The blinking shader is adjusted to use a shader trait vertex shader
and gets generated with a shader trait variant.
Overall the code is simplified to ensure that we always have a shader
bound with the correct mvp matrix when rendering the icon.
In addition it's required to keep the expandedGeometry alive until
the effects handled the deletion
BUG: 318322
BUG: 320892
BUG: 344359
REVIEW: 126323
FIXED-IN: 5.6
Most of those effects didn't need special screen locking handling on
X11 as they prevented screen locking. On Wayland though, we don't
want the lock screen shown on a cube.
REVIEW: 126122
the stackingOrderChanged() signal came before the window turned an effect window
usually this is no problem as the change shall not be caught anyway, but
the window may have changed its stack position
BUG: 353745
FIXED-IN: 5.5
REVIEW: 126134
window wasn't correctly represented in
zoom, present windows and desktop grid effects
on the fly code streamlining:
- logout and above windows were handled equally in different code paths
- windows were painted on 0 opacity instead of just aborting the paint
local semi-paintdata because PaintData isn't assignable
(const dptr, I think) and we're post freeze.
BUG: 355539
FIXED-IN: 5.5
REVIEW: 126133
Eg. if a window is closed while "timeline.currentValue() == 0",
slotWindowClosed() will shortcut exit and not be removed from the
managers, thus dangle around forever.
Maybe there're other ways for a window to be deleted w/o
the closed slot being entered.
CCBUG: 339970
REVIEW: 126034
stupid clients think it's relevant to withdraw other states
when going fullscreen, this means we trigger an
nmaximization animation and when that's done, the
window is fullscreen...
Unfortunately one of the stupid clients is QWidget and also
the window could get other, *legit*, resizes during the
animation, so we cancel it on occasion
BUG: 336467
FIXED-IN: 5.5
REVIEW: 125989
For the functions from GL_FOO_robustness we want to resolve it by
ourselves in order to add a custom implementation if it's not available.
Unfortunately once epoxy.h is included this breaks as epoxy defines the
names and so through the preprocessor epoxy always wins.
So we need different names: all functions from robustness get a "kwin"
prefix and the usage is changed everywhere in kwin source code.
REVIEW: 125883
While the main plasma desktop containment does one
desktop per screen, that does not necessarily hold
for other containments, let alone other desktop shells
To indicate that the window is "active" ie. can
be dragged or activated (like in present windows)
There was either a bug or a forum post complaining
about the inability to activate windows from DG
They're only traversed and QHash is unordered - so
the worst container. Also we have complete control
over the maintained class, so we can just keep the
mapped EffectWindow there
esp. the list-remove icon looks like "delete" and
can cause worries about what happens to the windows
as a bonus, the buttons now follow the DPI. are a
little smaller (presently might fit touch devices? but
is way to huge on "normal" desktops with ordinary DPI)
and the buttons have nice visible animations on pressing
BUG: 354131
FIXED-IN: 5.5
Removes the ElectricAction for ShowDashboard and the relevant code
in screenedges and the KCM.
Also a leftover in the glide effect.
REVIEW: 125700
BUG: 353928
Basically dead code given that Plasma doesn't have a dedicated dashboard
mode anymore and also doesn't set the required window role for it to
work.
By deleting we save one string comparison for each newly opened window.
REVIEW: 125686
Following the approach how it's done for X11: only create the interface
if the shader succeeded to compile and remove support again if it failed
to compile after a reconfigure.
REVIEW: 125444
before, hiding the buttons relied on the effect
seeing a paint event for the (with a low timeline
value) what's too wonky at least for instant animations
(certain failure) or when effect exits immediatey
(due to screen edge invocation issues or whatever)
BUG: 351869
FIXED-IN: 5.4.2
REVIEW: 124970
Theory:
----------
because PresentWindowsEffect::screenCountChanged() is shortcut
for "if (!isActive())", but the desktopgrid doesn't call
PresentWindowsEffect::setActive (or at least
PresentWindowsEffect::screenCountChanged), so the effect can
"miss" the increasing screen count change (it sees the signal,
but ignores it) and when desktopgrid calls it, it assumes the
m_gridSizes array is big enough (but it isn't)
Steps:
----------
1. effects are loaded, 1 screen present
2. 2nd screen gets added, but inactive effects ignore that
3. desktop grid gets activated, updates according to screen count,
calls presentwindows for screen #2
4. presentwindows data is only prepared for one screen from step 1
=> BOOM
BUG: 351724
CCBUG: 326032
FIXED-IN: 5.4.2
REVIEW: 124960
the (old) button effect windows used to be unreferenced
with the re-invocation of the effect. because we stopped
deleting/recreating the window, this approach failed and
the effect window was never deleted.
Unreferencing the window at proper occasion (see new hide
location) coked up - guess what - the exact same "texture is
junk" issue as for the QtQuick close button in present windows...
So we resort to the exact same stupid "hide by moving" solution
as we have there.
BUG: 348577
FIXED-IN: 5.4
REVIEW: 124136
Without the componentDisplayName the shortcut dialog takes the name of
the application e.g. Systemsettings or "KDE Control Module", but we want
it to be KWin.
REVIEW: 124706
If the user has only maximized windows or all
windows crammed into one region of the screen
the effect would not resemble an aperture at all
this ensures a maximum amount of corners is used
while seeking to still use shortest paths
BUG: 348709
FIXED-IN: 5.4
a) close to pointless
b) the target resolution texture is invalid if the resize effect is enabled (as the window was not actually resized)
REVIEW: 123901
transformed screens prevent hardware clipping
leading through filtering subquads
This requires to subquad in the painting pass
what is not supported for non rectangular
qauds ands presently aborts (in debug builds)
if any quad is transformed (as by eg. the wobbling)
REVIEW: 123733
BUG: 338972
Same issue as with present windows closers: qml is slow
Interestingly "hide" means "hide" here and there're also
no visual artifacts.
BUG: 347339
REVIEW: 123668
transformed screens prevent hardware clipping
leading through filtering subquads
This requires to subquad in the painting pass
what is not supported for non rectangular
qauds ands presently aborts (in debug builds)
if any quad is transformed (as by eg. the wobbling)
REVIEW: 123733
BUG: 338972
Same issue as with present windows closers: qml is slow
Interestingly "hide" means "hide" here and there're also
no visual artifacts.
BUG: 347339
REVIEW: 123668
This is now crucial, because while before
(the minimized) windows were conditionally
shown, but are now always behind the desktop.
Also, it makes the tabbox more consistent.
BUG: 344083
REVIEW: 122679
The closeview is not hidden because that causes
invalid textures (apparently QML related) and
deleting/recreating causes massive I/O overhead
on effect invocation.
As last resort, the window is "hidden" by moving
it out of the root window geometry.
Jonathan, please RESPIN kwin - sorry for the trouble
... again.
BUG: 345159
REVIEW: 123035
CC: riddell@gmail.com
Moves calculation of projection and modelview matrix
(aligned to new scene_opengl code) to effect start,
caches the matrices in private members and sets them
on WindowPaintData
BUG: 343509
REVIEW: 122355
FIXED-IN: 5.3
This change allows to include the effect_builtins.cpp into the build of
e.g. a kcm without having to link against all effects.
The use case for this is when one needs to resolve just the name of an
effect without wanting to hard code the name.
This is a kind of workaround for the flicker of fading out windows.
When a window is faded out it is a deleted and can by that be used
as a sufficient solution to work around the problem.
BUG: 307112
FIXED-IN: 5.2.0
REVIEW: 121909
Prior to this commit we didn't know if mipmaps were going to be used
when we created the GL texture, which meant that we couldn't tell the
driver whether to allocate storage for mipmaps or not.
This resulted in one of two things happening depending on the driver;
either it would allocate storage for mipmaps that in most cases would
never be used, or it wouldn't and would later be forced to reallocate
the texture when mipmaps were added.
By adding this parameter we can now explicitly tell the driver how
many mipmap levels will be used.
The parameter is only added to the non-image constructor for now. The
image constructor is changed to only allocate a single level, which
matches how textures created from images are used in kwin. This may
need to be revisited in the future.
This code is broken in a number of different ways; firstly by assuming
that the mipmaps need to be regenerated when the texture filter has
changed. Secondly by preventing mipmaps from being specified by other
means.
This commit removes the code from bind() and adds a generateMipmaps()
method instead.
The cube effect was relying on paintWindow() using the generic shader
without explicitly telling it to do so. This makes the cube effect
work with dynamic shaders.
Instead of getting size from displayWidth() and displayHeight() use
the information we have from Screens. This means there is only one
place to have the information and by that we can ensure that all
components use the same data to rely on. displayWidth/displayHeight
seem to provide the wrong information when unplugging an output
without disabling the output. This results in rendering artefacts.
But KWin::Screens has the correct information available.
On MacBooks the "Exposé" button is on XF86LaunchA. Let's use it
for activating Present Windows (ALL) as an alternative shortcut.
For non-macbook users this doesn't hurt, for MacBook users it's a
nice touch to have the key working as indicated by the pictogram.
Yes, it's realy Key_LaunchC:
"On X11 this key is mapped to XF86XK_LaunchA key for legacy reasons."
REVIEW: 118721
Scripts and scripted effects can provide a ui file which gets loaded at
runtime. Obviously such a ui file is not translated. This introduces
quite a hack to load the translated strings.
In the metadata.desktop file the scripted component can provide a key
X-KWin-Config-TranslationDomain. The genericscriptedconfig tries to
locate the metadata.desktop file and reads this entry. If it is present
we extract all string properties of the loaded UI and pass them through
ki18nd.
REVIEW: 118585
geometryChanged is always executed before maximizeChanged, so we
store the last changed geometry (through regular resize - the current
geometry) and the second last one (used exclusively so far) which is
the old geometry we want to use if the geometry change was caused
by the maximize change
CCBUG: 335023
Forward port of 3c8506992d3a05daa3f0752cc0f5e844acf2a8e2 from kde-workspace
DesktopGrid uses PresentWindows Effect as a proxy effect. Instead of
hard coding the name of the Effect let's use the name resolving provided
by BuiltInEffects.
xcbutils.h has quite a few inline only implementations such as Xcb::Atom,
the Wrappers, Xcb::Window and the convenient methods. Thus there is
nothing wrong with using it from the built-in Effects.
Xcb::Atom is used in Glide and Logout Effect to get the atom. To keep the
logic of the existing code it got extended by a bool isValid() which
gets the reply and returns true if the atom is set.
REVIEW: 117587
This removes all the hacks to add kwin4_effect_ to the name of the Effect
and adjusts the desktop files of the effect configuration's parent
component.
Note: the scripted effects still start with kwin4_effect_ prefix.
REVIEW: 117367
All KCMs and KWin core use the BuiltInEffects namespace to find and
interact with the effects. There is no information left in the desktop
file which are of usage. Thus they can be removed.
The EffectData in BuiltinEffects is extended by all the data needed for
the desktop effects KCM:
* display name
* comment
* category
* video-url
* exclusive group
* internal
This information is taken directly from the desktop files.
The Built-in effects are now also resolved through the BuiltInEffects
namespace and the KServiceTypeTrader query is adjusted to only find the
scripted effects.
Unfortunately this introduces another round of adding "kwin4_effect_" to
load and save the effects correctly. This will be removed once all KCMs
are adjusted to use the new BuiltInEffects.
Instead of having several hashes with data about the built-in effect,
we use a struct EffectData which contains the name, the enabled by
default state and function pointers to create, supported and check
enabled by default. There is one static vector with all the data
specified which is ordered by the BuiltInEffect enum. Thus an enum value
can be used as an index to the data.
In addition it's no longer resolved around QByteArray, but uses QString.
REVIEW: 117354
So far the effects could just use the connection() and rootWindow()
provided by kwinglobals. Thus an internal detail from KWin core is
accessed directly.
To be more consistent with the rest of the API it's wrapped through the
EffectsHandler and with a convenient method in Effect.
The connection() is provided as xcbConnection() to free the very generic
name connection which could create confusion once we provide a wayland
connection to the Effects.
The rootWindow() is provided as x11RootWindow() to indicate that it is
for the X11 world.
REVIEW: 117597
If the property changes the animation data is updated. As the last
property change event will definately be delivered before an unmap notify
we can be sure that the animation data is up to data. Reading the
property again does not have any advantages.
By removing it we get rid of at least one roundtrip to the X Server and
it fixes one failed GetProperty error when it tries to read the property
for an already destroyed unmanaged.
REVIEW: 117431
This method replaces the X-KDE-ORDERING property in the Effect's desktop
files. This change is a preparation step for integrating the new Effect
Loader which doesn't read the ordering information. Thus it needs to be
provided by the Effect itself so that the EffectsHandler can properly
insert it into the chain.
Also for the built-in Effects on the long run it doesn't make much sense
to install the desktop files. And binary plugin effects will migrate to
json metadata which also doesn't have the KService::Ptr. Thus overall it
simplifies to read this information directly from the Effect.
For each effect the value from the desktop file is added. This will
allow to query all effects which need to be loaded without looking for
the service files.
The functionality is to check whether the effect should be enabled by
default. It's not about whether the effect is enabled by default. This
is also needed as it's currently still taken from the plugin data.
For each effect there is an enum value. The internal mapping of effect to
create, supported and checkDefault is switched to the enum value, though
the methods for the name based variants are still available.
XRenderUtils are split out of kwineffects and are an own library just
like kwinglutils is an own library.
The library gets always build and is linked in KWin core unconditionally
(as it's used in outline) and conditionally in kwineffects (PaintClipper)
and the built in effects depending on XRender build option.
Affected effects
* Blur and Contrast Shader to get atom name
* Glide effect for the slide atom
* startupfeedback for cursor size - read from config now and cached
* showfps used an xsync, replaced by flush
* logout effect for creating hack items
REVIEW: 116828
Most is just switched to the ::read(). That should be enough for all the
Effects which have a KSharedConfig::Ptr underneath. If not we just need
to find a good place to put the reload.
Instead of using EffectsHandler::sendReloadMessage we generate the dbus
interface in each plugin and call the reconfigure slot directly. That way
it's more type safe and we don't need to link kwineffects from the
configs.
REVIEW: 116875
There are no advantages for the effects KCM to have all the effect
config modules in one plugin.
By having a plugin per effect we can use the KPluginTrader to easily
find the configuration plugin for a given effect and load it.
To make this possible the following changes are done:
* config_builtins.cpp is deleted
* add_subdirectory is used for all effects which have a config module
* toplevel CMakeLists.txt contains the sources again for the effects
which have a config module, but effects which don't have a config
module are still included and thus the macro is still used
* plugin created for the config module, name pattern is:
kwin_effectname_config
* plugin installed to ${PLUGIN_INSTALL_DIR}/kwin/effects/configs
* desktop file adjusted to new plugin name and keyword removed
* desktop file converted to json as meta data and no longer installed
* Uses K_PLUGIN_FACTORY_WITH_JSON
* Macros for config are dropped from kwineffects.h
REVIEW: 116854
Implemented in KWin core to forward to new global shortcut system. This
method should be extended/changed once we go to Qt5/KF5 to make the usage
easier (no more KAction).
Each global shortcut in the effects makes use of this new method.
Painting was already disabled in the effect inside ksmserver, thus it
was more or less dead code. Let's remove it completely.
This also allows to remove the "temporary" hack inside KWin's logout
effect.
REVIEW: 116869
X-KWin-Internal should be used for effects which are either internal
to KWin and are configured at a different location in KWin. Examples
are:
* CoverSwitch - configured in WindowSwitcher KCM, but not FlipSwitch
as that one has a non WindowSwitcher mode
* window geometry - configured in moving
And it should also be used for helper effects to the overall
kde-workspace module. Examples are:
* Dashboard - required by Plasma
* KScreen - required for KScreen
* Screenshot - required by KSnapshot
Why a new category and not the already existing NoShow property?
The idea is to just filter on the effects in the list. Thus it should
be possible to show them. But NoShow is clearly intended to not be
shown at all.
REVIEW: 116754
Lambda connections are tricky: they do not get automatically
disconnected when the captured [this] gets destroyed, but they get
disconnected when the sender object gets destroyed.
Thus the slide effect was crashing if it got disabled and a window
got deleted, because the connection was still present and tried to
access the destroyed this pointer.
To circumvent we need to pass this as context object.
REVIEW: 116712
All effects with
X-KWin-Exclusive-Category=category
are considered as mutual exclusive for specified category. So far this is
just a hint for the GUI representation. Such effects should get a
radio button while effects in the same category without that property
or with it set to false should get a checkbox.
As a first step the desktop change animation effects use this property
and are put into a new category "Virtual Desktop Switching Animation".
REVIEW: 116710
Inspired by the recent slidingpopups patch this aims to make the scalein effect feel a lot quicker.
It only scales the window from 0.75 to 1. Animation duraton is shortened to match the separate "fade" effect.
REVIEW: 115763
KWin already has a de facto OpenGL 2 dependency through QML. Combined
with the fact that the OpenGL 1 backend is basically unmaintained and
also unused, it's better to remove it for the new major release.
This change includes:
* Removal of cmake option KWIN_BUILD_OPENGL_1_COMPOSITING
* Removal of KWIN_HAVE_OPENGL_1 compile option and all code
ifdef'ed with it (partially removal of if-else constructs)
* Removal of CompositingType::OpenGL1Compositing (flags are kept
as a core flag should get introduced)
* Driver recommendation for OpenGL1Compositing changed to XRender
(should be evaluated whether the drivers can provide GL2)
* Removal of configuration option "GLLegacy"
* Removal of fooMatrix function in kwinglutils
* Removal of ARBBlurShader
* Removal of legacy code path in GLVertexBuffer
* Removal of GLShaderManager::disable
* if-blocks with ShaderManager::instance()->isValid() removed
REVIEW: 116042
Rational behind this change is that displayWidth and displayHeight are
X specific API calls in kwinglobals. For the future it's easier to only
rely on functionality which goes through the EffectsHandler API which
allows easier adjustments in KWin core.
displayWidth() and displayHeight() are only used to get the size or the
complete rect of all screens. This is also provided by:
effects->virtualScreenGeometry() or
effects->virtualScreenSize()
REVIEW: 116021
It's no longer needed as there won't be a KWin version specific to
Plasma Active. We want to have runtime adjustments of all parts in
the next version of Plasma.
REVIEW: 116563
instead 1/16 of the area, 1/4 of the bigger and 1/32
of the smaller screen/window ratio
(ie. a shaded window would typically zoom to 1/32 of the screen height)
BUG: 298402
FIXED-IN: 4.11.7
REVIEW: 115455
This seems like a more proper fix for the flickering issue in the
sliding popups effect. The problem is that slidingpopups grabs the
window in windowClosed, the fade effect checks it there, which makes
it racy.
In my tests, I've not seen this problem with the WindowAddedGrab, but
as far as I understand, the problem may well be present there as well.
(And my proposed trick doesn't work.) I've not seen this happening in my
debugging, however. The problem there is also less visible since the
transparency curves go into the same direction, and are more "in line
with each other".
So, fix: Move the setData(WindowClosedGrabRole, ...) call from
windowClosed into windowAdded, which makes sure it's set whenever the
window goes away.
REVIEW:115903
BUG:329991
This fixes the sliding popups losing their contrast effect when
animating, less flicker.
In this patch, we temporarily force the contrast effect on, but only if
it hasn't been explicitely disabled. As soon as the animation stops, the
force flag is disabled again. For disappearing windows, we just set the
flag in the same way, but skip over the bookkeeping, since the window is
going to be deleted, anyway.
REVIEW:115902
Without setting the property, Plasma's panel and dialogs lose the
backgroundcontrast effect during slides, which makes them flicker.
As the panel is shown on screen all the time, this is quite a visible
bug. To fix this, when the slide effect is started, we check for window
types and properties of each window, and force the blur flag on if it's
unset.
If the background contrast flag is set to false, we leave the window
alone assuming that there's a reason to force it off. Windows that
are newly added during the slide get the same treatment, so something
popping up while sliding (such as the desktop switch OSD) also gets
the background effect applied.When the effect stops or is interupted, we
unset what we've set, and clean up our internal bookkeeping.
Thanks Martin and Thomas for the thorough review!
REVIEW:115857
As all effects have always been compiled into the same .so file it's
questionable whether resolving the effects through a library is useful
at all. By linking against the built-in effects we gain the following
advantages:
* don't have to load/unload the KLibrary
* don't have to resolve the create, supported and enabled functions
* no version check required
* no dependency resolving (effects don't use it)
* remove the KWIN_EFFECT macros from the effects
All the effects are now registered in an effects_builtins file which
maps the name to a factory method and supported or enabled by default
methods.
During loading the effects we first check whether there is a built-in
effect by the given name and make a shortcut to create it through that.
If that's not possible the normal plugin loading is used.
Completely unscientific testing [1] showed an improvement of almost 10
msec during loading all the effects I use.
[1] QElapsedTimer around the loading code, start kwin five times, take
average.
REVIEW: 115073
We need to use the varying_in/out variables, the code was a little
bit too modern.
At the same time remove the identity matrix and replace it by mat4(1.0).
Note: the shader should probably go into glsl files as they are not
really generated.
* this effect is way cheaper than blur, don't cache it
* use its own atom
* also pass the matrix in the x property
* remove remnants of the cache
* do just a single pass
* get rid of config ui remnants
* a copy of the blur shader to become a copy of the background contrast effect
* contrastshader actually doing the light modification
* don't expand/shrink the area
It's 2014 and we don't have to wait half a minute for an application
to start. In fact we mostly get false positives due to applications
not handling correctly startup notifications for already running
instances (e.g. click on link in email).
So let's reduce to a default which doesn't look like a broken setup.
REVIEW: 115046
Videos for the following effects are added:
* Dim Inactive
* Dim Screen for Administration Mode
* Invert
* Looking Glass
* Magnifier
* Mouse Click
* Track Mouse
* Zoom
Link for Present Windows video fixed.
Most effects had a "collection" for one action. We don't need the
action collection, all it was used for is setting the object name.
With the removal of KActionCollection the effects do not need to link
XmlGui any more, though the dependency is still pulled in through
plasma.
Client used to have dedicated methods for different icon sizes instead
of combining all pixmaps into one QIcon. This resulted in various parts
of KWin having different access to the icons:
* effects only got one pixmap of size 32x32
* decorations only got the 16x16 and 32x32 pixmaps combined into a QIcon
* tabbox could request all icon sizes, but only as pixmap
Now all sizes are available in one QIcon allowing to easily access the
best fitting icon in a given UI.
With QtQuick2 it's possible that the scene graph rendering context either
lives in an own thread or uses the main GUI thread. In the latter case
it's the same thread as our compositing OpenGL context lives in. This
means our basic assumption that between two rendering passes the context
stays current does not hold.
The code already ensured that before we start a rendering pass the
context is made current, but there are many more possible cases. If we
use OpenGL in areas not triggered by the rendering loop but in response
to other events the context needs to be made current. This includes the
loading and unloading of effects (some effects use OpenGL in the static
effect check, in the ctor and dtor), background loading of texture data,
lazy loading after first usage invoked by shortcut, etc. etc.
To properly handle these cases new methods are added to EffectsHandler
to make the compositing OpenGL context current. These calls delegate down
into the scene. On non-OpenGL scenes they are noop, but on OpenGL they go
into the backend and make the context current. In addition they ensure
that Qt doesn't think that it's QOpenGLContext is current by calling
doneCurrent() on the QOpenGLContext::currentContext(). This unfortunately
causes an additional call to makeCurrent with a null context, but there
is no other way to tell Qt - it doesn't notice when a different context
is made current with low level API calls. In the multi-threaded
architecture this doesn't matter as ::currentContext() returns null.
A short evaluation showed that a transition to QOpenGLContext doesn't
seem feasible. Qt only supports either GLX or EGL while KWin supports
both and when entering the transition phase for Wayland, it would become
extremely tricky if our native platform is X11, but we want a Wayland
EGL context. A future solution might be to have a "KWin-QPA plugin" which
uses either xcb or Wayland and hides everything from Qt.
The API documentation is extended to describe when the effects-framework
ensures that an OpenGL context is current. The effects are changed to
make the context current in cases where it's not guaranteed. This has
been done by looking for creation or deletion of GLTextures and Shaders.
If there are other OpenGL usages outside the rendering loop, ctor/dtor
this needs to be changed, too.
