we return the contents of the value of
the `value-text` or `value-now` properties
- for any other role, we return a textual representation of
the GtkAccessibleRole enumeration value
When we create the first AT-SPI context we also need to register the
accessible root on the accessibility bus. The accessible root object is
the main entry point of an accessible application, and it holds the
global state to present to the ATs that connect to the bus.
Since we need to check at run time what kind of AT context to use, we
need a hook into the whole GDK backend machinery. The display connection
seems to be the best choice, in this case, as it allows us to determine
whether we're running on an X11 or Wayland system, and thus whether we
should create a GtkAtSpiContext.
This requires some surgery to fix the GtkATContext creation function, in
order to include a GdkDisplay instance.
Does not do anything, at the moment, but it's going to get filled out
soon.
The backend is selected depending on the platform being compiled in;
since we're using AT-SPI on X11 and Wayland, and we don't have other
accessibility implementations, we currently don't care about run time
selection, but we're going to have to deal with that.
And generate the code for the DBus interfaces.
We don't want the full object manager experience, here, because we're
going to have a single object responding to various interfaces and
remote method calls. For this reason, we're not using the gnome module
in Meson to call gdbus-codegen for us: we need to use the interface info
command line arguments, and those are not available from Meson.
Like we do for GdkX11. We can't use all of the public C API, but we can
expose enough type information to allow non-C developers to actually
check if they are running the Wayland GDK backend or not—plus some
additional Wayland-specific API.
The GdkWayland API takes generic GDK types and performs a run time
check, which means we need to properly annotate the actual expected
type in order to have methods recognised as such.
For the various uses of GDK_WINDOWING_QUARTZ, we need to use
alternatives from GDK_WINDOWING_MACOS.
Some minor loss of functionality is here, such as icons sent with
application menus. That can certainly be added back at a future
point.
No point in showing an apologetic tab for a blob of binary data.
gtk4-demo shows the resources under /DEMONAME/ for each demo,
so move the data to /DEMONAME_data/.
We are not propagating focus change events, and that is the only
place where we are listening for focus change events. If GtkWindow
does not see focus-in events for its popovers, we end up with
inadvertendly inactive windows.
Fixes: #3240
No users of gdk_display_peek_event, gdk_display_has_pending
_gdk_display_event_data_copy or _gdk_display_event_data_free,
so drop all of these, and related vfuncs.
It is a little annoying that this demo will not show up
if we don't find librsvg, but I think showing how easy
this paintable is outweights the annoyance.
We were inadvertedly setting the windows min size
to the default size, making it so that you can never
shrink a window below its default size.
Fixes: #3235
Go back to installing our debug message callback
unconditionally if G_ENABLE_CONSISTENCY_CHECKS is
defined, and allow opting into it using GDK_DEBUG=gl-debug
otherwise.
The GtkTreeListRowSortKeys implementation doesn't
know how it wants to cache its keys, and just crashes.
Since that is not cool, add a bandaid fix that forces
it to recreate its keys instead. Extra work, but hey,
no crash.
Related: #3228
Prevents GDK Popups from stealing focus from the parent window when
using Server Side Decorations on win32.
It uses `ShowWindow` and the `SW_SHOWNOACTIVATE` flag.
Don't call gtk_root_get_focus when we already have
the GtkWindowPrivate struct at hand. And use
gtk_window_set_focus to update the focus, like the
old code did.
When a widget is hidden, check harder for the keyboard focus being
contained in that widget, in order to reset it. Portions of the
focus child hierarchy may be outdated at the time, so it is more
reliable to check GtkRoot::focus (i.e. the property we intend to
update here).
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3214
We were not updating the remembered size at all when
the window is interactively resized, causing it to
snap back to its default size the next time we call
gdk_toplevel_present().
This is a bandaid fix to prevent very broken resizing
behavior, until we have properly redone toplevel sizing.
Fixes: #3076
When using the saved size because the compositor
told us to, we were forgetting to readd the margins.
The visible symptom of this was the window getting
smaller every time we went to tiled state and back.
Don't remember the surface size when we are in tiled
state either. This matches the 'fixed_size' condition
in gdk_wayland_surface_configure_toplevel.
This change fixes an issue where moving a window first
to tiled, then to maximized state and back would lead
to the unmaximized window having the tiled dimensions.
We should not emit configure events before we are realized - size
changes at this point are not relevant.
This gets rid of a mysterious emission of GdkSurface::size-changed
with a size of 52x52, that is happening when GtkWindow sets the
shadow_width before the window is mapped.
Always install the debug message callback when we can
and GDK_DEBUG=gl-debug is specified. Previously, we
were only installing the callback when the build was
a non-optimized debug build.
Move the Unicode names to a separate source file,
and only build the demo if we have harfbuzz (since
we use script names, and those are only available
with harfbuzz).
Also, fix a forgotten type name.
Trying a new scheme - when updating NEWS outside of doing
a release, keep the version out, to make the it clear that
this is just about the tip of the branch, not a complete
release.
This is rarely what you want, so lets turn it off
by default.
Update the one place in our demos where we want to
draw a value, add support for this to gtk-builder-tool,
add a test and mention this change in the migration
guide.
Use the data files from https://github.com/milesj/emojibase.git
as source for our Emoji data. Slightly change our data format by
adding a group to each item, in both the Emoji data and in the
setting for recent-emoji.
Install translated versions of the data as separate resource
bundles in $prefix/gtk-4.0/emoji, and load them when appropriate.
Currently, we have data for de, en, es, fr, zh, with data taken
from Unicode 13 and CLDR 13.
Fixes: #950#1511
Make these functions return FALSE if they did not
return the exact position that was requested.
Adapt tests.
Based on a patch by Sebastien Wilmet
Fixes: #506
Make sure that every object property in GTK has accessors for getting
its value (if the property is readable) or setting it (if it is
writable).
Since we are still missing accessors, the test is allowed to
fail for now. Sadly, mesons xfail support is busted, so we just
disable the test entirely :(
Based on initial work by Benjamin Otte.
Related: #2440
When encoding big blobs of data in base64, insert newlines.
Base64 allows it, CSS allows it, so not need to make GtkTextView
struggle with multi-megabyte lines.
Update nodeparser tests to reflect this change.
Rename _gtk_css_print_string to strip the _ and add
an insert_newlines argument to it. Update all callers,
and make the render node serializer insert newlines.
Don't close the Emoji chooser when the Control
key is held while clicking. So you can insert
multiple Emoji without having to reopen the
chooser every time.
Fixes: #1002
This reverts commit d761e3cf2c.
I am seeing PPD_CUSTOM_UNKNOWN in the cups headers in our
ci images, and that is cups 2.2.12. So this commit was
mistaken.
Yielding option means that if pango is built as a subproject, it will
take the value of that option from the parent project (e.g. gst-build).
For that to work it must be of the same type, which is "feature" instead
of "boolean" in all GStreamer modules.
Yielding option means that if pango is built as a subproject, it will
take the value of that option from the parent project (e.g. gst-build).
For that to work it must be of the same type, which is "feature" instead
of "boolean" in all GStreamer modules.
We're caching two things, either a node itself being rendered, or a
parent storing a cached version of a child as rendered to an offscreen
the size and location of the parent.
If both the parent and child uses the cache this will cause a conflict in
the cache as it is currently use keying of a node pointer which will have
the same value for the node-as-itself and the child-node-of-the-parent.
We fix this by adding another part to the key "pointer_is_child" which means
we can have the same node pointer twice in the cache.
Additionally, in the child-is-rendered-offscreen case the offscreen
result actually depends on the position and size of the parent viewport,
so we need to store the parent bounds in that case.
The GtkGears widget is a bit too chatty, especially when used inside
demos like the fishbowl. Let's use g_debug() instead of g_print() for
the GL debugging message.
This adds a bunch of snazz to the gltransitions demo. It is perhaps
a bit overloaded now, but it demos everything that we can do.
Changes:
* The fire shader is now not a bin, it just renders an animating
background with no textures involved.
* The stacks don't all start on the same page.
* The shaderbin passes the mouse coordinate to the shader.
* The shaderbin allows specifying a "border" so that you can
cause effects outside the bin child (something that is new to gtk4).
* All the buttons and the stacks are now in shader-bins that runs
a wobbly-widget effect based on the mouse position that
wobbles outside the child allocation.
Now that the functions that wrap them have gone away from the public
API, we need proper annotations for the virtual functions, otherwise
languages will not have enough information on nullable arguments and
ownership transfer.
Most of the time the snapshot is less than 16 levels deep (did some testing
in gtk-demo), so lets pre-allocate 16 levels of state stack to avoid the
extra allocation most of the time.
If all your callers already initialize the array element as needed,
then we don't need to memset it to zero first.
This is pretty useful for the snapshot state stack, because due
to the per-node-type data area the elements on the stack are
quite large, but often a lot of it is not used.
This inlines the splice and reserver GdkArray calls. These are
typically only called from the gdk_array_(append/set_size) functions
anyway, and inlining the caller means we can constant propagate the
constant arguments in those calls. Its hard to get exact numbers, but
in fishbowl i noticed a significant decrease in the time spent in
the array code when pushing and poping states.
This allows us to avoid updating uniforms if that is not necessary. This
in turn allows us to sometimes reuse the same draw op by just extending the
vertex array size we draw rather than doing a separate glDraw call.
For example, in the fishbowl demo, all the icons added at the same
time will have the same time and size, so we emit single draw calls
with 100s of triangles instead of 100s of draw calls with 2 triangles.
Add adds a demo showing off GskGLShaderNode in various ways.
It has a transistion widget, using some examples from
gl-transitions.com, with child widgets being both images, a GL area
and real widgets (that let you edit the transition shaders
themselves.
It also has a fancy fire effect on hove on the buttons.
For vulkan/broadway this just means to ignore it, but for the gl
backend we support (with up to 4 texture inputs, which is similar to
what shadertoy does, so should be widely supported).
A GskGLShader is an abstraction of a GLSL fragment shader that
can produce pixel values given inputs:
* N (currently max 4) textures
* Current arguments for the shader uniform
Uniform types are: float,(u)int,bool,vec234)
There is also a builder for the uniform arguments which are
passed around as immutable GBytes in the built form.
A GskGLShaderNode is a render node that renders a GskGLShader inside a
specified rectangular bounds. It renders its child nodes as textures
and passes those as texture arguments to the shader. You also pass it
a uniform arguments object.
Print out the full assembled shader sources when
GSK_DEBUG=shaders is given. This is very verbose,
but may be useful to see what we actually pass
to the compiler.
This way the child widgets can rely on the renderer (for example what
type it is) to decide details about how they render (such as if they
should use OpenGL shaders).
This adds a gsk prefix to the stuff in the preamble, as we want to
avoid it conflicting with things in the main shader. Especially once
we start allow some customization of shaders.
Almost always the source is created by combining various sources, which
means the line numbers in the error messages are hard to use. Adding
the line numbers to the source in the error message helps with this.
There is no real reason to have this on the side indexed via the
index, as it is stored next to each other anyway. Plus, storing them
together lets use use `Program` structures not in the array.
This property was only used until now when
there was neither an icon nor a label set,
for arrow direction and popover placement.
Starting with Gtk4, a GtkMenuButton with a
label shows an arrow at the right (in LTR)
of the label. Allow disabling the arrow or
changing its direction using the direction
property, to have a way to restore a Gtk3-
like look or to improve popover placement.
Fixes#2811.
Kinetic scrolling (and begin/end tracking) broke with commit cab1dcb696
since the pointing device used on X11 does not get as much GdkInputSource
granularity as the source device used to have in GTK3.
Actually this is kinda pointless, devices incapable of smooth scroll
should send discrete events, without those devices in the picture, we
want kinetic scroll to apply on every other device capable of smooth
scroll, so just do that.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3210
The texture that produce is upside-down, compared to what
GSK expects, so flip things around with a transform.
This fixes the shadertoy demo being upside-down after a
recent fix to avoid downloading and reuploading the texture.
... or gradients or borders or shadows. Instead, ensure that affines
have non-negative scale factors. Otherwise add a transform node.
The only place where this check is not necessary is color nodes, but
special casing them seems not worth it.
I added libcloudproviders to the base image, but
now I see that we already enabled libcloudproviders
in the build anyway, so I'm a bit confused.
Fixes: #2480
"inout" for the parameter ITER passed. This means that bindings would misjudge what
the function does. In the case of guile-gi, it would be misjudged for a predicate,
see gulie-gi bug 87.
I found that the gears demo was spending 40% cpu
downloading a GL texture every frame, only to
upload it again to another context.
While the GSK rendering and the GtkGLArea use different
GL contexts, they are (usually) connected by sharing data
with the same global context, so we can just use the
texture without the download/upload dance. This brings
gears down to < 10% cpu.
Currently, only if PangoFT2 is present and used it is supported
to retrieve the languages that are supported by a particular font.
If we don't have PangoFT2, remove the language filtering and the
sample text selection.
Based on earlier work by Chun-wei Fan, see
https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/2614
Copy the format conversion code from GdkMemoryTexture
so we can produce all formats, and test them all.
The upload fast paths assume that the stride is a
multiple of four, so some of the padding values cause
it to fail. Apart from that, things seem to work for
all combinations.
Create textures with various characteristics (alpha, premultiplication,
stride) that trigger different code paths in the gl texture upload
function, and show the resulting images. If all goes well, they all
should look the same.
On my system, this tests texture upload for memory formats
GDK_MEMORY_B8G8R8A8_PREMULTIPLIED, GDK_MEMORY_R8G8B8A8, and
GDK_MEMORY_R8G8B8, and it works with both gl and gles.
With the exception of gtk_buildable_get_id(), those are only used
to construct objects from XML descriptions, which is functionality
internal to GTK.
The API is therefore unlikely to be missed, and keeping it internal
means they can no longer unintentionally shadow object methods in
bindings with less namespacing; for example it's currently ambiguous
whether `infoBar.add_child()` refers to gtk_info_bar_add_child() or
gtk_buildable_add_child().
https://gitlab.gnome.org/GNOME/gtk/-/issues/3191
GtkBuildable's get_name()/set_name() methods may shadow
GtkWidget's methods. Avoid that by renaming the API to
get_buildable_id()/set_buildable_id(), which also reflects
the name of the XML attribute the API refers to.
https://gitlab.gnome.org/GNOME/gtk/-/issues/3191
Drop gtk_column_view_column_new_with_factory and
just make gtk_column_view_column_new accept a
nullable factory. This follows what we've been
doing elsewhere.
Update all callers.
The priv->in_button state that used to be relied upon for pointer
events has been reduced over time to a broken state, since the button
does not track crossing events anymore.
Make the coordinate-based checks apply for pointer events too, besides
touch events. This fixes GtkButton mistakenly emitting ::clicked with
pointer button releases outside the widget.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3121
Claiming early makes the contents unable to react to the touch press
event. Do this on GtkGestureDrag::update past a threshold, so the
child widget(s) can claim before the scrolledwindow does.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3125
This API is kinda stuck in the GdkEvent days, we now negotiate ownership
of the input sequence via GtkGestures. Remove it as it reflects a way to
work that was not exactly accurate and it will turn plainly wrong soon.
There is nothing really special about this code, its just a helper for
uploading pixel data to opengl, and we're not really in the business
of doing opengl-specific helper functions.
Do custom uploads rather than using gdk_cairo_surface_upload_to_gl(),
because this way we avoids a roundtrip (memcpy and possibly conversion)
to the cairo image surface format.
The gdk-pixbuf non-rgba format can be directly uploaded without
conversion.
The rgba format needs alpha premultiplication though, which is not
supported by GL during upload.
GLES doesn't support the GL_BGRA + GL_UNSIGNED_INT_24_8 hack that
we use on desktop OpenGL to upload textures directly in the cairo
pixel format. This adds the required conversions to all the places
that currently need it.
We also add a data_format to the internal gdk_gl_context_upload_texture()
function to make it clearer what the format are. Currently it is always
the cairo image surface format, but eventually we want to support other
formats so that we can avoid some of the unnecessary conversions we do.
Also, the current gdk_gl_context_upload_texture() code always converts
to a cairo format and uploads that like we did before. Later commits
will allow this to use other upload formats that gl supports to avoid
conversions.
This is the default OpenGL format, and in fact the only pixel format
that GLES supports uploading as. Actually, the premultiplied part is
really just about how we use the textures, but all textures in GTK
are premultiplied.
Just always tell the title and cell widgets to
clip their children to the right size. Otherwise
we risk things getting out of sync and unintended
overdraw.
Fixes: #3179
If we just parse a color, like image(#FFF), avoid allocating the
GPtrArray to store images. This happens in Adwaita for background images
of backdrop buttons. We save around 70 GPtrArrays this way.
The centered layout of the font previews don't look appealing
and make it harder to judge the relative width and weight of
the individual styles.
Fixes: #3188
In gtk_tree_view_build_tree with recurse=TRUE, the TEST_EXPAND_ROW
signal might invalidate the child iterator. Getting the iterator after
the signal (instead of before) fixes the issue.
Fixes https://gitlab.com/inkscape/inkscape/-/issues/1879
- Move padding from parent row to child cell.
- Align horizontal sizing of cell with header button.
- Properly support GtkColumnView:show-column-separators.
- Change cell height with and without .data-table.
This was showing up as tweak buttons being visible
when they should not. The code probably relied on
widgets being hidden by default (as they were in
GTK3).
Failure to do so makes the old pointer focus target
'sticky', because we end up ignoring the result of
picking the pointer focus until a motion event comes
in.
Fixes: #3172
Redo the tag insertion function to avoid quadratic
behavior, and at the same time, fix handling of
alpha for color attributes.
Update the copy of this function in gtk4-demo
as well.
Most of the surface api we have in the Wayland backend
only makes sense for toplevels, so reshuffle things to
take a GdkToplevel instead of a GdkSurface.
Update all callers and the docs.
Look at the languages supported by a font, and pick
a suitable sample text from the pango list of sample
texts. We can only implement this on platforms using
fontconfig, since it relies on pangofc apis.
This bumps the pango dependency to 1.47.1.
We must wl_surface.commit after xdg_surface.ack_configure to make it
have an effect. We failed to do so when a configure event didn't result
in new updates, so make sure we fall back on an simple
wl_surface.commit if there was no new actual frame painted.
Closes: #2910
In order to make the cairo renderer/context behave more similar to how
the OpenGL and Vulkan renderer/context behaves, request a frame callback
and commit in the end frame vfunc.
This means the end frame vfunc in cairo does
* attach buffer
* request frame callback
* sync surface state
* commit
Where as e.g. the OpenGL version of the same flow does
* attach buffer
* request frame callback
* sync surface state
* eglSwapBuffers()
where eglSwapBuffers() indirectly calls wl_surface_commit().
When we remove anchors with widgets from the text
buffer, we used to call gtk_widget_destroy(), which
indirectly called gtk_container_remove() which cleared
the child properly. When gtk_widget_destroy() was
removed, we replaced the calls with gtk_widget_unparent(),
but that is not enough. Explicitly call
gtk_text_view_remove() instead - we know the parent
is a text view.
Since the changes to GDK to use surface subtypes, CSD windows were
broken because we did not set the window styles properly. Fix this by
first acquiring whether decorations are used by the GtkWindow, and based
on that result we set the decorations that we want to use accordingly
and so apply them.
Thanks to Matt Jakeman for investigating into the issue and providing
pointers to a proposed fix.
Fixes issue #3157, besides the part where window sizes are not correct
since that is likely caused a separate issue.
Quickly clicking rows should always activate the row if
single-click-activation is enabled. Before, only the first click
(n_press == 1) would activate the row.
Using gtk_widget_insert_before on a complex container
is a *bad* idea; it will mess up the containers bookkeeping
of its children and can easily lead to failure and crashes.
While it's a bit dubious whether array+length annotations should be
marked as "nullable", we do this elsewhere in the API, so might as well
be consistent.
In practice, the array argument is only ever allowed to be NULL iff the
length argument is 0; annotations are static, so if somebody decides to
pass a NULL argument with a non-zero value, they will get a run time
critical error, instead of a compile time one, which is somewhat counter
to the point of annotating the API in the first place.
Fixes: #2923
When claiming a sequence in a gesture signal handler,
the expected result is that GtkGesture::handle-event
returns TRUE, causing the event to not be propagated
further.
This doesn't work for button release events, since
gtk_gesture_handle_event does the following:
add point
emit ::update
remove point
check claimed status
The ::update signal is where the application code
claims the sequence. But removing the point purges
the sequence from the gestures memory, so checking
the claimed status returns FALSE.
This patch fixes things to behave as expected, by
checking the claimed status before removing the point.
Add the names of the main widgets as keywords to
our demos, but also things like "game". This helps
finding relevant demos in our growing list. You
can now for example type "label", and find the
"error states" and "links" demos showing GtkLabel
features.
Filter the sidebar on keywords that can be provided
by the demos. We extract keywords from the doc comment
at the top of each demo source by looking for words that
look like class names. We also allow to specify keywords
explicitly.
In the video player demo, we have a button to make
the window fullscreen, but no easy way back. Add
the usual F11 keybinding, to make things at least
somewhat recoverable.
With csd, we are handling external widgets when
there is an entry in the headerbar. Use a weak ref
to prevent that pointer from going stale. This fixes
a crash when cancelling a save dialog.
Fixes: #3110
We were doing more iter comparisons than necessary in the
inner loop of gtk_text_layout_snapshot(), in the presence
of a selection. Rewrite the code to compare line numbers
instead, which is faster than full iter comparisons.
Don't pass 0x0 as size when calling gdk_surface_new().
The Wayland backend takes us literally, and we end
up with a surface that (temporarily) has these
dimensions, confusing other APIs that we pass the
size to, such as Vulkan.
We end up with a surface that has size 0x0 at the
time we create the Vulkan context, and that is a
size that Vulkan doesn't like, so ensure we request
at least 1x1.
Fixes: #3147
When the text says it has handled the event,
trust it. We don't want to emit ::search-started
if the content hasn't changed, but we still
should not propagate e.g. an Insert key press
if it has already toggled overwrite mode in
the text.
Fixes: #2874
We need to include both the scale and the filtering
in the key for the texture cache, since those affect
the texture.
This fixes misrendering in the recorder in the inspector
whenever transforms are involved. An example where this
was showing up is testrevealer's swing transition.
Assume that the fully expanded revealer will likely get an allocation
that matches the child's minimum or natural allocation, so we
special-case these two values.
So when - due to the precision loss - multiple sizes would match the
current allocation, we don't pick one at random, we prefer the min and
nat size.
The preference of nat size over min sie was decided after an IRC vote,
we don't actually have an idea what's more likely to happen in the real
world.
Should we ever get better data, we might want to switch.
We use ceil() in measure(), so using it again will increase the
child's size whenever there is even a tiny rounding error.
This should also not make the size too small, because:
min = ceil(child_min * scale)
min / scale >= child_min
floor (min / scale) >= floor (child_min) = child_min
The last equality is because child_min is an integer.
Fixes#3137
After commit 7e77afe94c moved the deletion
of text into the signal handler, in order to make undo work, we need to
override the GtkEntryBuffer::deleted-text class closure when subclassing
GtkEntryBuffer, as well as overriding GtkEntryBufferClass.delete_text,
otherwise the default class closure will be invoked, and will try to
delete an empty buffer.
Fixes: #3140
Use the Windows API CryptProtectMemory() to encrypt the data that we want to
secure, and use CryptUnprotectMemory() to de-crypt the secured data that we
want to access, since mmap() and mlock() are not available on Windows.
We have a widget for password and passphrase entries, but we have no way
to handle the data securely. This is usually performed by a separate
GtkEntryBuffer—for instance, the one in GCR. While we have API for
setting a new entry buffer on GtkText, we don't have API for
GtkPasswordEntry, though, so the options are:
- expose additional API for GtkPasswordEntry to allow setting a secure
text buffer on the internal GtkText widget
- provide a secure text buffer out of the box
Given that an insecure-by-default GtkPasswordEntry is basically
pointless, might as well have a secure buffer built in.
We don't really need to make the password entry buffer public out of the
box, but we can re-evaluate at a later date.
Fixes: #2403
This adds a small demo of using OpenGL shaders, it renders a quad
over the entire widget with a custom fragment shader. The coordinates
and the uniform names are compatible with the ones on shadertoy.com
(although some features, like texture inputs are missing currently).
The default shader in the demo is
https://www.shadertoy.com/view/wsjBD3 which is CC0, so it is
redistributable by Gtk+ (most other shaders are CC-BY-NC-SA which
isn't obviously compatible). I also added a set of buttons loading
a few other CC0 shaders I found.
The only likely place where this is going to happen
is if a renderer was explicitly requested with the
GSK_RENDERER environment variable, and in that case,
it is misleading to silently use a different renderer.
When we start a dnd of the selection in the drag-update handler,
set the gesture state to denied. Otherwise, we get more drag-update
signals, and things get really confused, leading to no dnd and
sadness.
Removed sentence that claimed the view will wrap the model in a
GtkSingleSelection, as it's no longer true. Fixed the code example in
GtkListView for the same reason. Fixed a small typo in GtkDropDown docs.
We were connecting signal handlers to the display
and seats here, and never cleaning them up, leading
to crashes after the inspector is closed. This is
fairly easy to reproduce under Wayland, where the
scroll device is only created the first time we
create a scroll event.
When rendering to an offscreen because of transforms,
check if transforming the bounds of the node results
in a non-axis-aligned quad. If it doesn't, we want
GL_NEAREST interpolation to get sharp edges. Otherwise,
we use GL_LINEAR to get better results for things
that are actually transformed.
This is a projecting version of the corresponding
graphene api. While we are at it, rewrite
gsk_matrix_transform_bounds() to use
gsk_matrix_transform_rect().
We want to include the gtk-doc subproject in release
tarballs, using --include-subprojects, but that only
works if we've actually built the subproject. And
enabling gtk-doc for dist builds is problematic -
it tends to break meson dist.
So declare the gtk-doc dependency independent of
-Dgtk_doc, and use --force-fallback-for for it.
Declarations, definitions, and gtk-doc stanzas should use the same name
for arguments. Otherwise both g-ir-scanner and gtk-doc will complain
that they can't find the argument.
The dependency block was completely wrong. It was:
1. Not searching for the lib manually when -Dvulkan=enabled (default).
The else block was only hit when -Dvulkan=auto.
2. Unconditionally searching for the vulkan library in the else block
when -Dvulkan=disabled
The manual searching is also not required because Meson has a custom
'vulkan' dependency class that already supports Windows, and is more
correct than the code here. Specifically, the current code does not
support picking up the Vulkan SDK from a custom path.
Fixes#3108
My previous change here was too hasty - this code is not
actually transforming points - it is just a convoluted
way to transform the z axis from child coordinates to
the parent.
Now that both arguments to the _new_with_factory() constructors
are nullable, there's no good reason to keep a separate _new()
around. Just make gtk_list_view_new() and gtk_grid_view_new()
take both a model and a factory.
Replace our uses of graphene_matrix_transform_point,
_point3d and _bounds by our own versions that handle
projective transforms correctly.
This fixes render node bounds being incorrect for widgets
involving projective transforms (e.g. testrevealer swing
transformations), and also fixes picking on such widgets.
Change the apis in GtkListView, GtkColumnView and
GtkGridView to be explicitly about GtkSelectionModel,
to make it obvious that the widgets handle selection.
Update all users.
Check buttons lost their ability to hold general
content. And while that is maybe sad, the tiny
images here are not really useful anyway, and
should just go away.
When using the gdk_display_close(), the handle to the Wayland compositor was not released. This could cause the consumption of all available handles, preventing other processes from accessing the display.
Fixing this by calling wl_display_disconnect() when releasing the GdkWaylandDisplay object.
Signed-off-by: Julien Ropé <jrope@redhat.com>
For --3to4, replace GtkRadioButton by either GtkCheckButton
or GtkToggleButton, depending on the value of :draw-indicator.
Update the testsuite to cover this.
Keep calling them radiobutton, since that is what they are.
And make the insensitive second group of three match what
we have in gtk3-widget-factory, and be parallel to the
insensitive checkbuttons next to it.
We trigger the paper dialog when the "manage" item
in the dropdown is selected. But the selection also
changes due to internal changes, such as reloading
the custom paper list when the paper dialog is
closed. We need to be extra careful to avoid triggering
another paper dialog when that happens.
Fixes: #3098
We trigger the paper dialog when the "manage" item
in the dropdown is selected. But the selection also
changes due to internal changes, such as reloading
the custom paper list when the paper dialog is
closed. We need to be extra careful to avoid triggering
another paper dialog when that happens.
The parser got its chars mixed up while parsing numbers
like 2.3e-04. While it is unlikely to meet such numbers
in human-generated css, we do have them e.g. when saving
render node trees with transforms.
Also add some css parser tests for number parsing.
A radiobutton without indicator is really just a togglebutton with a
group.
A radiobutton with indicator is really just a checkbutton with a group.
Make checkbutton its own widget not inheriting from GtkButton.
GtkRadioButton could be removed but it stays for now.
Radiobutton && !draw-indicator => Togglebutton
Checkbutton && !draw-indicator => Togglebutton
Radiobutton && draw-indicator => CheckButton + group
We want to ensure that the pointer position is reflected
when widget geometry changes, so add a function that tells
GDK "please create a motion event at the current position
on this surface, if one doesn't happen already".
Handle both these settings, and the older settings-daemon ones for
backwards compatibility. The keys are already checked for existence
in the schema, so it will just use the existing ones.
Prefer this location, but also look for the old location in
settings-daemon for backwards compatibility. This applies to both
direct settings lookups and via the settings portal.
A year ago, we make this function not return the child
surface anymore. But the information whether the device
is actually over the surface is still useful, and we
should not loose it.
We are reusing the GtkCrossingData struct for multiple
calls here, so we need to make sure that the targets
stay alive from beginning to end.
Fixes: #3090
If some node is fully outside the clip region we don't send it to the daemon.
This helps a lot in how much data we send for scrolling viewports.
However, sending partial trees makes node reuse a bit more tricky. We
can't save for reuse any node that could possibly clip different depending on
the clip region, as that could be different next frame. So, unless the
node is fully contained in the current clip (and we thus know it is not
parial) we don't allow reusing that next frame.
This fixes#3086
If alpha is 255, we use rgb() instead of rgba(), not if alpha is 0.
This makes the title bar gradient go from fully transparent to blue
rather than black to blue..
By adding 20 fonts / frame to the font list, we can
get the font chooser dialog to show up much faster.
This change gets the font chooser up in 265ms here.
We cannot pass all the data we pass to the virtual function, because the
types are private, but the class and the signal are public API.
The signal is just a notification, so we can decouple the virtual
function (which stays the same, for internal types that implement the
ATContext API contract) from the signal.
If the ATContext state hasn't changed—for instance, if the accessible
attributes have been set to their default value, or have been set to the
same value—do not emit an accessible state change. State changes can be
arbitrarily expensive, so we want to ensure that they are meaningful.
Bail out early, instead of going deep into the GtkAccessibleValue type
equal() implementation, where we expect both accessible values to have
the same type.
We want to ensure that the pointer position is reflected
when widget geometry changes, so add a function that tells
GDK "please create a motion event at the current position
on this surface, if one doesn't happen already".
Handle both these settings, and the older settings-daemon ones for
backwards compatibility. The keys are already checked for existence
in the schema, so it will just use the existing ones.
Prefer this location, but also look for the old location in
settings-daemon for backwards compatibility. This applies to both
direct settings lookups and via the settings portal.
If some node is fully outside the clip region we don't send it to the daemon.
This helps a lot in how much data we send for scrolling viewports.
However, sending partial trees makes node reuse a bit more tricky. We
can't save for reuse any node that could possibly clip different depending on
the clip region, as that could be different next frame. So, unless the
node is fully contained in the current clip (and we thus know it is not
parial) we don't allow reusing that next frame.
This fixes#3086
This commit adds border-spacing to actionbar, searchbar, .toolbar and
.app-notification, so their child widgets won't appear to be connected
even without margin.
A year ago, we make this function not return the child
surface anymore. But the information whether the device
is actually over the surface is still useful, and we
should not loose it.
We are reusing the GtkCrossingData struct for multiple
calls here, so we need to make sure that the targets
stay alive from beginning to end.
Fixes: #3090
If alpha is 255, we use rgb() instead of rgba(), not if alpha is 0.
This makes the title bar gradient go from fully transparent to blue
rather than black to blue..
By adding 20 fonts / frame to the font list, we can
get the font chooser dialog to show up much faster.
This change gets the font chooser up in 265ms here.
We cannot pass all the data we pass to the virtual function, because the
types are private, but the class and the signal are public API.
The signal is just a notification, so we can decouple the virtual
function (which stays the same, for internal types that implement the
ATContext API contract) from the signal.
If the ATContext state hasn't changed—for instance, if the accessible
attributes have been set to their default value, or have been set to the
same value—do not emit an accessible state change. State changes can be
arbitrarily expensive, so we want to ensure that they are meaningful.
Bail out early, instead of going deep into the GtkAccessibleValue type
equal() implementation, where we expect both accessible values to have
the same type.
As per GNOME mockups.
Since GtkFrame now sets GTK_OVERFLOW_HIDDEN, we can round the frame
without corner overlapping.
This also adds some margin to the child label of GtkFrame to ensure it
will not be clipped by the rounded corners of the frame.
To discriminate between is-focus and contains-focus,
we need to use notify::is-focus. This makes sure
we don't get annoying warnings when the blink_cb
gets triggered on an unfocused entry.
Fixes: #2979
When adding a custom palette, we need to arrange
for the custom section to stay at the bottom.
Maybe there should be a way to turn off custom
colors, too.
We are adding click gestures on the up/down buttons, but can't let
the GtkButton built-in ones prevent ours to run.
As the saying goes, if you can't beat them, join them. Group the
spinbutton and GtkButton gestures together, so it's irrelevant which
gets called first and ends up winning.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/3085
The code in gtk_widget_real_css_changed assumes that
queue_resize > queue_allocate > queue_draw, but the
second one is not really true. These days, we happily
keep reusing the same render node even when the child
allocation is changed.
So, if a css change has flags that tell us we should
redraw, we need to queue a draw, otherwise we might
end up reusing an outdated render node.
This fixes spinners staying visible when they stop
spinning, despite the theme setting their opacity
to 0.
Use feature options for things that are optional features,
update the docs.
Visible changes here is that the 'print-backends' option
got renamed to 'print' to go better with 'media', and the
'tracker3' option got renamed to 'tracker'.
For options that have been changed into features, the
syntax now is -Dfeature=enabled or -Dfeature=disabled
or -Dfeature=auto.
We don't support any profilers other than sysprof,
so name the option accordingly. While we are at it,
change it to a meson feature option, so
-Dprofiler=true becomes -Dsysprof=enabled
Instead of using sysprof-cli to profile subprocesses, this uses
libsysprof's SysprofProfiler directly so that we can avoid an indirect
subprocess as well as disabling the polkit nag.
To do this, we have to link against libsysprof instead of
libsysprof-capture. This is limited to the -Dbuild-tests=true and
-Dprofiler=true case.
Drop support for the org.gnome.Sysprof3.Profiler
D-Bus interface. It is not really used, and if
we don't expose it, we can simplify our profiler
infrastructure.
Rewrite expand/fill properties on GtkBox to
hexpand/halign/vexpand/valign on the child widget.
Rewrite GtkVBox and GtkHBox to GtkBox, setting the orientation
property.
Added a test for boxes.
CUPS uses resource paths in the form of "printers/printer_name"
or "classes/class_name" so it is enough to remove the "printers/"
or "classes/" prefix and use the string behind it as a name.
There was recently introduced a wrong check for the prefix.
This commit fixes it in the way it was originally intended.
Add a pango call to create the fontmap already in gtk_init.
This will let us hide the cost of FcInit() (which on font-heavy
systems can be ~100ms) in a thread, on the pango side.
The placessidebar gets this from somewhere else, but
it wasn't working in the gtk-demo sidebar, so add it
explicitly. placessidebar specific styles should be
dropped from Adwaita as much as possible, but not
doing that here.
The slowest step of highlighting our buffers is
inserting the markup into the buffer. Do that
incrementally, to avoid blocking the UI for
extended periods.
Run highlight asynchronously. It isn't a problem for
most demos, but the cursor demo has an unusually large
ui file, which takes highlight a little bit of time
to produce.
Instead of blindly creating new tags for every attribute,
reuse existing tags. For the syntax highlighting of the
ui file of the cursors demo, this gets us down from
20.000 tags to 6.
Commit 658719a205 moved the call to
gtk_window_compute_default_size() outside the
if (priv->needs_default_size)
but unfortunately, that function cleared that field,
so we never entered the branch.
Keep the setting of priv->needs_default_size in the
branch, where it belongs.
Commit 658719a205 moved the call to
gtk_window_compute_default_size() outside the
if (priv->needs_default_size)
but unfortunately, that function cleared that field,
so we never entered the branch.
Keep the setting of priv->needs_default_size in the
branch, where it belongs.
When we send an anchor rectangle with a width or
height of 0, mutter reponds with "Invalid anchor
rectangle size". So, don't do that.
This was seen as sudden disappearance of gtk4-demo
when you click the fishbowl benchmark all the way
through to the menubuttons.
Fixes: #3027
We only want the list .separators class to affect its
immediate children - otherwise, we end up with separators
in dropdowns that are places into button strips. As a side
benefit, restricting this to immediate children makes for
faster matching.
When running on a non-composited, non-rgba X server
(such as Xnest), force the drag icons for text selections
to have a background, so we don't end up with black
text on black background.
Fixes: #3048
As suggested by Matthias Clasen on gtk!2408. This keeps the
Apache-licensed stuff together.
As per the discussion on gtk!2408 and gtk!2409, replacing this with
upstream CRoaring is specifically not supported by the GTK maintainers.
Signed-off-by: Simon McVittie <smcv@debian.org>
4(a) requires giving recipients of the work a copy of the license, and
in any case it's best for source code distributions to be self-contained
(including the full text of all applicable licenses).
Signed-off-by: Simon McVittie <smcv@debian.org>
OS distributions that take copyright and licensing seriously will tend
to become concerned about source files that appear to come from a
third party and do not come with copyright/licensing information,
and stating that these files are modified is required by clause 4(b)
of the Apache license.
Signed-off-by: Simon McVittie <smcv@debian.org>
Surpisingly, when a monitored file gets atomically replaced
with g_file_set_contents(), the file monitor reports a
DELETED event for the temp file, which is not the file
I'm monitoring.
Just ignore DELETED events.
Fixes: #3036
G_DEFINE_DYNAMIC_TYPE declares these functions as static, and
-Werror=redundant-decls won't let us redeclare them. This is the
equivalent of 72c72d0b, but for a different backend.
Signed-off-by: Simon McVittie <smcv@debian.org>
Currently there is no way to alter the offset of the popup when positioning
with GdkPopupLayout. This makes using the popup difficult for scenarios
like completion windows where you may need to offset the window by a given
amount for aligning text.
gtk_popover_set_offset() allows setting these values and are analagous to
the function of the same name for GdkPopupLayout.
When we invalidate a y_range using the common pattern of y==0 and
old_height==new_height, we are generally invalidating the entire buffer.
This short-circuits that case to just invalidate the buffer in a faster
and more complete form. The problem here appears to be that we can't
always calculate the ranges properly to invalidate because validation
has not run far enough.
The colors demo was leaking a reference to its model.
This was showing up as crash when you manage to close
both the demo window and the main window while the
sorting is ongoing.
Drop the homegrown highlighting code, and just use highlight
to produce Pango markup.
When using an external highlighter, we can also highlight css,
xml, headers, at least.
When asking for a password, the message string is split on primary
and secondary if it contains a newline character. However, the newline
character is currently part of both strings, which creates weird
spacing between the GtkLabels. I suppose this is bug, which was not
visible as in most cases (if not all) the message string hasn't
contained the new line characters so far. But we are going to change
that now, see GNOME/gvfs!82. Let's drop the new line character similarly
as it is done when asking for a question, or showing processes in order
to fix the weird spacing.
We might break the loop early, e.g. if we're unmapped before the round
trip finishes, and to avoid the callback to write to invalid stack
memory, destroy the callback so it won't be invoked.
Fixes: #3026
We were not removing the pending_update idle
in dispose, which is at least suspicious, if
not deadly. Move the idle cleanup code into
unset_widget(), which we are already calling
in dispose().
As things currently stand, we get events for focus changes
before the widget is allocated, and try to scroll in response.
Therefore, leaving n_columns at 0 until size-allocate leads
to plenty of division-by-zero. Just set it to 1 initially
to avoid that. This is a workaround for #3025.
The testsvg test uses a method in librsvg that was introduced in
2.46.0. The test is now skipped if the librsvg version is too old.
(It was previously already skipped if librsvg wasn't found.)
The regular paths just emit ::end, which isn't the whole thing.
This gives an opportunity to gestures that are stolen the sequence
to clean themselves up.
The win toolchain has problems linking meson subprojects to libgtk.
To work around that build glib/pango from git and install them first
as long as the installed version is too old at least.
Fixes#3002
GtkAspectFrame code assumes that its child will be in
self->child, but that is only the case if we arrange
for <child> in ui files to end up calling
gtk_aspect_frame_set_child(). Therefore, implement
GtkBuildable.
Fixes: #3020
We were playing fast-and-loose with private GIO data
when showing settings bindings in the property editor,
and this was causing crashes.
We can show this information again if GIO ever gets
api to introspect it.
Fixes: #3015
GTK will not up front know how to correctly calculate a size, since it
will not be able to reliably predict the constraints that may exist
where it will be mapped.
Thus, to handle this, calculate the size of the toplevel by having GDK
emitting a signal called 'compute-size' that will contain information
needed for computing a toplevel window size.
This signal may be emitted at any time, e.g. during
gdk_toplevel_present(), or spontaneously if constraints change.
This also drops the max size from the toplevel layout, while moving the
min size from the toplevel layout struct to the struct passed via the
signal,
This needs changes to a test case where we make sure we process
GDK_CONFIGURE etc, which means we also needs to show the window and
process all pending events in the test-focus-chain test case.
gdk_gl_context_has_framebuffer_blit() and gdk_gl_context_has_frame_terminator()
were only used by by GDK/Win32, and they do not provide performance advantages
in GTK master, so clean up the code a bit by dropping them.
Check whether we really have x11 and wayland enabled before we try to setup the
tests to use these respective GDK backends, and only attempt to setup tests
running with the Broadway backend if it has been enabled.
Also, add a setup for running tests with the GDK-Win32 backend on Windows, for
builds that target Windows.
Use gdk_surface_get_geometry() to get the correct x and y coordinates of the
window that we are resizing, so that the window does not reposition itself
automatically at the top-left corner at resizing as we to used hard-code the x
and y coordinates to 0.
By doing so, we ensure that resizes of windows will work on Vulkan renderer, by
first calling gdk_win32_surface_handle_queued_move_resize() before we proceed
as usual
Use the shared function that was added in the previous commit, to simplify
things.
Also make gdk_win32_surface_get_queued_window_rect() and
gdk_win32_surface_apply_queued_move_resize() back into static functions, since
they are now used only by the code in gdksurface-win32.c
Since we need to deal with queued moves and resizes in the Cairo, GL and Vulkan
draw contexts, and the logic involved in all three of these are largely
similar, add a function gdk_win32_surface_handle_queued_move_resize() that will
handle this, which will be shared between these three types of draw contexts.
Move gdk_win32_surface_get_queued_window_rect() and
gdk_win32_surface_apply_queued_move_resize() to gdksurface-win32.c, since these
functions are not only used for Cairo draw contexts, but is also used for GL
draw contexts, and will be used for Vulkan draw contexts.
Don't get the default display when we compute the Aerosnap region, but instead
get it from the underlying GdkSurface that we are using for the computation.
Also, don't unref the monitors that we obtain from the display in the wrong
place, which was why we had crashes whenever we triggered AeroSnap code (and we
are actually not supposed to do that as they are owned by the GdkDisplay that
is owned by the GdkSurface we are using), and this will eliminate lots of
criticals that are spewed as a result.
gtk-doc assumes Docbook4, with <ulink> and so on.
Without this, all the links in markdown are converted
to <link xlink:href=...> and then lost in the docbook->html
conversion.
We can't use the ::destroy signal anymore; use
a weak ref instead, and make ensure the entry
stays around long enough for us to finish the
cleanup.
Fixes: #3004
No need for an intermediate box widget here;
we can just use a box layout. As a side-effect,
this fixes the theme to apply to the selection.
Fixes: #3005
ss00 doesn't exist, and we use xxxx as placeholder
for 'default' choices in alternatives. Add a warning
in case we run across invalid OpenType feature tags
in fonts.
Fixes: #2962
If we have a y==-1 then we are generally invalidating the whole textview.
For this case, we can just discard the entire GtkTextLineDisplay cache.
Fixes#2975
We don't want the completions to pop up after we call
gtk_file_chooser_set_current_name(). This used to be
handled by gtk_entry_set_text() blocking the completion
signal handler. We don't have that anymore, so block
popup completion around the call to gtk_editable_set_text()
instead.
Fixes: #2995
We can only insert tags in the buffer if they come
from the same GtkTextTagTable as the buffer uses.
If that is not the case, paste the text without tags.
Fixes: #2991
While we guarantee that the widget that a controller
is attached to stays around while it is handling an event,
the same is not true for the root that the widget belongs
to. In corner cases (such as clicking "Close" in the
fallback window menu), it may already be gone.
Avoid a critical in that case.
Fixes: #2998
We don't want to select on focus-in when the focus
comes from a child. The case where this does harm
is when you activate copy or paste actions from the
context menu. We close the menu before triggering the
action, and if that causes the text in the label to
be selected, unexpected things happen, since the action
applies to the current selection.
This is the equivalent of cd9f5733b3 for GtkLabel.
The code used to do
if (parent_class->clicked)
parent_class->clicked (...)
That is pointless because the parent_class never changes,
so there' no need for that if and commit 415946eb0f
took it out. Unfortunately, p arent_class->clicked is NULL
though, so the whole call needs to go.
This check used to read if (grab || device_type != GDK_DEVICE_TYPE_PHYSICAL),
the grab check was only reserved to physical devices, which the current
pointer device definitely doesn't act like. So the condition was "fixed" the
wrong way around, and the latter check is now moot, so the condition should
really go away. We always want to check the new toplevel under the pointer
here.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/2970
This allows us to use DPI_AWARENESS_CONTEXT_PER_MONITOR_AWARE_V2 for the
DPI awareness mode, which will help us to better support use cases with
multiple monitors. This is actualy a more advaned version of the
current PROCESS_PER_MONITOR_DPI_AWARE via using SetProcessDpiAwareness().
Note that this is not enabled by default, but also enabled via using
GDK_WIN32_PER_MONITOR_HIDPI, as in the PROCESS_PER_MONITOR_DPI_AWARE
case.
Note also, that appliation compatibility settings and DPI-awareness
manifests takes precedence over this API call, as before.
Since we have now made the Win32 GL contexts share the global context as
the other backends have, we are more ready to use the GL renderer by
default on Windows as well.
Note that currently we can only enable this when not running on
OpenGL/ES as the OpenGL/ES shaders are not ready at this point, and the
OpenGL/ES support that we have from libANGLE does not support full
desktop OpenGL operations.
Like the other backends, we ought to create our WGL/EGL GL contexts like
the following:
"Create a global GL context that connects all GL contexts on a display
and lets us share textures between them."
Use the new GTK_CROSSING_ACTIVE crossing events to update
the im context focus when the window becomes active or
inactive. IBus requires this, since it has only a single,
global focus location.
Fixes: #2390
Emit crossing events when the active window changes.
We don't want to emit GTK_CROSSING_FOCUS events, since
every window has its own focus location (focus does not
jump from window to window), so we use the new
GTK_CROSSING_ACTIVE type of crossing event for this.
Document the different crossing event kinds that we use,
and add GTK_CROSSING_ACTIVE. We are going to use it in
the future when the active window changes.
If GLES support is enabled on Windows, force GLES mode if we are running
on a ARM64 version of Windows (i.e. Windows 10 for ARM).
This is required as ARM64 versions of Windows only provide a software
implementation of OpenGL 1.1/1.2, which is not enough for our purposes.
Thus, we could make instead use the GLES support provided via Google's
libANGLE (which emulates OpenGL/ES 3 with Direct3D 9/11), so that we
can run GtkGLArea programs under OpenGL/ES in ARM64 versions of Windows.
Note that eventually we could update the libepoxy build files for Windows
to not check nor enable WGL when building for ARM64 Windows, as the WGL
items do not work, although they do build.
We need to use GL_BGRA instead of GL_RGBA when doing glReadPixels() on
EGL on Windows (ANGLE) so that the red and blue bits won't be displayed
inverted.
Also fix the logic where we determine whether to bit blit or redraw
everything.
Some implementations of the ES 1.00 shader (such as Google's ANGLE) do
not like the 'f' suffix for floats, so just drop it, as it should be
harmless to drop.
This is for adding a EGL-based renderer which is done via the ANGLE
project, which translate EGL calls to Direct3D 9/11. This is done as a
possible solution to issue #105, especially for cases where the needed
full GL extensions to map OpenGL to Direct3D is unavailable or
unreliable, or when the OpenGL implementation from the graphics drivers
are problematic.
To enable this, do the following:
-Build ANGLE and ensure the ANGLE libEGL.dll and libGLESv2.dll are
available. A sufficiently-recent ANGLE is needed for things to
work correctly--note that the copy of ANGLE that is included in
qtbase-5.10.1 is sufficient. ANGLE is licensed under a BSD 3-clause
license.
-Build libepoxy on Windows with EGL support enabled.
-Currently, prior to running GTK+ programs, the GDK_DEBUG envvar needs
to be set with gl-gles as at least one of the flags.
Known issues:
-Only OpenGL ES 3 is supported, ANGLE's ES 2 does not support the needed
extensions, notably GL_OES_vertex_array_object, but its ES 3 support is
sufficient.
-There is no autodetection or fallback mechanism to enable using
EGL/Angle automatically yet. There are no plans to do this in this
commit.
...EGL support needs to be explicitly enabled during the build of
libepoxy on Windows as it is not enabled by default on Windows.
With this, we can add an EGL renderer for Windows that make use of
Google's libANGLE, which is a library that translates OpenGL/ES calls
to Direct3D 9/11, which will provide better hardware compatibility
on Windows and would act as one of the foundations to resolve issue #105.
Set the accessible role for GtkLinkButton to button.
We don't use the 'link' role since ARIA says "if it
behaves like a button, use 'button'".
Update docs and add a test.
This changes should not be neccessary, since
GtkLinkButton derives from GtkButton, see #2965.
It's not a portable API, so remove it. The corresponding backend
specific functions are still available, if they were implemented, e.g.
gdk_macos_monitor_get_workarea() and gdk_x11_monitor_get_workarea().
pandoc insists on using the xlink namespace for hrefs,
and the namespace setup doesn't carry over xi:includes.
My first fix was to tell pandoc to generate standalone
docbook documents, which makes it insert the xlink
namespace. But it also makes it wrap all sections and
chapters in articles, and that messes up our toc structure.
So, patch things up differently by stripping the xlink:
from hrefs via regex.
Yay for XML!
Make GdkEvents hold a single GdkDevice. This device is closer to
the logical device conceptually, although it must be sufficient for
device checks (i.e. GdkInputSource), which makes it similar to the
physical devices.
Make the logical devices have a more accurate GdkInputSource where
needed, and conflate the event devices altogether.
Besides the implicit x/y assumptions, devices don't have axes. Those
are actually provided by the GdkDeviceTool driving the device, and
different tools may have different axes.
It does not make sense to offer this API that can change beneath
someone's feet, we now have gdk_device_tool_get_axes() which is static
to the tool.
Use the label accessible role for GtkLabel. ARIA has some
ominous wording about it going way, but while we have it,
GtkLabel is the obvious candidate for carrying it.
Update the documentation and add a test.
Sysprof has moved to a new ABI which removes GLib from the capture library
so that GLib itself can link against sysprof-capture.
This bumps the library ABI so we can keep things coordinated between all
the new tracing layers in the stack.
In some cases we explicitly want to unset an accessible attribute; for
instance, an accessible property is gated on a widget property, and if
the widget property gets unset, the accessible property should be reset.
We're currently overloading NULL to mean both "this value is undefined,
and should be reset to its default" and "the value collection failed".
Let's do error reporting right, by using GError to mean "the collection
failed, for this specific reason"; then, we can use a NULL return value
to signal that the accessible attribute should be reset to its default
value.
This is only relevant for pointer-sized attribute values: strings,
references, and reference lists; numeric, boolean, tristate, and token
values either cannot be undefined, or have a specific "undefined" value.
Looking at the xf86-input-wacom driver code, this is not even a thing
anymore. Drop this device type, in modern days there's
GDK_DEVICE_TOOL_TYPE_MOUSE for this.
Show a tab for accessibility information.
This shows the role and the accessible attributes
(states, properties, relations).
For now, changing the values is not possible, and
we only show the explicitly set values. In the future,
we want to show the attributes that are relevant for
the role, regardless of whether they are set or not,
and allow changing some of the attributes (the ones
that are not fully managed by GTK itself).
We're stopping activity mode when finalizing, which will change
accessible state; this will create a GtkATContext, and since GtkWidget
drops its GtkATContext on dispose(), we're going to end up leaking it on
the floor:
```
2,007 (64 direct, 1,943 indirect) bytes in 1 blocks are definitely lost in loss record 36,242 of 36,944
at 0x483977F: malloc (vg_replace_malloc.c:307)
by 0x5222105: g_malloc (gmem.c:106)
by 0x523E222: g_slice_alloc (gslice.c:1025)
by 0x523E261: g_slice_alloc0 (gslice.c:1051)
by 0x534B398: g_type_create_instance (gtype.c:1849)
by 0x53302EE: g_object_new_internal (gobject.c:1937)
by 0x53312AF: g_object_new_valist (gobject.c:2262)
by 0x532FEE8: g_object_new (gobject.c:1780)
by 0x4B3F942: gtk_test_at_context_new (gtktestatcontext.c:107)
by 0x491CC50: gtk_at_context_create (gtkatcontext.c:380)
by 0x4BFEDA0: gtk_widget_accessible_get_at_context (gtkwidget.c:8127)
by 0x4906079: gtk_accessible_get_at_context (gtkaccessible.c:83)
by 0x490618F: gtk_accessible_update_state (gtkaccessible.c:137)
by 0x4ACBA6D: gtk_progress_bar_act_mode_leave (gtkprogressbar.c:690)
by 0x4ACB4F8: gtk_progress_bar_finalize (gtkprogressbar.c:564)
```
We're also unparenting widgets and changing styles, which is another
potential source of leaks and side effects.
Track what we really need to send for inset shadows, which are used
as a border replacement in many cases.
Fishbowl says I can draw around 200-300 more switches per frame like
this too.
We must initialize the properties in init, since they
are expected to be there ab initio, and we can't call
gtk_accessible_update_property from finalize - it causes
us to recreate an at context and badness follows.
When converting DisplayLink frame presentation times, we need to take into
account the arch-specific types. This tracks changes in GNOME/GLib!1566 so
that precision is not lost.
Accessible values storing references and reference lists can be unset by
using NULL to mean "undefined"; since we cannot return the NULL value,
we need to intercept it when collecting a value, and replace it with an
undefined GtkAccessibleValue, which is also the value used as the
default for GtkAccessibleRelation values that store a reference or a
reference list.
Fixes: #2955
Make both gtk_grid_view_new and gtk_grid_view_new_with_factory
take a model as first argument, and make all arguments
allow-none and transfer full.
Update all callers.
Make both gtk_list_view_new and gtk_list_view_new_with_factory
take a model as first argument, and make all arguments
allow-none and transfer full.
Update all callers.
A dropdown without a model is useless, so accept a model
and expression in the constructor. Allow them to be NULL,
but consume them if given. This makes chained constructors
convenient without breaking language bindings.
Drop gtk_drop_down_set_from_strings() and instead add
gtk_drop_down_new_from_strings().
Update all users.
We don't pay attention to item-type anymore, so
drop the item-type property and the _for_item_type()
constructor, and allow passing NULL to the regular
constructor.
We don't make this constructor transfer-full, since
the selection filter model is not a wrapping model
like the others. It is more like fork than a wrap.
This is for consistency with other wrapping list constructors.
We want them all to be transfer full, allow-none.
Also make the constructor return GtkMultiSelection *.
Update all callers.
Make gtk_tree_list_model_new() take the root model
as first argument, and make it transfer full, for
consistency with other wrapping list constructors.
Update all callers.
Still missing here: Make the model property writable,
and allow passing NULL in the constructor.
We don't need as many functions to print out the property, relation, and
state of an accessible. Additionally, we should allow comparing the
accessible attributes with an expected value, and print out the real
accessible value if they do not match.
Some widgets have different roles after they are constructed, so we need
to allow changing the role defined by the class. We should still avoid
setting a role after the GtkATContext has been created.
This is a bit unfortunate, since the aria modelling
doesn't quite agree with ours, so we have to listen
for the togglebutton property change, and we inherit
the pressed state from the togglebutton accessible.
Only GtkWidget should use GTK_ACCESSIBLE_ROLE_WIDGET as its default
accessible role; the default for GtkAccessible and GtkATContext should
be GTK_ACCESSIBLE_ROLE_NONE, meaning "an element whose implicit native
role semantics will not be mapped to the accessibility API", according
to the WAI-ARIA specification.
We want to test the accessibility API, as well as the implementation
inside each widget. For that, we should expose an API that lets us
verify that a GtkAccessible has a given role, as well as a given
property.
The API follows the pattern of other GTest API:
- a macro to assert that a condition is respected
- a function that prints out the error message in case of failure
While we have split the various attributes for convenience, there's no
reason why we should have specialised data types for the attributes
container object.
Reduce the amount of subclassing, by handling collection of fundamental
types directly from the generic code paths. We now handle boolean,
tristate, integer, number, string, and relation values in the generic
code path; if an attribute supports the "undefined" value, we return the
undefined value singleton.
Drop roles and properties that were deprecated in WAI-ARIA 1.1, and add
new roles and properties defined in WAI-ARIA 1.2 and later.
We also split the relationship properties into their own enumeration, so
we can keep the GtkAccessibleProperty type more compact.
It's pointless, we can use an explicit value of `-1` everywhere.
Additionally, it complicates all code that uses the state enumeration as
an array index, since now we need to guard against a negative offset.
Some widgets have different accessible roles depending on some
parameter, so we cannot set the role at class init time. For those
widgets, we add an "accessible-role" property to GtkAccessible, and we
allow setting it (only) at construction time.
Each widget type has an accessible role associated to its class, as
roles cannot change during the life time of a widget instance.
Each widget is also responsible for creating an ATContext, to proxy
state changes to the underlying accessibility infrastructure.
An Accessible implementation must create an ATContext object. UI
elements are supposed to interact with the GtkAccessible API, but we
expose GtkATContext to allow patterns like delegation.
The ATContext type is meant to be used as the base class for
implementations of the assistive technology API—the actual mechanism
needed to communicate to components like the screen reader, or any other
AT.
Every time the widget state changes, the ATContext is meant to broadcast
the state change; and every time the AT queries the state of a UI
element, the ATContext is meant to provide that information.
We also have a "test" ATContext implementation, which is meant to be
used to write tests to verify that changes are propagated without
requiring a whole desktop session.
All accessible properties and states may have one of the following
types:
- true/false
- true/false/undefined
- true/false/mixed/undefined
- reference (to another UI element)
- reference list
- integer
- number (real numerical value)
- string
- token (one of a limited set of allowed values)
- token list
See: https://www.w3.org/WAI/PF/aria/states_and_properties#propcharacteristic_value
The GtkAccessibleValue is a simple reference counted type that can be
"subclassed" to implement each value type.
This initial commit adds GtkAccessibleValue and the basic subclasses for
plain boolean, tristate (true/false/undefined), and token types,
including statically allocated values that can be shared instead of
allocated.
GtkAccessible is an interface for accessible UI elements.
Currently, it doesn't do much except exist as a type; in the future, it
will be the entry point for all accessible state in GTK.
The list of roles is taken from the WAI-ARIA 1.2 specification:
https://w3c.github.io/aria/
Some of these roles do not make entirely sense from a GTK application
perspective, but we can remove them before finalizing the API.
To build a better world sometimes means having to tear the old one down.
-- Alexander Pierce, "Captain America: The Winter Soldier"
ATK served us well for nearly 20 years, but the world has changed, and
GTK has changed with it. Now ATK is mostly a hindrance towards improving
the accessibility stack:
- it maps to a very specific implementation, AT-SPI, which is Linux and
Unix specific
- it requires implementing the same functionality in three different
layers of the stack: AT-SPI, ATK, and GTK
- only GTK uses it; every other Linux and Unix toolkit and application
talks to AT-SPI directly, including assistive technologies
Sadly, we cannot incrementally port GTK to a new accessibility stack;
since ATK insulates us entirely from the underlying implementation, we
cannot replace it piecemeal. Instead, we're going to remove everything
and then incrementally build on a clean slate:
- add an "accessible" interface, implemented by GTK objects directly,
which describe the accessible role and state changes for every UI
element
- add an "assistive technology context" to proxy a native accessibility
API, and assign it to every widget
- implement the AT context depending on the platform
For more information, see: https://gitlab.gnome.org/GNOME/gtk/-/issues/2833
We are using floats for rgb, and we don't need more precision
for hsl colors either. We use hsl for computing color expressions
like shade(), lighter() and darker(), which are not precisely
specified anyway.
This commit updates the one test where the output changes a
tiny bit due to this.
Setting a width request is not quite enough, since
gtk_widget_set_size_request() only queues a resize
when the widget is visible. Explicitly force one
here. Without this, the popup sometimes shows up
too small.
This is truly a russian doll of documentation formats:
a string containing <> inside an xml fragment in an |[ ]|
gtk-doc example in markdown in a doc comment.
Sadly, something gets escaping wrong, so the <> end up
literally in the docbook and mess up the last step of
our document formatting, even after turning them into
entities.
Work around this with an extra level of entities that
really shouldn't be necessary.
This flag causes pandoc to emit a proper doctype
declaration and, crucially, namespace declarations
for the xlink namespace that it insists on using
for href attributes. Without this, putting external
links in md documents doesn't survive the journey
through xml.
Commit 0145809a94 replace the response-requested
signal with an action, but didn't actually remove the emission
of that no-longer-existing signal.
Fixes: #2942
Add a table mapping event signals to their event controller
replacements, and a table mapping former GtkContainer
subclasses to their gtk_container_add replacement.
SSave the missing keys as a bitset and iterate over that bitset in the
step function.
Solves the problem with a large UI block at the beginning of a sort
operation when all the keys were generated, in particular when key
generation was slow.
Benchmarks for maximum time taken by a single main loop callback:
initial sort with complex GFileInfo keys
old new
32,000 items 137ms 3ms
128,000 items 520ms 31ms
initial sort with string keys
old new
32,000 items 187ms 1ms
128,000 items 804ms 3ms
When updating a (partially) sorted model, take the known runs for the
existing sort and apply them to the new sort. That way, we don't have to
check the whole model again.
Benchmarks:
appending half the items to a model of strings
old new
512,000 items 437ms 389ms
1,024,000 items 1006ms 914ms
appending 10% of the items to a model of strings
old new
512,000 items 206ms 132ms
1,024,000 items 438ms 301ms
appending 1 item to a model of strings
old new
64,000 items 1.8ms 0.00ms
512,000 items --- 0.01ms
Previously, the sort was not stable when items were added/removed while
sorting or the sort algorithm was changed.
Now the sort looks at the item position (via the key's location in the
keys array) to make sure each comparison stays stable with respect to
this position.
This massively speeds up sorting with expensive sort functions that it's
the most worthwhile optimization of this whole branch.
It's slower for simple sort functions though.
It's also quite a lot slower when the model doesn't support sort keys
(like GtkCustomSorter), but all the other sorters do support keys.
Of course, this depends on the number of items in the model - the number
of comparisons scales O(N * log N) while the overhead for key handling
scales O(N).
So as the log N part grows, generating keys gets more and more
beneficial.
Benchmarks:
initial sort of a GFileInfo model with display-name keys
items keys
8,000 items 715ms 50ms
64,000 items --- 554ms
initial sort of a GFileInfo model with complex keys
items keys
64,000 items 340ms 295ms
128,000 items 641ms 605ms
removing half a GFileInfo model with display-name keys
(no comparisons, just key freeing overhead of a complex sorter)
items keys
512,000 items 14ms 21ms
2,048,000 items 40ms 62ms
removing half a GFileInfo model with complex keys
(no comparisons, just key freeing overhead of a complex sorter)
items keys
512,000 items 90ms 237ms
2,048,000 items 247ms 601ms
GtkSortKeys is an immutable struct that can be used to manage "sort
keys" for items.
Sort keys are memory that is created specifically for sorting. Because
sorting involves lots of comparisons, it's a good idea to prepare the
data relevant for sorting in advance and sort on that data.
In measurements with a PropertyExpression on a string sorter, it's about
??? faster
Instead of one item keeping the item + its position and sorting that
list, keep the items in 1 array and put the positions into a 2nd array.
This is generally slower while sorting, but allows multiple improvements:
1. We can replace items with keys
This allows avoiding multiple slow lookups when using complex
comparisons
2. We can keep multiple position arrays
This allows doing a sorting in the background without actually
emitting items-changed() until the array is completely sorted.
3. The main list tracks the items in the original model
So only a single memmove() is necessary there, while the old version
had to upgrade the position in every item.
Benchmarks:
sorting a model of simple strings
old new
256,000 items 256ms 268ms
512,000 items 569ms 638ms
sorting a model of file trees, directories first, by size
old new
64,000 items 350ms 364ms
128,000 items 667ms 691ms
removing half the model
old new
512,000 items 24ms 15ms
1,024,000 items 49ms 25ms
1. Run step() for a while to avoid very short steps
This way, we batch items-changed() emissions.
2. Track the change region accurately
This way, we can avoid invalidating the whole list if our step just
touched a small part of a huge list.
As this is a merge sort, this is a common occurence when we're buys
merging chunks: The rest of the model outside those chunks isn't
changed.
Note that the tracking is accurate: It determines the minimum change
region in the model.
This will be important, because the testsuite is going to test this.
... and use it in the SortListModel
Setting runs allows declaring already sorted regions so the sort does
not attempt to sort them again.
This massively speeds up partial inserts where we can reuse the sorted
model as a run and only resort the newly inserted parts.
Benchmarks:
appending half the model
qsort timsort
128,000 items 94ms 69ms
256,000 items 202ms 143ms
512,000 items 488ms 328ms
appending 1 item
qsort timsort
8,000 items 1.5ms 0.0ms
16,000 items 3.1ms 0.0ms
...
512,000 items --- 1.8ms
Simply replace the old qsort() call with a timsort() call.
This is ultimately relevant because timsort is a LOT faster in merging
to already sorted lists (think items-chaged adding some items) or
reversing an existing list (think columnview sort order changes).
Benchmarks:
initially sorting the model
qsort timsort
128,000 items 124ms 111ms
256,000 items 264ms 250ms
The model now tracks the original positions on top of just the items so that
it can remove items in an items-changed emission.
It now takes twice as much memory but removes items much faster.
Benchmarks:
Removing 50% of a model:
before after
250,000 items 135ms 10ms
500,000 items 300ms 25ms
Removing 1 item:
4,000 items 2.2ms 0ms
8,000 items 4.6ms 0ms
500,000 items --- 0.01ms
This is the dumbest possible sortmodel using an array:
Just grab all the items, put them in the array, qsort() the array.
Some benchmarks (setting a new model):
125,000 items - old: 549ms
new: 115ms
250,000 items - new: 250ms
This performance can not be kept for simple additions and removals
though.
This is fairly substantial rewrite of the GDK backend for quartz and
renamed to macOS to allow for a greenfield implementation.
Many things have come across from the quartz implementation fairly
intact such as the eventloop integration design and discovery of
event windows from the NSEvent.
However much has been changed to fit in with the new GDK design and
how removal of child GdkWindow have been completely eliminated.
Furthermore, the new GdkPopup allows for regular NSWindow to be used
to provide popovers unlike the previous implementation.
The object design more closely follows the ideal for a GDK backend.
Views have been broken out into subclasses so that we can support
multiple GSK renderer paths such as GL and Cairo (and Metal in the
future). However mixed mode GL and Cairo will not be supported. Currently
only the Cairo renderer has been implemented.
A new frame clock implementation using CVDisplayLink provides more
accurate information about when to draw drawing the next frame. Some
testing will need to be done here to understand the power implications
of this.
This implementation has also gained edge snapping for CSD windows. Some
work was also done to ensure that CSD windows have opaque regions
registered with the display server.
** This is still very much a work-in-progress **
Some outstanding work that needs to be done:
- Finish a GL context for macOS and alternate NSView for GL rendering
(possibly using speciailized CALayer for OpenGL).
- Input rework to ensure that we don't loose remapping of keys that was
dropped from GDK during GTK 4 development.
- Make sure input methods continue to work.
- Drag-n-Drop is still very much a work in progress
- High resolution input scrolling needs various work in GDK to land
first before we can plumb that to NSEvent.
- gtk/ has a number of things based on GDK_WINDOWING_QUARTZ that need
to be updated to use the macOS backend.
But this is good enough to start playing with and breaking things which
is what I'd like to see.
This was preventing any sort of building on macOS, even though the quartz
backend is currently non-functional. Fixing this is a pre-requisite to
getting a new macOS backend compiling.
Run the gdkkeysyms-update.pl script to pick up several
new keysyms:
GDK_dead_lowline
GDK_dead_aboveverticalline
GDK_dead_belowverticalline
GDK_dead_longsolidusoverlay
GDK_Keyboard
GDK_WWAN
GDK_RFKill
GDK_AudioPreset
Changing the selection in the object tree is
not a useful action if we are already in the
object details. Most likely, a user who picks
an object wants to inspect its details, so
just always show them.
Fixes: #1876
Bring back the actions tab; we don't receive
changes anymore, since GtkActionMuxer lost
the GActionGroup signals for this, and the
action observer machinery has no way to listen
for all changes.
Instead of implementing the GActionGroup interface
and using its signals for propagating changes up
and down the muxer hierarchy, use the GtkActionObserver
mechanism. This cuts down on the signal emission
overhead.
We should not rely on GtkWindow to have global
"activate-default" key bindings that happen to
fall back to activating the focus widget. This is
unreliable, since the bubbling up from the button
to the toplevel may run across other widgets that
may want to use Enter for their own purpose, and
then the button loses out. By adding our own
key bindings, the button gets to handle it before
its ancestors.
This fixes check buttons in the inspector property
list not reacting to Enter despite having focus.
If we don't, an ancestor (such a GtkListItemWidget)
may interpret the click as "I should grab focus!",
and still our focus away. This was causing hard-to-focus
entries in the property list in the inspector.
We were hiding the inspector when the window
is closed, but that has the side-effect of
keeping references to application windows,
so we would keep them artificially alive,
which can have side-effects.
So, make the inspector go away when closed.
This fixes the widget factory rendering too much.
In the widget-factory, we generally have a pretty small update area (two
spinners and a progressbar). We take the extents of that as a update
area and inital clip.
However, the first clip node we see is from the toplevel window, which
essentially increases the clip again to almost the entire window.
Fix that by ignoring such cases.
Porting code from GTK 3 without the ability to subclass GtkTreeView
directly can cause an extreme amount of pain on application developers.
It can also complicate performance when it comes to dealing with
encapsulation as the outer widget would also encapsulate the GtkScrollable
implementation from GtkTreeView, typically through GtkViewport.
Fixes#2936
That way, demo windows can be maximized and multiple demos can run at
once.
It's especially useful when using --run because the main window is
invisible then.
* GDK_ARRAY_BY_VALUE
#define this to get GArray-like behavior
* gdk_array_splice (v, 0, 0, NULL, 25)
Adding items but passing NULL as the items will zero() them.
* gdk_array_set_size()
A nicer way to call gdk_array_splice()
* constify getters
This is a scary idea where you #define a bunch of preprocessor values
and then #include "gdkarrayimpl.c" and end up with a dynamic array for
that data type.
See https://en.wikipedia.org/wiki/X_Macro for what's going on.
What are the advantages over using GArray or GPtrArray?
* It's typesafe
Because it works like C++ templates, we can use the actual type of
the object instead of having to use gpointer.
* It's one less indirection
instead of 2 indirections via self->array->data, this array is
embedded, so self->array is the actual data, and just one indirection
away. This is pretty irrelevant in general, but can be very noticable
in tight loops.
* It's all inline
Because the whole API is defined as static inline functions, the
compiler has full access to everything and can (and does) optimize
out unnecessary calls, thereby speeding up some operations quite
significantly, when full optimizations are enabled.
* It has more features
In particular preallocation allows for avoiding malloc() calls, which
can again speed up tight loops a lot.
But there's also splice(), which is very useful when used with
listmodels.
This feature was previously only supported on DBus compositors, such as
Mutter, this adds support for other compositors such as all of those
based on wlroots.
This implementation prefers the idle-inhibit Wayland protocol to the
DBus version if it is available, since the inhibitor is per-surface
instead of global it allows a finer control over which displays get
dimmed for instance. For every case not supported by this protocol, a
fallback to the DBus version is used.
It can’t do anything if the GtkWindow isn’t passed, which might warrant
some documentation change to encourage users to not use NULL for this
argument.
This has been tested on Sway.
Fixes#2202.
This uses the idle-inhibit protocol from wayland-protocols, to attach an
inhibitor to the GdkSurface. The inhibit function can be called as many
times as the user wants, but the uninhibit function MUST be called as
many times to unset the idle inhibition.
This has been tested on Sway.
When this code was ported from gtk_container_get_children
to the dom api, we inadvertendly inverted the order of the
list in one place. With the dom api, we can just avoid
reversing lists altogether, so do that.
Fixes: #2928
Not all compositors support _NET_WM_FRAME_DRAWN. In cases
where the compositor doesn't support _NET_WM_FRAME_DRAWN we don't
need to do all the fancy damage tracking and fence watching.
Furthermore, if the compositor doesn't support _NET_WM_FRAME_DRAWN,
it's possible that one frame will start before the previous frame has
made it through the pipeline, leading to a blown assertion.
This commit side-steps the unnecessary code and associated assertion
when _NET_WM_FRAME_DRAWN isn't supported.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/2927
The press gesture does not need to set claimed state just for n-press
tracking, it will however reset all other click gestures on each press
if the event gets propagated all the way up.
Fixes tracking of multi-press in gestures in widgets that are children
of the window handle. E.g. the headerbar listview in the "File browser"
gtk4-demo demo.
When the search entry disappears on Escape,
reset the search string to "", so we don't
end up with a filtered list and no obvious
way to remove the filtering.
Instead of an array of arrays, let's use an array of dictionaries; it's
easier to add optional keys without requiring to remember where to put
empty arrays.
The leak sanitizer causes on average 3-4 tests
to segfault during a testsuite run. Disable it
for now to see if we can get a successful
testsuite run with asan alone.
Run our testsuite under asan. We do this in a
separate build, since we need to turn off
introspection to make the build succeed.
As Michael Catanzaro pointed out, this requires a
privileged runner in order to use kernel apis.
When running the testsuite with the address sanitizer,
many of our dependencies cause it to report cause it
to report memory leaks, causing tests to fail.
Therefore, point the leak sanitizer at a list of
suppressions. The list is kept in the lsan.supp
file in git.
Add libasan and libubsan to fedora-base:v20,
and build fedora:20 on top of it.
This is so we can build and run the tests with
the address and undefined behavior sanitizers.
The gtk_gesture_group() call is not a commutative operation, it
takes two gestures, maybe detaches the first one from its current
group, and adds it to the same group than the second gesture.
With the flipped argument order here, GtkRange was actually detaching
the same gesture in order to group it with a second one two times, so
the desired effect to group all 3 gestures was not achieved.
Fixes autoscroll as the drag gesture is now actually grouped with the
click one, so drag offsets can be accessed from the autoscroll
timeout.
If we create an implicit grab on a surface, leave the surface, and
release the button, we would get 2 XI_Leave events, one with mode
XINotifyNormal when the pointer leaves the surface, and another with
mode XINotifyUngrab when the button is released.
Meanwhile, the upper layers rely on crossing events being paired,
and particularly in no crossing event being sent until the implicit
grab is dismissed (either by releasing it, or via more pervasive
grabs).
Ignoring the set of XINotifyNormal events while an implicit grab
is active adapts the X11 backend to this behavior. If the grab were
released or taken away by another grab, a crossing event with one
of the other XINotify*Grab/XINotify*Ungrab will be generated.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/2879
Check correctly that the captured motion events are emitted towards the
content or one of the scrollbars, in order to have it set the expected
"over" state depending on whether the drag begins from the scrolledwindow
content or one of the scrollbars.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/2879
This signal does not work on native file choosers,
and it exposes internals of the widget that should
not be public. And it is just not very interesting.
This signal does not work on native file choosers,
and it exposes internals of the widget that should
not be public. And it is just not very interesting.
This signal does not work on native file choosers,
and it exposes internals of the widget that should
not be public. And it is just not very interesting.
Add a GtkDirectoryList:monitored property, and
keep a file monitor if it is set to TRUE. To ensure
that the list reflects reality, we reload the directory
when monitoring is turned on after the fact. This means
that turning monitoring is expensive, while turning it
off is cheap, so we default to monitoring being on.
The stack wasn't updating its visible-child when
the stack pages visible property changes. This
showed up in the inspector, when showing the details
for a list model item.
Make mime-type rules hold multiple types. Store the
content types, so we don't have to do mime-type -> content-type
conversion in the match function. Store content types for the
pixbuf-formats rule as well, so we can avoid memory allocation
in the match function altogether.
char ** arrays are null-terminated everywhere, so make sure they are in
splice(), too.
Also fix the argument to be a const char * const * like in the
constructor.
Simplify all view model APIs and always return G_TYPE_OBJECT as the
item-type for every model.
It turns out nobody uses item-type anyway.
So instead of adding lots of APIs, forcing people to think about it and
trying to figure out how to handle filter or map models that modify item
types, just having an easy life is a better approach.
All the models need to be able to deal with any type of object going
through anyway.
Verify that the selection filter changes mirror
the selection changes of the underlying model,
as expected. These tests verify the fixes in
the previous commit.
When the position is 0, we can't check for unchanged
elements below with gtk_bitset_size_in_range. And
we don't need to, either.
And be careful when translating [start,length]
intervals to [first,last] ones. Off-by-one errors
lurk everywhere.
GtkTextLayout is private now and therefore we can drop all of
the indirection through the class vtable. Instead, just call the
implementations directly and remove the unused vtable entries
for default signal handlers.
Rename the DataList object to TreeData, in preparation
for adding a ListData object for list models. While
we are touching it, modernize it a bit (drop the Private
struct, use a layout manager, etc).
Since commit 972134abe4 a frame getting
drawn has three states (with the vendor nvidia driver at least):
1. drawn by gtk waiting on the GPU
2. drawn by GPU waiting on the compositor
3. drawn by compositor
Those three states are encoded in two flags: frame_pending and
frame_still_painting.
frame_pending means step 1 is done, but step 2 and 3 are still
in progress. frame_still_painting means step 2 is still in progress.
After step 1 is finished the surface is frozen until step 3 is finished.
When the compositor notifies gtk it's done with step 3, with a
_NET_WM_FRAME_DRAWN client message, the toolkit thaws the surface to
allow the next frame to proceed.
The compositor sometimes sends gtk a _NET_WM_FRAME_DRAWN client message
between steps 1 and 2. This message should be ignored because it's not
a reply to the current frame.
Unfortunately, gtk currently assumes if it gets a _NET_WM_FRAME_DRAWN
client message while waiting for step 2 that it's actually at step 3,
and proceeds to draw a new frame while the existing frame is still
pending, leading to a blown assertion.
This commit addresses the problem by ignoring _NET_WM_FRAME_DRAWN
client messages from the compositor unless actually expecting one.
Fixes: #2902
Since commit 972134abe4 we now call
glClientWaitSync for the vendor nvidia driver, to know when a frame
is ready for the compositor to process.
If a surface is hidden while a frame is still being rendered by the GPU,
the surface will never produce the damage event the code relies on to
trigger the call to glClientWaitSync. This leaves the fence dangling,
and the next time the surface is shown, it will start a fresh frame
and blow an assertion since the fence from the last frame is still
hanging around.
This commit ensures a frame gets fully wrapped up before hiding a
surface.
This makes the inspector lock up when used with any production
size list model, and blocks access to properties of the model
itself. Instead, we'll make the model available as an object
and add a data tab for list model contents, like we already
do for tree models.
In particular, it will NULL-ified the current global context if this is
the finalized one, avoiding dangling invalid pointers.
Would have been a cherry-pick from branch gtk-3-24 of commit
b592ded80a, but files moved.
Add a Help item to the gear menu that opens the
node-format.md file in a new window. This could
be improved if we could parse markdown and apply
tags, similar to how we can load pango markup.
One of the widget-factory focus tests is flaky in ci,
perhaps due to font changes causing size computations
to go slightly differently.
Drop this for now.
Include docstrings and format the list of supported
values better.
Also, add the same warning we have for GTK_DEBUG when
the environment variable is ignored.
Even if `gtk_expression_watch()` will do the same, we're calling public
API, so we should perform a check at the point of use, to ensure that
warnings are easily debuggable.
The pkg-config variables have been added in GLib 2.62.0. Let's fallback
to default names for these tools in such case (`pkg-config` still
returns a 0 return value, but with empty output for absent variables).
Some distributions are renaming Glib/GIO utilities for multi-arch
reasons so pkg-config variables have been added to find the correct name
of a tool. GTK+ should use these variables instead of searching in PATH.
See glib#1796.
Use GTK_DEBUG=builder-objects to make GtkBuilder warn
if a named object from a ui files doesn't get claimed
by gtk_builder_get_object(). This is useful for finding
dead wood in .ui files.
Always keep the order:
- [value]
- [marks.top]
- [marks.bottom]
- trough
Which makes sense given the rendering order. Slider should be drawn
after the marks.
Makes it possible to simply remove the custom snapshot implementations
in scale and range. And Adwaita does not depend on the node order
anyway.
Take ordering of cursor_position and selection_bound
into account when copying text to the clipboard, and
ensure that both orders work the same.
Fixes: #2898
We don't want to select on focus-in when the focus
comes from a child. The case where this does harm
is when you activate copy or paste actions from the
context menu. We close the menu before triggering the
action, and if that causes the text in the entry to
be selected, unexpected things happen, since the action
applies to the current selection.
Fixes: #2869
This reverts commit 67c2665028.
The splicing we do here has the important side-effect
of shifting the preedit attributes to the right position.
Without it, we end up always underlining the first chars
in the entry, regardless where the preedit happens.
This makes sure that we do actual key input right
in the middle between all the capture and bubble
event controllers, and are not dependent on the
ordering of those controllers.
The bug that triggered this change was that the
shortcut for activation (Enter) was getting triggered
before the key input, causing Ctrl-Shift-u hex
to stop working, since it never received the enter
to commit the sequence.
The gesture should claim the sequence after triggering uncancellable
actions, like pasting, showing a menu or selecting words/lines. A
single first button press initiating a drag does not trigger
anything yet, so it should avoid claiming the sequence.
The gesture should be accepted whenever it triggers uncancellable
actions in the widget. This means it should be accepted if the
click does result in toggling the switch.
This leaves the pan gesture room to handle dragging the handle.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/2895
... and do the right things:
nothing: selection = rubberband
ctrl: selection = selection OR rubberband
shift: selection = selection AND (NOT rubberband)
ctrl+shift: selection = selection XOR rubberband
(not sure this one makes sense, but toggling is fun)
Instead of storing the active items as we go, compute the affected items
whenever the rubberband changes and in particular when the rubberband
ends.
That way, the rubberband is guaranteed to select a rectangle even
after scrolling very far.
This is achieved by having a get_items_in_rect() vfunc that selects all
the items in the rubberbanded rectangle and returns them as a bitset.
The rubberband is now handled on the list coordinate system.
When starting the rubberband, we track the item under the pointer and
follow it when it is moving.
This may lead to the rubberband start position changing position and
while this may be confusing, it alerts users to the fact that something
crazy is going on.
In particular, track which items remain in ::items-changed
signal emissions.
But the main use case is sorting, which causes items-changed(0, n, n)
to be emitted.
We only want to send grab-notify to widgets that might have been
interacting with devices via events. Instead of going through all
widgets in all toplevels, we have the window/pointer focus information,
so we can just traverse the widget stacks for every involved foci.
Move away from grab_notify, and use the set_child_focus() vfunc to track
child widgets being set the keyboard focus. This is not 1:1, but seems
good enough at the moment.
Fix various issues that prevented inline completion
and inline selection from working reliably. We were
passing byte counts to gtk_editable_select_region in
one place, but that function expects char counts.
We were listening for GtkEditable::insert-text on
the GtkText widget, but that does not emit those signals,
so listen for GtkEntryBuffer::inserted-text instead.
Finally, we were not clearing the stored completion_prefix
enough, leading to situations where the stored prefix
does not match the text in the entry anymore.
In 99.9% of all cases, these are just NULL, NULL.
So just do away with these arguments, people can
use the setters for the rare cases where they want
the scrolled window to use a different adjustment.
While it's worth thinking about bringing the "windows can be dragged
with open popovers" behavior back, this does not kick in anymore, nor
should be the way to handle this given all the autoclose surface
semantic changes.
This kind of transient state sets the expectative that events update
devices, while it's more accurate to say that devices generate events.
It does not make to expose this function anymore.
This got stuck in ancient times when widgets were windowed, so the devices
in a window to know the devices in that widget would pan out. We do only
want here the devices that are inside the widget, not spread over the
surface, so rewrite this helper function to poke the toplevel foci, and
look they are contained inside the widget.
Crossing events are now detached from widget state, all tricky consequences
from getting multiple crossing events are now somewhat moot. Resort to sending
all generated crossing events, and drop this barely (ever?) used API.
When a gesture (group) claims a sequence, all other gesture groups
in the same widget should get cancelled. Not just previously claimed
ones, that shouldn't happen actually.
This is a list model holding strings, initialized
from a char **. String lists are buildable as well,
and that replaces the buildable support in GktDropDowns.
These sources are using GtkListStore apis,
but were replying on indirect includes to
get the header. Make this explicit, to prepare
for GtkEntryCompletion losing its tree view
dependencies.
Commit a0f6ff101e made sure that a
context was bound before calling glClientWaitSync, but it doesn't
check that the context shares objects with the context that created
the fence.
This commit does a little more validation before deciding the current
context is good enough.
Since commit 972134abe4 we now call
glClientWaitSync for the vendor nvidia driver, to know when a frame
is ready for the compositor to process.
glClientWaitSync can be called regardless of which context is currently
bound, but if no context is bound at all, it returns 0 without
doing anything.
This commit checks for that edge case, and ensures a context gets
made current in the event no context is already current, before calling
glClientWaitSync.
This api has not really been kept up with current
user experiences in popups, and we're better off
just dropping it and letting people do their own
popups if they need custom UI.
Use gtk_widget_prepend_controller to supersede entry keynav
while the popup is open. This fixes selecting completions
with the keyboard - the Enter keypress was ending up
triggering GtkText::activate instead of inserting the
selected completion into the entry.
Add a variant of gtk_widget_add_controller that
inserts the controller at the beginning, instead
of the end. This will be used in entry completion
to make sure the entry completion key event handling
supersedes the entry one while the popup is open.
Keep this private for now, until we determine if
it needs to be public api.
We were adding event controllers at the end, but
announcing a change at the beginning, in
gtk_widget_add_controller. Fix that by emitting
::items-changed for the position where we actually
inserted the controller.
When given a 0 timeout, glClientWaitSync is only supposed to return one
of three possible values:
- GL_ALREADY_SIGNALED - fence fired
- GL_WAIT_FAILED - there was an error
- GL_TIMEOUT_EXPIRED - fence hasn't fired yet
In addition, it can also return GL_CONDITION_SATISFIED if a non-zero
timeout is passed, and the fence fires while waiting on the timeout.
Since commit 972134abe4 we now call
glClientWaitSync (with a 0 timeout), but one user is reporting it's
returning some value that's not one of the above four.
This commit changes the g_assert to a g_error so we can see what
value is getting returned.
May help with https://gitlab.gnome.org/GNOME/gtk/-/issues/2858
According to [1], '_timezone' is already used for a global variable in the
time.h system header that is supplied by Microsoft, so using that for our
variable name when we are including time.h either directly or indirectly
will cause trouble.
This renames such variables to '_tz' to avoid that
[1]: https://docs.microsoft.com/en-us/cpp/c-runtime-library/daylight-dstbias-timezone-and-tzname
Due to an oversight, when multi filters were split into
any and every, any ended up with the listmodel and buildable
implementations, and every didn't get any.
Move the implementations up.
Do it via the new size setting on the color list.
Also add an option to set the maximum size of the color list that it'll
be filled to so rudimentary performance testing is possible.
Hint: Do not select 16777216.
This reinstates commit c16848c2bb683f7ddf51571aa5951f8a4a1ea622.
This showed up as a test failure when we get NULL as
selected item and then try to unref it. Luckily
get_selected_item is transfer none, so we don't have
to worry about it.
gtk_weather_info_new was consuming the info, but not
the timestamp, which confused me. Make it not consume
either, and be more explicit about where the unrefs
happen.
We are currently not robust against model changes or
widget invalidations, so we can actually end up in
situations where we run out of items here. Handle
the failure a bit more gracefully, by returning NULL.
This is good enough to make scrolling work okish most
of the time. We still need a proper fix to handle
other situations.
Use a drop motion controller to autoscroll horizontally
while a drag operation is hovering over the list. The
vertical scrolling is handled by the listview.
Break out an update_autoscroll() function that can
be used for other things than rubberbanding. It will
be used for autoscroll during DND in the future.
All widgets cache their render node already. Just allocate the
last_visible_child always at 0/0 and then move its rendernode around
during snapshot.
Fixes#2678
We already know that a widget will have literally 1 node, not more.
Avoid doing the GtkSnapshot state stack dance and just append a new
transform node instead.
Seems to give me around 400 more icons in the fishbowl
First, almost all widgets are either visible and mapped, or visible and
unmapped because they are !child_visible. In both cases, we do care
about the size. If they are indeed invisible, gtk_widget_measure() will
take care of that by returning all 0.
And gtk_widget_unparent() already queues a resize on the parent if
necessary, so that is unneeded as well.
Juneteenth (a portmanteau of June and nineteenth)[2] (also known as
Freedom Day,[3] Jubilee Day,[4] and Liberation Day,[5]) is an unofficial
American holiday celebrated annually on the 19th of June in the United
States.
-- https://en.wikipedia.org/wiki/Juneteenth
The GtkIMMultiContext uses a delegate pattern to proxy an existing input
method context. Let's not use loaded terms like "slave" when we have
perfectly fungible terms like "delegate".
Libcloudproviders has a single header entry point, so we shouldn't
include a sub-header.
Additionally, the include path provided by the pkg-config file is:
-I${includedir}/cloudproviders
So the include directive should be:
#include <cloudproviders.h>
The fact that it worked until now was an accident caused by the blanket:
-I${includedir}
we get for free; it broke the build when using libcloudproviders as a
subproject.
GtkSingleSelection does not give us a notification when
autoselect kicks in to select the first item. This was
barely noticeable, since the notebook ends up correctly
populated anyway, but the window title is not.
Currently we have to hunt all the `image` keys when updating the image
in use for a job. We should centralise them all at the top of the file,
so they are easy to update in one simple change.
This patch implements the openFiles delegate which is required
to open files which are associated with an application via the
Finder or via open on the command line. The patch has been
proposed by jessevdk@gmail.com.
See: https://gitlab.gnome.org/GNOME/gtk/-/issues/463
I tested the patch with the GNU pspp application on MacOS with
the quartz backend.
Shift-Tabbing was getting stuck in GtkSearchEntry and
GtkPasswordEntry, since they grab the focus to a child
of theirs. Copy the same fix that we are already using
in GtkEntry.
Fixes: #2842
If we leave the valign to be the default fill, then
the images pick up baseline alignment from the environment,
which can occasionally lead to misplaced -/+ icons in
spinbuttons, for example in the "Page Setup" tab of
the print dialog.
Fix scheduling of the frame clock when we don't receive "frame drawn"
messages from the compositor.
If we received "frame drawn" events recently, then the "smooth frame
time" would be in sync with the vsync time. When we don't receive frame
drawn events, the "smooth frame time" is simply incremented by constant
multiples of the refresh interval. In both cases we can use this smooth
time as the basis for scheduling the next clock cycle.
By only using the "smooth frame time" as a basis we also benefit from
more consistent scheduling cadence. If, for example, we got "frame
drawn" events, then didn't receive them for a few frames, we would still
be in sync when we start receiving these events again.
When an animation is started while the application is idle, that often
happens as a result of some external event. This can be an input event,
an expired timer, data arriving over the network etc. The result is that
the first animation clock cycle could be scheduled at some random time,
as opposed to follow up cycles which are usually scheduled right after a
vsync.
Since the frame time we report to the application is correlated to the
time when the frame clock was scheduled to run, this can result in
uneven times reported in the first few animation frames. In order to fix
that, we measure the phase of the first clock cycle - i.e. the offset
between the first cycle and the preceding vsync. Once we start receiving
"frame drawn" signals, the cadence of the frame clock scheduling becomes
tied to the vsync. In order to maintain the regularity of the reported
frame times, we adjust subsequent reported frame times with the
aforementioned phase.
Use better matching format modifiers/specifiers, initialise some things
which in theory wont be written to because of getters using g_return_if_fail(),
a cast, and gsize as input for malloc because gsize!=glong on 64bit Windows.
We lost this when GtkSpinButton was first ported
to the new editable regime, and then the GtkBoxLayout.
Bring it back, but without text measurement, by overriding
width-chars for the GtkText inside, and only do it if
GtkSpinButton::width-chars is unset (ie -1).
Also adjust the documentation slightly to point out
how auto-sizing can be turned off.
GTK 4.0 was currently using GL_EXT_framebuffer_object, which is
deprecated as the ARB version has been merged into OpenGL 3.0 as well as
OpenGL ES 2.0, and provides laxer requirements.
This is a port of !2076 for 4.x.
This gesture handles both individual touch events and touchpad gesture
events, and was checking the touchpad phase in generic code paths. This
is dubious since event methods error out on the wrong GdkEventTypes.
Check the touchpad gesture phase within the branch handling touchpad
events, and make it clear which is the gesture phase of all that we are
ignoring.
Fixes: https://gitlab.gnome.org/GNOME/gtk/-/issues/2825
This makes 'gtk_file_chooser_get_filter' work for the
portal native file chooser by handling the corresponding
'current_filter' argument in the response retrieved via
D-Bus.
In order to try to map the retrieved 'current_filter' to one
of the existing list of filters, use the retrieved filter's name,
similar to how xdg-desktop-portal-gtk does it when evaluating the
'current_filter' input parameter in 'options'.)
Note: This depends on the following merge/pull requests
which fix the filter handling in gtk for native file choosers
and introduce the 'current_filter' handling for FileChooser portal.
* https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/1959
* https://github.com/flatpak/xdg-desktop-portal/pull/493
* https://github.com/flatpak/xdg-desktop-portal-gtk/pull/311
This fixes#1820 for desktop portal case.
Fixes: #1820
'gtk_file_chooser_get_filter' did not work for GtkFileChooserNative,
since the previous way did not properly handle the delegate dialog,
s.a. commit a136cbae8f
("filechoosernative: forward current_filter to delegate dialog",
2018-11-29) for details, wich basiscally fixed the same thing for
the 'gtk_file_chooser_set_filter' case.
This fixes#1820 for the fallback dialog. A solution for the portal
one (which also requires changes to xdg-desktop-portal and
xdg-desktop-portal-gtk as well) will be suggested in a subsequent step.
Bug: #1820
Scroll events can have history too, so make a
getter that works for both. This drops the
gdk_scroll_event_get_history getter that was
added a few commits earlier, since we now
store scroll history in the same way as
motion history.
Update the docs, and all callers.
There is really no need to store 128 doubles as axes,
ever. We can do just fine with 10. At the same time,
add a GdkAxisFlags member, so we can interpret the
values without having to go chasing the right device
for this information.
Only return one accumulated scroll event per frame.
Compress them by adding up the deltas.
Still missing: a way to capture history, like
we do for motion events.
Fixes: #2800
This is not just about consistency with other functions.
It is about avoiding reentrancy problems.
GtkListBase first doing an unselect_all() will then force the
SelectionModel to consider a state where all items are unselected
(and potentially deciding to autoselect one) and then cause a
"selection-changed" emission that unselects all items and potentially
updates all the list item widgets in the GtkListBase to the unselected
state.
After this, GtkListBase selects new items, but to the SelectionModel and
the list item widgets this looks like an enitrely new operation and
there is no way to associate it with the previous state, so the
SelectionModel cannot undo any previous actions it took when
unselecting.
And all listitem widgets will now think they were just selected and
start running animations about selecting.
Replace a previous fix with a more correct one: Update the
selected state from the model instead of reusing the old state, the
model might have updated the selected state.
When translating coordinates from an event, we need to
take the surface transform into account. This makes
double-clicking on editable cells in treeviews work
again.
Fixes: #2831
This is a selection model that stores the selection
state in a boolean property of the items, and thus
persists across reordering and similar changes.
Fixes: #2826
The selection model will only emit one of items-changed
or selection-changed, so when we handle an items-changed,
we must assume that selection state may have changed for
any of the newly added items.
This fixes lingering selection state in the visible range
if you change the sorting of the colors demo.
Fixes: #2827
Move the selection changes to button release, to
avoid conflict with the drag gesture for rubberbanding.
This avoids peeking at the parent, and is generally
nicer.
Among other things, you can now shift-click to select
a range in the colors demo in gtk4-demo.
When I changed things to only collect the set
in the stop() function, I overlooked that this
has the side-effect of only handling items which
are backed by a widget at the time stop() is called.
If we make a big rubberband and autoscroll down too
far, we loose the items that go out the visible range
at the top. Fix that by maintaining the set as we go.
It's finally unused.
Accessible types should either watch properties they are interested in
directly, or should have (private) API to allow widgets to update the
accessible state directly.
Now that we don't have any additional subclasses of GtkEntryAccessible
in GTK, we can drop all the conditional fluff in the base class.
We still need to subscribe to the global notify signal, because of the
sheer amount of properties watched by GtkEntryAccessible.
GtkPasswordEntryAccessible is not a GtkEntryAccessible any more, so it
will need a proper implementation of various interfaces and
functionality in order to work like any other entry.
We're already listening to the adjustment property on the spin button,
there's no need to reset the adjustment on widget set/unset, since the
accessible instance is always tied to the same widget.
Drop the GtkWidgetAccessibleClass.notify_gtk and the
AtkObjectClass.initialize overrides: they don't do anything relevant.
Instead, have GtkProgressBar update the accessible state when the
fraction changes.
Do not use a generic "notify" signal handler.
Additionally, clean up the GtkIconViewAccessible implementation to bring
it up with modern idiomatic GObject.
It feels slightly wrong to have GtkOrientable operate on widgets, but at
least what happens when an orientable widget changes orientation should
be part of GtkWidget.
This will allow to add more state changes without accessing widget state
from the outside of gtkwidget.c.
We expect widgets to use their own derived GtkWidgetAccessible type,
these days, and given that we hard code the default accessible type of a
GtkWidget to GtkWidgetAccessible, and that we enforce the dependency of
the type passed to gtk_widget_class_set_accessible_type(), the registry
code path is clearly unused.
The tooltip handling in GtkWidget is "special":
- the string is stored inside the qdata instead of the private
instance data
- the accessors call g_object_set() and g_object_get(), and the
logic is all inside the property implementation, instead of
being the other way around
- the getters return a copy of the string
- the setters don't really notify all the involved properties
The GtkWidgetAccessible uses the (escaped) tooltip text as a source for
the accessible object description, which means it has to store the
tooltip inside the object qdata, and update its copy at construction and
property notification time.
We can simplify this whole circus by making the tooltip properties (text
and markup) more idiomatic:
- notify all side-effect properties
- return a constant string from the getter
- if tooltip-text is set:
- store the text as is
- escape the markup and store it separately for the markup getter
- if tooltip-markup is set:
- store the markup as is
- parse the markup and store it separately for the text getter
The part of the testtooltips interactive test that checks that the
getters are doing the right thing is now part of the gtk testsuite, so
we ensure we don't regress in behaviour.
When exclusive is TRUE, we would not always emit a
::selection-changed signal that covers all the items
that were unselected.
This commit includes a test.
When exclusive is TRUE, we would not always emit a
::selection-changed signal that covers all the items
that were unselected.
This commit includes a test.
Tell reordered columns to reorder their cells to
the new position. This is necessary to get things
like separators right. The visible symptom of this
problem was the lack of the right border when the
last column is reorder to another position, since
the title widget was still the last in its container,
so :last-child applied.
When we are given a for_size as width for the whole
column view, we need to distribute it over the columns
as gtk_column_view_allocate_columns would, in order
to find out which for_size to give to each cell.
This is a bit recursive, but works. Since we are
doing this recursion for every row, we should consider
adding a cache for those distributed widths.
With the vendor provided Nvidia driver there is a small window of time
after drawing to a GL surface before the updates to that surface
can be used by the compositor.
Drawing is already coordinated with the compositor through the frame
synchronization protocol detailed here:
https://fishsoup.net/misc/wm-spec-synchronization.html
Unfortunately, at the moment, GdkX11Surface tells the compositor the
frame is ready immediately after drawing to the surface, not later,
when it's consumable by the compositor.
This commit defers announcing the frame as ready until it's consumable
by the compositor. It does this by listening for the X server to announce
damage events associated with the frame drawing. It tries to find the
right damage event by waiting until fence placed at buffer swap time
signals.
This commit moves some of the end frame sync counter handling
code to subroutines.
It's a minor readability win, but the main motivation is to
make it easier in a subsequent commit to defer updating the
sync counter until a more appropriate time.
commit 14bf58ec5d dropped support
for using the DAMAGE extension since there was no code that
needed it.
We're going to need it again, however, to address an NVidia
vendor driver issue.
This commit does the plumbing to add it back.
The columnview overrides the width of its content,
so if a label is not willing to wrap its content
below a certain width, it will just get cut off.
Avoid that by not setting width-chars on the wrapping
content.
When the application does not receive "frame drawn" signals we schedule
the clock to run more or less at intervals equal to the last known
refresh interval. In order to minimize clock skew we have to aim for
exact intervals.
We try to step the frame clock in whole refresh_interval steps, but to
avoid drift and rounding issues we additionally try to converge it to
be synced to the physical vblank (actually the time we get the
frame-drawn message from the compositor, but these are tied together).
However, the convergence to vsync only really makes sense if the new
frame_time actually is tied to the vsync. It may very well be that
some other kind of event (say a network or mouse event) triggered
the redraw, and not a vsync presentation.
We used to assume that all frames that are close in time (< 4 frames
apart) were regular and thus tied to the vsync, but there is really no
guarantee of that. Even non regular times could be rapid.
This commit changes the code to only do the convergence-to-real-time
if the cause of the clock cycle was a thaw (i.e. last frame drawn and
animating). Paint cycles for any other kind of reason are always
scheduled an integer number of frames after the last cycle that was
caused by a thaw.
When we get to a paint cycle we now know if this was caused by a
thaw, which typically means last frame was drawn, or some other event.
In the first case the time of the cycle is tied to the vblank in some
sense, and in the others it is essentially random. We can use this
information to compute better frame times. (Will be done in later
commits.)
Visual Studio does not allow decorating functions with '__declspec (dllexport)'
if a prototype exists and is not decorated with '__declspec (dllexport)' as
well, so we cannot just decorate g_io_module_[load|unload|query] in the various
module sources with G_MODULE_EXPORT because the prototypes of these functions
have been marked with _GLIB_EXTERN, which equates to 'extern' unless overridden
Fix this by overriding _GLIB_EXTERN with the appropriate visibility flag, as we
have used to define _GDK_EXTERN. Unfortunately, we can't just use _GDK_EXTERN
G_MODULE_EXPORT as they may have not been defined yet for our use
Do this across the board for all modules, even if they are not buildable on
Visual Studio nor Windows, for consistency's sake.
The included fribidi header is not used in gdkkeys-wayland.c and already
included in gdk.c which causes linker issues due to the header defining
a global variable.
If you add a widget to a parent, this will invalidate the css nodes
for parent/siblings. Afterwards, if the parent is mapped, we will
realize the new child. This calls gtk_widget_update_alpha() which
needs the css opacity, so it revalidates the css.
Thus, for each widget_add (while visible) will trigger a full
revalidation of each sibling. If you add N children to a parent that
leads to O(N^2) revalidations.
To demo this I changed gtk-demo to always double the count
(independent of the fps) and print the time it took. Here is the
results (after a bit):
Setting fishbowl count=256 took 3,4 msec
Setting fishbowl count=512 took 10,1 msec
Setting fishbowl count=1024 took 34,1 msec
Setting fishbowl count=2048 took 126,3 msec
Setting fishbowl count=4096 took 480,3 msec
Setting fishbowl count=8192 took 1892,7 msec
Setting fishbowl count=16384 took 7751,0 msec
Setting fishbowl count=32768 took 38097,7 msec
Setting fishbowl count=65536 took 191987,7 msec
To fix this we drop gtk_widget_update_alpha() and just
calculate it when needed (which is only in a single place).
It was really only necessary because we previously set
the alpha on the surface.
With this fix the above becomes:
Setting fishbowl count=256 took 1,0 msec
Setting fishbowl count=512 took 1,9 msec
Setting fishbowl count=1024 took 3,7 msec
Setting fishbowl count=2048 took 7,4 msec
Setting fishbowl count=4096 took 18,1 msec
Setting fishbowl count=8192 took 31,0 msec
Setting fishbowl count=16384 took 66,3 msec
Setting fishbowl count=32768 took 126,7 msec
Setting fishbowl count=65536 took 244,6 msec
Setting fishbowl count=131072 took 492,2 msec
Setting fishbowl count=262144 took 984,3 msec
The new names are
GtkListView - listview row
GtkGridView - gridview child
GtkColumView - columnview header
columnview listview row
Adwaita css has been updated to preserve
existing styles.
Fixes: #2818
This was done in a weird way where we always call reftest_uninhibit_snapshot()
on paint, and then re-inhibited it if it wasn't inhibited. To make this
work it also started with an extra inhibit.
This is very contorted and based on how this historically worked. This
changes it to just do:
if (inhibit_count > 0)
return;
And keep inhibit_count at its initial zero value unless it is actually
inhibited.
In https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/2027 i was getting
Bail out! ERROR:../testsuite/reftests/reftest-snapshot.c:212:reftest_uninhibit_snapshot: assertion failed: (inhibit_count > 0)
In (for example the box-shadow-changes-modify-clip reftest. I can reproduce this (on master) with:
```
$ xvfb-run -a -s "-screen 0 1024x768x24" meson test --suite gtk:reftest "reftest box-shadow-changes-modify-clip.ui"
...
1/1 gtk:reftest / reftest box-shadow-changes-modify-clip.ui ERROR 0.77s
``
Fix this by re-inhibiting if we didn't draw anything, or we will get an assert the next paint.
gtk_css_node_ensure_style() recurses over previous siblings to ensure
these have a style before its following sibling. As seen in
https://gitlab.gnome.org/GNOME/gtk/-/merge_requests/2027 this can
cause us to stack overflow and crash if we have a lot of children.
And even if we don't have *that* many children its still somewhat
bad to have stack depths of the same magnitude as the number of
children, both for performance reasons and debuggability.
This is a demo that measures performance, so keep the child
info in a hashtable instead of a list. This means adding or removing
a child is not O(n-children).
Rename the show-separators property to show-row-separators,
and add a matching show-column-separators property. It is
implemented by setting the .column-separators style class
on the column view.
This is an expected feature with rubberband selection:
as you get close to the edge while doing rubberband
selection, the list scrolls to extend your selection.
Implement the typical rubberband selection, including
autoscroll. This is only useful with multiselection,
and not very compatible with single-click-activate.
Therefore, it is not enabled by default, and needs
to be turned on explicitly.
Autoscroll when the pointer gets close to the
edge during column resizing or reordering. This
is similar to what the treeview does, but it is
implemented using a tick callback, and has
variable speed.
Allow rearranging columns by dragging, in the same
way the treeview does.
We add the "dnd" style class to the header while
it is dragged, and we move the header of the dragged
column to the end of its parents children, so that
it gets drawn on top.
Add helper functions that let us temporarily give
a different allocation to headers. These will be
used to implement interactive column reordering
in GtkColumnView.
The listview inside always thinks it gets its full size,
and updates its horizontal adjustment accordingly.
So keep our own adjustment, and update it when allocating.
Tweak the behavior slightly. We don't show
a scrollbar as long as we have at least
min-size available, but we still give the
entire size to the child, up to nat-size.
This matches how viewports handle scroll-minimum.
stat/fstat aren't compatible with GStatBuf on Windows in all cases.
Since we don't really need the file descriptor anyway here replace it with
a g_stat call.
This fixes a compiler warning with 64bit mingw.
And lets us remove this scary ifdeffery.
Since it's a type with sub-classes, we need to use GTypeInstance (at the
very least), otherwise we won't be able to address each sub-class as
such.
This is similar to how GskRenderNode and GdkEvent are handled, with the
added difficulty that GtkExpression is meant to be used in properties,
in order to be deserialised by GtkBuilder. This requires adding a
GParamSpec sub-class that we can match on from within GtkBuilder,
alongside some convenience API for storing a GtkExpression inside a
GValue.
We're printing out the file we're testing once we succeed, but it's hard
to know which file caused a failure. Let's add a g_test_message()
directive so we can look in our logs.
The print backends do some complicated dispose handling
where the implementations call gtk_print_backend_destroy().
Our tests (in particular, the templates test) trigger
situations where we use print backends after dispose,
and they can't handle the printers listmodel being
NULL at that time. So just remove the printers in
dispose, keep the empty liststore until finalize.
Scroll events do not have a position, so they shouldn't implement the
GdkEventClass.get_position() virtual function; nor they should have an x
and y fields that never get updated.
When we run the frameclock RUN_FLUSH_IDLE idle before the paint,
then gdk_frame_clock_flush_idle() sets
```
priv->phase = GDK_FRAME_CLOCK_PHASE_BEFORE_PAINT
```
at the end if there is a paint comming.
But, before doing the paint cycle it may handle other X events, and
during that time the phase is set to BEFORE_PAINT. This means that the
current check on whether we're inside a paint is wrong:
```
if (priv->phase != GDK_FRAME_CLOCK_PHASE_NONE &&
priv->phase != GDK_FRAME_CLOCK_PHASE_FLUSH_EVENTS)
return priv->smoothed_frame_time_base;
```
This caused us to sometimes use this smoothed_frame_time_base even
though we previously reported a later value during PHASE_NONE, thus
being non-monotonic.
We can't just additionally check for the BEGIN_PAINT phase though,
becasue if we are in the paint loop actually doing that phase we
should use the time base. Instead we check for `!(BEFORE_PAINT &&
in_paint_idle)`.
There is no agreement that a coverflow widget is
appropriate for GTK 4.
It would be ok as a demo if it could live in gtk-demo,
but that requires us to make GtkListBase public first.
The demo is also somewhat rough and needs more work
to look plausible.
Drop GtkCoverFlow and the related demo for now.
We require a C compiler supporting C99 now. The main purpose of
these fallbacks was for MSVC. From what I can see this is now all supported
by MSVC 2015+ anyway.
The only other change this includes is to replace isnanf() with the
(type infering) C99 isnan() macro, because MSVC doesn't provide isnanf().
currently when mouse clicking on a column header
to sort it it is grabbing keyboard focus, this
should not happen, keyboard focus should remain
where it was before. This can be seen on the
GtkFileChooser widget, when having the keyboard
focus on the file list items and clicking on a
column header to sort it the keyboard focus is
now on the header.
At least from GTK v3.22.30, the code disallows attempts to print PDF and
PS files using the LPR backend. Although it is not easy or possible for
GTK to determine if the printer can print these formats by querying lpr,
the print backend should allow lpr to try.
There are two uses cases:
- A filter is being used by lpr, specified in the printcap file, to process the
files prior to printing them.
- The printer can print PDF and PS files directly. There are many printers that
can print these formats, so GTK should allow for the possibility.
2018-11-17 13:04:35 -05:00
4112 changed files with 604350 additions and 541346 deletions
<property name="label">“Copy” will copy the selected data the clipboard, “Paste” will show the current clipboard contents. You can also drag the data to the bottom.</property>
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
