Two cases with no code between them don't need a fall-through marker.
The only code in the first case is subject to #if-ery; it either
returns, without falling through, or has no code there. Putting in an
overt Q_FALLTHROUGH() confused a rather dumb compiler ...
Change-Id: Icc034be08859cc1656401af1c62367aec302668b
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
This reverts commit 4bdd8d4eca.
It contains an error.
Change-Id: I51052029f001b9e82c2a53de15b4ba354aafdbae
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
For winrt we cannot rely on the fact, that QThread::current will be called
from the correct thread for the first time. The application's main entry
point creates a suspended thread and starts it right afterwards. At that
moment, other functionality (QLoggingRegistry for example) might have
called QThread::current, which set the wrong thread as the main thread. In
order to avoid this situation, the main thread is explicitly set in
QCoreApplication's constructor.
Task-number: QTBUG-66418
Change-Id: I8b6347357a80eb395ae758bd3d420adef0826751
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
This change applies to darwin. It adds each include path in the
QMAKE_DEFAULT_INCDIRS variable to the qdoc command line with the -I
flag for both the prepare and the generate phase. These include paths
provide access to the standard c++ and c headers, which clang needs
to see. This change should work on all platforms, but it increased
the qdoc warning count on the linuxsystem where it was tested, so
now it only applies to darwin.
Change-Id: I16e2e0d744e2cf68743dc12d39155dda2ece1536
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
That allows us to remove custom handling in QMetaType.
Change-Id: Ic09fb96e1a05c6897803811d70aebbc6ea6e4f2e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The matatype should not keep manually maintained list of stream
operators. The patch adds automatic detection for all builtin types
so load and save functions pick the right delegate automatically.
This change exposed some existing anomalies:
- char is enforced to be signed while it seems that just calling
the operator directly does not have that feature.
- [unsigned] long type is always upgraded to [unsigned] long long
- QCborSimpleType doesn't have the data stream operators while
metatype is able to stream it through casting
Change-Id: I51178d6acd97d0585a6089e30ddd6acb2a29af54
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There is not reason why it should be kept in secret. Not having public
accessor forces a user code to keep a copy of the id. Visible for
example in QML (QQmlValueType).
Change-Id: If0de65fb8d91bcd50880c66b5f87c68f6d888dd3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This change removes redundant 'virtual' from function declarations.
Clang Tidy's modernize-use-override check reports:
warning: 'virtual' is redundant since the function is already
declared 'override'
CppCoreGuidelines say:
C.128: Virtual functions should specify exactly one of virtual,
override, or final
Change-Id: I9a4bdd6cc041d46ae64b25597ba4f7268ac4c2b7
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Kevin Funk <kevin.funk@kdab.com>
This update fixes several qdoc warnings for things like incorrect
links to section titles, missing enum \value commands, and incorrect
\fn signatures.
Change-Id: I35ae18810d386a78ffa87d674489c7ad8c5fa215
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
The wrong return value didn't seem to result in observable bugs.
Change-Id: Iaba74d0acd4352af9be70498a7d5556bb1e02f5f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It was accessing the deleted current element in a potentially
reallocated container.
Change-Id: I81d86410907b7b0c20c2edc62ad256315606bc96
Reviewed-by: Andreas Hartmetz <ahartmetz@gmail.com>
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Commit 1c623bc6d1 introduced a new
QMetaObject revision, which change the size of the QMetaEnum data.
When looking up QMetaEnum in a QMetaObject, this size need to be
checked for every different QMEtaObject from the hierarchy, not just
the first one.
Change-Id: I6f0d3982329822e15e284aef9b141d4c9ab351b9
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Weak pointers are zeroed before destroyed() is emitted. This makes sense
because the signal is called "destroyed()", not "aboutToBeDestroyed()",
and the code has been like this for a long time.
Change-Id: Ic852fe354405e328d1922a75931e7eb7bf651b26
Reviewed-by: Christian Kandeler <christian.kandeler@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Some people suggested the later, so let's have a second look at it
before it's too late.
Although I was in favor of the former I'm now having second thoughts.
connectTo() is meant to only be used in classes which have a clear dominant signal,
but there are rare classes that have two (example: QAbstractButton::toggled, QAbstractButton::clicked).
QAbstractButton::connectTo() would be ambiguous if we ever wanted to add these shorthand connects
to QtWidgets
Change-Id: I198ad3326d31dcf89b9c47a299124572ee7b11b3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Replace by switch() where appropriate, remove else and unindent code
or simplify the return value.
Change-Id: Ie988b9068a9579ae5a899b3765e43aad480b564e
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
We no longer support macOS 10.11, iOS/tvOS 10, or watchOS 3.
Change-Id: Ide03d8fac06185ef4162ba75ee54a0adf6916905
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
This is the squashed diff from wip/webassembly to dev.
Done-with: Peng Wu <peng.wu@intopalo.com>
Done-with: Sami Enne <sami.enne@intopalo.com>
Done-with: Morten Johan Sørvig <morten.sorvig@qt.io>
Started-by: Andrew Knight <andrew.knight@intopalo.com>
Change-Id: I6562433c0a38d6ec49ab675e0f104f2665f3392d
Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Replace by reinterpret_cast or const_cast, respectively.
Use auto when initializing a variable to fix Clang warnings
about repeating the type name, do minor tidying along the way,
and a few conversions of 0 or NULL to nullptr.
Change-Id: Ieb271a87ddcf064f536e1ff05d23b1e688b1b56a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Synced QMetaObjectBuilder::MetaObjectFlag with the MetaObjectFlags enum
from qmetaobject_p.h. Also added a few comments for Qt 6.
Change-Id: Ieccd5cf8d512a6bf7256b2f4db88d45662774536
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This feature is needed to register complex types (e.g. Q_GADGETS) at
runtime, using a single constructor method and added type parameter.
Without having the type id available to the Constructor it is impossible
to specialize behavior, thus requiring separate constructors for each
type. Generating these separate constructors at compile time is easy,
but not at runtime.
[ChangeLog][QMetaType] QMetaType can now register constructor for a set
of dynamic types.
Change-Id: I6071271d0e157864594dd07b4bc3a0abbeb15a44
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Otherwise the dynamic loader will complain about missing symbols when
the binary is run on platforms below our supported deployment target:
dyld: Symbol not found: __os_activity_current
Referenced from: QtCore.framework/Versions/5/QtCore (which was built for Mac OS X 10.12)
Expected in: /usr/lib/libSystem.B.dylib
in /Users/torarne/build/qt/5.12/qtbase/lib/QtCore.framework/Versions/5/QtCore
Trace/BPT trap: 5
We want this to trigger our own logic in qt_apple_check_os_version(),
where we tell the user in more friendly terms what's going on.
An alternative to the targeted weak imports would be do import the
whole library as weak, using -weak-lSystem.B. This doesn't seem
to cause any performance issues at startup, but since we only need
the two global symbols we stick to the more targeted solution just
to be on the safe side.
Change-Id: I87c1f185f6dcf9df26c700d31bb5071ddf7685be
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The application name wasn't always printed, so we try try a few
more possibilities before falling back to the process name.
We also run the check as early as possible, instead of relying
on a QCoreApplication.
We do not have to provide a dialog to the user, as macOS will do
this for us if the application is launched from Finder.
Change-Id: Ifbec86946d60294806364e08964852fd4b74ff56
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
macOS 10.14 uses a new font smoothing algorithm that takes the fill color into
account. This means our default approach of drawing white on black to produce
the alpha map will result in non-native looking text when then drawn as black
on white during the final blit. As a workaround we use the application's current
appearance to decide whether to draw with white or black fill, and then invert
the glyph image in the latter case, producing an alpha map. This covers the
most common use-cases, but longer term we should propagate the fill color all
the way from the paint engine, and include it in the key for the glyph cache.
At the moment we do not react to changes in the application appearance,
as that seems to be buggy in general in Qt (palette/style, e.g.), and those
bugs need to be weeded before we can react to the theme change with confidence.
Task-number: QTBUG-68824
Change-Id: Ibbfd49fcf3a091e454009c08159f46b3499e2bd0
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Add it to configure.json and replace all occurrences of QT_NO_THREAD
with QT_CONFIG(thread). Add conditions for other features that depend
on thread support. Remove conditions where we can use the QMutex and
QThreadStorage stubs.
Change-Id: I284e5d794fda9a4c6f4a1ab29e55aa686272a0eb
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
... which is Qt's version of std::make_index_sequence / C++14.
The current Indexes generator is recursing O(N) times which is reaching
recursion limits level for long IndexesList.
This implementation has only O(log N) recursion.
Change-Id: I423ef05152d77b18f1f5a64c49b273f49aa958bb
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
For some reason, the GLib event dispatcher was lowering the priority for
timers compared to other events like for example socket events.
the IDLE priority is for "Do this later when you are not busy", and that
is wrong for generic timer events.
See also https://bugs.kde.org/show_bug.cgi?id=230184
Task-number: QTBUG-48344
Change-Id: Id337a8f0c7fad3ebd5e5fa49d7831a0001ea9010
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Adds an enumName to QMetaEnum to carry the name of the enum since for
flags that doesn't match the name of the Qt type, but is needed if the
flag is scoped.
Change-Id: I1c0f77eb9e40e6fd1eb6a59bea77caf0f33fcf43
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
We were using CFRunLoopGetMain() everywhere. Get the
correct run loop using CFRunLoopGetCurrent() during
initialization, and store it.
Event dispatcher initialization must now be delayed
until after the constructor has run, since event dispatchers
may be created on the main thread and then moved to
the target thread. Initialize on first call to processEvents()
where the current thread will be the correct thread.
Use the stored m_runLoop instead of CFRunLoopGetMain().
This is required for wakeUp() and interrupt() which
may be called from another thread.
Change-Id: I6fffcfd4394899c4a12f241c42781979aaf99d5e
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
We allocate the variant's PrivateShared for custom types with space for
the actual custom type right after that, using operator new to allocate
the memory and a new calls for the constructors. However to free the
memory we merely call delete on the private shared, which upsets ASAN
because the size passed to operator delete is the size of the private
shared, which is not the same as the size that was provided for the
returned address.
This is easily fixable by calling the destructor explicitly and calling
operator delete with just the pointer.
Change-Id: I50afbe0e8afc875c0876e85e02689dcbdc152633
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
"type qualifiers ignored on cast result type [-Werror=ignored-qualifiers]"
And preemptively fix Clang nagging about repeated type.
Change-Id: I2253b032d1867aa9976b4bdfc0f45023198f917c
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Edward Welbourne
Martin Smith
Liang Qi
Oliver Wolff
Jędrzej Nowacki
Qt Forward Merge Bot
Giuseppe D'Angelo
Alessandro Portale
Thiago Macieira
Andreas Hartmetz
Gatis Paeglis
Olivier Goffart
Paul Wicking
Ulf Hermann
Sergio Martins
Friedemann Kleint
Tor Arne Vestbø
Morten Johan Sørvig
BogDan Vatra
Mikhail Svetkin
Sune Vuorela
Allan Sandfeld Jensen
Simon Hausmann
Mårten Nordheim