We have a problem. Our types don't play well with the std unordered
containers, because they do not specialize std::hash. We therefore
force our users to come up with an implementation, hindering
interoperability, since any two developers are unlikely to come up
with compatible implementations. So combining libraries written by
different developers will result in ODR violations.
Now that we depend on C++11, and thus the presence of std::hash, we
still face the problem that the standard does not provide us with a
means to compose new hash functions out of old ones. In particular, we
cannot, yet, depend on C++17's std::hash<std::string_view> to
implement std::hash<QByteArray>, say. There's also no std::hash for
std::tuple, which would allow easy composition by using std::tie().
So piggy-back on the work we have done over the years on qHash()
functions, and implement the std::hash specializations for Qt types
using the existing qHash() functions, with a twist: The standard
allows implementations to provide means against predictable hash
values. Qt has this, too, but the seed is managed by the container and
passed to the qHash() function as a separate argument. The standard
does not have this explicit seed, so any protection must be implicit
in the normal use of std::hash.
To reap whatever protection that std library has on offer, if any, we
calculate a seed value by hashing int(0). This will be subject to
constant folding if there's no actual seed, but will produce a value
dependent on the seed if there is one.
Add some tests.
A question that remains is how to document the specialization. Can we
have a \stdhashable QDoc macro that does everything for us?
Task-number: QTBUG-33428
Change-Id: Idfe775f1661f8489587353c4b148d76611ac76f3
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Silence a warnings flood about qVariantFromValue(), QString::null
and QWeakPointer::data().
Change-Id: I73347190c0fa396b39b9efd00447cf24e48259a0
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Use QCOMPARE rather than QVERIFY an equality; ditch a stray blank line.
Change-Id: Ie828837919fb9d3cc774d82d0eebcf7728fed645
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QScopedPointer uses normal delete, but we need delete[].
Change-Id: Id62a2c55f75ef4aa60580f5e04c4bf306a6dd3c9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Such files do not exist (as per QFileInfo::exists), but on some
platforms that rely on realpath(), QFileInfo::canonicalFilePath did
not return the empty string.
Use the same logic on macOS as we already did on Android, and include
a test case. Remove the unnecessary dynamic memory allocation and
use a stack-allocated array instead, unless we use modern POSIX in
which case realpath() will alloc the memory for the result for us.
Change-Id: Ide987c68ebf00cbb7b1a66c2e9245a12c7807128
Fixes: QTBUG-44242
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
It's not clear why this test fails - and only does so sometimes - but
fail it does, so we ned to skip it to let development keep going. As
it happens, the same platform over-optimizes various computations
using qfloat16; which can at least be used to test for this platform,
since it wrongly distinguishes two qfloat16 values that theory and all
other platfomrs agree should coincide.
Fixes: QTBUG-75812
Change-Id: Ie9463d7dc21bca679337b475d13417b9f42bbf9b
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Add support for surrogate code points U+D800 through U+DFFF,
represent them with JSON escape sequences.
https://github.com/tc39/proposal-well-formed-stringify
Change-Id: I84fea53a8ef400beebefdba10ea82dc510fe7dda
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Conflicts:
src/corelib/tools/qlocale_data_p.h
(Regenerated by running the scripts in util/local_database/)
src/gui/opengl/qopengltextureuploader.cpp
Done-With: Edward Welbourne <edward.welbourne@qt.io>
Done-With: Allan Sandfeld Jensen <allan.jensen@qt.io>
Change-Id: I12df7f066ed0a25eb109f61c4b8d8dea63b683e2
If the deadline is far in the future, the conversions to nanoseconds
or internal arithmetic may overflow and give an invalid object, thus
the deadline may end up in the past. Added a test to the testlib
selftest for sleep.
[ChangeLog][QtCore][QDeadlineTimer] Fixed integer overflows
leading to immediate timeouts.
Task-number: QTBUG-69750
Change-Id: I9814eccdf9f9b3add9ca66ec3e27e10cd5ad54a8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The is a copy of the QString::SplitBehavior enum, but scoped
in the Qt namespace instead of inside QString, where it creates
problems using it elsewhere (QStringView, in particular).
Overload all QString{,Ref} functions taking QString::SplitBehavior
with Qt::SplitBehavior.
Make Qt::SplitBehavior a QFlags for easier future extensions (e.g.
a hint to use Boyer-Moore searching).
Added tests in QStringApiSymmetry.
[ChangeLog][QtCore] Added new Qt::SplitBehavior.
[ChangeLog][QtCore][QString/QStringRef] The split functions now
optionally take Qt::SplitBehavior.
Change-Id: I43a1f8d6b22f09af3709a0b4fb46fca61f9d1d1f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The code contained a sizeable chunk of string parsing along with
qDebug()s in the normal path of execution. That code, however, was
only used for Qt's own autotests.
The idea of this patch is, then, to not only move the autotest case
into the cold text section (using Q_UNLIKELY), but also to completely
exclude it, when QT_BUILD_INTERNAL is not set.
Unfortunately, the structure of the function did not really lend
itself to #ifdefing that part of the code out (production code was in
the middle of non-production code), so I transformed the engine
selection code into a lambda, replacing assignment with returns, and
swapping the branches of the central if around to yield a single block
of code that can be excluded from compilation with just one #ifdef.
As a consequence, the runtime code is almost unaffected, and the
function is much easier to read now.
Since the test-specific code is only compiled into Qt now in developer
builds, guard the tests that rely on this behavior with the same macro.
Change-Id: I9fd1c57020a13cef4cd1b1674ed2d3ab9424d7cd
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Setting the childMode property to ParallelStates will result in an
invalid state machine. This is never checked (worse, we explicitly
allow it and have a constructor to set it), but it results in
findLCCA failing, which then results in a failing assert or crash.
This fix in this patch is to handle this case separately. The proper
fix would be to remove completely the ability to set the childMode
on a QStateMachine, but that will have to wait until Qt6.
Fixes: QTBUG-49975
Change-Id: I43692309c4d438ee1a9bc55fa4f65f8bce8e0a59
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
The formatting of times in Norwegian has reverted to using dots in
place of colons, as it did before v31 (commit 82deb0ad1), so reverted
the tests to their state before that.
Change-Id: I8a09ce253731bb0f0f3caca117f06ad568940a81
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The file system appears to cache too aggressively, so if the reported
storage size doesn't change after flushing to disk, ignore the failure.
Change-Id: Iba7dce79591447fac296bfe92c2dc993d36d0c2a
Fixes: QTBUG-69868
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QDateTime's short names setUtcOffset() and utcOffset() have been
deprecated since 5.2, in favor of setOffsetFromUtc() and
offsetFromUtc().
QDate's shortDayName() and shortMOnthName() have been deprecated since
5.10, in favor of QLocale's dayName() and monthName(). Also, the
tests that were using them are testing methods only present when the
datestring feature is enabled; so condition them on that feature.
Change-Id: Ibfd4b132523ca8fbc1cb163353a44e0500877fd5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
[ChangeLog][QtCore][QMetaObject] Non-copyable lambdas can now be used
with invokeMethod(). For consistency reasons, the functor object is
now always moved.
Fixes: QTBUG-69683
Change-Id: I66ff5e21d6d1926f0f7c5f8c304bed1a90b69917
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Without this change, the target of a symbolic link that points to a UNC
share would include UNC in the target path, and not be correctly made
absolute.
Add a relevant test case, and use the opportunity to factor out the
helper code that creates NTFS symlinks into a function that takes
care of error handling.
The file created with the new test case only gets cleaned up correctly
when passing the file path into QDir::rmdir, which is either way the
right thing to do.
[ChangeLog][QtCore][QFileInfo] Fixed resolving of symbolic links to UNC
shares on NTFS file systems.
Change-Id: I9ba75d627aedf7c4cc289f0cb71088d795d30d8a
Fixes: QTBUG-73688
Task-number: QTBUG-63970
Task-number: QTBUG-30401
Task-number: QTBUG-20791
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
QMetaObject::newInstance returns a QObject, thus it's not possible to
create a new instance of a Q_GADGET using this function. Previously, we
returned a non-null QObject pointer for such scenarios, which then
leads to crashes when one tries to use it. Now, we check whether the
meta object inherits QObject's meta object, and error out early
otherwise.
Change-Id: I7b1fb6c8d48b3e98161894be2f281a491963345e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The test changes the current directory to the test directory
in fromUserInputWithCwd(), but did not restore it, causing:
Totals 898 passed, 0 failed, 1 skipped, 0 blacklisted, 368ms
********* Finished testing of tst_QUrl *********
QTemporaryDir Unable to remove "C:\\TEMP\\tst_qurl-ryVxqu" most likely due to the presence of read-only files.
Restore the old directory at the end to fix this.
Change-Id: I62669868f3c6d97dd38ebac76515428c14b7e1e7
Reviewed-by: David Faure <david.faure@kdab.com>
Includes a fixup for 03fadc26e7, which
removed the check on empty transition list, needed when no data are
available. Ensured that such a data-free zone would in fact be
noticed as invalid during init().
Fixed handling of times before the epoch (we still want to consult a
POSIX rule, if that's all that's available) while ensuring we (as
documented) ignore DST for such times.
Fixed handling of large times (milliseconds since epoch outside int
range) when looking up POSIX rules. Gave QTimeZonePrivate a YearRange
enum (to be moved to QTimeZone once this merges up to dev) so as to
eliminate a magic number (and avoid adding another). Moved
year-munging in POSIX rules after the one early return, which doesn't
need the year range.
Added test-cases for the distant past/future (just checking UTC's
offsets; SLES has a minimal version of the UTC data-file that triggers
the bugs fixed here for them).
Fixes: QTBUG-74666
Fixes: QTBUG-74550
Change-Id: Ief7b7e55c62cf11064700934f404b2fc283614e1
Reviewed-by: Tony Sarajärvi <tony.sarajarvi@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
A string can parse as a non-zero double that's smaller than the
smallest float yet be a faithful representation of the smallest float.
So rather than testing for non-zero doubles less than the smallest
float, test for non-zero doubles that cast to float zero; these
underflow. This means small values close below the smallest float
shall round up to it, rather than down to zero, requiring a tweak to
an existing test. Added a test for the boundary case (and tidied the
test data).
Fixes: QTBUG-74833
Change-Id: I4cb30b3c0e54683574b98253505607caaf88fbfb
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It can happen that QDataStream is fed a QVariant that contains a QPixmap
representation, that will make the application crash when trying to
restore it
This is specially important for cases in which applications expose dbus
interfaces with QVariantMaps
Change-Id: Ife4feaef30f30e7e27d88464bd6b2a247f743123
Reported-by: Fabian Vogt <fabian@ritter-vogt.de>
Reviewed-by: Fabian Vogt <fabian@ritter-vogt.de>
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Recurse down the sibling at column 0 of the index instead down the
index.
Change-Id: Ie78d8b28eab7438ca3f83ee0df177115ca82806e
Fixes: QTBUG-73864
Reviewed-by: David Faure <david.faure@kdab.com>
While an invalid time-zone shall have no transitions, so may various
constant zones, like UTC. The TZ data may include only the POSIX rule
for such a zone, in which case we should use it, even if there are no
transitions.
Broke out a piece of repeated code as a common method, in the process,
since I was complicating it further.
Added test for the case that revealed this; and made sure we see a
warning if any of the checkOffset() tests gets skipped because its
zone is unsupported.
Fixes: QTBUG-74614
Change-Id: Ic8e039a2a9b3f4e0f567585682a94f4b494b558d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Introduce a logging category for the qDebug()-output.
Add a meta type registration for QList<QPersistentModelIndex>, fixing
numerous warnings like:
WARN : tst_QItemModel::remove(QStandardItemModel:invalid start, valid count 5) QSignalSpy: Unable to handle parameter 'parents' of type 'QList<QPersistentModelIndex>' of method 'layoutChanged', use qRegisterMetaType to register it.
Fix a Clang warning about potential misuse of operator ,
Task-number: QTBUG-73864
Change-Id: I60998403a44f5df8767926951ee13d1ed1e93c37
Reviewed-by: David Faure <david.faure@kdab.com>
Use setPath() instead, as advised in the deprecation warning.
Change-Id: I2f22220885938808c8efb85720ad10f7e05801ff
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Some directories that depend on QtGui were being included without the
appropriate check for qtHaveModule(gui).
Change-Id: I7c348c74464d44cbd35a027f188f8a23bb2021d9
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Regular update in preparation for 5.13, adding tests for additions
since 5.9.4/5.10.1/5.11.0's update 7e946030 (the last to record its
upstream version sha1). Corrected the license header: it's now
published under MPL 2.0 (not 1.1); and our secondary licensing of it
is as LGPL3. Deferred full header over-haul until we've worked one
out in detail.
[ChangeLog][Third-Party Code] Updated DNS public suffix list
Task-number: QTBUG-72623
Change-Id: Iabdbbbfd79624830396c2a6fe0a73389bd6ce5b7
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
In preparation of deprecating QLinkedList.
This actually simplifies the code, since std::list has a ctor from size,
which QLinkedList lacks, and which the code worked around by using
initializer_list.
Change-Id: I07f9d590f863d9e4e00de73339cdfa27079f6e03
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The TZ environment variable can validly contain a POSIX rule, rather
than an IANA ID, as described here:
http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/V1_chap08.html#tag_08_03
However, if TZ were set to such a value, leading to it being used as
systemTimeZoneId(), it would be passed to QTzTimeZonePrivate::init(),
which is a no-op unless it manages to open a zoneinfo/ file with the
given ianaId as name. When the environment variable doesn't name a
zoneinfo/ file, we would thus get an invalid time-zone. We can,
instead, check whether the ianaId looks like a valid POSIX rule and,
if it does, use it as m_posixRule, enabling us to correctly handle
this case.
Tweak parsing of POSIX rules so that a zone using name "UTC" or "GMT"
with an offset other than 0 will be rejected as invalid. This avoids
parsing a zone name such as "GMT+17" or "UTC+00:01" as a POSIX rule,
where it should be understood as an offset from UTC (and only certain
well-established offsets are supported).
Added two test-cases to tst_QTimeZone::tzTest() for validity of a
POSIX zone value - a simple one constructed during discussion of the
bug, the other taken from an example in:
http://leaf.sourceforge.net/doc/buci-tz3.html
Task-number: QTBUG-75565
Change-Id: Ia5cb1cc56b13b0f6b56258e48be98d04d909e32a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
