This constitutes a fairly complete submission of an entirely new
floating point type which conforms to IEEE 754 as a 16-bit storage
class. Conversion between qfloat16 and float is currently performed
through a sequence of lookup tables. Global-level functions
qRound(), qRound64(), qFuzzyCompare(), qFuzzyIsNull(), and
qIsNull() each with a qfloat16 parameter have been included
for completeness.
[ChangeLog][QtCore] Added new qfloat16 class.
Change-Id: Ia52eb27846965c14f8140c00faf5ba33c9443976
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Under certain circumstances, VS2015 reported ambiguous options in using
the operator>(Enum,int) operator. This change adds a static_cast<qint64>
to remove any ambiguity. In the process of testing this change, a gap
in the existing logic was identified: the handling (just in the test
code) of large negative enum values. Consequently, and additional
test case was added, and additional if-conditions were added to account
for that case.
Change-Id: Ife2c471ba4caa4b9a0107722042114e58145c4d0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The first one was already suppressed for GCC, so also do that for clang:
/Users/erik/dev/qt5-dev/qtbase/tests/auto/corelib/tools/qstring/tst_qstring.cpp:1076:16: warning: format string is not a string literal (potentially insecure)
[-Wformat-security]
a.sprintf( zero );
^~~~
/Users/erik/dev/qt5-dev/qtbase/tests/auto/corelib/tools/qstring/tst_qstring.cpp:1076:16: note: treat the string as an argument to avoid this
a.sprintf( zero );
^
"%s",
The second one could also occur with other compilers, so fix it in a
generic way.
/Users/erik/dev/qt5-dev/qtbase/tests/auto/corelib/tools/qstring/tst_qstring.cpp:6382:5: warning: ignoring return value of function declared with
warn_unused_result attribute [-Wunused-result]
string.repeated(3);
^~~~~~~~~~~~~~~ ~
2 warnings generated.
Change-Id: Id999179e795580a37b5be673ee54d6fa1a006dd7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The uvfd flag is implicitly defined when -Olink is used.
This causes the compiler to generate a warning for every
file being compiled in release mode.
Change-Id: I75759151864da7cf2f6d9c812e466a52c1208444
Reviewed-by: Nikola Velinov <nvelinov@ghs.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Alloca() is not supported on all platforms, like
INTEGRITY on ARM, so adding a configure check for it.
This can be used when building QtQml and 3rd party
code, in particular PCRE2 and SQLite.
Change-Id: I9785e16c21f67d1a68fef567e18c3356170f027e
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
It's included by qglobal.h, so we get it for free in other headers.
Change-Id: I90072156e313271a5354a39cbf78a83a6885c431
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Instead of
result += QString(string builder expression);
forcing a QString creation incl. memory allocation, do
result += string builder expression;
using the overloaded QString += QStringBuilder operator.
Change-Id: I23023c76620fa6bb7bf9f2786c22f6a2ec0d87c2
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Having two implementations was nice for testing, but now remove
one of them.
Change-Id: I70dc7d16496427dd36ba2464c9f650ec865202b1
Reviewed-by: Sérgio Martins <sergio.martins@kdab.com>
The crc32() functions take a size_t length, but the hash() functions
wrapping them took int lengths.
That makes no sense and actively hurts adding hash functions for STL
types or QStringView, so port the hash() interface to size_t.
Change-Id: Id303d6df4b698560fce656cec8ed693b01daac1c
Reviewed-by: Milian Wolff <milian.wolff@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Manually vectorizing is significantly faster because we can optimize
for common cases like long stretches of opaque or transparent pixels.
This is both smaller and faster than the auto-vectorized version, it is
also much faster than the autovectorized version for AVX2 which then can
be removed.
Change-Id: I0fa80ce273a8387cc6cd084879822ad9bade385c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This fixes a regression introduced in the merge 318b5856.
Due to the removal of actual simulator_and_device in 5.8 (397f345a6),
conditions using it have become meaningless.
Task-number: QTBUG-58440
Change-Id: I9f874f9f85efa590c40602dbcd07793ff17d35f5
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Liang Qi <liang.qi@qt.io>
It was first introduced when we had to work-around old macOS SDK (<= 10.7) with
missing API; so we split the original code into two functions, one for
iOS (any supported version) and macOS > 10.7, another function (with all
ifdefs needed and old, deprecated API usage) for macOS <= 10.7. Now that
support for those versions was dropped and the second function gone, having
the remaining code in some external function looks illogical. It can be moved
into the member-function back.
Change-Id: Ib6355f225b5df3d92a70bd7679545cc89c450228
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Add an API for modes (that is screen size and refresh rate).
This will allow platform plugins to list modes available
for a screen.
[ChangeLog][QtGui][QPA] Add an API for modes.
Change-Id: I91851c51cc60a1544465dfa3b4d96cc667237a0a
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
In some situations, this allows for nicer code. It's also possible to
make this constexpr in C++11, whereas the mutable transpose() would
require C++14-style constexpr.
The new function should also be faster, since it just swaps the member
variables.
Because of constexpr-function limitations, the way the return value is
constructed needs to depend on the level of the compiler's C++11 support.
This is not the only class that requires uniform init to provide a fully
constexpr interface (QUuid and QBasicAtomic come to mind), so this should
probably be generalized across Qt at some point.
Added tests.
[ChangeLog][QtWidgets][QSizePolicy] Added transposed() method.
Change-Id: Ic1077a0d5a861e7c63bd1daeeb42b97c3a2f71ef
Reviewed-by: Sérgio Martins <sergio.martins@kdab.com>
Also add a basic test for constexpr expressions involving QSizePolicy.
GCC < 4.8.0 has problems with initializing variant members in constexpr ctors:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54922
Instead of revoking its Q_COMPILER_CONSTEXPR, which is a source-incompatible
change, simply note the fact so we can selectively provide constexpr for
QSizePolicy.
Change-Id: Iba8e5a7cdf847d73e8e2b6bb6211fb3c9846aa0a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
On some keyboard layouts it is possible that pressing SHIFT+letter does
not give the upper case version of the character. So we depend on QKeyMapper
here to give us the right keysequence as then it will compare against
a shortcut created with this combination then.
Task-number: QTBUG-57928
Task-number: QTBUG-57931
Change-Id: I9421f3ab4d3f8d1ee42f9680200d4b017d551057
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
We have an invalid pointer dereference in our callback.
Apple's docs are quite ambiguous, whatever we receive in our callback's
'client' parameter is our PACInfo, not clientContext's address.
That's what we pass to CFNetworkExecuteProxyAutoConfigurationURL:
"clientContext - A stream context containing a client info object and optionally
retain and release callbacks for that object."
This 'client info' is our PACInfo struct with 2 pointers.
Now in a callback, 'client':
"The client reference originally passed in the clientContext
parameter of the ..."
So apparently we should read this as "client info originally passed in ...."
Otherwise incorrect cast results in invalid pointer (apparently
the value of retain/release/copy pointers we carefully set to 0)
and a crash on dereference.
Task-number: QTBUG-56747
Change-Id: I89378f1582679638cd29a36c563e506d8f5af518
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
The compiler command line used to populate QMAKE_DEFAULT_INCDIRS must
include the sysroot in order to generate the correct paths list.
This fixes a regression introduced in afd8263 which in turn attempted to
fix an earlier regression making it impossible to override the
deployment target in user project files.
Task-number: QTBUG-58325
Change-Id: I93e6b7ef90b2744dd2f03c77da31c692cb194976
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Enables optimizing with -Og if GCC has the option available, this
should produce faster debug binaries without compromising debugability.
Is a privateConfig to limit it to the default Qt build.
Includes two fixes for false positives of maybe_uninitialized triggered
by -Og on gcc 4.9.
Change-Id: I466d7a4070295714189024369312e6cbd36cfacf
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
This reverts commit 556b2ee773.
The reason for this is that this change appears to stop
QtLocation from fetching tiles.
Task-number: QTBUG-58303
Change-Id: I64cd3a17d24f10652bb68cee0267cc7ff1f9d479
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
- port from inefficient QLists to QVector
- mark types held in Qt containers (incl. QVariant) as Q_MOVABLE/PRIMITIVE_TYPE
- remove pointless user-defined copy special members which prevent the class
from having nothrow move special members
Fixes errors reported by my local tree's static checks.
Change-Id: If3910484cea81a8e2c5ab737908c9443f75782c5
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The equality operator was supplied, but this one was missing.
[ChangeLog][QtNetwork][QHostAddress] Added op!=(SpecialAddress, QHostAddress).
Change-Id: Iad9c55fa0ee7a8e97d5e4ea4be0605b8b74649d1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When dragging a dock widget from a floating group to the main window
reparentWidgets() is supposed to be called. It's usually triggered
by some unrelated event, like a LayoutRequest.
Instead of relying on luck for reparentWidgets() to get called be
explicit, otherwise the dock widget that was dropped into main window
will still have as parent the floating group window.
The item.skip() condition seems overly restrictive.
Task-number: QTBUG-58036
Change-Id: I65b5699e1acb6ca9bedb10620daa055fa9d91943
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
[ChangeLog][QtWidgets][QColorDialog] Fixed long standing bug that prevented
custom colors in QColorDialog to be stored in the settings.
Task-number: QTBUG-58424
Change-Id: If3ee5eef75358d811f08e7ce52fb60622972ddd4
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
The "spinbox" feature was defined twice, and in two different ways.
Change-Id: Iabc8939036e2166c5c5a417181ae8fb244829123
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This makes -no-feature-printer and -no-feature-combobox work.
Change-Id: I1097ab0a81f1c92b808c3e36ced0db9117db559a
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Amends 7896ae052a. The previous change
focused only on ZWJ and ZWNJ, but there are many other formatting characters
that we need to support and that may be rejected by the German keyboard-hack.
This opens up for all characters in the Other_Format category.
Task-number: QTBUG-58364
Change-Id: Idd967a9ae5b12060c851f6030b7e019508561696
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Previously, the code queried the window to be grabbed for the geometry
and thus returned the entire virtual desktop for multi screen setups.
Use the QPlatformScreen's geometry in case of WID = 0.
Task-number: QTBUG-58110
Change-Id: I3a9c0b0b3ea057f5e58f272f5c3fd40fafc073ba
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
The comment says "don't send event if nothing changed", but the
condition it is attached to reads if (velocityOnly), which is
totally opaque.
Change the variable name to stationaryTouchPointChangedVelocity
and drop the then-branch of the if, because it is duplicated in
the following default case of the switch.
The comment makes sense again, and GCC 7 is happy, too, because
this whole issue was pointed out by its -Wimplicit-fallthrough,
which is why I added the fall-through attribute, too.
Change-Id: I0631a381095f8897c55d9440304f6aefbc021a9a
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Sérgio Martins <sergio.martins@kdab.com>
In tst_QDBusMetaObject::types(), hold a QMetaObject obtained from
QDBusMetaObject::createMetaObject() in a QScopedPointer instead of
leaking it. Use correct return value type.
This fixes the remaining errors in GCC 6.1 Linux ASan runs of
tests/auto/dbus.
Change-Id: I1df7f8e42d45f40ecf381fe7b684a8ab5ebee675
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The char* returned from QTest::toString() calls must be manually
delete[]ed.
Change-Id: Iad078e8741e3e97693b1a417693f414b3fb3ec09
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Taken QTreeWidgetItems need to be deleted, as do items
created without a parent, and widgets without parent.
Change-Id: I7ffa69903af9a1b92ba308f9f9416aec1d6d975f
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
... for calculated test data names.
That involved removing the leading ", " from test name literals
(folding it into the format string) and porting from some QString code
to QByteArray to make the result usable with addRow(), which does not
support %ls...
Change-Id: Icb2344778203f10939ae46b9e46872101f3878a9
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
This patch changes the way we ensure NoLessSafeRedirectsPolicy and
also the way we create actual redirect requests: https->http redirect
is now reported as InsecureRedirectError (under NoLessSafeRedirectsPolicy)
only if STS is disabled or we were redirected to a host whithout Strict
Transport Security policy.
Otherwise, we replace 'http' scheme with 'https' and explicitly set port
80 with port 443 as defined by HTTP Strict Transport Security policy.
This scheme/port replacement will affect both NoLessSafeRedirectsPolicy
and UserVerifiedRedirectsPolicy (SameOriginRedirectsPolicy does not allow
any scheme change and we continue to report such redirects as
InsecureRedirectError).
Change-Id: Ib370b830e5fb6a0fec503d6fa3a0dec771c4b741
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
There was an off-by-one error in the while loop for aarch64: we start
counting at 0 for the first position, so the last valid input position
is "a+7", not 8.
This wasn't covered by the tests, nor was the SSE2 version, so now there
are also tests for both versions.
Change-Id: I7eb8c5708e6179f45ea56885b0e66e1a37969c1d
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Use the new QImage method instead of operating on private API.
At the same time the code is improved to ensure the QImage is detached.
Change-Id: Ia015c0bb18d7bc62da38397594730254843e5a0d
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
It was automatically merged from 5.6 branch. Qt 5.8 does not support
Windows CE.
Change-Id: I6968f50ef568035c224851d595d6c057128491a7
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
