The functional style interface is nice, but does feel alien in some
contexts, so better also have explicit encode and decode methods.
Change-Id: Ic07ced15f65cdb3a7f1cf044041e341d2ef87f79
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Applied to headers only. Source file to be changed separately.
Omitted statemachine for now to avoid conflicts.
Omitted qmetatype.h for now - to be handled later.
Task-number: QTBUG-84469
Change-Id: I317376037a62467c313467d92955ad0b7473aa97
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Adding support for a static notifier within QProperty itself - through a
QProperty "sister" class - is more efficient in terms of memory
consumption and run-time performance.
The MemberChangeHandler permanently takes up at least three pointers,
while the notified properties only cost one pointer in the binding.
Change-Id: Ia1a8c2b66f1f3c2fe13ae0ad9f12cdb6bdcc35ef
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
A guard callback is a predicate which takes the new value set by
setValue or computed as the result of a binding expression. If it
returns false, the value is discarded and the old value is kept.
Note that due to lazyness, when setting a binding, we still notify
everyone as the binding is only evaluated on demand, and the guard can
thus only run when someone actually queries the value.
Note further that a guard is allowed to modify the value that is passed
to it (e.g. to clamp it to a certain range).
Task-number: QTBUG-85032
Change-Id: I3551e4357fe5780fb75da80bf8be208ec152dc2a
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Check at compile time whether the static callback takes an argument
(which has to be of the same time as the type of the property). If so,
retrieve the old value and pass it to the callback.
Change-Id: Ib1c4c9e05b826b6be492b03f66fa72ad015963ee
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
[ChangeLog][QtCore][QByteArray] Remove method overloads taking
QString as argument, all of which were equivalent to passing the
toUtf8() of the string instead.
Change-Id: I9251733a9b3711153b2faddbbc907672a7cba190
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Add overloads for qMin and friends where the arguments are of different
type, but one can be easily promoted to the other. Return the promoted
type. Promotions are only allowed if both types are either signed,
unsigned or floating point numbers.
This should simplify writing code in many case (as for example
qMin(myint64, 1)) and also help reduce source incompatibilities between
Qt 5 and Qt 6, where the return types for sizes of our containers changes
from int to qsizetype.
Change-Id: Ia6bcf16bef0469ea568063e7c32f532da610d1cd
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Fixes regression introduced in b4981f9d4c,
due to which it was possible to back-tab into a widget even though it or
its focusProxy had a NoFocus policy.
As a drive-by, split the complicated if-statement up a bit for improved
readability.
Change-Id: Ib0ac2604076e812e340b11534c23ae8ae958d082
Fixes: QTBUG-76924
Pick-to: 5.15 5.12
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Breaks the build if any global header has
using P = SomeType;
statement, and unnecessary to do so before the headers are included.
Change-Id: I3b8cc705bf662f768b8895e4b3ec1428ab39ef27
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
When I wrote the QCborValue to QJsonValue conversion, I used
QJsonValue::Undefined because it allowed to keep some level of
compatibility in CBOR, despite the function documentation saying that
CBOR undefineds became JSON nulls. Which they did.
But when we converted QJson{Array,Object} to be backed by CBOR classes,
that Undefined meant the insertion into the array/object actually
deleted the entry.
[ChangeLog][JSON] Fixed a regression from 5.14 that caused values of
default-constructed QVariants in QVariantLists, QVariantMaps and
QVariantHashes to disappear when converting to JSON via
fromVariant{,List,Map,Hash}.
Fixes: QTBUG-84610
Pick-to: 5.15
Change-Id: Ic0987177fe463f352db9bd84993f116e2bdacc75
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We have no doubt it does, because it's compiler-synthesized, but
we might want to implement the tuple protocol for QPair in the
future and then this will act as a safety net, emulating what
users are currently already doing with QPair.
Change-Id: Ie37f0214bb1aa64210d25be8a256606f4572febe
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This (and use of Q_SLOTS for the test slots) makes it possible to
enable QT_NO_KEYWORDS and QT_NO_FOREACH in all the corelib/time/
tests.
Change-Id: I85fd358f3d1a72c9269d5260d0224640c1751f2d
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Use the overload-with-template trick from P1423 to avoid ambiguities
when existing callers pass 0 or nullptr.
Add a qdoc-ignored macro to hide the fact that the overload is a
template.
[ChangeLog][QtCore][QString] Added char8_t overload of fromUtf8().
Change-Id: Iaa2d365bfa161ef36cc73fa3bad50aabf34d01db
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
All our tests were find() == end() or !=, which depends on the
evaluation order of the arguments to operator==(). If end() is called
first, then the detach happens before find() and all is well. But if
find() is called first, it may return end() before end() detaches.
[ChangeLog][QCborMap] Fixed a bug that could cause the iterator returned
from a failing key search with find() not to match end(). Now, every
call to find() will detach in shared QCborMaps; to avoid this, use
constFind() and constEnd().
Fixes: QTBUG-84583
Pick-to: 5.15 5.12
Change-Id: I552d244076a447ab92d7fffd161793496a8d03a8
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
A QVLA is copyable, so it should be movable, too.
Added a helper function a la P1144's uninitialized_relocate_n to deal
with the QTypeInfoQuery stuff. This way, the code is re-usable
everywhere it's needed. The same cannot be said for QArrayDataOps,
which only a parent can love...
[ChangeLog][QtCore][QVarLengthArray] Added missing move constructor
and move-assignment operator.
Task-number: QTBUG-39111
Change-Id: If0dc2aa78eb29062d73dcd3dc4647ba345ae39e6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
They were deprecated with replacements in QSslConfiguration proposed (and
some without alternative, which we'll provide if there is any demand
in such an API). Special thanks to M.N. for a nice hint on how to
amend the test without introducing a new API.
Change-Id: I7841a5b3f30469d8204b61cb65921c34275e0650
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The idea was to keep nagging us to update all the platform plugins to
do device registration. But besides being annoying, it would cause
test failures if we start adding QTest::ignoreMessage() all over,
and then some platforms start doing device registration properly.
Change-Id: Ia0fbb64cf86f33532be032ec9eebe6e4ad607f20
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
This enables the order of boundValues to be consistent as with a QMap
it could have been reordered which can be a problem for positional
bindings.
[ChangeLog][QtSQL] Changed signature of QSqlQuery::boundValues() to
return a QVariantList
Fixes: QTBUG-51609
Change-Id: I1c80fa8522fa7352723420b6fc9ec466406315fb
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Even it was not marked as deprecated the replacement function initFrom()
is available since Qt4 times (and init() is deprecated since then)
Change-Id: I09a4ebbf66b01fbe7aec67691dc68d2e42d1cd78
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
And name the main class QList. That's also the one we document.
This gives less porting pain for our users, and a lot less churn
in our API, as we use QList in Qt 5 in 95% of our API.
In addition, it gives more consistent naming with QStringList and
QByteArrayList and disambiguates QList vs QVector(2|3|4)D.
Fixes: QTBUG-84468
Change-Id: I3cba9d1d3179969d8bf9320b31be2230d021d1a9
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
* Assume UTF-8 on all Unix like systems
* Export some functions to be able to compile QTextCodec once
moved to Qt5Compat.
Task-number: QTBUG-75665
Change-Id: I52ec47a848bc0ba72e9c7689668b1bcc5d736c29
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
There is no reason for keep using our macro now that we have C++17.
The macro itself is left in for the moment being, as well as its
detection logic, because it's needed for C code (not everything
supports C11 yet). A few more cleanups will arrive in the next few
patches.
Note that this is a mere search/replace; some places were using
double braces to work around the presence of commas in a macro, no
attempt has been done to fix those.
tst_qglobal had just some minor changes to keep testing the macro.
Change-Id: I1c1c397d9f3e63db3338842bf350c9069ea57639
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This was causing some bogus failures in Qt Quick autotests.
Existing APIs like QQuickWindow::mouseGrabberItem() are not really
compatible with the idea of a mouse-less system; but perhaps we can
revisit this later.
Task-number: QTBUG-85114
Change-Id: Id1c2e5894e5cf13a79998aaea28d5f42fad920cf
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
We want every QInputEvent to carry a valid device pointer. It may be
some time until all QPA plugins are sending it, but it's necessary to
provide the functions for them to start doing that.
We now try to maintain the same order of arguments to all the functions.
handleTouchEvent(window, timestamp, device, the rest) was already there
(except "device" has changed type now), and is used in a lot of platform
plugins; so it seems easiest to let that set the precedent, and modify
the rest to match. We do that by adding new functions; we can deprecate
the older functions after it becomes clear that the new ones work well.
However the handleGestureEvent functions have only ever been used in
the cocoa plugin, so it's easy to change their argument order right now.
Modify tst_qwindow::tabletEvents() to test new tablet event API.
Task-number: QTBUG-46412
Change-Id: I1828b61183cf51f3a08774936156c6a91cfc9a12
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Documented QPromise. Added snippets under auto tests
to ensure they are compiled and run in CI.
Task-number: QTBUG-81586
Change-Id: I20084e38f9d2f6fc8540f95ee03ec3d2827177e8
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Change the meaning of Q_PRIVATE_QPROPERTY to imply that the property is
implemented using a QNotifiedProperty. That requires passing the owner
object instance to the value and binding setters.
Similarly, detect QNotifiedProperty members like QProperty.
Change-Id: If49bbb04c8ccd4a661973888c50d2d556c25034f
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
...starting with a new feature: registering different devices at
different seats and verifying their capabilities and that we can get
them back again.
Change-Id: I8e58a49080633753d02a76e5fdc4932f5c674e0a
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
- stop using deprecated accessors in QTabletEvent and QMouseEvent
- position() and globalPosition() provide high resolution so
we don't need to display hi-res global pos separately
- keep a pointer to the last-seen QPointingDevice instance
rather than copying fields out of it; it will only be deleted
if the platform plugin does that (usually only at exit or
when it's unplugged)
- show seat ID
- show platform-specific device ID
Change-Id: Ic34e1d1256d6956867992db831c2e66bdda7001e
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
We have seen during the Qt 5 series that QMouseEvent::source() does
not provide enough information: if it is synthesized, it could have
come from any device for which mouse events are synthesized, not only
from a touchscreen. By providing in every QInputEvent as complete
information about the actual source device as possible, we will enable
very fine-tuned behavior in the object that handles each event.
Further, we would like to support multiple keyboards, pointing devices,
and named groups of devices that are known as "seats" in Wayland.
In Qt 5, QPA plugins registered each touchscreen as it was discovered.
Now we extend this pattern to all input devices. This new requirement
can be implemented gradually; for now, if a QTWSI input event is
received wtihout a device pointer, a default "core" device will be
created on-the-fly, and a warning emitted.
In Qt 5, QTouchEvent::TouchPoint::id() was forced to be unique even when
multiple devices were in use simultaneously. Now that each event
identifies the device it came from, this hack is no longer needed.
A stub of the new QPointerEvent is added; it will be developed further
in subsequent patches.
[ChangeLog][QtGui][QInputEvent] Every QInputEvent now carries a pointer
to an instance of QInputDevice, or the subclass QPointingDevice in case
of mouse, touch and tablet events. Each platform plugin is expected to
create the device instances, register them, and provide valid pointers
with all input events. If this is not done, warnings are emitted and
default devices are created as necessary. When the device has accurate
information, it provides the opportunity to fine-tune behavior depending
on device type and capabilities: for example if a QMouseEvent is
synthesized from a touchscreen, the recipient can see which touchscreen
it came from. Each device also has a seatName to distinguish users on
multi-user windowing systems. Touchpoint IDs are no longer unique on
their own, but the combination of ID and device is.
Fixes: QTBUG-46412
Fixes: QTBUG-72167
Task-number: QTBUG-69433
Task-number: QTBUG-52430
Change-Id: I933fb2b86182efa722037b7a33e404c5daf5292a
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
The QPair changes trigger warnings about size_t vs. quint32.
We made offsets and sizes 32-bit in the QRhi API to emphasize that
some of the graphics APIs are using 32-bit sizes still. It's a bit
unfortunate that pairs now generate warnings when the size does not
match. Just cast as needed.
Change-Id: I88504eed8be6f4bdb2205b3671e2c2a9db9fcb1e
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
The Vulkan spec changed the behavior for VkApplicationInfo::apiVersion
in 1.1, conveniently breaking compatibility with all existing 1.0 logic.
We can no longer assume that the 1.0 behavior, which was failing instance
creation with VK_ERROR_INCOMPATIBLE_DRIVER for an unsupported version,
is always in place. So do not rely on this in the test, and add a
reminder in QVulkanInstance docs as well.
Fixes: QTBUG-85040
Change-Id: I8f5c7a7830877b72d106c444aebfaea191083ee0
Reviewed-by: Christian Strømme <christian.stromme@qt.io>
They are justified in using huge memory.
Pick-to: 5.15
Change-Id: Id16d2ea67cfac0e031d05258173391e222b41097
Reviewed-by: Albert Astals Cid <albert.astals.cid@kdab.com>
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
CL = Content-Length
The uploadByteDevice was kept after a redirect which caused the
internals to assume that we had to upload the data. Even if this was
not the case we still transmitted the Content-Length header from the
first request which was now stored in two places.
Fixes: QTBUG-84162
Pick-to: 5.15
Change-Id: Ic86b1ef0766ffcc50beeed96c1c915b721d40209
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
In Qt 5 we have to work with the esisting API (QSslCertificate::verify),
taking only the peer's chain and the name. We already have a private
API to have a CA's list as an additional parameter, the proper fix
in Qt6 will also introduce a public complement for this.
Pick-to: 5.15
Change-Id: Ib1f75c3056b135c6e6d42f977b5a7034fca658ee
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
The recently-added slice() method has the problem that it's a noun
as well as a verb in the imperative. Like std::vector::empty, which
is both an adjective and a verb in the imperative, this may cause
confusion as to what the function does. Using the passive voice form
of slice(), sliced(), removes the confusion. While it can be read as
an adjective, too, that doesn't change the meaning compared to the
verb form.
Change-Id: If0aa01acb6cf5dd5eafa8226e3ea7f7a0c9da4f1
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This changes the layout of the meta object data, so
also bump the meta object revision.
Original-patch-by: Lars Knoll <lars.knoll@qt.io>
Change-Id: I176fb16c207e8ebe59e358e69554be813406232f
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Property flags should be compile time booleans, not something to
be determined at runtime.
We've been using this to dynamically disable some properties in QWidget
based classes dependent on the state of a different property, but this
should better get implemented on top of our widgets.
Change-Id: I6296e8761303ecdf24d9e842142e8596304c015d
Reviewed-by: Simon Hausmann <hausmann@gmail.com>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Add QStringView overloads where they were missing. This keeps things
almost 100% source compatible.
Task-number: QTBUG-84319
Change-Id: Ica8ee773b2ce655e9597097786117959e34b236e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Amends comment 6e1d70ae12, which
introduced SE_PushButtonBevel so that QPushButton could ignore clicks
outside of the button's bevel.
In the macOS style, make sure that the framerect we pass to
NSButton::alignmentRectForFrame is the rect we receive from QPushButton
in the style options. The frame property of the shared NSButton* object
might not be initialized.
In the style sheet style, handle SE_PushButtonBevel the same ways as
Contents and FocusRect, as it is not a separately styleable property.
Change-Id: I12eb1b046c864a02b34d276e6352e2e16d44231e
Fixes: QTBUG-84852
Fixes: QTBUG-84879
Task-number: QTBUG-81452
Pick-to: 5.15
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Whoever wrote this test, was a PROPER hacker: trying to force a TLS implementation
not to properly compress some data, they generated a sequence of bytes in a very fancy manner,
something like 255 0 0 0 255 0 0 0 123 0 0 0 255 0 0 0 - yeah, it's really a random sequence
of bytes, surely, it's impossible to compress! Meh.
Pick-to: 5.15
Change-Id: Ia10ae18a40b5b8f006c45147b06fe5be6efcb129
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Amends 32b586864e, which added an override
of QStandardItemModel::roleNames(). It's good for it to have its own
QHash so that roles can be added, and because
QStandardItemModel::setItemRoleNames() remains un-deprecated; but before
that change, they were the roles that QAbstractItemModel had initialized
from QAbstractItemModelPrivate::defaultRoleNames(). In particular, we
need "display" to map to Qt::DisplayRole by default; several tests in
qtdeclarative depend on that.
Change-Id: I58b2d6aa6b6c78b1d618335ddc4ecb834af57274
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Make QPair an alias for std::pair, and qMakePair just a forwarder
towards std::make_pair.
Why? Fundamentally to ditch a bunch of NIH code; gain for free
structured bindings, std::tuple and std::reference_wrapper
compatibility, and so on.
Breakages:
* Some that code manually forward declares QPair.
We don't care about it (<QContainerFwd> is the proper way).
* Some code that overloads on std::pair and QPair. Luckily
it's mostly centralized: debug, metatypes, testing macros.
Just remove the QPair overload.
* Usages of qMakePair forcing the template type parameters.
There are a handful of these in qtbase, but only one was actually
broken.
* std::pair is NOT (and will never likely be) trivially copiable.
This is agreed to be a mistake done by practically all implementations
in C++11, can can't be fixed without breaking ABI.
Some code using QPair assuming it's trivially copiable may break;
exactly one occurrence was in qtbase.
* QMetaType logic extracts the type names in two different ways,
one by looking at the source code string (e.g. extracted by moc)
and one via some ad-hoc reflection in C++. We need to make
"QPair" (as spelled in the source code) be the same as "std::pair"
(gathered via reflection, which will see through the alias)
when compared. The way it's already done e.g. for QList is
by actually replacing the moc-extracted name with the name
of the actual type used in C++; do the same here.
On libc++, std::pair is actually in an inline namespace --
i.e. std::__1::pair; the reflection will extract and store
"std::__1::pair" so we need an ad-hoc fix to QMetaType.
[ChangeLog][QtCore][QPair] QPair is now an alias to std::pair,
and does not exist as a class in Qt any more. This may break
code such as functions overloaded for both QPair and std::pair.
Usually, the overload taking a QPair can be safely discarded,
leaving only the one taking a std::pair. QPair API has not changed,
and qMakePair is still available for compatibility (although
new code is encouraged to use std::pair and std::make_pair
directly instead).
Change-Id: I7725c751bf23946cde577b1406e86a336c0a3dcf
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
We deprecated several setters on QSslSocket, no need to test
them then (and I'm removing them in Qt 6 anyway, so changing
the test is the phase 1).
Pick-to: 5.15
Change-Id: I5807abfb280cbacd4fcc19468793f9d1f3b2ff20
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
[ChangeLog][General] The QtOpenglExtensions module has been
discontinued. It provided only convenience functionality for certain
direct OpenGL usage. With the new RHI graphics API abstraction, that
is no longer a primary use case. Applications that still need to
access that API have a number of alternative options, including
QOpenGLExtraFunctions.
Fixes: QTBUG-84085
Change-Id: I272af61c69ebcec207b576d67d08b59623d485ec
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
QString and QStringRef did bounds checking for left/right/mid, whereas
QStringView was asserting on out of bounds.
Relax the behavior for QStringView and do bounds checking on pos/n
as well. This removes a source of potentially hidden errors when porting
from QStringRef (or QString) to QStringView.
Unfortunately, one difference remains, where QByteArray::left/right()
behaves differently (and somewhat more sane) than QString and
QStringRef. We're keeping the difference here, as it has been around
for many years.
Mark left/right/mid as obsolete and to be replaced with the new
first/last/slice methods.
Change-Id: I18c203799ba78c928a4610a6038089f27696c22e
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
QPromise and QFuture created from it share the same internal state
(namely, QFutureInterface object). QPromise provides high-level
management of the shared resource, ensuring thread-safe behavior
on construction and destruction (also taking into account
QFuture::waitForFinished() semantics).
QFuture acts as a primary controller of QPromise via action
initiating methods such as suspend() or cancel(). QPromise is
equipped with methods to check the status, but the actual
handling of QFuture action "requests" is user-defined.
[ChangeLog][QtCore][QPromise] Added QPromise class to accompany
QFuture. It allows one to communicate computation results and
progress to the QFuture via a shared state.
Task-number: QTBUG-81586
Change-Id: Ibab9681d35fe63754bf394ad0e7923e2683e2457
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
MinGW plugins should not have the lib prefix in the shared library
names. Do this manually for a couple of libraries, and also for the
generic qt_add_cmake_library function.
Amends 9b0e23ef8a
Change-Id: I1cfaf8fc046f86edd3e755adfa599aa0aa854ee3
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Remove QDestopWidget public header, simplify the implementation that
maintains a Qt::Desktop type QWidget for each QScreen, and turn
QWidget's initial target screen into a QScreen pointer.
QApplication::desktop() now takes an optional QScreen pointer, and
returns a QWidget pointer, so that applications and widgets can get
access to the root widget for a specific screen without having to
resort to private APIs.
QDesktopWidgetPrivate implementations to look up a screen for an index,
widget, or point are now all inline functions that thinly wrap
QGuiApplication::screens/screenAt calls. We should consider adding those
as convenience APIs to QScreen instead.
Note that QWidget::screen is assumed to return a valid pointer; there is
code that handles the case that it returns nullptr (but also code that
trusts that it never is nullptr), so this needs to be defined, verified
with tests, and asserted. We can then simplify the code further.
Change-Id: Ifc89be65a0dce265b6729feaf54121c35137cb94
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Many of these were generated by clazy using the new qevent-accessors check.
Change-Id: Ie17af17f50fdc9f47d7859d267c14568cc350fd0
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The default font on Windows 10 (Segoe UI) will return the
wrong minimum right bearing at some sizes, which will cause
us to skip the textWidthVsWIdth() test at some scale factors,
since we cannot trust the layout to be perfect in this case.
Based on experiment, Arial is more accurate, so in order to
avoid skipping the test, we default to this on Windows instead.
(Note: The problem has not been observed with the default fonts
on Linux or macOS, so we only do this for Windows specifically.)
Task-number: QTBUG-84415
Pick-to: 5.15
Change-Id: I8cdb5d0d9922915a6ed1574d62a561dda0e1dc5d
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
These methods are scheduled as a replacement for left/right/mid()
in Qt 6 with a consistent, narrow contract that does not allow
out of bounds indices, and therefore does permit faster
implementations.
Change-Id: Iabf22e8d4f3fef3c5e69a17f103e6cddebe420b1
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
There are some slight differences between normal drawText()
and QGlyphRuns/QStaticText for decoration widths in certain
fonts. We decided to accept this and tried working around it
in the test by using ForceIntegerMetrics (since the difference
is < 0.5 pixel). This enum has been deprecated, so we move
the tests into Lancelot instead, since the idea here is to
test for regressions, not to compare the two painter commands.
Note that there is something off about decorations with
drawGlyphRuns() and drawStaticText() which is exposed
(not caused) by this, perhaps related to using a matrix
for positioning, since that was untested before.
This also takes the liberty of moving the emoji test string
from text.qps, since this was not in the statictext.qps yet.
Change-Id: Ib2d697095cbd11829cdd50b3c0268c85e9607c78
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Macros and the await helper function from qfunctions_winrt(_p).h are
needed in other Qt modules which use UWP APIs on desktop windows.
Task-number: QTBUG-84434
Change-Id: Ice09c11436ad151c17bdccd2c7defadd08c13925
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The code was reading from uninitialized memory when creating a
variant. Fixing that reveals that the test semantics
are broken: when dealing with datatypes without a registered
operator==, QVariant resorts to memcmp, so the two objects
would've actually compared equal. Amend that.
Change-Id: I36bad7ee6a45154d5d534b7dd8b618cc0a900126
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The QLayout::setMargin(x) call has been deprecated since Qt 5.13.
It was an alias for setContentsMargins(x, x, x, x) so we just
replace it with that.
Change-Id: I1c89eebfe776e0e3c8d0bdc19244bd452db4ef3c
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Previously only qstr*icmp() were tested and the test data was sent via
QString and {en,de}coding. Use a local data-type to package pointers
to actual string literals for passing to these functions that take
them. Fold the various tests involving null pointers and empty
strings into the general testing, removing from "singularity" tests,
and combine the remainders of those tests into a single test of
singular cases for QByteArray::ompare. Move all these tests to
alongside the existing tests for QByteArray::compare. Use nullptr
rather than 0 as the null string.
Change-Id: Ie6d01e839c330c2f960af4bcc95e5633539337d6
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
We don't want to have untested qtbase.
This reverts commit 3598ffcc26.
This change also skips two network tests on b2qt.
Change-Id: I5c6f5e19487775f0a1c0feafa5e085208cbf7e9a
Reviewed-by: Liang Qi <liang.qi@qt.io>
Replace the implementation of blockingMappedReduced():
after calling non-blocking version of mappedReduced() we are getting
the future object, so we may call in sequence result(), which
will block and return the result when the all tasks are done.
The same is done with blockigMapped(), which calls blockingMappedReduced()
with a custom reduce function.
Looks like with this pattern we can reuse the non-blocking version
for implementing blocking version of mapped / filtered methods.
Task-number: QTBUG-83918
Change-Id: I7f240cfbd04834d551ff79d717b72194a26996d7
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
The helper processes were not in the correct location.
Change-Id: I0a80a22a931625ea0c9370db38ff29c881b964cb
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
The helper executable needs to be a bundle, because that's
what the test expects.
Change-Id: I7c3f72f12b7a744478b4832685d032abfc679e5b
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
It's disabled in qmake too, due to being almost all blacklisted
and flaky. Plus locally when i tested, i had to give accessibility
permissions manually, otherwise it failed.
Change-Id: I55ce7a81c46ac28f41c43f5c159fecc22d489966
Reviewed-by: Cristian Adam <cristian.adam@qt.io>
Deprecated the pause-related APIs of QFuture* classes and
added alternatives having "suspend" in the name instead.
With 2f15927f01 new
isSuspended()/suspended() APIs have been added to QFuture* classes for
checking if pause/suspension is still in progress or it already took
effect. To keep the naming more consistent, renamed:
- setPaused() -> setSuspended()
- pause() -> suspend()
- togglePaused() -> toggleSuspended()
- QFutureWatcher::paused() -> QFutureWatcher::suspending()
Note that QFuture*::isPaused() now corresponds to (isSuspending() ||
isSuspended()).
[ChangeLog][Deprecation Notice] Deprecated pause-related APIs of
QFuture and QFutureWatcher. Added alternatives having "suspend" in
the name instead.
Change-Id: Ibeb75017a118401d64d18b72fb95d78e28c4661c
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Starting from Qt 5.11 QWindow::event is called after QDialog::closeEvent
which would cause a crash if "delete this" was called on closeEvent. The commit
that changed this was e0b5ff4ad5. Added
a check before QWindow::event call utilizing QPointer to prevent the
function call in case object is destroyed by a user in close event handler.
Change-Id: I64a4a0f3271714e55bf7e806177f0d8b39b67fa3
Fixes: QTBUG-84222
Pick-to: 5.15 5.12
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
QFileSystemModel is the documented replacement. It uses threads to
populate the model, which QDirModel doesn't.
Change-Id: I7818ecd8f849eb566ac176612f382e17a0471c47
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Use 'errorOccurred' instead of 'error' and 'sslHandshakeErrors'
instead of 'sslErrors'.
Fixes: QTBUG-82605
Change-Id: I19d4845b16c7b636af7b0658d4fbbba4eef9d029
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Including other headers, while defining very questionable macro
names, can and will result in trouble. Stop doing that.
While at it, move from including <QtTest> to include-what-you-use,
and clean up the code a bit.
Change-Id: Idb02ef2b612c0805baecac3ce6edd435609aca4c
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Previously it handled Latin-1, which made it incompatible with UTF-8,
which is now our preferred 8-bit encoding. For Qt6 it is limited to
ASCII. Adjusted tests to match. QLatin1String::compare() turned out
to be relying on qstrnicmp()'s Latin-1 handling.
Removed some spurious Q_UNLIKELY()s and tidied up code a little in the
process.
[ChangeLog][QtCore][Important Behavior Changes] Encoding-dependent
features of QByteArrray are now limited to ASCII, where previously
they worked for the whole of Latin-1. This affects case-insensitive
comparison, notably including qstricmp() and qstrnicmp(), and
case-transforming functions.
Fixes: QTBUG-84323
Change-Id: I2925d9908f8654599195a2860847b17083911b41
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
This class is designed as C++20-style generator / lazy sequence, and
the new return value of QString{,View}::tokenize().
It thus is more similar to a hand-coded loop around indexOf() than
QString::split(), which returns a container (the filling of which
allocates memory).
The template arguments of QStringTokenizer intricately depend on the
arguments with which it is constructed, so QStringTokenizer cannot be used
directly without C++17 CTAD. To work around this issue, add a factory
function, qTokenize().
LATER:
- ~Optimize QLatin1String needles (avoid repeated L1->UTF16 conversion)~
(out of scope for QStringTokenizer, should be solved in the respective
indexOf())
- Keep per-instantiation state:
* Boyer-Moore table
[ChangeLog][QtCore][QStringTokenizer] New class.
[ChangeLog][QtCore][qTokenize] New function.
Change-Id: I7a7a02e9175cdd3887778f29f2f91933329be759
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Since the variable names in QShaderDescription are later compared to
QByteArrays we can gain some performance from not having to convert them
to QByteArrays later.
Task-Id: QTBUG-83706
Change-Id: Iaf80d0966f45cbb09e7c1000b7854bc488e57bb3
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
The static observer can live in a union with the inline observers. We
only need to take care of calling the ctors and dtors manually then.
In order for any observers to be called in the presence of a static
observer, the static observer has to be called after the other
observers.
Change-Id: I2f56fa64f3fe6fcd7f06cc403929362da7d86f65
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Windows unexpectedly passes PM_NOYIELD flag in wParam parameter to the
hook procedure, if ::PeekMessage(..., PM_REMOVE | PM_NOYIELD) is called
from the event loop. So, to ignore undocumented flag, we should
interpret wParam as a bit field.
Thanks to Robin Lobel for research.
Pick-to: 5.15
Fixes: QTBUG-84562
Change-Id: Ib16d7d747aebc9a3628e4ee67478c4d3edeb96f1
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
