It is finite and normal; it classifies as a zero; and it should not be > qfloat16(0).
Added tests to match.
Change-Id: I7874fb54f622b4cdf28b0894050ad3e75cf5d77c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Its limits() test was rather large and had some overlap with an older
qNan() test, that needed some clean-up (it combined qfloat16 values
with double and float values in ways that caused qfloat16 to be
promoted to another type, so we weren't testing qfloat16).
Renamed the qNan() test to qNaN(), separated out the parts of it that
actually tested infinity. Moved various parts of limits() to these and
rationalised the result. Split out a properties() test from limits()
for the properties of the qfloat16 type that are supplied by its
numeric_limits. Split out a data-driven finite() test to cover some
repeated code that was in limits() and extended it to test more
values. Added more tests of isNormal().
Fixed my earlier UK-ish spelling of "optimise", in the process, and
identify the processor rather than the virtualization as the context
where the compiler errs.
Change-Id: I8133da6fb7995ee20e5802c6357d611c8c0cba73
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
One of our compilers for emscripten coerces all signaling NaNs to
quiet ones, so won't do any actual signaling. Anyone relying on them
to do so shall be disappointed, so it's better that they know about it
at compile-time - or, at least, have the ability to find it out.
Put the signaling NaN producers (and remaining (test) code using them)
under the control of a feature that's disabled when numeric_limits
claims double has no signaling NaN. Assume the bootstrap library
doesn't need signaling NaNs. Sadly, until C++20 <bit>, there's no
contexpr way to test that alleged signalling and quiet NaNs are
actually distinct.
Added some auto-tests for signaling NaN, including that it's distinct
from quiet NaN. Any platform on which the last fails should disable
this feature.
Task-number: QTBUG-77967
Change-Id: I57e9d14bfe276732cd313887adc9acc354d88f08
Reviewed-by: Joerg Bornemann <joerg.bornemann@qt.io>
Some AMD CPUs (e.g. AMD A4-6250J and AMD Ryzen 3000-series) have a
failing random generation instruction, which always returns
0xffffffff, even when generation was "successful".
This code checks if hardware random generator generates four consecutive
equal numbers. If it does, then we probably have a failing one and
should disable it completely.
Change-Id: I38c87920ca2e8cce4143afbff5e453ce3845d11a
Fixes: QTBUG-69423
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Separate quiet NaN from infinity and expand the nan-with-payload test
to a general test that bits outside the exponent don't break qIsNan().
Generally test more thoroughly and systematically.
Tests for signalling NaN shall follow.
Change-Id: Ib35dabacc8ebcc9a0761df38f6f419f0398d0e20
Reviewed-by: Erik Verbruggen <erik.verbruggen@me.com>
Somehow missed these during my first iteration.
Change-Id: Iaef0ab84d9320a98f49ec071c93cd6f2907d92c3
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
- Replaced the usages of deprecated APIs of corelib by corresponding
alternatives in the library code and documentation.
- Modified the tests to make them build when deprecated APIs disabled:
* Made the the parts of the tests testing the deprecated APIs to
be compiled conditionally, only when the corresponding methods are
enabled.
* If the test-case tests only the deprecated API, but not the
corresponding replacement, added tests for the replacement.
Task-number: QTBUG-76491
Task-number: QTBUG-76539
Task-number: QTBUG-76541
Change-Id: I62ed4a5b530a965ec3f6502c6480808f938921aa
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Semi-automated, just needed ~20 manual fixes:
$ find \( -iname \*.cpp -or -iname \*.h \) -exec perl -pe 's/(\.|->)load\(\)/$1loadRelaxed\(\)/g' -i \{\} +
$ find \( -iname \*.cpp -or -iname \*.h \) -exec perl -pe 's/(\.|->)store\(/$1storeRelaxed\(/g' -i \{\} +
It can be easily improved (e.g. for store check that there are no commas
after the opening parens). The most common offender is QLibrary::load,
and some code using std::atomic directly.
Change-Id: I07c38a3c8ed32c924ef4999e85c7e45cf48f0f6c
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Not removing QT_EMULATED_ALIGNOF logic from qglobal.h at this point, as
it might be used elsewhere.
Change-Id: Ie78922bb604a54aed03ab5b88e31a7f29a3a4de0
Fixes: QTBUG-73561
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Without this, a local build of this test on macOS fails.
Change-Id: Ie03fa47ff0a54db752af47f223fbe5724cd9c976
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It should have been qfloat16(1)/qfloat16(infinity) in any case.
Sadly, that behaves no better than qfloat16(1.f/qfloat16(infinity)),
which was promoting the infinity back to float. So retain the check
for over-optimization (but make the comment more accurate).
This is a follow-up to d441f6bba7.
Change-Id: Iec4afe4b04081b0ebfbf98058da606dc3ade07f4
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
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>
... except where they are actually the component under test.
Java-style iterators are scheduled for deprecation.
Change-Id: If4399f7f74c5ffc0f7e65205e422edfa1d908ee8
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
[ChangeLog][qmake] A new feature "cmdline" was added that implies
"CONFIG += console" and "CONFIG -= app_bundle".
Task-number: QTBUG-27079
Change-Id: I6e52b07c9341c904bb1424fc717057432f9360e1
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
Some tests were fixed and others were skipped/blacklisted.
Task-number: QTBUG-63152
Change-Id: Ica7df555f8d152ee589865911130525101d4b941
Reviewed-by: Liang Qi <liang.qi@qt.io>
This way there is only one for loop, which is more optimizer friendly
Change-Id: Iaa02026627d5259c3eea1ff5664e8f22664eef73
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Since Qt 5.10, qTo/FromBig/LittleEndian<float/double> stopped working.
It may be confusing, but big endian floats do exist, so not to break old
code, we should support them.
Change-Id: I21cdbc7f48ec030ce3d82f1cd1aad212f0fe5dd0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This test triggered a compiler warning for good reason, it made no
sense, trying to change it to what it was probably meant to be.
Change-Id: I01a848272b42dae2aaa58a4f5bed998644d864da
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
tst_QtEndian's transformRegion_template() was getting a
signed/unsigned comparison warning when T was unsigned in a
QCOMPARE(T-value, 0); so use T(0) instead.
Change-Id: I78cb2ab96f79393def65ed2c020aa3039017ab92
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
[ChangeLog][QtCore][QByteArray] QByteArray::toInt(),
QByteArray::toDouble() and the other number conversion functions
now ignore leading and trailing whitespaces, as their QString
counterparts already did. For consistency reasons, the same
behavior was added to qEnvironmentVariableIntValue() also.
Task-number: QTBUG-66187
Change-Id: I8b5e478ea8577b811d969286ea9e269f539c1ea4
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The compiler was generating some vectorized code for qresource.cpp but
it wasn't very efficient. So improve upon it and make use in other
places where we read UTF-16BE strings.
[ChangeLog][QtCore] Added an overload of q{To,From}{Big,Little}Endian
that operates on a memory region.
Change-Id: I6a540578e810472bb455fffd1531fa2f1d724dfc
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The "app" subfolder was already excluded in the .pro-file but Android
supports QProcess, so lets include it in the build. Unfortunately it
currently has trouble and crashes (the child process or both processes).
So we skip those tests.
Task-number: QTBUG-68596
Change-Id: I2e6d0869c408bf08b22c02145db8ce522c64c617
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Not putting executables into debug/release subdirectories leads to the
WinRT AppxManifest being overwritten by the wrong configuration. When Qt
is configured with -release for example, it was possible that the debug
manifest (Manifest files are always created next to the target) is
written last and thus contains debug VCLibs as a dependency.
Additionally the test was changed in that way, that the resulting file
system structure (having helper and test application in a "top level"
debug and release folder) is the same structure as in tst_qobject.
Change-Id: I034752b4e5d22b98f6def95fb53c2b1947dded03
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
This only enables compilation, it doesn't fix any test.
Qt on Android supports process, but not TEST_HELPER_INSTALLS. See also
acdd57cb for winrt.
android-ndk-r10e is used to compile, see
http://doc-snapshots.qt.io/qt5-5.11/androidgs.html .
corelib/io/{qdir,qresourceengine} need to be fixed later.
Done-with: Frederik Gladhorn <frederik.gladhorn@qt.io>
Done-with: Mårten Nordheim <marten.nordheim@qt.io>
Change-Id: I34b924c8ae5d46d6835b8f0a6606450920f4423b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
This is a long overdue change so we don't break ADL of operator|.
I think will not break source or binary compatibility.
The problem is code like this:
namespace Foo {
struct MyStruct;
MyStruct operator|(MyStruct, MyStruct);
void someFunction() {
fooLabel->setAlignement(Qt::AlignLeft | Qt::AlignTop)
}
}
This would be an error before as ADL would find only the Foo::operator| and not
the global one since the arguments are not in the global namespace.
After this change, ADL works fine and this code compiles
This bites people with misterious error, see questions on
https://stackoverflow.com/questions/10755058/qflags-enum-type-conversion-fails-all-of-a-suddenhttps://stackoverflow.com/questions/39919142/broken-bitwise-or-operator-in-a-qt-project
[ChangeLog][QtCore] QFlags's operator| for enum types in the Qt namespace are
now declared in the Qt namespace itself.
Change-Id: I021bce11ec1521b4d8795a2cf3084a0be1960804
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
If _Thread_local is used on a block-scope declaration, it must be
combined with either static or extern to decide linkage.
Change-Id: I228b3520767197c6cdf5134ff5a666ab2aca33ea
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Commit 29bc68cf16 added support for
unsigned and commit 5ff7a3d96e later added
support for int. This commit adds support for qsizetype, which isn't int
on 64-bit platforms.
We do this by reorganizing the code and using the generic version of
__builtin_{add,sub,mul}_overflow from GCC 5 and Clang 3.8, which ICC 18
seems to support now too on Linux. That leaves older versions of GCC and
Clang, as well as MSVC, ICC on Windows, and the GHS compiler, to use the
generic implementations, as I've removed the assembly code those
versions of GCC and Clang on x86 are now uncommon.
Note: any older version of ICC probably breaks. We only support the
latest.
Change-Id: I9e2892cb6c374e93bcb7fffd14fc11bcd5f067a7
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Also drops a few instances where the dependency was purely runtime,
especially for examples.
Change-Id: I2a0476f79928143596bdb3b8f01193af90574ae8
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Unary ~ is not defined for enum classes, so we need a cast.
Change-Id: I79d495ebcc24ab960da8dae3be08eb307a9de448
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Both ARM and x86 can convert fp16 much faster in bulk than one at a
time. This also enables hardware accelerated conversion on x86, when
F16C isn't unconditionally available at compile time.
This code is implemented in C to ensure that there's no leakage of
inline symbols from the .obj file that was compiled by Visual Studio
with AVX support. Unfortunately, simd.prf uses $(CXX) instead of $(CC)
for all its sources, which means the file gets interpreted as C++ by
g++, clang++ and icpc. Those compilers at least don't leak any symbols.
Done-with: Thiago Macieira <thiago.macieira@intel.com>
Change-Id: I9d26d99e83392861fb09564e0e8e8d76cd8483b3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Tested with Clang, GCC 4.5 & up, ICC 17 and MSVC 2017. No current
version of MSVC supports C11 and GCC implemented the features slightly
later in C than in C++.
Change-Id: I57a1bd6e0c194530b732fffd14f45c5074c9a052
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
