This allows to toss a large amount of custom rendering code.
Furthermore, it removes the build-time dependency on Plasma Framework
for FrameSvg and Theme from KWin core as it's pulled in through QML
imports now.
It also cleans up the API and removes functions that are effectively
unused or no-op after this change.
For instance, effects often destroy their effect frames
in pre/postPaintScreen, which would now destroy an `OffscreenQuickView`,
which changes GL context. This is alleviated by delaying detruction
of the internal view.
Support for the features of text cross-fade and selection frame,
which are not used by any of the built-in effects, is dropped.
Signed-off-by: Eike Hein <eike.hein@mbition.io>
The lanczos filter depends on the effect system. It makes very difficult
changing painting code from SceneWindow to Item.
Given that the last big users of the laczos filter - present windows and
desktop grid effects were re-written in QML. The only remaining user of
the lanczos filter is the thumbnail aside effect. Given that it's a
really obscure usecase, switching to the linear filter won't be very
noticeable.
As a backup plan, one can reimplement the thumbnailaside effect using
QML. The lanczos filter is already implemented in plasma-framework.
The first time the GBM backend's EGL context is made current after
creation, both the read and draw surfaces are set to EGL_NO_SURFACE.
This will set the GL read and draw buffers to GL_NONE in accordance with
the EGL spec.
When a real surface is later made current, however, the spec is arguably
unclear on whether the read and draw buffers should remain set to
GL_NONE or whether they should be restored to the default GL_BACK. The
Mesa driver does the latter, the NVIDIA driver does the former.
To work around this difference, Kwin has an explicit call to
glDrawBuffer in GbmSurface::makeContextCurrent. It does not have a
corresponding call to glReadBuffer, though, which can cause some desktop
effects such as background contrast to render incorrectly with the
NVIDIA driver. This change adds that missing call.
Workspace::findAbstractClient(QUuid) overload is used only by the
windows krunner plugin.
This change fixes a potential null dereferencing and removes the
findAbstractClient() overload in favor of a more generic findToplevel()
overload.
Releasing the buffers is necessary for example in the case of a GPU reset,
to make sure that the gbm surface is still properly destroyed and all buffers
with invalid content freed.
Instead of buffers being both drm framebuffers and gbm / dumb buffers, these
responsibilities are now split, which makes it possible to do zero copy
screen casting in the future.
Both framebuffers and gbm / dumb buffers also now always hold a valid underlying
resource, which simplifies code a bit.
since it has been replaced by windowview, and Desktop Grid
is ported as well, remove present windows which is effectively dead code
now
BUG:447001
BUG:362844
BUG:450487
BUG:453426
BUG:374481
BUG:185381
BUG:413342
BUG:451150
BUG:283333
BUG:315314
BUG:397500
BUG:321236
BUG:436572
BUG:335782
* give windowview the ability to show windows from current desktop, all
desktops or current class
* invokable from shortcuts screen edges or gestures
* add a search field to quickly filter
* current present windows effect still present but only internal as is used by desktop grid, which should eventually be ported as well
This can be done either from windowview or overview, tough note that some of the duplication from overview added in windowview is necessary regardless, as WindowHeap, the shared qml part needs some properties exposed from the c++ part.
Implementation wise I'm ok for it being in either place, but i think the functionality of present windows needs to be preserved, otherwise would be a completely unacceptable regression, namely:
* Behavior of what now are the ctrl+f7,f10 and f11 global shortcuts, showing windows of current app (invokable by shoortcut instead of having to use the pointer), windows of current desktop (the only thing overview does atm) or all desktops
* filter on typing, as opposed to invoking krunner. main use case of present windows is quickly switching, and filtering is the most helpful feature, some people do like krunner instead, but is completely out of place for my use case, and i suspect for many other users as well
* also clicking on a taskbar group should be possible to filter
* the view that opens by clicking on a taskbar group should follow the same layout strategy used elsewhere and currently being an effect used only there and not accessible form the list can't even be configured
Since WindowItem::visible is kept in sync with the effective visible
status of the window, window items that are not lockscreen greeter or
input methods can be hidden when the lock state changes.
With this, the WindowItem will know whether it's actually visible. As
the result, if a native wayland window has been minimized, kwin won't
try to schedule a new frame if just a frame callback has been committed.
EffectWindow::enablePainting() and EffectWindow::disablePainting() act
as a stone in the shoe. They have the final say whether the given window
is visible and they are invoked too late in the rendering process.
WindowItem needs to know whether the window is visible in advance,
before compositing starts.
This change replaces EffectWindow::enablePainting() and
EffectWindow::disablePainting() with EffectWindow::refVisible() and
EffectWindow::unrefVisible(). If an effect calls the refVisible()
function, the window will be kept visible regardless of its state. It
should be called when a window is minimized or closed, etc. If an effect
doesn't want to paint a window, it should not call effects->paintWindow().
EffectWindow::refVisible() doesn't replace EffectWindow::refWindow() but
supplements it. refVisible() only ensures that a window will be kept
visible while refWindow() ensures that the window won't be destroyed
until the effect is done with it.
Item::isVisible() is true if either the item has been marked hidden or
one of its ancestors.
In some cases, kwin may render invisible windows, for example for window
thumbnails.
This change makes rendering code use explicit visibility status when
rendering to ensure that it's still possible to render invisible windows.
If an effect is reloaded while it holds deleted references, it's
possible that the closed windows will get stuck in the "zombie" state.
This change introduces EffectWindowDeletedRef helper that can be
used to keep the closed window alive as long as the reference is valid.
If you lift fingers but not swipe them enough to switch to another
virtual desktop, the slide effect will play an animation to move from
the current position in the virtual desktop grid to the current desktop.
However, that animation doesn't feel right, there's something missing.
The slide effect uses a TimeLine to animate switching between virtual
desktops, it's great if the amount of sliding is constant.
This change makes the slide effect use the mass-spring-damper model to
simulate the motion of a spring in order to animate switching between
virtual desktops.
The mass-spring-damper equation is integrated using RK4. If the delta
interval is not multiple of the integration step precisely, the
SpringMotion will perform integration as many times as the integration
step fits into the delta. The leftover will be used for LERP between the
previous and the next integration results.
With the spring animation, the slide animation feels more natural when
you lift fingers. If you switch between virtual desktops without using a
gesture, the slide animation should look almost the same as if it were
implemented with the TimeLine.
It helps to contextualise the method as it's using several x11-isms,
some of them possible ot abstract.
In any case, the method is only called with X11Window and it's the only
case where it makes sense doing so.
The scripting API is extended to support custom fragment shaders. To
support this a new method addFragmentShader is added taking ShaderTraits
and fragment shader. The GLShader is not exposed, instead a uint id is
provided which maps to the GLShader.
This shader id can be used in the animate and set calls to specify the
shader. The animation object is extended by the "fragmentShader" property.
The shader sources are located in the "shaders" directory of the package
contents.
E.g. the scale effect extended by the shader of the invert effect will
have the following layout:
package/contents/
-> code/
-> main.js
-> shaders/
-> invert_core.frag
-> invert.frag
The adjustment in code are:
* in constructor to load the shader
this.shader = effect.addFragmentShader(Effect.MapTexture, "invert.frag");
* in animations objects of the slots the additions
fragmentShader: this.shader
* using the type Effect.Shader
or in full:
window.scaleInAnimation = animate({
window: window,
curve: QEasingCurve.OutCubic,
duration: this.duration,
animations: [
{
type: Effect.Scale,
from: this.inScale
},
{
type: Effect.Opacity,
from: 0
},
{
type: Effect.Shader,
fragmentShader: this.shader
}
]
});
The animation settings object supports a "uniform" value which takes the
string name of the uniform. For this uniform the location is resolved
and stored in the meta data of the AnimationEffect. This requires the
type Effect.ShaderUniform.
An example animation:
window.scaleInAnimation = animate({
window: window,
curve: QEasingCurve.Linear,
duration: this.duration,
animations: [
{
type: Effect.ShaderUniform,
fragmentShader: this.shader,
uniform: "uForOpening",
from: 1.0,
to: 1.0
}
]
});
Furthermore a new setUniform scriptable method is added to the
ScriptedEffect. This allows to update uniforms when the configuration
changes.
The call takes a generic QJSValue which supports:
* float
* array of 2, 3 or 4 components
* string as color
* variant as color
An example usage to read a color from the configuration and set it as a
uniform:
effect.setUniform(this.shaderId,
"uEffectColor",
effect.readConfig("Color", "white"))
The animate and set calls are extended for an optional GLShader* to
allow specifying a custom shader to use during the animation.
To properly support rendering a complete window in the effect the
AnimationEffect gets based on the DeformEffect. If a shader is used
during the animation the window gets redirected.
For the animation with shaders two new enum values are added to the
AnimationType enum:
* Shader
* ShaderUniform
The Shader animation type is for specifying that the animation uses a
shader. During the animation a uniform "animationProgress" is set on the
shader.
The ShaderUniform animation type behaves exactly like the Shader type,
but also animates a user provided uniform. The meta data of the
animation is interpreted as a uniform location for a float uniform and
during the animation this uniform is updated with the interpolated
animation data.
For a redirected window a custom shader can be specified to draw the
redirected texture to screen. This is useful for inheriting effects to
customize the rendering.
Two new int uniforms TextureWidth and TextureHeight are added and set
from DeformEffect when rendering the texture.
Currently, fullscreen geometry restore is computed from maximized
geometry restore. However, the latter is set only when the window is
maximized.
Also, updateGeometryRestoresForFullscreen() can be called when the
window has not been moved. Avoid updating geometry restore if the output
has not been changed.
If the animation reaches the end, desktop grid may render the screen
incorrectly. Make sure that PAINT_SCREEN_BACKGROUND_FIRST and flags as
such are set even if animation has reached the end.
Also, while on this, simplify the paintWindow() method by removing
redundant effect status checks.
