They were never referenced outside the classes' ctor and, worse,
remained uninitialized if the specification string contained devices.
Change-Id: I977a156acf10190428da00fe128fee70cff8f98d
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
All four manager classes contained roughly the same code in their
ctors that parsed out devices from a colon-separated string.
Extract shared code, and port the parsing to QStringRef (later to be
ported to QStringView).
Saves ~2.4KiB on optimized Linux GCC 9.1 AMD64 builds across all
.so's that link to libQtInputSupport.a.
Change-Id: I3db826ee2b422cfc02f8d49bd21985a03b6c0935
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Make create() return, and m_mice/m_keyboards/etc store, handlers by unique_ptr.
In most cases, we can't use qt_make_unique(), since the ctor we're calling is
marked as private.
Since QHash can't hold move-only types, use a std::vector<{QString, unique_ptr}>
instead. As this pattern repeats in all four QEvdev*Manager classes, create a
small class template.
Saves almost 6KiB on optimized Linux AMD64 GCC 9.1 builds across all .so's that
link to QtInputSupport.a.
Change-Id: I8f62b6b629d6e1855314c0a4fb4fc069db9ae0ce
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The code is noisy and repeats, so wrap it in a function.
Change-Id: I5e6e924e22b0bc631eb8176de96c49066b1c9029
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Also use qUtf16Printable() and qErrnoWarning (removing explicit errno, where
present).
Saves 6.6KiB in text size on optimized Linux AMD64 GCC 9.1 build across
all .so's that link to QtInputSupport.a.
Change-Id: I1def2cfabd2eed65390099cd1d06f8061a9355be
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Extra difficulty: the lock is conditional. Not a problem with std::unique_lock, which is movable.
Change-Id: Ib5515838ccb10100d5aa31163ab7f171591c04c4
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
This is required to get Qt EGLFS in qemu working, as qemu mouse
device issues only absolute mouse events.
Fixes: QTBUG-69172
Change-Id: Ia9985dc01d7871fdcd856c14160d54d3ce684396
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
This automatically searches for some default devices. This not only
reduces the amount of code, it also makes sure that Qt will by default
find the same input devices that tslib (e.g. ts_calibrate) will find.
Fixes: QTBUG-74680
Change-Id: Ied3fff7e8ed9d6637d2b8cf32119660133d251a3
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
The release coordinates may be simply initialized to 0 in lower levels of the
tslib. They are then passed through the calibration stage, which applies
offset and multiplication to them, so they may end up even as positive numbers.
Since we can't know if we can trust them or not simply ignore all release
coordinates and just always use the ones from the previous event.
Change-Id: Ib845b5ab1c5b81967cc089d601576893f433fa4a
Reviewed-by: Martin Kepplinger <martink@posteo.de>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
... because API from qxkbcommon_p.h includes xkbcommon headers.
When we use "QT += xkbcommon_support-private" in *.pro files, we
should not explicitly require "QMAKE_USE += xkbcommon" in those
projects.
Change-Id: I21049034ce93bee13a1107723f26498c221f8ea4
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Our implementation of compose table parser was added on Mar, 2013.
libxkbcommon added APIs for the same thing in Oct, 2014 (ver: 0.5.0).
After removing RHEL 6.6 from the list of supported platforms we were
able to move the minimal required libxkbcommon version to 0.5.0. Now
we can use the xkbcommon-compose APIs on all supported platforms.
With this patch we can drop nearly 1000 lines of maintenance burden.
This patch fixes user reported issues with our implementation.
Known issues:
- Testing revealed that xkbcommon-compose does not support non-utf8 locales,
and that is by design - https://github.com/xkbcommon/libxkbcommon/issues/76
Our implementation did work for those locales too, but it is unclear
if anyone actually uses non-utf8 locales. It is a corner case (work-arounds
existing) and likely a configuration error on the users' system.
- Looking at the release notes for versions above 0.6.1, only one issue
that stands out. Compose input does not work on system with tr_TR.UTF-8
locale, fixed in 0.7.1. Compose input works fine when using e.g. en_US.UTF-8
locale with Turkish keyboard layout.
Note:
With Qt 5.13 we have removed Ubuntu 16.04 and openSUSE 42.3 from CI:
Ubuntu 16.04 - 0.5.0
openSUSE 42.3 - 0.6.1
CI for Qt 5.13 has:
Ubuntu 18.04 - 0.8.0
RHEL-7.4 - 0.7.1
openSUSE 15.0 - 0.8.1
Currently the minimal required libxkbcommon version in src/gui/configure.json
is set to 0.5.0, but we could bump it to 0.7.1 to avoid known issues from above,
but that is a decision for a separate patch.
[ChangeLog][plugins][platforminputcontexts] Now using libxkbcommon-compose
APIs for compose key input, instead of Qt's own implementation.
Fixes: QTBUG-42181
Fixes: QTBUG-53663
Fixes: QTBUG-48657
Change-Id: I79aafe2bc601293844066e7e5f5eddd3719c6bba
Reviewed-by: Giulio Camuffo <giulio.camuffo@kdab.com>
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
xcb/eglfs/wayland - all use XKB keyboard configs and APIs. There
is a lot of duplicated and naturally a diverging code. This patch
adds a helper library to avoid all the mentioned problems and unify
feature set between these platforms.
qlibinputkeyboard:
Added a fixup for 2803cdf758. From
spec: "keysyms, when bound to modifiers, affect the rules [..]",
meaning we can't look at keys in isolation, but have to check if
bounding exists in the keymap. This is done by using
xkb_state_mod_name_is_active() API, but that API has its limitations -
https://github.com/xkbcommon/libxkbcommon/issues/88
I will fix this separately in the LTS (5.12) branch.
We need to read the modifier state before the key action. This
patch fixes a regression introduced by aforementioned patch, which
caused modifiers being reported wrongly in QKeyEvent::modifiers().
qtwayland:
Moved toKeysym(QKeyEvent) from qtwayland repository into this library.
For this and other key mapping functionality wayland was duplicating
the key table. All of that will be removed from qtwayland, and calls
will be replaced to use this lib. Adjusted toKeysym() to fix QTBUG-71301.
Qt keys don't map to ASCII codes, so first we need search in our key
table, instead of mapping from unicode.
lookupStringNoKeysymTransformations():
fixed off-by-one error, where we were including terminating NUL in
QString.
Fixes: QTBUG-71301
Task-number: QTBUG-65503
Change-Id: Idfddea5b34ad620235dc08c0b9e5a0669111821a
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
I need to change keymap under platforms eglfs and linuxfb (without wayland).
I use the function QEglFSFunctions::loadKeymap(const QString&), but there's
no way to switch between english and another language than to press AltGr as
a modifier. I added the function that allows to change the language. And also
added the ability to switch the keymap and language for the platform linuxfb
and also added the ability to switch the keymap and language for the platform linuxfb
Task-number: QTBUG-72452
Change-Id: I37432cf60d375555bea2bf668ec1387322b4964f
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
The only reason why we bundled this library ~6 years ago was because
it was not available on distributions that we supported at the time,
but library was a hard dependency for XCB plugin. See:
2122e731ab
Later more and more projects started to depend on it (compose input
context plugin, libinput, mir, wayland). The configuration had become
too complex, because some projects used bundled and some used the
version from the system.
Having libxkbcommon in 3rdparty sources is not necessary anymore, after
RHEL 6.6 was removed from the list of supported platforms for Qt 5.12.
Ubuntu 16.04 - 0.5.0
Ubuntu 18.04 - 0.8.0
openSUSE 42.3 - 0.6.1
RHEL-7.4 - 0.7.1
This will also simplify further development, e.g. QTBUG-42181
Bumped the minimal required version 0.4.1 -> 0.5.0.
The patch also contains a code marked with "TRANSITION HACK", which
is temporary needed so we can update the dependent wayland module.
[ChangeLog][Third-Party Code] Removed xkbcommon from bundled sources.
This library is present on all supported platforms. The minimal required
version now is 0.5.0.
Task-number: QTBUG-65503
Change-Id: Iec50829bb6f8fbb19f3c4e4ad62e332beb837de5
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Also clean up QTextCodec usage in qmake build and some includes
of qtextcodec.h.
Change-Id: I0475b82690024054add4e85a8724c8ea3adcf62a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Newer mouse devices support a so-called "free-scroll" mode, which
unlocks the wheel and allows for faster scrolling. For such input
devices, evdev will report an event value greater than 1, when free
scrolling is active. Examples for such devices would be the Logitech
G700 or MX Master series.
However QEvdevMouseHandler interpreted such an event's data incorrectly
and triggered wheel events for the opposite scrolling direction. This
also often resulted in jittery & jumpy scrolling, when the events'
values alternated between 1 and 2.
Change-Id: Ibb23ed4419d647fff9b90d371d5bb4037cf2bd9b
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
If a device fails to open during the plugin startup, we exit the handler
constructor early and leave the d member as nullptr.
However the recently added m_handler->isFiltered() call after m_handler
is instantiated assumes that d is always valid, which triggers a crash
in the forementionned situation. Fix it to check for d's validity first.
This can occur when a device is connected then disconnected right after,
so that it's gone by the time we get the notification from udev.
Change-Id: Ia755868338f92b91c181be8557e06e087d70fcc6
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Modifier keys are saved in qinputdevicemanager such way that both evdev
and libinput can use it the same way, it is also handling the repeating
modifier key events. Evdev support is important for VxWorks support
because it is using it.
Task-number: QTBUG-60694
Change-Id: I49038cb7fe2ad5134b3a37167c19953867ea31c3
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Qt was missing symbols for many dead keys defined in <X11/keysymdef.h>.
These dead keys were thus ignored by the "compose" input module. This
commit adds the missing dead key symbols.
[ChangeLog][Linux/XCB] Added missing dead key symbols, enabling their
use with the "compose" input module.
Task-number: QTBUG-56452
Change-Id: Ib5c37168990c9d9fa99fdd50f63b934c793e8dc4
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
Reviewed-by: Matt Whitlock <qt@mattwhitlock.name>
... and deprecate QWSI APIs that accepts mouse event without mouse button/
type data.
In the early days of Qt5 it was decided to centralize mouse button/type
handling in QGuiApplication (because of limitation of some now unknown
platform). This has proven to be problematic as mouse handling details
differ across platforms (e.g on X11 we do not receive mouse release event
when closing popup windows or ordinary windows that are closed from the
mouse press event). Instead of hacking around platform specific behaviors
in Qt Gui, we should move this task back to platform plugins (similar to
how this was done in Qt4 with native APIs sending mouse details directly
to QApplication). There are even cases where it simply is not possible
to deduce (from QGuiApplication) which button caused the event (e.g. when
more than one button is involved and some event goes missing). Besisdes,
throwing away information which is already available at QPA level (for free)
and trying to deduce it again at Qt Gui level seems impractical, fagile
(as probably noticed by people fixing all the unexpected issues) and adds
unnecessary complexity.
Note:
Removing the deprecated QWSI APIs from offscreen plugin depends on fixing
autotests that rely on QOffscreenCursor::setPos() logic.
For the convenience of testing use QT_QPA_DISABLE_ENHANCED_MOUSE to restore
to the old code path where QGuiApplication does the mouse state deducing.
Other platforms have similar issues. I do not have all supported platform
available on my desk, so other platform maintainers will need to take care
of porting those platforms to the new APIs. And mainly, I don't want to deal
with all the hacks that other platforms have added to workaround this broken
mouse logic.
In Qt6 we need to remove deprecated code path from QGuiApplication.
This patch:
- Extends QWindowSystemInterfacePrivate::MouseEvent ctor with QEvent::Type
and Qt::MouseButton. We use this extra data when processing mouse events in
QGuiApplication. This actually is similar to KeyEvent, where we do pass the
type (press or release) to QtGui.
- Refactors QGuiApplicationPrivate::processMouseEvent and qtestlib to use
the new APIs.
Task-number: QTBUG-59277
Task-number: QTBUG-62329
Task-number: QTBUG-63467
Change-Id: If94fd46a7cccfea8264dcb1368804c73334558b8
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
Remaining uses of Q_NULLPTR are in:
src/corelib/global/qcompilerdetection.h
(definition and documentation of Q_NULLPTR)
tests/manual/qcursor/qcursorhighdpi/main.cpp
(a test executable compilable both under Qt4 and Qt5)
Change-Id: If6b074d91486e9b784138f4514f5c6d072acda9a
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Remaining uses of Q_DECL_OVERRIDE are in:
src/corelib/global/qcompilerdetection.h
src/corelib/global/qglobal.cpp
doc/global/qt-cpp-defines.qdocconf
(definition and documentation of Q_DECL_OVERRIDE)
tests/manual/qcursor/qcursorhighdpi/main.cpp
(a test executable compilable both under Qt4 and Qt5)
Change-Id: Ib9b05d829add69e98a86238274b6a1fcb19b49ba
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Default scroll increment on other platforms (e.g. XCB, Windows) is 120.
With libinput it was 15 * 120 = 1800, which results in non-smooth scolling
experience. This patch also replaces deprecated versions of
QWindowSystemInterface::handleWheelEvent().
Change-Id: I363f13a2922fd871a93dbd1bd611778fa18f6122
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
I guess this signal pattern was copy-pasted from evdev input integration
code, where we use similar signals in combination with QDeviceDiscovery.
In libinput, device notifications are delivered via libinput APIs, so
these signals are not necessary. In evdev we do some logging from
slots that are connected to these signals. In libinput this is handled
via libinput_log_set_handler. As we can see, we can simply drop this code.
Change-Id: I50b519784e89b59ceeb3cf296780705fe98a8cbf
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
This results in less boilerplate code, among other benefits that
come with functor-based connections. Simple expressions have been
converted to use lambda.
Change-Id: I6887980524027eada24beed95e6f9ba43f0fc8d5
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Marc Mutz
Liang Qi
Alexander Kanavin
Rolf Eike Beer
Qt Forward Merge Bot
Allan Sandfeld Jensen
Friedemann Kleint
Gatis Paeglis
Kevin Funk
Robert Loehning
Elena Zaretskaya
Christian Muehlhaeuser
Romain Pokrzywka
Karim Pinter
Ville Voutilainen
Jarek Kobus
Matt Whitlock
Lars Knoll
Laszlo Agocs
Jake Petroules