The call to CGImageCreateCopyWithColorSpace took a naked toCGImage(),
which left the resulting CGImageRef without a release, causing the
extra ref by toCGImage() to never be derefed, and a subsequent detach
of the image data on the next paint event.

Wrapping the call in a QCFType<CGImageRef> solves the problem. The code
has also been moved directly into QCocoaBackingStore::flush(), as there
is no need to keep the CGImageRef a member.

A local autorelease pool has been added to QCocoaBackingStore::flush(),
so that the NSImage used for blitting the backingstore is released upon
exit of the function, thereby releasing the corresponding CGImageRef.

Note that for layered mode, the QImage will still detach, as the view's
layer.contents property keeps a reference to the image data until being
replaced in a subsequent flush.

Task-number: QTBUG-63559
Change-Id: I06b9298f65a84deae7cc2eff617ba75c92ec3b87
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>

[ChangeLog][iOS] The minimum deployment target for applications is now
iOS 10.0.

Change-Id: Icb37e4eaecbf6f62fd3c9293b2abf19a0954a02d
Reviewed-by: Jake Petroules <jake.petroules@qt.io>

A "compound widget" is a widget that has a focus proxy set to an inner
child. This is normal for complex black-box components where focus handling
is delegated to the children. Since the compound can have several
children, a local tab order might exist between them.

The current implementation of setTabOrder had no idea about
compound widgets. As such, when connecting two compounds in the
tab chain, it would just break up their inner tab order and
cause tabbing to ignore children other than the proxy.

The new implementation recognizes compound widgets, and add some
extra code to figure out the correct tab targets. This way, the
local tab order between the children will be preserved.

This implementation was inspired by the patches of Marek Wieckowski posted
in the linked bug report, and later modified by Nikita Krupenko.

[ChangeLog][Widgets] QWidget::setTabOrder() will now preserve the local
tab order inside a widget if it has a focus proxy set to an inner child.

Task-number: QTBUG-10907
Change-Id: I0673d39d70ec8c6bf64af30bf978d67c651b2f3c
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>

That's no longer a thing on macOS. We still keep it around for
QMdiSubWindow, but we should support 4-edge resizing there as
do native windows.

In practice, because of WA_MacOpaqueSizeGrip, we were already
rendering the size grip manually instead of relying on HITheme.

Change-Id: If613a14f03dc650b457688f3f2f57631b5fb30a3
Reviewed-by: Jake Petroules <jake.petroules@qt.io>

NSBox is good enough for this and pixel equivalent with
the previous HITheme calls.

Change-Id: I32b2a1cc8a9cd2a6ef17d1e01032540344808ebc
Reviewed-by: Jake Petroules <jake.petroules@qt.io>

That switch following the if looked a bit corny.

Also made cocoaCells hash mutable to get rid of the
const_cast(this) in QMSP::cocoaCell().

Change-Id: I4facec827409314cf1214152c19efb9688715eb2
Reviewed-by: Jake Petroules <jake.petroules@qt.io>

Window icons on macOS are tied to document windows, and should not show
up unless a represented filename has also been set according to the HIG.

We follow this, and no longer create a document button based on the window
title if one did no exists. We also fall back to using the filename of
the file if a title has not been set, including being able to restore
this default title after setting a custom title.

The icon is no longer reset to nil after setting a QIcon(), which would
remove the icon completely, but instead we restore the default behavior
of showing the default filetype icon of the file.

Finally, the two callbacks in QNSWindowDelegate dealing with the document
icon/title popups and drags have been taught to look for spaces in the
represented filename. This allows clients who really want the old
behavior of setting an icon without caring about the filename to set
the filename to a single space character, which will prevent the
dropdowna and drag from occurring.

The reason for not tying this behavior to the existence of the file in
the filesystem is that being able to represent a file that is yet to
be saved is a valid usecase.

Task-number: QTBUG-63338
Change-Id: I7d4377c6358226fb6058f8c2b3c28588eec77ce5
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>

It's the one that is more complex - with a text-based 'progress-bar'
and queueing (for some, probably educational, reason) of requests.

Changes:

- update the C++ syntax (mem-initializers, range for, etc.)
- new-style headers
- redirects should not result in creating an empty file.  Since we
  have no UI, and this example is already complex enough, settle
  for just reporting the redirect and removing the empty file.

Task-number: QTBUG-60628
Change-Id: I0b69cd77414ecac7c0bc6b2f8f787befc978de28
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>

Mainly 'modernizing' - use <c...> c-library includes (<stdio.h> -> <cstdio>),
add appropriate using directive; minor fixes in formatting + removal of a
hated double negation (aka ifndef QT_NO_NOTHING). Also, as our rules
('how to write examples') suggest - replace too many inclusion directives
with module-level headers. Basic redirects handling - do not create empty files
for redirected requests (or even files with some useless html).

Task-number: QTBUG-60628
Change-Id: Ia4398d39126313e6213bc7244d11a55958e64dec
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>

The PID cannot change.

Change-Id: I0b48fc8e90304e0dacc3fffd14e9324bfcda4412
Reviewed-by: Jake Petroules <jake.petroules@qt.io>

MSVC is too stupid for them. Not only will it not inline those functions
if they are in a __declspec(dllimport) class, it will also print an
annoying Level 4 warning:

 function 'function' marked as __forceinline not inlined

Task-number: QTBUG-57811
Task-number: QTBUG-55042
Change-Id: I0b48fc8e90304e0dacc3fffd14e8a346d18a9e0c
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>

The highlighted examples list for QQC2 is maintained separately.

Task-number: QTBUG-60647
Change-Id: Ib64ff7a44c6fcaf99058422d76ef6680df310911
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
Reviewed-by: J-P Nurmi <jpnurmi@qt.io>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>

Change-Id: I5fe5c8d3214876520b5259d478ac9ef887c61e95
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>

Change-Id: Ic07b45491a13bc4136e5259d45232297b7b25d8f
Reviewed-by: David Faure <david.faure@kdab.com>

This reverts commit 9e268185

.
Reasons: must go into 5.9 + the logic in this patch needs fixing, not good as
it is now.

Change-Id: I92516ddb1d2661d85ce77c54a1307f45df1ab635
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>

Task-number: QTBUG-60635
Change-Id: Ic1aef003be9cd08dc38021e5b4a4add8cabbfa98
Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>

Task-number: QTBUG-60635
Change-Id: I4485502510aca2797ead28f3b0085d1060ea3d49
Reviewed-by: Venugopal Shivashankar <Venugopal.Shivashankar@qt.io>
Reviewed-by: Nico Vertriest <nico.vertriest@qt.io>

Task-number: QTBUG-60635
Change-Id: I41d9af6836dc80b5c5c5303d313d15daff72b37f
Reviewed-by: Venugopal Shivashankar <Venugopal.Shivashankar@qt.io>

- new screenshot with more up-to-date graphical look
- deleted itemviews-editabletreemodel.png from old folder
doc/src/images
- deleted unnecessary item in qtwidgets.qdocconf

Task-number: QTBUG-60635
Change-Id: I9b82fa027347fb0fd7adb86038abd32853e68c3a
Reviewed-by: Venugopal Shivashankar <Venugopal.Shivashankar@qt.io>

Task-number: QTBUG-60635
Change-Id: I458cb09101c3423fb2abd3011e6f340e31cd0b42
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>

access\qnetworkdiskcache.cpp: In member function 'virtual qint64 QNetworkDiskCache::expire()':
access\qnetworkdiskcache.cpp:540:44: warning: 'QDateTime QFileInfo::created() const' is deprecated: Use either birthTime() or metadataChangeTime() [-Wdeprecated-declarations]
             cacheItems.insert(info.created(), path);
                                            ^

Change-Id: I5705de5cb4b490ef2bd788fb94ceac0e5b6c0d05
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>

This guarantees that we have proper version checks in place for APIs on
Apple platforms that are not necessarily available on the deployment
target.

Change-Id: I10060f8b910f2bb790aa4a9c6f8c5cdc14d7cf06
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Convert QSysInfo/QOperatingSystemVersion to __builtin_available where
required or possible, or to QOperatingSystemVersion where
__builtin_available cannot be used and is not needed (such as negated
conditions, which are not supported by that construct).

Change-Id: I83c0e7e777605b99ff4d24598bfcccf22126fdda
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Unlike client, this needs a bit more changes:
- remove redundant (and outdated) stdlib.h includes
- use QRandomGenerator (instead of qsrand/qrand pair)
- replace QStringList with QVector<QString>
- Q_NULLPTR->nullptr, ExplicitType * -> auto
- fix some weird indentation

Task-number: QTBUG-60628
Change-Id: I12eed12711b1e622407bd8ecd1afdf56a2cf2097
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>

For HTTP/1 it's done when no data expected and response headers received -
protocol handler emits channel->allDone which handles the status code and
sets (if needed) a redirectUrl. HTTP/2 protocol handler cannot emit allDone
(it has many requests multiplexed and actually cannot say allDone yet).
So we set a redirect url if we have the corresponding status code and
found 'location' header. Before it's too late.

Task-number: QTBUG-63471
Change-Id: Icfa871fb999e676650d62c45a20e689596da4028
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>

symLinkTarget should always return absolute path.

Change-Id: Id0078924d2f0befdea91e652a6c7775707c1b3c8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Change-Id: I4a123484ea1f5e53af85a52e86d4a80cef1cd7bc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Otherwise QWindow::setFilePath() before create() would have no effect.

Change-Id: Ia2171054aa13e21316d525193d330efaf4653dbf
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>

Copy into the intended array elements, not into the start of the
array.

Change-Id: I05c4ece2450583c9358aa1eddfd702262e995146
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Most linux distribution uses libjpeg-turbo as a replacement for libjpeg.
Since Qt will be linked against -turbo more frequently than the "original"
libjpeg we should use it for the bundled copy, too.

Also add an import script.

[ChangeLog][Third-Party Code] Replaced bundled libjpeg by libjpeg-turbo 1.5.2

Change-Id: I41081db74f194cbc288820fc783c9fef16545efd
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>

Minimal changes: Q_NULLPTR->nullptr, Type->auto where initializing expression
self documents the type actually.

Task-number: QTBUG-60628
Change-Id: I2fa784aeb30ac40c6b78e34dd58ad837ad607180
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>

... as soon as we updated multicastsender. Changes are minimal and mostly
cosmetic - use 'explcit' and 'nullptr' where appropriate, make a socket
data-member and not a pointer, move the 'datagram's declaration outside
of a loop.

Task-number: QTBUG-60628
Change-Id: Icfa46e6d2844c40a605f2f4066847594943a8ea8
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>

Problem was for 32 bit macOS, or if QT_COORD_TYPE was configured
to be float.

Change-Id: I0a3d3d305f2e8a0f9be7c40f4fef7eabf4643d05
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>

This is mostly relevant for Apple platforms, where we can use the new
unguarded availability warnings to guarantee that proper version checks
are present when using APIs that are not necessarily available on the
deployment target.

Change-Id: Ie408704b2924e1220491a9ea30f0141dfa4867d9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

... as soon as we updated the 'sender' example. Changes are purely cosmetic
though - 'explicit', nullptrs and local variables instead of redundant
data-members (QPushButton).

Task-number: QTBUG-60628
Change-Id: I572219da9d2a0ced07d94efb6f52f00b5a9c546d
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>

GCC 7 and other compilers compile this code just fine. GCC 6 fails, so
this is clearly a compiler bug. So revert the implementation of the
qWait() function back to the original, to make the problem disappear.

 qtestsystem.h:58:42: error: ‘bool QTest::qWaitFor(Functor, int) [with Functor = bool (*)()]’ used but never defined [-Werror]

Change-Id: I0b48fc8e90304e0dacc3fffd14e8110138cc9f45
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>

If you're using GCC 7 or Clang 4, Q_REQUIRED_RESULT expands to
[[nodiscard]] that must appear in as specific place on the line.

 qcompilerdetection.h:1182:29: error: attribute ignored [-Werror=attributes]
 qtestsystem.h:58:12: note: in expansion of macro ‘Q_REQUIRED_RESULT’
     static Q_REQUIRED_RESULT bool qWaitFor(Functor predicate, int timeout = 5000)
            ^~~~~~~~~~~~~~~~~

Change-Id: I0b48fc8e90304e0dacc3fffd14e80ede168d5c6b
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>

Change-Id: I62d98e4463003a1da8f28318d82635f6e0b2cc8c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

This avoids looping prior to the main 16-byte loop, by performing one
load that may include bytes prior to the start of the string. This is
guaranteed not to fault, since str points to a valid character, but it
may cause Valgrind to print warnings.

Change-Id: I6e9274c1e7444ad48c81fffd14dcae854bba24b2
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>

Child QWindows (or in the case of QWindows embedded in native applications:
top level QWindows where the corresponding NSView is a child of another
view, so not being the contentView of its window), still need some of the
NSWindow notifications to e.g. update their exposed state when the window
becomes visible.

We make sure to send the notification to all QCococaWindow children of
the relevant NSWindow, and let each callback decide if it should only
apply to content views.

This fixes an issue where a QWindow would never be exposed if the window
was a child NSView and added to a NSWindow that was yet to be shown.

Change-Id: I7f7df8bc5f4ca3ac553a2c146f8c3229b197c059
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>