Core Foundation Framework forwards notifications about socket activity
through a callback function which called from the run loop.

The default behavior of Core Foundation is to automatically re-enable the
read callback after each notification, and we explicitly enabled the same
behavior for the write callback.

With this behavior, if the client did multiple recv() calls in response to
the first notification in a series of read notifications, the client would
still get the QSocketNotifier notifications for the data that was already
read.

To get rid of these extra notifications, we disable automatically re-enabling
the callbacks, and then manually enable them on each run loop pass.

Task-number: QTBUG-48556
Change-Id: I0b060222b787f45600be0cb7da85d04aef415e57
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>

This reverts commit b8e0f7cf

.
Needs a more testing.

Change-Id: Iff0b2741922cfa8f16fbc3f4ce0f83869d6cd8b6
Reviewed-by: Alex Trotsenko <alex1973tr@gmail.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>

Commit 7943d4f7

 tried to fix this with a switch/case, but the feature
levels need to be in descending order so this failed. So, follow the
same style used for feature levels 10/11.

Change-Id: Ia1c22981bf8b99eb53df13833aba452482398295
Task-number: QTBUG-38481
Task-number: QTBUG-48571
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>

Core Foundation Framework forwards notifications about socket activity
through a callback function which called from the run loop. Previous
implementation sets kCFSocketReadCallBack, kCFSocketWriteCallBack to be
automatically re-enabled after they are triggered. With these semantics,
an application need not read all available data in response to a read
notification: a single recv in response to each read notification is
appropriate. If an application issues multiple recv calls in response to
a single notification, it can receive spurious notifications.

To solve this issue, this patch disables automatically reenabling callback
feature. Now, callback gets called exactly once, and is not called again
until manually re-enabled by calling CFSocketEnableCallBacks() just before
entering to wait for the new events.

Task-number: QTBUG-48556
Change-Id: Ia3393c2026230c7b3397cc614758dec1d432535f
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>

Due to additional validation not covered in previous patches, the Windows
Store certification compatibility had regressed. These changes ensure that
the required D3D behaviors are met.

Change-Id: I0a74f0d2fecaa87d4a9409da3a7a194254609759
Task-number: QTBUG-38481
Reviewed-by: Samuel Nevala <samuel.nevala@intopalo.com>
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@theqtcompany.com>
Reviewed-by: Jani Heikkinen <jani.heikkinen@theqtcompany.com>

Change-Id: I42e7ef1a481840699a8dffff1405e4154dd6b5dc
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>

Like MSVC, ICC on Windows in debug mode always makes calls to
dllexported functions instead of inlining them. Since MSVC 2013 doesn't
know about ref-qualification of member functions, this creates an
incompatibility between DLL creation and DLL use.

Task-number: QTBUG-48349
Change-Id: I42e7ef1a481840699a8dffff14053b594810fb42
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>

Change-Id: I310f9eff657e9730830cd5c302bdfc46fe67febc
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>

Opening a connection to an e.g. inactive server will return true
regardless of the server accessibility.
This patch aims to fix the current checks done.

The first one is an allocation check which should only fail if there's
not enough memory but is currently wrote as if the connection failed
there.

The second check that is "failing" is the connection setup. The return
value should either be NULL or the same value provided as first
parameter. That is now verified.

[ChangeLog][QtSql][QSqlDatabase] Fixed a bug where opening a
connection to a MySQL database using the QMYSQL plugin would always
return true even if the server was unreachable. This bug could
also lead to crashes depending on the platform used.

Task-number: QTBUG-47784
Task-number: QTBUG-47452
Change-Id: I91651684b5a342eaa7305473e26d8371b35396c4
Reviewed-by: Andy Shaw <andy.shaw@theqtcompany.com>

Wrong calculation of flip.

This reverts commit 1a5cc2a8

.

Task-number: QTBUG-47030
Change-Id: Ide178eb5e027c4ecec1e3952c973fb64987eb7ce
Reviewed-by: Jake Petroules <jake.petroules@theqtcompany.com>

Change-Id: I3a78e7228aa21b7c83cf7f07dc6e1c8ad0d0ebff

We should carry this in all 5.5.x and 5.6.x releases, to make sure
everyone gets the message. But for this release, we need to note the
changes from what was announced for 5.5.0:

Clang 3.2 cannot compile std::atomic properly, so it will not be
supported. That means the minimum XCode version we'll require will be
5.1, which fortunately does not raise the minimum OS X version that can
compile Qt. See http://llvm.org/bugs/show_bug.cgi?id=12670

 for more
information.

ICC 14.x and 15.x on Windows miscompiles pointer-to-members across DLL
boundaries, so signal-slot connections fail. See QTBUG-40281.

Change-Id: I42e7ef1a481840699a8dffff1400a4377fcf4589
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>

If the popup will show too close to the screen bottom,
we need to help Cocoa a bit. The horizontal positioning
hasn't shown any problems.

Change-Id: I5f298529fbf4a902e39f686f368046a8d1c11760
Task-number: QTBUG-45063
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>

Android does not support GLES 3.0 before 4.3 (API level 18). However some
4.2.2 devices are reported to return 3.0 in the version string and therefore
choose the GLES 3+ code paths in Qt. This blows up sooner or later because
the 3.0 specific functions are apparently not present at all.

Task-number: QTBUG-46831
Change-Id: Ic3eeb7c55829cf36c6d142c01ff8a1e18e9ecc1a
Reviewed-by: Kati Kankaanpaa <kati.kankaanpaa@theqtcompany.com>
Reviewed-by: Christian Stromme <christian.stromme@theqtcompany.com>

On some network conditions, WSARecvFrom() may return WSAECONNRESET
or WSAENETRESET error code to indicate that the connection has been
broken and should be reset. According to MSDN documentation,
WSAECONNRESET mean that the virtual circuit was reset by the remote
side executing a hard or abortive close. Also, it would indicate
that a previous send operation resulted in an ICMP "Port Unreachable"
message. For a datagram socket, WSAENETRESET indicates that the time
to live has expired.

Previously, hasPendingDatagram() discarded datagrams with these
errors and reported no data available. This behavior is incorrect
and can lead to infinite "freezing" of the socket.

This patch allows to handle these notifications as a result of the
readDatagram() call.

Task-number: QTBUG-46552
Change-Id: I7d84babe22d36736b928b4dd4841e30be53d16bd
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>

To handle network events, QEventDispatcherWin32 uses I/O model
based on notifications through the window message queue. Having
successfully posted notification of a particular event to an
application window, no further messages for that network event
will be posted to the application window until the application
makes the function call that implicitly re-enables notification
of that network event. With these semantics, an application need
not read all available data in response to an FD_READ message:
a single recv in response to each FD_READ message is appropriate.
If an application issues multiple recv calls in response to a
single FD_READ, it can receive multiple FD_READ messages
(including spurious).

To solve this issue, this patch always disables the notifier
after getting a notification, and re-enables it only when the
message queue is empty.

Task-number: QTBUG-46552
Change-Id: I05df67032911cd1f5927fa7912f7864bfbf8711e
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>

There was a small amount of time between the last readDatagram() call
and disabling a read notifier in case the socket had a pending
datagram. If a new datagram arrived in this period, this qualified as
absence of a datagram reader. Do not change the read notifier state
because it is disabled on canReadNotification() entry and always enabled
by the datagram reader.

Thanks to Peter Seiderer, who investigated the same: "Querying
hasPendingDatagrams() for enabling/disabling setReadNotificationEnabled()
is racy (a new datagram could arrive after readDatagam() is called and
before hasPendingDatagrams() is checked). But for unbuffered sockets the
ReadNotification is already disabled before the readReady signal is
emitted and should be re-enabled when calling read() or readDatagram()
from the user."

However, this patch does not completely solve the problem under Windows,
as the socket notifier may emit spurious notifications.

Task-number: QTBUG-46552
Change-Id: If7295d53ae2c788c39e86303502f38135c4d6180
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Implement a dummy method to silence the compiler.
After testing, this method seems to never be called. Which is
good, since the current IM API in Qt have little to offer to resolve
what is being asked. Until a need arise, we just return
an empty array.

Change-Id: I573eb8205a7e635a46d487ae175fb46e3a602001
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>

Disable input and cursor for QGuiApplication instances used in
autotest to initialize it properly.

Change-Id: I78dc9b776269c082c20f244a51f858289129275d
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>

Change-Id: I895b9c12de8734c20ec87ac30a9a9cca8f4242d7
Reviewed-by: Pasi Petäjäjärvi <pasi.petajajarvi@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Generate error for network requests when connection gets disconnected.
Documentation states that QNAM requests will fail if network is not
accessible, so we need to track session state.

Task-number: QTBUG-47482
Change-Id: I2c2d348637f72b2a908b438a66aa543a878de1e5
Reviewed-by: Timo Jyrinki <timo.jyrinki@canonical.com>
Reviewed-by: Richard J. Moore <rich@kde.org>

While generating Visual Studio 2015 solution files for a project using
the subdirs template qmake writes out both the header for version 2015
and version 2013. The problem is a case fall-through.

Task-number: QTBUG-48110
Change-Id: Ib6ddc1ceb306be9b3098d7b7c66a8ffabbd86481
Reviewed-by: J-P Nurmi <jpnurmi@theqtcompany.com>
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>

When merging the accessibility plugin into the widgets library, the
headers were just moved. They should have gotten the _p at that time.

Task-number: QTBUG-47569
Change-Id: I0a2290dae3a8187596e9d7541ccf69beeb603296
Reviewed-by: Dimitar Dobrev
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

We're now at 3.5 level:

- Fixed changed filenames and topic titles
- Removed links to obsolete pages
- Added links to new pages

Task-number: QTBUG-47901
Change-Id: I3393555c86bf3cf58770f230e7a9c7fd7e258476
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>

Since 10.9, System Preferences->Mission Control->Displays have separate
Spaces settings needs to be followed.

Task-number: QTBUG-47030
Change-Id: I1c1cf326246bd5609090ce5ac3212d963d562593
Reviewed-by: Jake Petroules <jake.petroules@theqtcompany.com>

The hotspot is defined in device independent coordinates, so
the default coordinates need to be divided by device pixel ratio.

Also, modify the scaling of cursor's pixmap to use SmoothTransformation to
generate cleaner looking lodpi cursors from hidpi cursors.

Change-Id: Ia938fd1e476e19e796f30712e23b06a5efed9964
Task-number: QTBUG-34116
Reviewed-by: Jake Petroules <jake.petroules@theqtcompany.com>

Using predictive text will not refresh the selected files until a 
word is completed or the keyboard is hidden.

Task-number: QTBUG-44337
Change-Id: I9e9f1e760fe5d5a69abd6e112af55b217ae6a16d
Reviewed-by: Shawn Rutledge <shawn.rutledge@theqtcompany.com>

Change-Id: I868e8ecdff6a503ee891a257121bf14a7da77fec
Reviewed-by: Markus Goetz (Woboq GmbH) <markus@woboq.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>

An invalid mime magic definition could lead to an assert. Replaced with
a qWarning. Move all checking to the QMimeMagicRule constructor, and do
keep invalid rules since they are need to parse child rules.

Unit test added, with QTest::ignoreMessage when using the XML backend
(there's no warning from update-mime-database when using the cache).
Also make it easier to add more shared mime info files for tests.

Task-number: QTBUG-44319
Done-with: Eike Ziller <eike.ziller@theqtcompany.com>
Change-Id: Ie39a160a106b650cdcee88778fa7eff9e932a988
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Change-Id: I794259f28c7adbaad3cfb40f92a0ad2dc512e5b4
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Change-Id: Ie20bd30328ae353ace82fbeee5808546f0568703
Reviewed-by: Christoph Cullmann <cullmann@kde.org>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>

Follow up from b494d859

. We need to filter the responder
actions after a sync as well, otherwise they will be
added back if a menu item e.g changes text.

Change-Id: I2ecbcc292400ada97a8e29d4b97f087349d8a061
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>

Instead of hard-coding the key and modifier that should trigger a
shortcut, we read it out from the QKeySequence that corresponds to
a StandardKey instead.

Change-Id: I6325534d3ff91c788d7e660d9009954e437b8534
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>

Moved from qtsvg to qtbase

Task-number: QTBUG-47201
Change-Id: Iab185ea2e270893c0937d1ff87fdb544d226d603
Reviewed-by: Martin Smith <martin.smith@digia.com>

Change-Id: Ie5fdfd6aa3866c14b6fbfabf4533327f5c73c5a3
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@theqtcompany.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>

Change-Id: Ia44edbac3a1bd2da92ee8c92956abfe49d8763b8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Task-number: QTBUG-43810
Change-Id: I0a019becc53b222cb6a7df1fafdccd57aca5b598
Reviewed-by: Martin Smith <martin.smith@digia.com>

Task-number: QTBUG-43810
Change-Id: Ib756382833fabecaae2526a413d046646f3e443e
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
Reviewed-by: Martin Smith <martin.smith@digia.com>

Added link in C++ documentation

Task-number: QTBUG-46285
Change-Id: I0f330829f7df713d4f5292b2a300c5c9d3732bda
Reviewed-by: Venugopal Shivashankar <venugopal.shivashankar@digia.com>

When the state of the paint changes then the one used for the picture
paintengine needs to be kept in sync. Otherwise the rendering will be
incorrect.

Task-number: QTBUG-43145
Change-Id: Ia55a4e940d109bedb7c2eff4d985d3b212da75a4
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>