This allows building the QVariant tests without the QtWidgets module.
Change-Id: I7cd7e78a60c7bc7614ec16df1abe1e93e45d4923
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
WinAPI GetTempPath() sometimes returns short names
for C:/Users/<user>/AppData/Local/Temp.
Change-Id: I33f991acc06e652ccd484d36a5a384eb776f8395
Reviewed-by: Joerg Bornemann <joerg.bornemann@nokia.com>
This makes it possible to add API for setting the restore policy
per state, or even per property assignment (QTBUG-17861).
This change is fully source compatible with Qt4.
Change-Id: I53628546b070f6fc84891f86e7ad7bd8ef5ba285
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Back when QStateMachine was changed to inherit QState, this
constructor was conveniently left out because setting the state
machine (root state) to be a parallel state group didn't actually
work. But as of commit d281aa6936,
it does work, so add the missing constructor.
Task-number: QTBUG-15430
Change-Id: I68c599baa0ef1bfc869195140cf5daf645e75b8b
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
under extremely rare circumstances this would have actually failed
Change-Id: I4132d0f82e9f924e92e9e96f6d34451c94a67201
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Since Qt's connections are thread-safe, QStateMachine's plumbing
around them should be thread-safe too.
Change-Id: I8ae91c2edc2d32ca4ed4258b71e5da22de30ed91
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
By default, QStateMachine lazily registers signal transitions (i.e.,
connects to the signal) when the transition's source state is
entered. The connections are established in Qt::AutoConnection mode,
which means that if the sender object lives in a different thread,
the signal processing will be queued.
But if a sender object's signal is used in an out-going transition
of the target state of the queued transition, it's possible that a
second signal emission on the sender object's thread will be
"missed" by the state machine; before the machine gets around to
processing the first queued emission (and registering the
transitions of the new state), a sender object on the other thread
could have emitted a new signal.
The solution employed here is to eagerly register any signal
transition whose sender object is on a different thread; that is,
register it regardless of whether the transition's source state is
active.
Conversely, when a machine's transitions are unregistered (i.e.,
because the machine finished), signal transitions with sender
objects on other threads should be left as-is, in case the machine
will be run again.
This doesn't solve the case where the sender object is moved to a
different thread _after_ the transition has been initialized.
Theoretically, we could catch that by installing an event filter
on every sender object and handle the ThreadChange events, but
that would be very expensive, and likely useless in most cases.
So let's just say that that case isn't supported for now.
Task-number: QTBUG-19789
Change-Id: Ibc87bfbf2ed83217ac61ae9401fe4f179ef26c24
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Some of the transition constructors didn't call the maybeRegister()
function, causing the transitions to be ignored if they were created
when the state machine was running and the transition's source state
was active.
Added tests that cover all possible cases.
Change-Id: If1b593b127bd719e3be4e5a2e6949a780c4e97c3
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
The originalSignalIndex member was not set if the signature had to be
normalized. This caused the SignalEvent passed to onTransition() to
report a signal index of -1.
Improve the signal transition tests so they check both the event
passed to eventTest() and onTransition().
Change-Id: I5331fd1944d53310b6d11eb2fd8713b80faa53a1
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Use the same trick as used for private signals in the models.
Change-Id: I4235788490cae0e3d554565621d145652dc5b0ca
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
The private signals can not be used as function pointers, as
required by the new syntax, so we introduce a parameter which
can only be created privately.
Change-Id: I3d7bb8a163e764d685e8007cba831fb77e3c6855
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Type traits can not be changed durring Qt5 life time.
Change-Id: If69f65ff2113c901580afee91b11ae1b11c13a4f
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Replace a list of QFuture<void>s and a loop that calls waitForFinished()
on each of them with a QFutureSynchronizer<void>, which does exactly that.
Change-Id: I1f2e90169a5b2949bd8cb9d1009a5a7af1500139
Reviewed-by: David Faure <faure@kde.org>
The SCXML spec had a bug that would cause the initial state of a
compound state within a parallel state group to be entered even if
the transition specified another (non-initial) state of the compound
state as its target. This only happened if the transition had
multiple target states.
The bug has been fixed in recent revisions of the SCXML spec. This
commit implements the fix, which is to walk the ancestors of the
transition's target states only after all the target states
themselves have been added, so that the default initial states are
correctly overridden/ignored.
Task-number: QTBUG-25958
Change-Id: Iac532047678c483a4a3996e24dacf30e00f6bbe0
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Remember to register the metatype where we use it, so we don't depend
on another test being run previously.
And skip the setWorkingDirectory test completely on Unix. I don't know
why it needs to be skipped, but if we're not going to verify anything,
don't even try to do anything. This saves us one memory leak at least.
Change-Id: I22e151cc3fa7b4e976972aca8978b88b263d9bee
Reviewed-by: Mitch Curtis <mitch.curtis@nokia.com>
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
This is accomplished by introducing dependencies to catalogs.
This requires one API change:
QTranslator::load(const uchar *, int);
changes to
QTranslator::load(const uchar*, int len,
const QString &directory = QString());
Since now, even the load from memory might need a directory if
the memory block contains a qm file with dependencies.
Change-Id: I781f333d07f53bb431d0a7b5fa1abe282dc4d338
Task-number: QTBUG-26138
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
As requested by Winfried Schenke:
"QPoint should have an unary operator+ (the unary operator- exists).
Classes with arithmetic operators should provide a complete set of
operators, because some template code relies on it."
Task-number: QTBUG-22913
Change-Id: Ib0c5105975f56c15f00bb48d83c8d911f5a204ac
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Change-Id: I19d3b2e9a5180b13deb828b55195404ef20be295
Reviewed-by: Daniel Teske <daniel.teske@nokia.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Improve test coverage for QPoint and QPointF.
Separate QPointF tests into their own project.
Change-Id: Id28dc5b85aba9fc179d87b2bca1d99854f27a5ab
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QStateMachine inherits from QState, so it should be possible to set
its childMode to ParallelStates, and it should behave as expected
(the machine should emit the finished() signal when all its child
states are in final states).
Task-number: QTBUG-22931
Change-Id: Ic436351be0be69e3b01ae9984561132cd9839fa7
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
It's legal to set a QFinalState as the initial state. The state
machine should correctly emit the finished() signal upon entering
such a state in the initial transition, and don't do any further
processing.
Change-Id: Ica8d3fadbbde604512ea1136624af54eb3b13b11
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
In the old implementation, property assignments
(QState::assignProperty()) were "second-class citizens".
Assignments were not really integrated into the state machine
algorithm, but rather done as a separate step
(QStateMachinePrivate::applyProperties()). While that was
convenient for SCXML spec transcription purposes, it resulted
in some pretty poor semantics on the user side:
* Properties were not assigned until _after_ both the
QAbstractState::onEntry() function had been called and the
QState::entered() signal had been emitted.
* Automatic property restoration (QStateMachine::RestoreProperties)
did not play nice with nested states (and parallel states, in
particular).
The proper fix is to refactor the implementation to make
property assignments first-class in the core state machine
algorithm (QStateMachinePrivate::microstep()).
In practice, this meant splitting some steps. Instead of calling
exitStates() straight away, we now first only compute the states
to exit (without actually exiting them), and use the resulting set
to compute which properties are candidates for restoration.
Similarly, instead of calling enterStates(), we first only compute
the states to enter (without actually entering them), and use the
resulting set to compute which properties are assigned by the
entered states.
With that in place, the rest was a matter of moving the various
chunks of the old applyProperties() logic to the place where they
belong in the per-state entry/exit.
All existing autotests pass. Added several tests that verify the
desired semantics in more detail.
Task-number: QTBUG-20362
Change-Id: I7d8c7253b66cae87bb0d09aa504303218e230c65
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
This test used requires(contains(QT_CONFIG,private_tests)) in its
.pro file, but did not subtract itself from its parent project
SUBDIRS when private_tests weren't enabled.
Change-Id: Idcd0893c4804a8217e4dd33ba9838ff67e996f58
Reviewed-by: David Faure <faure@kde.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
If the transition has no target states, that means the current state
won't change; hence, property assignments should not be performed.
In particular, properties should not be restored to the values they
had before the state was entered.
Change-Id: I237bbb541f939c272777e70c5f26c886ec457a17
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Previously, a registered restorable property would only be
unregistered if the property was animated (see
QStateMachinePrivate::_q_animationFinished()).
But if a property is set directly, it should also be unregistered;
otherwise, the state machine would use the previously saved (stale)
value the next time that property should be restored.
Change-Id: I5d246aa5355ddd0ba5f81b0186a9f0e4f3bbaa3f
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
Do like QPropertyAnimation and store the QObject in a QPointer.
Purge the assignments list upon state entry and property restore.
Change-Id: I54a56885a2905178ab6aa5cf292b3d25c86b7a97
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
In qdbus_cast(), qMetaTypeId<QDBusArgument> as well as
qvariant_cast<QDBusArgument> are used. They don't depend
on any template argument of qdbus_cast(), so their
definitions need to be available at function template
definition instead of instantiation time.
But the necessary Q_DECLARE_METATYPE(QDBusArgument)
was at the end of the header, after the defintion of
qdbus_cast(), which is too late for conformin compilers.
Fixed by moving it up just after the QDBusArgument
definition.
Similarly, in tst_qdatetime and tst_qvector, the
Q_DECLARE_METATYPE() for Qt::DateFormat and QVector<int>,
and with it the specialisation of QMetaTypeId<>, were
issued after the first use of meta typing; too late for
conforming compilers.
Change-Id: I25ca0b06e68d5184597a22708404a8f2040b2de1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The name is the unique identifier. Code such as
if (oldItem.mimeType() == newItem.mimeType())
really wants to detect whether the item has a new mimetype (name),
not compare static mimetype data such as comments and icons.
Change-Id: I5fe56443295c91e1024c066ad6e7f93d842ae507
Reviewed-by: Wolf-Michael Bolle <wolf-michael.bolle@nokia.com>
Add more test data, merging operator!= tests in with operator==
to take advantage of added data.
Change-Id: If0426a3d01b8800cb7363385dbf3bcb21af5ed8f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Now that Q_PROPERTY with a QObject derived type is more powerful.
This property can be used in QML so that wrappers for proxy models
do not need to be created, such as in the example
at https://codereview.qt-project.org/#change,13007
Change-Id: I6ba676549d2135585d429a28e214fef0b2a6b1f9
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Make test projects declare TEST_HELPER_INSTALLS rather than calling a
function exported by testcase.prf. load(testcase) may be unsafe, as
testcase.prf should be processed after default_post.prf.
Fixes silent disabling of various autotests.
Change-Id: I56b35ffd653a637ad5ab18d64dd1a1edadfac59f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Reviewed-by: Kurt Korbatits <kurt.korbatits@nokia.com>
Just like for the QChar::ByteOrderMark, `ch == QChar::SoftHyphen`
is much more readable than `ch == 0x00ad // (soft-hyphen)`, etc.
Change-Id: I9c85f14cfd979037d35103c3259a435fd729b869
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
This will make it possible (in Qt 6) to remove the enums listing
metatype ids. As it is constexpr, it can be used in switch statements
just like enums, as enum values, and as template specialization values.
Change-Id: I51293674c403714e34cb8a8b8953522fc97a740a
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
canConvert() and convert() use the metaobject to convert such types.
Change-Id: Ic05e74c5c2423b4b9682b88adc856a16dcba4cff
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
This will allow conversion between pointers to compatible QObject
derived types.
Change-Id: I19e08934571fb3f1b91e594892214041fe5f6a11
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
The dependencies on QFont, QBrush, QIcon are all in QtGui, so there's
little sense to still have these classes in QtWidgets.
This also copies and pastes a version of QWidgetItemData as
QStandardItemData inside qstandarditemmodel_p.h.
Change-Id: Ibafc5a30748e7ce0b54753309ae6dc4a797fc20e
Reviewed-by: Kent Hansen <kent.hansen@nokia.com>
Store the signal index in QObjectPrivate::Connection, thereby making
it available in "implicit" disconnect contexts (i.e., receiver
deletion).
This change does not cause the size of QObjectPrivate::Connection
to grow (still 40 bytes on 32-bit Linux, 72 bytes on 64-bit Mac).
Valgrinding the new benchmark indicates that the percentage of the
time spent in the QObject destructor increased from 7.8% to 8.4%
on ia32, for that particular stress test; the increase is the
combined cost of calling metaObject(), QMetaObjectPrivate::signal(),
and disconnectNotify() for one connection. In practice, the measured
wallclock time increased by about 3ms for a 500ms run (which
repeatedly constructs, connects, and destroys an object).
Task-number: QTBUG-4844
Change-Id: I1beb01c753f31542fc0acb62edb4c6d165fcc5b4
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
When opening a QFile on stdout, for example,
we must not call seek as it is a sequential device.
This has been flagged as a warning since commit Ie3a96d3a
and has resulted in spurious warnings being emitted.
In the case of opening a QFile in Append mode, QIODevice::open
already sets the position marker, so calling seek is redundant.
This is also true for the file engine's open function (called
through openExternalFile()), which also ensures the handle or
descriptor is repositioned appropriately.
Task-number: QTBUG-26104
Change-Id: I71040c399efe54e7538f54433368b432e959e08d
Reviewed-by: Kent Hansen <kent.hansen@nokia.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@nokia.com>
Updated removeGroupSeparators(QLocalePrivate::CharBuff *num) so that it
removes also positive sign ('+') at the start of the string. Auto test
included.
Task-number: QTBUG-26035
Change-Id: I8e0e071d6c682d9192a8c6bb2f282510e21b3c48
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
QVector::erase shouldn't try to make sense of iterators it doesn't own,
so the validation being done here is bogus and dangerous. Instead, it's
preferrable to assert, the user needs to ensure proper ownership.
The case of erasing an empty sequence is not checked for preconditions
to allow
QVector v;
v.erase(v.begin(), v.end());
, while being stricter on other uses.
Autotests were using ill-formed calls to the single argument erase()
function on an empty vector and were fixed. This function erases exactly
one element, the one pointed to by abegin and require the element exist
and be valid.
Change-Id: I5f1a6d0d8da072eae0c73a3012620c4ce1065cf0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This makes things easier for developers touching QtCore and running
all QtCore unit tests.
Change-Id: I7aa832a6a1be07d90cacad2eecb2364285ff3818
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
This allows QVariant/QMetaType software (such as QtDeclarative) to
deal with smart pointers in a similar way to how they can deal with
naked pointers (accessing properties etc).
This also adds a requirement that T be fully defined when
QSharedPointer<T> is inserted into a QVariant.
Change-Id: I29e12b8a6aa5f4aadbd62f92b89bc238f64b5725
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
The T must be derived from QObject, or it will fail to compile.
This will allow scripting or other 'wrapping' and runtime environments like
QtDeclarative to handle QSharedPointers to types derived from QObject
properly. A QSharedPointer<T> can be inserted into a QVariant, and
where T derives from QObject, a QSharedPointer<QObject> can be
extracted from the QVariant, and its properties are then accessible.
Change-Id: I68d6d89aceceb019267bd7301baa2047f9c09b90
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Currently QDateTime::fromString and QTime::fromString do not correctly
handle fractional minutes and, in some cases, fractional seconds.
In the case of reading fractional minutes, it has been decided to
ignore invalid characters outside of the 5 character portion that
we're interested in (see code comments in fromStringImpl() for
info on why we read 5 digits). The motive is that there is a
performance penalty for calling mid to get the portion of surplus
string and also for converting to it to a float. This is also in
line with what QDate does with surplus characters, for example.
Task-number: QTBUG-14418
Task-number: QTBUG-25387
Change-Id: Ib742fe80686aff3c3770b995678cf838fb4e3bb4
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Clean up and consolidate different tests in tst_qdatetime.cpp
(that seem to be doing the same thing) into single tests.
Change-Id: Ib6ceb1cb7fb4c6eca672495f96d9cfd907853c85
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QDate::toString(Qt::ISODate) lacks prefixed 0's on years below 1000.
The ISO 8601 standard dictates that this should be the case.
Task-number: QTBUG-16476
Change-Id: I7e73152bba0f5894bcbaa3f4418732b74ce86bc5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Verify that this produces a warning containing the relevant typename.
Change-Id: I046c02585e410a211e9175600b1027dda83bdd9c
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
this cleans up a lot of hacks supporting the build of qt, including the
last bits of $QTDIR.
Change-Id: Id119886ed8097967dad6cf86ebd4e71d90c42841
Reviewed-by: Joerg Bornemann <joerg.bornemann@nokia.com>
The standard streams can be redirected to a file, so don't assume
anything, but try to get the actual size and position from the OS and
from the C library (stdout is usually buffered, so the result of lseek
might be different from ftell).
Change-Id: Ice4a0aa21726671928f56a13cc07cc0e4b52091d
Reviewed-by: Richard J. Moore <rich@kde.org>
I didn't do this earlier since the current test data doesn't contain any SMP code points,
the Unicode 6.2 test data does - so, I can confirm this code really works.
Change-Id: Ieae35e8480a89e22d846fd038e79592fefbbf2ee
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Instead of expanding the scripts table with script values for the code points
>= 0x10000, it has been merged with the properties table in order to
increase perfomance of the script itemization code (not affected yet).
(Stats: the properties table grew up in 97428-89800 = 7628 bytes;
the old scripts table was of size 7680 bytes)
The outdated ScriptsInitial.txt and ScriptsCorrections.txt file has been removed
(they were just empty, the "corrigendum" script corrections should be applied
to Scripts.txt directly, *no customization allowed*!).
More script testcases has been added - at least one per supported script.
Task-number: QTBUG-6530
Change-Id: I40a9e76f681e2dd552fd4c61af0808d043962e79
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
This extension doesn't work for e.g. default arguments
in function declarations.
Change-Id: I32b7afa6e01b6af55fb2409179b4fd94cb04cd8d
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Completing the work of the previous commit: we don't need separate
classes. Merge into the main class's body.
Change-Id: I2f89b34cb6b7f5f9e8d8b809bebd86656f458644
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
This allows a QSharedPointer to be used in contexts where the class in
question is still forward-declared. This produced a warning in Qt 4 due
to the expansion of the template, even if there was no chance of the
pointer being deleted there (because the reference count could not drop
to zero).
Now, not only is the warning removed, but you can actually have the
reference count drop to zero in a forward-declared class and it will
do the right thing. That's because the deleter function is always
recorded from the point of construction and we're sure that it wasn't
forward-declared.
The unit test for forward-declarations had to be rewritten. The
previous version was passing only because the QSharedPointer object
was created under the "tracking pointers" mode, which causes a custom
deleter to be used in all cases.
Task-number: QTBUG-25819
Change-Id: Ife37a4cea4551d94084b49ee03504dd39b8802c1
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
QSharedPointer doesn't work like the other automatic template metatype
declarations because in some cases T* is declared as a metatype, but we
are interested in QSharedPointer<T> (eg QObject*). In other cases, T is
declared as a metatype and we are interested
in QSharedPointer<T> (eg char).
In particular the macro used before this patch was attempting to get the
metatype id of the element_type using for example qMetaTypeId<QObject>()
instead of qMetaTypeId<QObject*>(), which did not work.
Similarly, the variadic macro driven test is no good, because it was
testing QSharedPointer<QObject*> instead of QSharedPointer<QObject>,
so that is removed.
In the end, the only thing we can sensibly automatically declare as
metatypes are QSharedPointers to QObject derived types. That is also
the type that makes the most sense in a QML context anyway.
Change-Id: I13dd40147e2e6bedf38661f898102abaaaa96208
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Some test functions that only test QDate and QTime were in
tst_qdatetime.cpp. Upon moving these into tst_qdate.cpp and
tst_qtime.cpp, there were already some similar tests so I
consolidated them.
Change-Id: I5f8758bf8b4804ae9d3a482f49d21de9f7a1dc03
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Most fixes are simple and quite obvious. The ones more involved are
the ones to QArrayData, which had probably not been compiled with
strict iterators thus far.
Change-Id: Ic4ff84c34fd9a04fd686fecaa98149b1c47c9346
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
The QSet<int> is a more expensive container to use and create, so
it should be avoided.
This is source incompatible compared to earlier Qt 5 for
QAbstractItemView subclasses which reimplement dataChanged, but this
patch changes nothing compared to already-present SiC compared
to Qt 4.
Change-Id: Id95391dfd62a0a7f487a8765790b007badefb937
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Containers are auto-registered and use normalized names.
Change-Id: Id65c3940401f69436929220e1f6a971135e147ed
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
This already worked, but let's have a test so we can be sure it doesn't
regress.
Change-Id: I358b436d216e3ec4310f05ccf4f70f9e7aad3281
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
* The Line Breaking Algorithm implementation conformance tests has been added;
* The Grapheme, Word, and Sentence Breaking Algorithm implementation
conformance tests has been updated.
Change-Id: Ia1a6eef6272d580964cb23788ddf30dfd5f4a5a3
Note: the Line Break test data contains some extended cases we don't currently support;
just skip them for now.
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Update NormalizationTest.txt data file with one from UCD 6.1;
Add few more QChar::unicodeVersion() testcases;
Add some line break class mapping testcases;
Add some exceptional case mapping testcases;
Add script class mapping test;
Change-Id: I164394984abb2b893c8db62fb77e7bd87aa0850b
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
According to ISO 8601 (section 4.2.2.3), seconds can be omitted
from a string representing time.
Task-number: QTBUG-2813
Change-Id: I2578f290845e46a8f49be489f1d7427984ae7f08
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
SoftHyphen enum value was added to specify such a boundary reason
Change-Id: I4248909eed6ab8cbca419de4dcf9fe917620a158
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Because QSystemLocale::fallbackLocale() is about UI languages,
it makes sense to check LANGUAGE as well if appropriate.
Adapt tst_qlocale.cpp accordingly.
Suggested by Oswald Buddenhagen.
Change-Id: Ib2c9674081809e3251be4e34456b05210eebc010
Reviewed-by: Oswald Buddenhagen
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
postDelayedEvent() and cancelDelayedEvent() are marked as thread-safe
in the documentation. Unfortunately, they didn't actually work when
called from another thread; they just produced some warnings:
QObject::startTimer: timers cannot be started from another thread
QObject::killTimer: timers cannot be stopped from another thread
As the warnings indicate, the issue was that postDelayedEvent()
(cancelDelayedEvent()) unconditionally called QObject::startTimer()
(stopTimer()), i.e. without considering which thread the function
was called from.
If the function is called from a different thread, the actual
starting/stopping of the associated timer is now done from the
correct thread, by asynchronously calling a private slot on the
state machine.
This also means that the raw timer id can no longer be used as the
id of the delayed event, since a valid event id must be returned
before the timer has started. The state machine now manages those
ids itself (using a QFreeList, just like startTimer() and
killTimer() do), and also keeps a mapping from timer id to event
id once the timer has been started. This is inherently more complex
than before, but at least the API should work as advertised/intended
now.
Task-number: QTBUG-17975
Change-Id: I3a866d01dca23174c8841112af50b87141df0943
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
The function returns mutable iterator on the object that can later be passed to
e.g. erase(), hence it should detach() to be consistent with
QJsonObject::begin() which also detaches.
Change-Id: Id79e8e012fd5469e06b68fbc9eecb7c6848ce9c1
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
The use of QWeakPointer for tracking QObject pointers is to be
deprecated.
Change-Id: If460ca7f515db77af24030152f4bd56e1a5fae7c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
If a QTemporaryFile is constructed using a template file path,
the path is generated in QTemporaryFileEngine::open() and then
filePathIsTemplate is set to false. If remove() and then open()
are called on the same QTemporaryFile, the path is not regenerated.
This change ensures that if the file path was generated, it will be
generated again in the scenario above.
Task-number: QTBUG-2557
Change-Id: I718ceb89daa9a9d46fdbe811fecc3d57d6dc08c2
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Currently, QString::toFloat() returns 0 (and sets ok to false) if you
try to convert "inf". This is because inf is greater than QT_MAX_FLOAT
and there is currently no check to handle inf.
Task-number: QTBUG-8629
Change-Id: I498daf4a7a6f880f928461fca628fcaf7d1d6d08
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
The existing tests has been retained, the new ones has been added.
Change-Id: I12ae1b4e63dde46f3b14a7c1423c13d5881d4507
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
ISO 8601 section 4.2.3 states that "The end of one calendar day [24:00]
coincides with [00:00] at the start of the next calendar day", so
fromString() was updated to account for this.
Task-number: QTBUG-25387
Change-Id: I391db0da755dbc822ba0820c302a2c10391e1f3b
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The documentation says that started() "is emitted when the state
machine has entered its initial state", but the implementation
didn't adhere to that.
The consequence is that if you e.g. emitted a signal from a slot
connected to started(), and that signal was used by a transition
from the initial state, the signal would effectively get ignored and
the state machine would remain in the initial state.
Task-number: QTBUG-24307
Change-Id: Ibbeb627d517eaff821d88e256a949eacf6aae350
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
A QObject can't be a child of itself, so the comparison always
returned false. In practice, this was causing the entry/exit order
of parallel states to be random.
QObject::children() is documented to contain the children in the
order in which they were added, so this fix actually achieves
deterministic behavior.
Task-number: QTBUG-25959
Change-Id: Id3f12d6bfbc249f1d4fed0bafb7d0217093e458e
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@nokia.com>
tryRun() already implies tryLink() and tryCompile(), so there is no
point in executing the stages separately ...
Change-Id: Id7321efaca474e8c5db2bc246ac26323d8a99e58
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Internally, QObject and QMetaObject already leave out non-signal
methods when working with signals. This is possible because the
signals always come before other types of meta-method in the
meta-object data. Ignoring irrelevant methods is faster and can
save memory.
QMetaObject provides internal indexed-based connect() and
disconnect() functions. However, these functions currently take an
absolute method index as the signal specifier, instead of an
absolute _signal_ index. Hence, QMetaObject and friends must convert
from the method index range to the signal index range.
By providing an API that only considers signal indices, clients of
the index-based QMetaObject::connect()/disconnect() can provide the
proper signal index directly. Similarly, for the qtdeclarative
integration (QDeclarativeData hooks) the signal index can be passed
directly. This will eliminate most of the conversions back and forth
between signal index and method index, and some other redundant work
done by qtdeclarative's custom connection implementation.
There are some places where the behavior can't be changed; for
example, QObject::senderSignalIndex() will still need to return an
index in the method range, since that function is public API.
Changing QMetaObject::connect()/disconnect() to take an index in
the signal range will be done in a separate commit; this commit is
only an enabler for porting existing usage of those functions to
the new behavior.
Change-Id: Icb475b6bbdccc74b4e7ee5bf72b944b47159cebd
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Chars that have a case conversion that converts
them into several characters can't be handled
by QChar::toUpper() etc and should get ignored. The code
didn't do that correctly.
Change-Id: I281d122e90bf49187b6449088d2fccef2ef75e86
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Initialiser lists were not tested before in the QVector rewrite, so
the older malloc call was left behind.
Also, std::initializer_list has const iterators returning const data
and broke the build in a few places where const qualifiers were
missing.
Change-Id: I3c04e58361989aa7438621cda63c7df457d7dad8
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
This autotest appears to be parallel-safe.
Change-Id: I12f9202633941e9339de0709353efb2b41df4fa1
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
Previously, the append functions in QConcatenable in the QStringBuilder
dereferenced the data() pointer of the argument QLatin1String without
performing null check.
Change-Id: I629f19fbce3113f1f80f4272fa7ae34e1dbc6bee
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
tst_QStateMachine::postEventFromOtherThread() assumed that 10,000
iterations of the event loop in a separate thread could always be
completed successfully within the default QTRY_COMPARE timeout.
This may be false if the CPU load is already high - such as when running
other tests concurrently.
Decrease the iteration count.
Change-Id: I780523f73c0c16fa0b2ab3201b2b0affdebc198d
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
Write to a QTemporaryDir instead of the test's build directory.
Change-Id: Ib65a0d58fbdf8caf8f2cb7002aeed1ce34742183
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
This test failed a parallel stress test, but seemingly only because it
writes to its own build directory. This should not interfere with other
running autotests.
Change-Id: I27a2f31e32a5b8157ef1082cf0e939bcc0c61c70
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
This test failed a parallel stress test, but seemingly only because it
writes to its own build directory. This should not interfere with other
running autotests.
Change-Id: I80e548fdb0e915ebe86dcd2205537cb6fee09cff
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
This autotest appears to be parallel-safe. It fails our parallel stress
test, but only because it writes to its own build directory. This
should not interfere with other autotests.
Change-Id: Ie99dde24edc0fda0c8ec4352a6e44abb7cbc54f8
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
This autotest seems to be parallel-safe. It was not marked as such due
to an issue which rarely causes the test to hang on exit on Windows, but
that appears unrelated to whether or not the test is run in parallel.
Change-Id: I30bac75be3ddc14139594605481eb6af3f6795e7
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
These tests have passed a parallel stress test on all three of Linux,
Mac, Windows. Mark them with CONFIG+=parallel_test to allow CI to run
them in parallel, saving time.
Change-Id: I19fd333c3c645a67374ca998f6c8530dd236b0f8
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
tst_qprocess::lockupsInStartDetached sometimes locks up on mac.
Mark this as a known issue.
Task-number: QTBUG-25895
Change-Id: I08b1bcf39f2bf373e74509a06415d9ba514b8993
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
MSVC2010 32-bit (with and without service pack 1) takes about 1 hour to
compile this file in some builds, since
1c7421ad14.
Avoid the relevant portion of the code just for these compilers.
Change-Id: Icbb4fa12a6d563a7cdc882c30cdb5705675bedb0
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
When given an invalid url, the output shouldn't be a valid url.
KDE's kurltest detected this regression compared to Qt4, where
all invalid urls were empty in toString() -- but we don't want that,
to give as much feedback as possible to the user.
Change-Id: Ie53e6e1c0a1d4bb9e12b820220dfb7e2f7753959
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
These tests have failed a parallel stress test and may contribute to
instability in test runs.
Change-Id: Ibbbe01f7d9550b953fc9fbd6ed52fc99fdb5f5d7
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
QChar is actually a ushort and passing it via const-ref is suboptimal
Change-Id: Ib806b90397de6a816142ed130a22c0fe10a85d79
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
This allows the QUrl component getters to return fully decoded data,
like they did in Qt 4. This is necessary for some use-cases where the
component like the user name, password or path are used outside the
context of a URL. In those contexts, the percent-encoded data makes no
sense, and the loss of data of what could be represented in a URL is
acceptable.
Also take the opportunity to expand the documentation of those getter
methods, explaining what the options argument does.
Discussed-on: http://lists.qt-project.org/pipermail/development/2012-May/003811.html
Change-Id: I89f743cde78c02f169c88314bff0768714341419
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: David Faure <faure@kde.org>
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
This allows one to instruct QUrl to ignore the percent-encodings and
interpret the data exactly as provided. This is useful in certain
use-cases where the data comes from a non-URL context.
The strict-mode checking of the components is not implemented
yet. Currently, the behaviour is equal to that of TolerantMode.
Discussed-on: http://lists.qt-project.org/pipermail/development/2012-May/003811.html
Change-Id: Ia5abe045a8ce7f9b50cbce3b5a7e3735e068d03a
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
Since we're about to introduce QUrl::FullyDecoded, this
QUrl::MostDecoded value would be confusing. Replace its uses with what
was intended at the point in question.
Change-Id: Iefd87bc33d37bace507c5cb0f206fa902e08e2df
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: David Faure <faure@kde.org>
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
This was trying all the possibilities by brute force, but it turns out
that some combinations are not valid so they should not be
tested. What's more, it was using old values of the flags, so this was
actually testing nothing.
Change-Id: I6c2f5230d240fc23418df2d3a1ca905dbc47dd10
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: David Faure <faure@kde.org>
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
These tests have failed a parallel stress test and may contribute to
instability in test runs.
Change-Id: I2c4456ad7d3846c2262a0ba714ab8f0c9a05c597
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
A (probable) typo was causing the code dealing with anchors
to use uninitialized values. This used to work by chance, but was
indeed detected by Valgrind f.i. when running tst_qregexp --
the indexIn test on anc11 data reported:
==3015== Conditional jump or move depends on uninitialised value(s)
==3015== at 0x514B4EA: PeppeQt::QRegExpMatchState::testAnchor(int, int, int const*) (qregexp.cpp:1813)
[...]
==3015== Uninitialised value was created by a stack allocation
==3015== at 0x514B3EB: PeppeQt::QRegExpMatchState::testAnchor(int, int, int const*) (qregexp.cpp:1803)
Fixing the code also makes the aforementioned test to succeed.
Change-Id: If7b3e518c1bbfcf12573d2637c33ef2eca27c4d5
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
QIcon has been moved back from QWidget to QtGui, so the QIcon QVariant
and QMetaType handler can now be moved back to QtGui.
Also we can give back QIcon its old number, allowing to get rid of some
compatibility hack when unstreaming QVariant
Change-Id: I439d5c2987c06ecd619f394407850f678164afb8
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
The CI system is now using the parallel_test flag to run tests in
parallel. This test has become flaky, or at least more flaky than it
was previously. Mark it to no longer run in parallel.
Change-Id: I47bca3be620a8f648a0eb9c9b9f26d2d925efc01
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The parser is recursive and too deeply nested json would
cause it to exhaust the available stack space leading to
crashes.
We now abort parsing with a DeepNesting parse error if the
document is too deeply nested. The current nesting limit
is set to 1024, which should be more then enough for any
real JSON data set.
Change-Id: I4adea3fd727149f7342536d73cf4530361a0a3a1
Reviewed-by: Jamey Hicks <jamey.hicks@nokia.com>
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
+ QChar::LastValidCodePoint enum value that supercede the UNICODE_LAST_CODEPOINT macro
replace uses of hardcoded values with the new API; remove leftovers
Change-Id: I1395c9840b85fcb6b08e241b131794a98773c952
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
inline all non-static members to a static ones (declared with QT_FASTCALL),
ushort converts automatically to uint and the conversion cost is minimal.
Task-Number: QTBUG-13052
Change-Id: I189a6f205736766adcd3de2d61cee71f30cc64f3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When qt_ntfs_permission_lookup is used, QFile::permissions failed
for files with long filenames.
Also created a test case for this API, which revealed another bug.
Task-number: QTBUG-25629
Change-Id: I73b7676a9d059c0e782b3f701b2e6bbc92f671ed
Reviewed-by: Prasanth Ullattil <prasanth.ullattil@nokia.com>
When a file is specified on a path that includes a drive letter
followed by a colon but no slash then it didn't always account
for the fact that this refers to the current path on that drive.
This fixes the problems in completeBaseName(), baseName() and
path(). Tests are also added for these three cases and some
others too.
Task-number: QTBUG-25353
Change-Id: I47a197c6af066f532442ad269be57597ec61303a
Reviewed-by: Irfan Omair <irfan.omair@gmail.com>
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
On some systems prior to this change the code
would not build with the following error reported
by gcc:
error: integer constant is too large for 'long' type
Change-Id: I778bce9a72ccf3a41cdf17883d734082ed3fb4b3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Name is about mimetype names, while FileName is about, well, file names.
Task-number: QTBUG-25748
Change-Id: I34a9ac1a5fc06dc3e3855365e19c4dc7a1aa2671
Reviewed-by: Wolf-Michael Bolle <wolf-michael.bolle@nokia.com>
This completes the transition from connectNotify(const char *) and
disconnectNotify(const char *) to the new QMetaMethod-based
functions.
Removed the old connectNotify autotests and renamed the
connectNotifyMethodXXX autotests to connectNotify, since there is
no longer any ambiguity about which overload is being tested.
Change-Id: Icf108a80177155f21bb73c165fb8ab5d4e997bc2
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
MinGW installations on case-sensitive filesystems expect
lowercase names of include-libraries and (usually) include
files.
When crosscompiling on Debian 6 (targeting MS Windows) linking
fails because mingw is looking for non-existent include-libraries.
Using lowercase names solves this.
Change-Id: Id3454f4ed8ba42b6ea93d65d9c0ce567db6712df
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
QUrl::fromLocalFile("/foo") doesn't set Host, but QUrl("file:///foo")
does (to remember that it saw a Host section, even if empty, which is
useful for urls like "remote://"). So ignore the Host flag in operator==.
Change-Id: I4322b4a75420c4e42766c0d65c1b121f28028a76
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Fix operator== and operator< so that a URL with an empty fragment
or query, is not treated as equal to a URL without any fragment or query.
This restores the Qt4 behavior on this particular issue.
Change-Id: Ie989f37353fb13c791b1d558d638d2e8a5b5d1b8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Null bytearray means no query, and QString::fromLatin1(QByteArray())
doesn't give a null string, but an empty string.
Same for setEncodedFragment(QByteArray()).
Change-Id: I992e9253e35941d66886456872ea06aa2ae92450
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
A leading byte order mark is valid in utf-8 and we should
parse documents starting with those correctly.
Change-Id: Id85398ff6e05b93ceefbaf4a6de5571d5e61ca13
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
there are several reasons to do this:
* text breaking is not a shaper's job;
* since the text breaking rules are bound to a specific Unicode version,
updating Qt's internal unicode data would require updating the data in HB as well;
* makes porting to HurfBuzz-NG some easier
Change-Id: I0bbf8e8a343bc074696f4ddf2ae4e7fa32a61629
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
This was crashing because the ':' was found past the end of the
username, causing the recoder to run from position 22 to 11, via the
long way around the memory.
Change-Id: Ic1ae596f34f7db857fb4210294974fb5a6adf691
Reviewed-by: Alexis Menard <alexis.menard@openbossa.org>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
If the url we pass as parameter already have percentage encoded data,
we don't want to decode it and call fromPercentEncoding. The test
coverage is not complete for qdataurl.cpp file but it is better than
previously and it will also protect us from future regressions.
Change-Id: I79f709f44bed1b7f274a3de639c7e291fa91a193
Reviewed-by: Thiago Macieira
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
19d160b72b broke it temporarily, and this wasn't detected by
tst_headersclean, because it sets QT_NO_CAST_FROM_ASCII too,
which disabled the faulty code.
So this adds a new unittest for QT_NO_CAST_FROM_BYTEARRAY alone.
Change-Id: Iaf7a36a1378e77188bcc636e5dc9a1f9b84f70a7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
C++ distinguish between "char", "signed char" and
"unsigned char", they are three independent types.
Fix QVariant behavior on ARM. On ARM "char" may mean
"unsigned char", but we depends on the sign during
a numerical conversions.
Change-Id: I610ce3fb88ed5964b67f3ae442d264fe16b2d261
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The implicit cast to QJsonValue was being ignored probably because the
compiler was generating a default QJsonValueRef assignment operator
Change-Id: I3a041595497308868dd7e4aab71027ce21bf8f0b
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
This also tests by consequence that the behaviour of QByteArrays
containing NULs is consistent. Right now, that means the QByteArray
processing stops at the NUL, which is the same behaviour as if a
pointer to the byte array's data were used. (it's what happens if
there's no QByteArray overload and the const char* one is called)
Change-Id: If56a822f95866e8cb5b153d07b48198bb83fb386
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
This commit completes the previous commit so that both QString and
QStringBuilder now operate on UTF-8 input.
A small fix was required in QStringBuilder: an if clause isn't enough
to separate the two append versions. Since there are no QString
functions that append to char*, if we're converting to a QByteArray,
we need to go through a QString first in a separate function.
Change-Id: Ic503340c5d0c32d420c90c91cc2e0fc1ae9230f3
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
> http://www.unicode.org/versions/Unicode5.2.0/
D. Character Additions:
There are three new characters in the newly-encoded Kaithi script that will
require changes in implementations which make hard-coded assumptions about
composition during normalization. Most new characters added to the standard
with decompositions cannot be generated by the operations toNFC() or toNFKC),
but these three can. Implementers should check their code carefully
to ensure that it handles these three characters correctly.
U+1109A KAITHI LETTER DDDHA
U+1109C KAITHI LETTER RHA
U+110AB KAITHI LETTER VA
UCD 6.1 adds two more of them:
U+1112E CHAKMA VOWEL SIGN O
U+1112F CHAKMA VOWEL SIGN AU
Change-Id: I781a26848078d8b83a182b0fd4e681be2a6d9a27
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Show that nothing is changed either way, regardless of the encoding
flags used.
Change-Id: I31fba5f87eae777d4b708ab789b32169004bcbcc
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
It should only strip one slash (as the name indicates), and not if the
path is just "/".
Change-Id: I133a81977241de77a49d1d1559143d30e0bd52f8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This operation should be a no-op anyway, since at this point in time,
the fromAscii and toAscii functions simply call their fromLatin1 and
toLatin1 counterparts.
Task-number: QTBUG-21872
Change-Id: I38f97ad379deafebef02c75d611343ca15640c8a
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
This way, QWinEventNotifier will work on all Windows systems, not just
with the default event dispatcher. Other dispatchers (other than
QWin32EventDispatcher) are permitted, so the class should not abort just
because of that.
If a dispatcher really doesn't want to implement this, they need to
implement the virtuals to do nothing, possibly print a warning.
Change-Id: I2c132bcde95b9d5941c8906a0fcd2ad964087772
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
This is much more performant than calling QObject::receivers(const char*)
Can be used instead of connectNotify in some cases.
Change-Id: I19e0933f678f171f515d9a0f69f0ad4fb7d894b4
Reviewed-by: Kent Hansen <kent.hansen@nokia.com>
qVariantValue and qVariantCanConvert are Compatibility members, while in
Qt4.8 they are marked as Qt 3 Support Members.
qVariantFromValue and qVariantSetValue are Obsolete members.
Change-Id: Ie8505cad1e0950e40c6f6710fde9f6fb2ac670fd
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@nokia.com>
Reimplementations of connectNotify() and disconnectNotify() can
assume that the signal argument is in normalized form, but after the
introduction of the Qt5 meta-object format, it could happen that it's
not.
The problem is that the internal QArgumentType class, which attempts
to resolve a typename to a type id, was calling QMetaType::type().
QMetaType::type() falls back to trying the normalized form of the
typename if the original argument can't be resolved as a type (this
behavior isn't documented, but that's how it works). This means that
e.g. QMetaType::type("const QString &") returns QMetaType::QString.
Since QMetaObjectPrivate::indexOfMethodRelative() (more specifically,
the methodMatch() helper function) prefers to compare type ids
over typenames (since the type ids are stored directly in the meta-
object data for built-in types), the method lookup would *succeed*
for signatures with non-normalized built-in typenames as parameters.
QObject::connect() would then think that it did not have to
normalize the signature (see "// check for normalized signatures").
The consequence was that the original, non-normalized form got
passed to connectNotify().
This commit introduces an internal typename-to-type function that
is the same as QMetaType::type(), except it doesn't try to normalize
the name. This way, the only place where normalization can occur in
the signature-to-meta-method processing is through the calls to
QMetaObject::normalizedSignature() in QObject::connect() itself.
The implication is that there are now cases where the method
signature will be decoded and processed twice, where processing it
once was sufficient before. On the other hand, it is consistent with
the pre-Qt5-meta-object behavior, where we predict that the
signature is already normalized, and only perform (comparatively
costly) normalization if the initial lookup fails.
Change-Id: Ie6b60f60b0f9a57ebd378d980329dac62d57bbd9
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
static_metacall was never set on the metaobject written by
QMetaObjectBuilder::fromRelocatableData, sometimes causing a crash. It
should be initialized to 0.
Change-Id: I79373d895e131f0cc2ff1af6d2177a0c1a282be7
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
This API will fully replace the const char *-based connectNotify()
and disconnectNotify() in Qt5; the old functions will be REMOVED
before Qt 5.0 final.
The new implementation fixes the long-standing issue of
connectNotify() not being called when using the (internal)
index-based QMetaObject::connect() (e.g., from QML).
As with the old API, there are still two "unintuitive" behaviors
concerning disconnectNotify():
- disconnectNotify() is not called when the signal is disconnected
using the QObject::disconnect(QMetaObject::Connection) overload.
- disconnectNotify() is not called when a receiver is destroyed
(i.e., when a connection is implicitly removed).
The old versions of connectNotify() and disconnectNotify() are kept
for now, and they are still called. They will be removed once known
existing reimplementations (e.g., QtNetwork, QtDBus) have been
ported to the new API.
Change-Id: I8b4f007f3c6d89199c1ba04a3e23c8ca314e0896
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Turns out that we've had some old unit tests commented out that did not
compile. QString does not have a std::string constructor nor overloads
to many other methods. And std::string does not cast to char* on its
own. So these tests need to be removed.
Change-Id: I22df66fc3ccc68bc2840f2d83747234418e480f5
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Remove all non-UTF-8 sequences from source code in Qt.
Change-Id: I46d9cb23ef2199894896f171d553b3144822f36c
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Given a member function that's a signal, returns the corresponding
QMetaMethod. Inspired by the implementation of the template-based
QObject::connect().
The primary use case for this function is to have an effective and
exact (not subject to shadowing) way of checking whether a known
signal was connected to in reimplementations of
QObject::connectNotify(QMetaMethod), avoiding string comparisons.
Example:
void MyObject::connectNotify(const QMetaMethod &signal)
{
if (signal == QMetaMethod::fromSignal(&MyObject::mySignal)) {
// Someone connected to mySignal ...
}
}
Change-Id: I5e4de434275fe543c004d569dcaa9ceda3442f03
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Known failures in this test are now handled by QEXPECT_FAIL.
Task-number: QTBUG-24796
Change-Id: I12ba57370cf3df1a85a108fbbcdc9db2222491c1
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
Unlike path(), toLocalFile() isn't reporting a URL component, so it
should decode the percent-encoded characters fully. This extra
decoding pass is meant to catch %00 to %1F, %7F and %25 (the percent
sign itself).
It also catches %80 to %FF, which aren't decoded because they don't
form UTF-8 sequences. That means QUrl::toLocalFile() has undefined
behaviour if the path contained non-UTF8 sequences.
Task-number: QTBUG-25459
Change-Id: Iab5a0ba6afcfc4510e297984f2ffc208cedd752b
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
QMetaObjectExtraData was added when support for QMetaObject::newInstance
was added. One needed a place to put the pointer to static_metacall in
the QMetaObject.
But as we break binary compatibility, one can change the size of
QMetaObject, and put everything back inside QMetaObject's own structure.
Meaning it is not required anymore to have one QMetaObjectExtraData
instance per QMetaObject anymore.
Change-Id: If0b8f586cbaf633eed10045adee3ba3366826c86
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Kent Hansen <kent.hansen@nokia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
This is done in preparation of introducing the
QObject::connectNotify(QMetaMethod) function. Together with the
forthcoming QMetaMethod::fromSignal() function, which returns the
QMetaMethod corresponding to a Qt/C++ signal (member function), the
comparison operators provide an effective way of checking which
signal was connected to.
Change-Id: I2de48628c4884a7174fb8574895f272cb3fe5634
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Instead of trying to return whether the URL is relative to something
undefined, let's instead follow what the documentation was saying all
along and what the RFC says about "Relative References".
Change-Id: I32722321a6b36c6e3480669ad769390e4c6f7d1c
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Previously, the CI system has not been testing on Windows with the
-developer-build configure flag. Mark known failures for this
configuration so that tests can be run in enforcing mode.
Change-Id: I5fbbbe09a7b400d626107c66dcbd5c5469a45b20
Reviewed-by: Sergio Ahumada <sergio.ahumada@nokia.com>
QLatin1Literal is just a typedef of QLatin1String.
Change-Id: If20ca225e57a7fb45a7775f0fc81aedb6da88c96
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
QUrl::path() already decodes almost everything, but let's pass the
formatting flag to be sure.
Note: decoding of control characters from U+0001 to U+001F is not
implemented. Non-UTF8 sequences are also not representable.
Change-Id: I9a0ae2282ec3d48cc0e70e5b2d3824fb120709ed
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
It's perfectly valid to have a path of /c:/a.txt on Unix, so don't
strip the leading slash unless we're on Windows.
Task-number: QTBUG-20322
Change-Id: I721bd0a65b41048bc735d4eaa0d536174164fe64
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
Since this test began to run again on Mac, it has been passing
consistently.
Task-number: QTBUG-22748
Change-Id: Ia81cf60b11d45fb331b5eca5d13df00556c18e07
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Add qSetMessagePattern() to configure the default
message pattern. This one can still be overwritten by setting the
QT_MESSAGE_PATTERN environment variable.
Without this method, there's actually no way to change the
default output programatically. Since QT_MESSAGE_PATTERN is
evaluated when the first message arrives, setting it via e.g. qputenv
might have no effect/be too late.
Change-Id: I115e0c30606f128fdbf5c169a951ffa2a6a48517
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Up until now, the macros would return an internal type that contained
the pointer to the data. This breaks code that tried to use the macros
with operators, like QStringBuilder but also when writing:
QStringList() << QStringLiteral("a") << QStringLiteral("b");
This change seems to work fine now and I can also verify that this
works:
const auto str = QStringLiteral("Hello");
Even though it creates a QString, which is non-POD and non-constexpr.
Change-Id: Iaf82af9bea4245513a1128ea54f9d2d3d785fb09
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
These tests were disabled when trying to get CI working on Mac OS
because they asserted or crashed. Now that CI is working well on Mac
OS, start running these tests again, initially as insignificant tests.
CI results will then be used to determine whether the tests can be made
significant.
Change-Id: Ife411e6b8c84ade45c865ef35f3ae0071d6f8d2b
Reviewed-by: Sergio Ahumada <sergio.ahumada@nokia.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
These warnings are expected and correct, so ignore them.
Change-Id: I43931950e46bd3c931db869902574ee7219efa1d
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
Adjust the test because we don't read past the end anymore.
Task-number: QTBUG-25108
Change-Id: I8243f1d5ae79d1256aab2cb1132598a716a7eeeb
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
Introduce a new QtMessageHandler that takes QString instead of
char *: This avoids converting to local8bit , only to convert it back
to utf16 for Windows.
The old QMessageHandler is kept for a transition period, but will
be removed before Qt 5.0.
Also fix qEmergencyOut (that is called in OOM situations) to not rely
on the default message handler.
Change-Id: Iee0ce5838f97175c98788b847964273dd22d4a37
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The constructor is wrong, it creates instance of QVariant encapsulating
a QColor instance. QVariant should not implicitly convert data, never.
Change-Id: Idc794ecdecb42d8b53fee3f993bf51ddd43f595d
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
This test hangs ~2.6% of the time in CI.
The previous commit which attempted to mark this test as insignificant
did not work due to this .pro file doing a load(testcase) prior to the
line which set CONFIG += insignificant_test. testcase.prf must be
loaded _after_ insignificant_test is set.
Task-number: QTBUG-25342
Change-Id: I20470337fda8235e2fd0b6e8d5f564c8c57d167e
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
That change also fix moduleForType() which was wrongly recognizing
negative ids as belonging to Core.
New tests were added.
Change-Id: I40a5819effb32489a45937011980457387c9f8be
Reviewed-by: Kent Hansen <kent.hansen@nokia.com>
results are now equals to results of ICU's u_isprint() for the entire set
of the Unicode code points
Change-Id: I763f4b37cccd285eb01543d486f25bd7ea011241
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Archived build logs on http://testresults.qt-project.org/ci/ show that
the QPluginLoader test has failed in CI only once in the last 1000
builds (in build 1786), and on that occasion the test was legitimately
blocking a regression.
The QFactoryLoader test was derived from the QPluginLoader test and has
not failed since its first run on March 29, 2012.
Task-number: QTBUG-22765
Change-Id: I866b4b8e30e393e0c7e7292119c072b27008ab43
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
This test is like qguieventdispatcher, it duplicates a corelib test in
the gui test suite, since the QtGui library often gets a different event
dispatcher implementation from the platform plugin.
Change-Id: Ifd724066950bc3b98a804bc2e5d40ce7b0429af4
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
QTimeLine appears to have very poor timing characteristics. Historical
CI logs show roughly one failure in every twenty-five test runs on
Windows, and less frequent failures on Mac and Linux.
The root of the problem seems to be that QTimeLine's currentTime
counter appears to run at a variable speed and the only guarantee is
that it is slower than wall time. The frameChanged() test
function waited for double the expected duration of the timeline and
still found that the timeline had failed to finish in about one in every
thirty test runs. The interpolation() test function also failed for the
same reason, though less often.
This commit makes the frameChanged test more strict so that the poor
timing will be demonstrated more often, waiting only 1.5 times the
duration instead of double the duration. It also makes the test fail
gracefully so that this known issue won't disrupt CI when the test is
made significant in a later commit.
Task-number: QTBUG-24796
Change-Id: If469d43abb662e24445a9da619052eea9cf7c581
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
QTimeLine::currentTime() is an integer in the range [0..duration], not a
float in the range [0.0..1.0]. The aim of the test appears to be to
verify that currentTime() is at least 90% of the way to duration() when
the timeline is almost due to finish, so verify that and give the
corresponding 10% tolerance on reaching the end state.
Change-Id: I38646947c3b9189a4e8e91a450c6071430ddc66a
Reviewed-by: Rohan McGovern <rohan.mcgovern@nokia.com>
This test hangs ~2.6% of the time in CI.
Task-number: QTBUG-25342
Change-Id: I2c3531140e15edfe2dc2524e101b84e3206a4e61
Reviewed-by: Kalle Lehtonen <kalle.ju.lehtonen@nokia.com>
Otherwise, the order of updating of the indexes will cause
inconsistent results because it will rely on ordering within a
QHash (which is indeterminate).
Task-number: QTBUG-25325
Change-Id: I7d99578c8ee2954b8562dc5aff7dc32e74d41fb5
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
This test has not failed in the last 500 Continuous Integration runs.
Task-number: QTBUG-22769
Change-Id: Ib2e95bb2291757941baa0ea46d568816eef20b09
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
The testChildrenLayoutsChanged fails randomly. This happens rarely,
f.i. wasn't spotted by CI when QHash randomization itself was merged;
but is indeed reproducible by running the test a few times in a row.
This is now blocking api_merges integration, and I have no idea
how to fix it.
This patch marks the test as insignificant for now (the bug
tracking this test failure is QTBUG-25325), and switches the failing
tests from QVERIFY(a == b) to a proper QCOMPARE (so that the
expected values do show up in the build logs).
Change-Id: I16f0e28bcbb06dbac2e7169f4676a19ccf626a92
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
http://unicode.org/versions/corrigendum6.html:
> in Unicode 5.0, the list of characters with the Bidi_Mirrored property
> was made consistent for brackets and quotation marks, in preparation for
> new constraints on bidi mirroring. However, after publication of
> Unicode 5.0.0 it was discovered that this change adversely affected
> several quotation mark characters in deployed data.
Task-number: QTBUG-25169
Change-Id: Id49caf401af2d5a1e6dbcc32b2f350aa20b7f901
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
The key returned by QHash::key is an arbitrary one that maps to the
given value. The test instead relied on it being a specific one.
Change-Id: I090351797e8b52036d78160fd810518a11e8107d
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@nokia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The hash autotest is wrong: it assumed that the iterator on the hash
would reach the end after iterating on two elements with identical key.
But three elements were added to that hash, and the third one
can appear after the other two.
That code path is left for the map test only.
Change-Id: I51de7987e2b132b6caff7bb4bac6a57fb7fcb530
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
That was a regression introduced in 1c5db1aff
Example:
signals: int *someSignal();
would produce this code:
int* _t0 = int*();
which does not compile
So have special handling for pointer to change it to '= 0'
Change-Id: Ie695e15e309d15c3cfd5c5a69ac8bf6d61ae9915
Reviewed-by: Stephen Kelly <stephen.kelly@kdab.com>
This test hangs 2-3% of the time.
Task-number: QTBUG-25284
Change-Id: I32e01696262be2de7e015b8f811d1666551426cc
Reviewed-by: Toby Tomkins <toby.tomkins@nokia.com>
The new QUrl is able to distinguish a URL component that is empty from
one that is absent. The previous one already had that capability for
the port, fragment and query, and the new one extends that to the username,
password and path. The path did not need this handling because its
delimiter from the authority it part of the path.
For example, a URL with no username is one where it's set to QString()
(null). A URL like "http://:kde@kde.org" is understood as an
empty-but-present username, for which toString(RemovePassword) will
return "http://@kde.org", keeping the empty-but-present username.
Change-Id: I2d97a7656f3f1099e3cf400b199e68e4c480d924
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
This test no longer fails, so we can remove CONFIG+=insignificant_test
Task-number: QTBUG-22767
Change-Id: If3ca194fc982ad8fdc3e9a7f62fc346190ff01ea
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
This test no longer fails, so we can remove CONFIG+=insignificant_test
Task-number: QTBUG-22766
Change-Id: I379873d5c483157e414201e5f8a13c3f4407f9fd
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
This does not fail anymore, remove CONFIG+=insignificant_test
Change-Id: I4f98cfad563adfa460910976317c91e852db6872
Reviewed-by: Jason McDonald <jason.mcdonald@nokia.com>
Most of the tests were removed while QUrl::toEncoded or fromEncoded
were deprecated in the development process. Since they aren't
deprecated in the end, bring them back.
Change-Id: Ibdb6cd3c4b83869150724a8e327a03a2cd22580d
Reviewed-by: Robin Burchell <robin+qt@viroteck.net>
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
By having the default value equal to zero, we follow the principle of
least surprise. For example, if we had
url.path()
and we refactored to
url.path(QUrl::DecodeSpaces)
Then instead of ensuring spaces are decoded, we make spaces the only
thing encoded (unicode, delimiters and reserved characters are
encoded).
Besides, modifying the default can only be used to encode something
that wasn't encoded previously, so having the enums as Encode makes
more sense.
As a side-effect, toEncoded() does not support any extra encoding
options.
Change-Id: I2624ec446e65c2d979e9ca2f81bd3db22b00bb13
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
This allows things like http://example.com/{1234-5678}?id={abcd-ef01}.
But do not allow it in other parts of the URL. I could allow it in the
fragment, but in the username and password it would be too ugly.
In order to do that, make DecodeReserved use two bits and have
PrettyDecoded set only one of them. That way, toString(PrettyDecoded)
can be distinguished from toString(PrettyDecoded | DecodeReserved),
just as path(PrettyDecoded) can be distinguished from
path(PrettyDecoded & ~DecodeDelimiters).
Also, take the opportunity to avoid decoding the reserved characters
in the query. Keep them encoded as they should be.
Change-Id: I1604a0c8015c6b03dc2fbf49ea9d1dbed96fc186
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
DecodeReserved applies to all characters between 0x21 and 0x7E that
aren't unreserved, a delimiter, or the percent sign itself.
Change-Id: Ie64bddb6b814dfa3bb8380e3aa24de1bb3645a65
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
There's little value in having the DecodeUnambiguousDelimiters option
since neither QUrl nor QUrlQuery can return values that are ambiguous
in that particular context, ever.
This option could be used to encode a character if, when placed
in a URL, it would need to be encoded. Such cases are hash (#) or
question marks (?) in the path component, or slashes (/) and at signs
(@) in the userinfo.
However, we don't need two enums for that, since there are no
other characters that can appear in either form. Still, leave two bits
for this enum. In the future, if we want to split the gen-delims from
the sub-delims, we are able to.
Change-Id: If5416b524680eb67dd4abbe7d072ca0ef7218506
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
This tests how QUrl encodes and decodes certain characters and leaves
some other ones alone. It also tests that the output of toString() (in
whichever encoding was being tested) is also parsed again to be
exactly the same as the previously decoded form.
Change-Id: Ie358d001f8b903409db61db48bde1ea679241a60
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
This is the same fix as the previous commit did for the other
components of the URL. But we're also changing how we handle the "[]"
characters in a query: previously the handling was like for other
sub-delims; now, they're always decoded, assuming that the RFC had a
mistake and they were meant to be decoded.
Change-Id: If4b1c3df8f341cb114f2cc4860de22f8bf0be743
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
Refactor the way that QUrl stores and returns the components of the
URL so that ambiguous delimiters (gen-delims that could change the
meaning of the parsing) are interpreted correctly. Previously, QUrl
called "unambiguous" the form found in a full URL, even though each
item in isolation could have more characters decoded.
Now, instead, store only the fully decoded form. To recreate the
compound forms (the full URL, as well as the user info and the
authority), we need to do more processing.
This commit applies to the user name, password, path and fragment
only. The scheme, host and port do not need this work because they are
special; the query is handled separately.
Change-Id: I5907ba9b8fe048fff23c128be95668c22820663a
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
0xfdef-0xfdd0 is definitely 31 and not 15 :)
also fix all copy-pastes of this code (greping for '0xfdd0' helps ;)
Change-Id: I8f3bd4fd9d85f9de066f0f5df378b9188c12bd48
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Denis Dzyubenko <denis.dzyubenko@nokia.com>
Just like qMalloc/qRealloc/qFree, there is absolutely no reason to wrap these
functions just to avoid an include, except to pay for it with worse runtime
performance.
On OS X, on byte sizes from 50 up to 1000, calling memset directly is 28-15%
faster(!) than adding an additional call to qMemSet. The advantage on sizes
above that is unmeasurable.
For qMemCopy, the benefits are a little more modest: 16-7%.
Change-Id: I98aa92bb765aea0448e3f20af42a039b369af0b3
Reviewed-by: Giuseppe D'Angelo <dangelog@gmail.com>
Reviewed-by: John Brooks <john.brooks@dereferenced.net>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@nokia.com>
Some FTP implementations (currently not including QNAM) strip the first
slash off the path in an FTP URL so that the path in the URL is relative
to the login path (the user's home directory). To reach the root
directory, another slash is necessary, hence the double slash.
In anticipation of future URL normalisation, which Qt 4 could do, "//"
could be rendered to "/", so this extra slash should be "%2F".
This operation is done only in QUrl::fromUserInput.
Change-Id: If9619ef6b546a3f4026cb26b74a7a5a865123609
Reviewed-by: Shane Kearns <shane.kearns@accenture.com>
QT_NO_STL is now no longer defined, so remove the conditionals and
select the STL side.
Change-Id: Ieedd248ae16e5a128b4ac287f850b3ebc8fb6181
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
Two equal QByteArrays must return the same hash.
Change-Id: Iddd45b0c420213ca2b82bbcb164367acb6104ec8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Two equal strings / stringrefs must return the same hash.
Change-Id: I2af9a11ab721ca25f4039048a7e5f260e6ff0148
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It was confusing DataLocation and GenericDataLocation, and the same
for CacheLocation and GenericCacheLocation. The test was passing in
the api_changes branch because these were giving the same result
(empty app name), but the QCoreApplication::applicationName fix in master
makes these different, so the bug in the test showed up after merging.
Change-Id: I80ef6883c96cfd02b8c277d9d686717028d396bb
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This uses an alternative approach to the testing formerly introduced
in 4ef5a626. Zero-termination tests are injected into all QCOMPARE/QTEST
invocations. This makes such testing more thorough and widespread, and
gets seamlessly extended by future tests.
It also fixes an issue uncovered by the test where using a past-the-end
position with QString::insert(pos, char), could move uninitialized data
and clobber the null-terminator.
Change-Id: I7392580245b419ee65c3ae6f261b6e851d66dd4f
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
The approach used to verify for zero-termination is too intrusive and
requires additional maintenance work to ensure new zero-termination
tests are added with new functionality.
Zero-termination testing will be re-established in a subsequent commit.
This reverts commit 4ef5a6269c.
Change-Id: I862434a072f447f7f0c4bbf8f757ba216212db3c
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@nokia.com>
This enables easier updating of those structs, by reducing the amount of
code that needs to be fixed. The common (and known) use cases are
covered by the two macros being introduced in each case.
Change-Id: I44981ca9b9b034f99238a11797b30bb85471cfb7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There were two constuctors offering essentially the same functionality.
One taking the QStatic*Data<N> struct, the other what essentially
amounts to a pointer wrapper of that struct. The former was dropped and
the latter untemplatized and kept, as that is the most generic and
widely applicable. The template parameter in the wrapper was not very
useful as it essentially duplicated information that already maintained
in the struct, and there were no consistency checks to ensure they were
in sync.
In this case, using a wrapper is preferred over the use of naked
pointers both as a way to make explicit the transfer of ownership as
well as to avoid unintended conversions. By using the reference count
(even if only by calling deref() in the destructor), QByteArray and
QString must own their Data pointers.
Const qualification was dropped from the member variable in these
wrappers as it causes some compilers to emit warnings on the lack of
constructors, and because it isn't needed there.
To otherwise reduce noise, QStatic*Data<N> gained a member function to
directly access the const_cast'ed naked pointer. This plays nicely with
the above constructor. Its use also allows us to do further changes in
the QStatic*Data structs with fewer changes in remaining code. The
function has an assert on isStatic(), to ensure it is not inadvertently
used with data that requires ref-count operations.
With this change, the need for the private constructor taking a naked
Q*Data pointer is obviated and that was dropped too.
In updating QStringBuilder's QConcatenable specializations I noticed
they were broken (using data, instead of data()), so a test was added to
avoid this happening again in the future.
An unnecessary ref-count increment in QByteArray::clear was also
dropped.
Change-Id: I9b92fbaae726ab9807837e83d0d19812bf7db5ab
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Algorithmic complexity attacks against hash tables have been known
since 2003 (cf. [1, 2]), and they have been left unpatched for years
until the 2011 attacks [3] against many libraries /
(reference) implementations of programming languages.
This patch adds a qHash overload taking two arguments: the value to
be hashed, and a uint to be used as a seed for the hash function
itself (support the global QHash seed was added in a previous patch).
The seed itself is not used just yet; instead, 0 is passed.
Compatibility with the one-argument qHash(T) implementation is kept
through a catch-all template.
[1] http://www.cs.rice.edu/~scrosby/hash/CrosbyWallach_UsenixSec2003.pdf
[2] http://perldoc.perl.org/perlsec.html#Algorithmic-Complexity-Attacks
[3] http://www.ocert.org/advisories/ocert-2011-003.html
Task-number: QTBUG-23529
Change-Id: I1d0a84899476d134db455418c8043a349a7e5317
Reviewed-by: João Abecasis <joao.abecasis@nokia.com>
It is an extension coming from the use case when you, for instance, need to
implement a countdown timer in client codes, and manually maintain a dedicated
variable for counting down with the help of yet another Timer. There might be
other use cases as well. The returned value is meant to be in milliseconds, as
the method documentation says, since it is reasonable, and consistent with the
rest (ie. the interval accessor).
The elapsed time is already being tracked inside the event dispatcher, thus the
effort is only exposing that for all platforms supported according to the
desired timer identifier, and propagating up to the QTimer public API. It is
done by using the QTimerInfoList class in the glib and unix dispatchers, and the
WinTimeInfo struct for the windows dispatcher.
It might be a good idea to to establish a QWinTimerInfo
(qtimerinfo_win{_p.h,cpp}) in the future for resembling the interface for
windows with the glib/unix management so that it would be consistent. That would
mean abstracting out a base class (~interface) for the timer info classes.
Something like that QAbstractTimerInfo.
Test: Build test only on (Arch)Linux, Windows and Mac. I have also run the unit
tests and they passed as well.
Change-Id: Ie37b3aff909313ebc92e511e27d029abb070f110
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Bradley T. Hughes <bradley.hughes@nokia.com>
While QArrayDataPointer offers generic detach() functionality, this is
only useful for operations that may modify data, but don't otherwise
affect the container itself, such as non-const iteration, front() and
back().
For other modifying operations, users of the API typically need to
decide whether a detach is needed based on QArrayData's requirements
(is data mutable? is it currently shared?) and its own (do we have
spare capacity for growth?).
Now that data may be shared, static or otherwise immutable (e.g.,
fromRawData) it no longer suffices to check the ref-count for
isShared().
This commit adds needsDetach() which, from the point-of-view of
QArrayData(Pointer), answers the question: 'Can contained data and
associated metadata be changed?'.
This fixes QArrayDataPointer::setSharable for static data (e.g.,
Q_ARRAY_LITERAL), previously it only catered to shared_null.
SimpleVector is also fixed since it wasn't checking Mutability and it
needs to because it supports fromRawData().
Change-Id: I3c7f9c85c83dfd02333762852fa456208e96d5ad
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This enables a truncating resize() to be implemented. It is similar to
destroyAll(), but updates the size() as it goes, so it is safe to use
outside a container's destructor (and doesn't necessarily destroy all
elements).
The appendInitialize test was repurposed and now doubles as an
additional test for QArrayDataOps as well as exercising SimpleVector's
resize().
Change-Id: Iee94a685c9ea436c6af5b1b77486734a38c49ca1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This follows QArrayData::detachFlags's lead. Given the (known) size for
a detached container, the function helps determine capacity, ensuring
the capacityReserved flag is respected.
This further helps aggregating behaviour on detach in QArrayData itself.
SimpleVector was previously using qMax(capacity(), newSize), but there's
no reason to pin the previous capacity value if reserve() wasn't
requested. It now uses detachCapacity().
Change-Id: Ide2d99ea7ecd2cd98ae4c1aa397b4475d09c8485
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
