| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
- Remove the casting operator of QFuture<T> to T. It calls
QFuture::result(), which may lead to undefined behavior if the user
has moved the results from QFuture via QFuture::takeResult() before
trying to do the conversion.
- Disable implicit conversion of QFuture<T> to QFuture<void>, by making
the constructor explicit. If the users really intend to do the
conversion, they should do it explicitly.
[ChangeLog][Source-Incompatible Changes][QFuture] Implicit conversions
of QFuture<T> to T and to QFuture<void> have been disabled. Use
QFuture::result() or QFuture::takeResult() where you need to convert
QFuture<T> to T. Use the explicit QFuture<void>(const QFuture<T> &)
constructor to convert QFuture<T> to QFuture<void>.
Change-Id: I153d4137d36365b1611ac934fb3ac2eb667fdd6c
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
These operators don't do what the user might expect and may lead to
confusing results.
[ChangeLog][Source-Incompatible Changes][QFuture] The comparison
operators of QFuture have been removed. They were comparing the
underlying d-ptrs instead of comparing the results (as the users
might expect), which is not very helpful for the users point of view.
Change-Id: I80a887610eac38b60329128cca52cdb5fb515207
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
| |
Change-Id: I6eb95aa66ff847e8bb9aac348fded3a5d55015b6
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For historical reasons Qt Concurrent reports QUnhandledException in
case if an exception that is not derived from QException is thrown
from a worker thread. Changing this behavior may not be a good idea,
since the existing user code may rely on it. Changed QUnhandledException
to wrap the std::exception_ptr to the actual exception, so that the
users can obtain the information about the thrown exception if needed.
[ChangeLog][QtCore][QUnhandledException] Improved QUnhandledException to
store the std::exception_ptr to the actual exception thrown from a
QtCocnurrent worker thread.
Change-Id: I30e7c1d3e01aff6e1ed9938c421da0a888f12066
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
|
|
|
|
|
|
| |
Task-number: QTBUG-83331
Change-Id: I572f68f6d3be4a50970d8d77d070f175be3ec785
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
| |
Change-Id: I4c416f52c7102750a77c3f91274dd0a235569d6e
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
| |
Done with selective application of clang-format
Change-Id: Iee6bf2426de81356b6d480629ba972f980b6d93d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
| |
Change-Id: I25faab16ad2df3682e6c6b55d4aaff1c20402995
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
| |
As requested by a ### Qt 6 comment. This then implied a few other
functions weren't constexpr, which broke some tests.
Task-number: QTBUG-85700
Change-Id: I6522a9b2d7a74e117442121400a1d7198d323967
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
| |
Also mark the helper function.
Change-Id: I1469abf22cd132dbb1afe680121b6c928ffbe41e
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
|
|
|
|
|
|
|
| |
Avoid creating a a d pointer for recursive mutexes.
Change-Id: I28af15a416ee17de346e2ea5b1442279d9d3e159
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
| |
Change-Id: Ia4868259bfa25e4b929a51de6de580df7277c282
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
These classes should not inherit from each other
anymore in Qt 6. The reason is that this makes
the 95% case of using a non-recursive mutex
much slower than it has to be.
This way, QMutex can now inline the fast path
and be pretty much as fast as QBasicMutex is
in Qt 5. They actually use the same code paths
now. The main difference is that QMutex allows
calling tryLock() with a timeout, which that
is not allowed for QBasicMutex.
[ChangeLog][QtCore][QMutex] QMutex does not support
recursive locking anymore. Use QRecursiveMutex for that
purpose. QRecursiveMutex does not inherit QMutex anymore
in Qt 6.
Change-Id: I10f9bab6269a9181a2e9f534fb72ce65bc76d989
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
| |
Since we're going to split QMutex and QRecursiveMutex into
separate classes, make sure QMutexLocker is prepared for that.
Change-Id: Id5e9a955d1db7c8ee663dd3811ad6448dad0aeae
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
|
|
|
|
|
|
|
|
|
| |
qfuturewatcher.h includes qfuture.h, which includes EnableForNonVoid
through qfuture_impl.h header. Thus, there is never a need to keep
the same alias in QFutureWatcher as it always can use one from QFuture
Change-Id: I293fd087aea1a21ef5bcfdf50cdefc176a9703d0
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Accidentally found out that we enable/disable QPromise::addResult based
on type deduced from input argument, instead of using "value_type" of
QPromise itself, which is wrong
Simplified the checks to a single one -
EnableIfSameOrConvertible<InputType, StoredType> as this is sufficient
to account for both cases: QPromise<void> and QPromise<T> with input,
convertible to T
Change-Id: I657998c0e26241b0fc5e70988622984ece8871df
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
|
| |
Proposed during API review
Change-Id: I9c43e1915c50803ab69bfe07a91c05d2224b86c4
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Changed QPromise::addResult() to return bool value. True is returned
when result is added and false is returned when e.g. promise is in final
state (canceled or finished) or when addResult() is called twice with
the same index as argument (in which case new value is rejected)
Updated QFutureInterface::reportFinished() that accepts optional result
as argument to align with other result adding methods. This function
is "internal" only (as of now), so no documentation update is needed
Change-Id: I2d63069246e5e5c8cf04529c22bb296faaaae53d
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
One can call addResult(value, index) twice and consequently set the
value twice by the same index. This seems rather strange and probably
should not be allowed. This commit rejects setting results when there's
already a valid result by that index. Consequently, this fixes memory
leaks caused by N-times-called addResult(..., index)
Fixes: QTBUG-86828
Change-Id: I77494f2cb73ce727ffad721cfcdcaa420899eb25
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
QFuture::takeResult() currently returns std::vector instead of QList,
because QList does not support move-only types. Disable this method
until QList is fixed to work with move-only types in Qt 6.1.
Also did minor doc-fixes.
Change-Id: I87feaf75d9433a3b540edd00039c3e21d6994985
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
|
|
|
|
|
|
|
|
|
| |
There is no <sanitizer/tsan_interface.h> header when building with gcc,
at least on Ubuntu 18.04.3.
Fixes: QTBUG-87317
Change-Id: Ie933f6fa478f11b5062c665007e91be68e31ebe3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
| |
Moved the logic of finding a result in ResultStore to separate function
and parameterized it with QMap<...>. This is a pre-step to make find
procedure uniform regardless of the storage we are looking in (either
visible or pending as of now)
Change-Id: I41641d70751925f223e992f52fbc7814085c452d
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
| |
Pending results were never cleared by result store. This led to memory
leaks when the results never transitioned to "visible" results
Change-Id: I674302eb51542ad5f4d918da68d616428c73ae9f
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
|
| |
This time based on grepping to also include documentation, tests and
examples previously missed by the automatic tool.
Change-Id: Ied1703f4bcc470fbc275f759ed5b7c588a5c4e9f
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
At the moment we have two main strategies for dealing with move
assignment in Qt:
1) move-and-swap, used by "containers" (in the broad sense): containers,
but also smart pointers and similar classes that can hold user-defined
types;
2) pure swap, used by containers that hold only memory (e.g. QString,
QByteArray, ...) as well as most implicitly shared datatypes.
Given the fact that a move assignment operator's code is just
boilerplate (whether it's move-and-swap or pure swap), provide two
_strictly internal_ macros to help write them, and apply the macros
across corelib and gui, porting away from the hand-rolled
implementations.
The rule of thumb when porting to the new macros is:
* Try to stick to the existing code behavior, unless broken
* if changing, then follow this checklist:
* if the class does not have a move constructor => pure swap
(but consider ADDING a move constructor, if possible!)
* if the class does have a move constructor, try to follow the
criteria above, namely:
* if the class holds only memory, pure swap;
* if the class may hold anything else but memory (file handles,
etc.), then move and swap.
Noteworthy details:
* some operators planned to be removed in Qt 6 were not ported;
* as drive-by, some move constructors were simplified to be using
qExchange(); others were outright broken and got fixed;
* some contained some more interesting code and were not touched.
Change-Id: Idaab3489247dcbabb6df3fa1e5286b69e1d372e9
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
|
|
|
|
|
|
|
| |
It could never be higher than 1 anyway.
Change-Id: If33c7978a4397a08e9eb091926726725d8bd3ea6
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
Since we drop the lock while deleting threads, we need to handle
the queue possibly being accessed and changed by the pool threads
while clear() is running.
Pick-to: 5.15
Fixes: QTBUG-87092
Change-Id: I7611edab90520454278502a58621e299f9cd1f6e
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
|
| |
None of this code is even compiled in qt6.
Change-Id: I5891cc9459320083ad3908fcbf646f3ba75b8a4d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
| |
Change-Id: I559f2fd73a9aae3d126be18cb259f8a9abe0efaf
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
|
|
|
|
|
|
|
| |
It was already used many places directly making the code inconsistent.
Change-Id: I3b14bc6c333640fb3ba33c71eba97e78c973e44b
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
| |
Remove the check since a QMutex can no longer be recursive, fixing:
qwaitcondition_win.cpp:164:28: warning: 'bool QMutex::isRecursive() const' is deprecated: Use QRecursiveMutex instead of a recursive QMutex [-Wdeprecated-declarations]
Task-number: QTBUG-85700
Change-Id: Ic1631c1e671cf3234b5823d6d20121d746304c8e
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
| |
Use QRecursiveMutex instead.
Pick-to: 5.15
Change-Id: I862fc2b3143deeb5c96dc8d445be5f9fa2535670
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
| |
Updated QFuture docs to be more precise about QtFuture::Launch::Sync
policy.
Change-Id: Ic267c71f858e04a47ea1fc0996ea342d5eae7744
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
| |
Change-Id: I3af019b65835b2f82161d6f1c24feb0523a32c11
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
| |
These were now always defined, hence redundant.
Leave the #define in place so that we can verify we actually do always
define it, in a #else of an existing #if check on it.
Change-Id: Iea4c3dbc8f9982268bcf81da5ef17fe2ebf5c462
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
| |
Change-Id: I4bf0ddf4ddf4044a60d881a57ef63b96d4bac262
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
| |
As directed by ### Qt 6 comment.
Task-number: QTBUG-85700
Change-Id: Iae4179b017840efe4902de2b1529cf7ec0606865
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The implementation can be more efficient with two 32-bit fields instead
of just one, but it depends on whether operations on 64-bit atomics are
locked or not.
Quick check for the following architectures does not reveal any
problems, but someone needs to investigate more. Notably, GCC and Clang
generate a call to libatomic's __atomic_load_n / __atomic_store_n / etc.
32-bit std::atomic<uint64_t>
Arch ::is_always_lock_free .is_lock_free()
ARMv7-A false true
i386 false true
MIPS false true
PPC false true
RISC-V false true
SPARC false true
Change-Id: Ia2273af1172d493092d6fffd163251a99064c51b
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit ff69227a49119c00f703cf89c9d72db7eeeeaa66.
Reason for revert: 64-bit atomics on 32-bit systems are often
(but not always) worse than the 32-bit semaphore as it was
implemented. Plus the High32 and Low32 functions are returning
the same thing, forgetting the endianness check.
Change-Id: I5d5ade6e9bc7086600ff2302546385151e32142b
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
|
|
|
|
|
|
|
|
| |
This simplifies the implementation so that we don't
need separate code paths for 32 and 64 bit anymore.
Change-Id: I823612865e7d648fb0bd1632385ce67b5a452b8a
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
|
|
|
|
|
|
|
| |
Task-number: QTBUG-85700
Change-Id: I5ce368e8edca2b9483a0f0ef34bc9eb6b4e44574
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Currently QFuture::waitForFinished() exits as soon as the future is not
in the running state. If the user calls it before
QPromise::reportStarted() is called, it will exit immediately, because
nothing is running yet. Fix the behavior to wait for the finished state.
[ChangeLog][Important Behavior Changes][QtCore] Fixed the behavior of
QFuture::waitForFinished() to wait until the future is actually in the
finished state, instead of exiting as soon as it is not in the running
state. This prevents waitForFinished() from exiting immediately, if at
the moment of calling it the future is not started yet.
Task-number: QTBUG-84867
Change-Id: I12f5e95d8200cfffa5653b6aa566a625f8320ca8
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Apparently the Linux asm assumptions are not correct for the
x86 Android emulator and this caused Android apps not to work
properly: signal connections being queued instead of being direct,
crashes in QPropertyAnimation, etc.
Using currentThreadIdImpl on the Android emulator works fine though.
Optimizing the code for the Android emulator case can be done in
another change.
Amends 5e9b2ade678f37e43bfc2e3484f54cbbb5844d2e
Fixes: QTBUG-85640
Change-Id: I3b3ba76ea143aed949a6e50678c850b6ba231476
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Add missing \threadsafe command.
* Add missing note for methods callable only from the started thread.
* Expand note on excerting care when interacting with objects across
threads in QThread's class overview documentation.
Fixes: QTBUG-86112
Pick-to: 5.15
Change-Id: I8f181d92ad6196ff0c13f5a866a36793209a75ab
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The differences to run() method:
1. The passed function should have additional
argument QPromise<T> &, declared as a first argument.
2. The return value of the function must be void.
Result reporting should be done through
passed QPromise<T> &promise argument.
3. By default, runWithPromise() doesn't support functors
with overloaded operator()().
In case of overloaded functors the user
needs to explicitly specify the result type
as a template parameter passed to runWithPromise,
like:
struct Functor {
void operator()(QPromise<int> &) { }
void operator()(QPromise<double> &) { }
};
Functor f;
runWithPromise<double>(f); // this will select the 2nd overload
Task-number: QTBUG-84702
Change-Id: Ie40d466938d316fc46eb7690e6ae0ce1c6c6d649
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
| |
Both normal and relaxed constexpr are required by our new minimum of
C++17.
Change-Id: Ic028b88a2e7a6cb7d5925f3133b9d54859a81744
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
|
|
|
| |
Deprecated in 5.14 in favor of loadRelaxed() and storeRelaxed().
Caught one surviving use of load() in the ios platform plugin.
Change-Id: I9518064a948e5d26ccb956490cbb0561bed5d8b5
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
| |
A few new files were added with old-school defines.
Change-Id: Ieb2c71e094e55102f3f39fb9551823f36863f5f4
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
Made QPromise::swap public, added free standing swap() for
QFutureInterface and QPromise. Updated QPromise special member
functions. Extended tests
Task-number: QTBUG-84977
Change-Id: I5daf6876df306d082441dbcdf5ae4dee3bfc0ead
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|
|
|
|
|
| |
Change-Id: Ic8db7dc252f8fea46eb5a4f334726d6c7f4645a6
Reviewed-by: Sona Kurazyan <sona.kurazyan@qt.io>
|