Fixed parsing version string for PostgreSQL.
PostgreSQL versioning changed since version 10, see link:
https://www.postgresql.org/support/versioning
Extended QPSQLDriver::Protocol enum for PostreSQL 9.x and later,
added underscore to item names to separate major and minor version.
Changed long switch-case statements to if-else.
Change-Id: Ib19ae7ba426f262e80c52670e7ecb3532ff460a0
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
... by moving them in QTestPrivate namespace (qtesthelpers_p.h). This
header file is a convenient staging area for helper APIs, eventually
some could be moved to public QTest API.
This header file utilizes the same pattern as other qtestlib header
files - wrapping functions with QT_${LIBNAME}_LIB to automatically
enable certain APIs based on what is in the projects dependencies,
e.g. QT += widgets.
Change-Id: Ic0266429939c1f3788912ad8b84fc6e0d5edd68b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Looks like I never even tested this. There were two problems:
1) when we asked for the recvmsg and sendmsg functions, we used the
wrong variable (socketDescriptor was still -1)
2) we extracted the destination addresses, but never set them in the
QIpPacketHeader object
The added tests confirm that this works on Windows, Linux, Darwin,
FreeBSD. There also seems to be a problem, obtaining the destination
address on an IPv4 socket with a dual-stack sender (I can reproduce that
on FreeBSD, macOS and Windows, plus an old version of Linux).
Task-number: QTBUG-63605
Change-Id: I638cf58bfa7b4e5fb386fffd14ea732bddbc0c42
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
If you had never used QHash before, this function returned -1. That's
not useful if you're trying to implement your own QHash that uses Qt's
global seed.
Change-Id: Ib0e40a7a3ebc44329f23fffd14b2e875b970a55c
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Do not attempt to copy directory items, which will fail and result
in empty result lists. Amends 5865e582fd.
Task-number: QTBUG-57070
Task-number: QTBUG-63645
Change-Id: I59efce196b28099ec8aff5a802ef0a4d9a098453
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Qt cannot handle places like 'Network', etc.
Task-number: QTBUG-63645
Change-Id: I53d0eedc2996af6a1ec3230e3d65a3e272aa3710
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Previously, the native event filter listening on removable drivers
was installed on QCoreApplication::eventDispatcher() which led to
a mismatch when launched from a non-GUI thread since
~QAbstractNativeEventFilter() removes itself from
QAbstractEventDispatcher::instance().
Amends 45580aa925,
e612fe8d47.
Task-number: QTBUG-64171
Change-Id: Icbe289bd585f124d66989d0cd574040b986e680c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When the dataChanged() signal is handled by QIconModeViewBase, the size
of the items are recalculated. During this operation the optional
grid size is not taken into account which leads to a screwed up layout.
This patch adds the missing check similar it is done in
doStaticLayout()/doDynamicLayout().
Task-number: QTBUG-45427
Change-Id: Iba7adb44b1510c511a69c289ccb4f168992a6871
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
If a window is blocked by a WindowModal dialog, it should not be
possible to trigger window shortcuts on that window if it receives
a WindowActivate event.
This currently happens if the blocked window gets clicked, because the
window becomes the active_window and then QApplication sends it a
WindowActivate event (this doesn't happen with application modal dialogs).
The correctWidgetContext() function calls QApplicationPrivate::tryModalHelper()
only if the shortcut context is ApplicationShortcut. This patch makes it
call even if the shortcut context is WindowShortcut.
Change-Id: Iff87d85bcae603a6a24128e0cedfa9d33b6485fd
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
udev encodes the labels for /dev/disk/by-label/ with ID_LABEL_FS_ENC
which is done with blkid_encode_string(). This function encodes some
unsafe 1-byte utf-8 characters as hex (e.g. '\' or ' ')
Task-number: QTBUG-61420
Change-Id: If82f4381d348acf9008b79ec5ac7c55e6d3819de
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Nothing (except the *global*.h headers themselves) is supposed to include
the generated *config*.h files directly. The QtConfig forwarding header
should not exist in the first place.
Change-Id: I5e9edd25d905582381cdc6386804c1f3ef8716ae
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Currently, when two colors are equal except for their alpha values,
QCOMPARE produces the following failure message:
FAIL! : tst_Test::test() Compared values are not the same
Actual (colorA): #ff0000
Expected (colorB): #ff0000
By using the HexArgb format instead of the default HexRgb, we can
see the full hex string, with alpha values included:
FAIL! : tst_Test::test() Compared values are not the same
Actual (colorA): #88ff0000
Expected (colorB): #ffff0000
Task-number: QTBUG-55574
Change-Id: Id82c60a1b473ac6025a6f6ac560fce95a910d782
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
QPlatformWindow::mapToGlobal() should also be used in case
a window is embedded.
Task-number: QTBUG-64116
Change-Id: I1fbdf3d185659d0faea13a593db901e36ab27d8d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
...in order to allow applications built against Vulkan-enabled pre-built packages
to include <QtGui> on systems without Vulkan headers.
This has the downside of not being able to pull in qvulkan* headers via
the master header. This is an acceptable compromise for now.
Task-number: QTBUG-64073
Change-Id: I63c5834dcec60e66aba34c003d4bfe8e7d31607f
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Change-Id: Iaa6460d4eb96632f3cb03bc29b57934c53cbf88e
Reviewed-by: Jesus Fernandez <Jesus.Fernandez@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
We have logic to deal with self.platformWindow and frame strut mouse
events which happens after delivering the event to the NSWindow, but
delivering it might dealloc the window, e.g. when closing it, so we
need to explicitly retain it for the duration of [QNSWindow sendEvent:]
Task-number: QTBUG-64023
Change-Id: I223fe3e3ac36a309da375522ba11f20f1ad6fc4f
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
A regression was introduced with a3d59c7c7f
which caused dragging to fail within a modal dialog on the XCB platform.
By adding an exception for the QShapedPixmapWindow, which is the window
used for the drag, we can allow that to continue to work whilst blocking
to the other newly created windows.
Task-number: QTBUG-63846
Change-Id: I7c7f365f30fcf5f04f50dc1a7fff7a09e6e5ed6c
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
The RENAME_NOREPLACE flag is supported for all Linux local filesystems,
since that can be easily checked by the VFS layer (it knows which files
exist and which ones don't). For non-local filesystems, the backend
needs support and that might need server-side support too. So we may get
EINVAL errors for those, in which case we fall back to link/unlink,
which in turn can fall back to rename().
EINVAL can also happen if we attempt to make a directory a
subdirectory of itself. In that case, we will attempt to link() it,
which will result in EPERM as we can't hardlink directories. Then we try
rename() again, which should result in the expected EINVAL.
Task-number: QTBUG-64008
Change-Id: Icaa86fc7b54d4b368c0efffd14f09ca23602dd2e
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
This warning used to be part of -Wconversion, but that generates too
more noise than we're willing to fix now (like conversion from qint64 to
int). The float conversion does trigger for conversion from double to
float, as shown in all the QVectorND uses of float, but more
importantly, it triggers on passing floats to ints.
Change-Id: I69f37f9304f24709a823fffd14e69cfd33f75988
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Discussed during Qt Contributor Summit 2017. We concluded that we don't
want to make these functions public, as they do not follow Qt coding
style API. Specifically,
qStartsWith(a, b)
is not easily understood which argument is the needle and which argument
is the haystack (same problem memcpy() has). Compare that to
a.startsWith(b)
which can clearly be read in English as a subject-verb-object sentence.
This commit removes the unit tests that called compare().
Discussed-on: http://lists.qt-project.org/pipermail/development/2017-October/031060.html
Change-Id: Icaa86fc7b54d4b368c0efffd14ee6205eb9043fb
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Right now,this does really nothing. This commit is just to allow us to
transition the other modules (besides qtbase) to use the syntax that
will become the API.
I've marked three places to use the system CSPRNG:
1) the QHash seed
2) QUuid
3) QAuthenticator
I didn't think the HTTP multipart boundary needed to be
cryptographically safe, so I changed that one to the global generator.
Change-Id: Ib17dde1a1dbb49a7bba8fffd14ecf1938bd8ff61
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
I'll need it in the AF_NETLINK implementation of QNetworkInterface.
Change-Id: Icaa86fc7b54d4b368c0efffd14ef5ce895d0ed5b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
A follow-up patch for 8e70241dcc:
Make 'shareContext' conditionally included/compiled
(protected by QT_CONFIG(opengl)).
Change-Id: Ieaeea6b146b47fba7a77e2576ae0a33042280199
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
Print warning only for visible windows or in debug mode.
Task-number: QTBUG-63661
Change-Id: I742c86afcb40455074a6de753b0b1ce6a11d55af
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
It's perfectly benign, but I spent a lot of time debugging this and
trying to figure out how to solve something that didn't need solving. So
document for posterity.
For an adopted thread, the TLS destructors or the adopted thread watcher
on Windows will call QThreadData::deref():
- QThreadData::deref(), count drops to zero
-> delete this;
- ~QThreadData() deletes the QAdoptedThread
-> delete t;
- ~QThreadPrivate() calls deref() again
-> data->deref();
- QThreadData::deref(), count drops to -1, no action taken
- ~QObjectPrivate() calls deref() yet again
-> threadData->deref()
- QThreadData::deref(), count drops to -2, no action taken
Change-Id: Icaa86fc7b54d4b368c0efffd14ee448e0796e8d7
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Otherwise, the plain socket might be blocked indefinitely on waiting for
socket activity.
This issue is tested by tst_QSslSocket::disconnectFromHostWhenConnected()
which was unstable in CI.
Task-number: QTBUG-64016
Change-Id: I6a1a111dea4d1d1adaf55e6a90c0c5f995a270af
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
This file could not be included by user code in an installed Qt.
Change-Id: Id222d56dda9ef47d010ab947efa01bf63ecac050
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The manager and managerPrivate member variables are set in the
constructor and never changed after that.
Change-Id: I7cc2fd2eb3f50bdc529ed29485e74bf9c016b0c0
Coverity-Id: 185272
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
[ChangeLog][Third-Party Code] Sqlite was updated to version 3.20.1
Change-Id: I538a4de9b915fd1655479e44aa7ca8bae7b9a0b3
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
These are third party licenses that are part of
freetype.
Change-Id: I8c54feb6b5537ccfb0d0a4ffc24bc830c3c530e4
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
The LICENSE.txt we showed so far only mentioned the two mutual
licenses (FTL and GPLv2), whithout actually spelling them out.
This is required though.
[ChangeLog][Third-Party Code] Improve documentation about
Freetype 2 licenses.
Change-Id: I3af84b732aff58b12218cb7e8bb8e8de00c86307
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
There were a couple of corner cases where doing setPort() would result
in QUrl thinking that an authority was not present. Since the full URL
parsing implies that a host is always present if the authority is
present, then we also imply that setting the port number makes the host
be present too.
Change-Id: I69f37f9304f24709a823fffd14e67c12da18d69f
Reviewed-by: David Faure <david.faure@kdab.com>
1. Recently we have updated our receive window size to a larger value.
Unfortunately, this also results in auto-test pumping through
more data, which probably takes more time on CI.
At the moment we do not have any public API on QNAM's level to
customize HTTP/2 parameters (aka 5.10/FF and so on). So we use the fact
that QNAM is QObject and we can set a property on it. This property
is our Http2::ProtocolParameters object that allows us to configure:
- HPACK parameters (in 5.10 - noop)
- session receive window size
- different SETTINGS as described by RFC 7540, 6.5.2.
2. Undocumented environment variable to set ENABLE_PUSH is not needed
anymore.
3. In 5.11 Http2::ProtocolParameter will become a public API
and we'll introduce a new setter in QNAM.
Change-Id: If08fd5e09e7c0b61cf9700b426b60b5837b6b2e6
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Make an explicit mention of the fact that modifying a device while it
is being held by a QImageReader is undefined.
Task-number: QTBUG-61121
Change-Id: Ie0a016255c2614c5b8b415c8cd9602169153c8f8
Reviewed-by: Paul Olav Tvete <paul.tvete@qt.io>
When embedding foreign windows, we won't be able to return a valid child
accessible interface, so do not report it at all.
Supporting foreign windows properly is platform specific and something
to consider, but at least we shouldn't crash.
Task-number: QTBUG-63451
Change-Id: I19350cf97dc8d0c3f3052411eba0eee5f750dbab
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
It has several problems:
1. It could potentially create an intArray with uninitialized elements.
This could happen because the index for getting interfaces were the
same as the storage index. This was not correct, because they could
diverge if iface->child() returned an invalid interface.
2. The count of accessible child elements could change while iterating.
This could cause out-of-bounds condition when calling
SetIntArrayRegion as described in QTBUG-45855. Instead now, we call
SetIntArrayRegion only once, after we have gathered all the child
interface ids.
Task-number: QTBUG-45855
Change-Id: I77e813158df5f563d04931ac4e296e3fc2a16e67
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Currently on Windows, the Next button's shortcut doesn't work, because
QWizard overrides it with an empty key sequence.
The key sequence should be changed only if isVistaThemeEnabled() returns
true.
Task-number: QTBUG-46894
Change-Id: I54f26388b167973cc8065a867d9e771c1e6a2a72
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
In Qt 5.7, it was possible to call moc "-bfoo.h" or "-b foo.h" and it had the
same effect. With the port to QCommandLineOption, we broke the -b option as it
was not annotated as a short option.
(Regression in a7e3c17e75)
Task-number: QTBUG-63706
Change-Id: I161d0f1a4e65d129063b5e8431802257677da19d
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
c4cf90b1f7 made POST requests be
redirected properly, but this wasn't enough and should have included
every method/verb.
Change-Id: I37b12dc9fdffcbf2aadbd2360d4fc2584c024939
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
QAction::setData() always emits changed() even without actual data change.
Original code lacks a guard to check if the data changes.
According to http://doc.qt.io/qt-4.8/signalsandslots.html,
adding guard also benefits to prevent infinite looping in case
of cyclic connections.
Task-number: QTBUG-62006
Change-Id: I776369b668082f9f02e4502a36b1ae234ee7e079
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
"global/qrandom.cpp", line 155: error #2000-D: attribute "destructor" is not implemented and will be ignored
Task-number: QTBUG-63948
Change-Id: Icaa86fc7b54d4b368c0efffd14efa35381d4e797
Reviewed-by: Liang Qi <liang.qi@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The QTextCharFormat documentation said that the used style is based on
QStyle::SH_SpellCheckUnderlineStyle style hint, however in fact the
implementation (drawTextItemDecoration in qpainter.cpp) uses
themeHint(QPlatformTheme::SpellCheckUnderlineStyle) instead since Qt 5
(see commit 1f9ae50457).
Make the documentation match that behavior, and update QPlatformTheme
to use the correct default value.
Also, switch Cocoa theme to use DotLine, as that is what native macOS
applications use.
Change-Id: I2a6bb3da6c7b0686dca87ed2c251b6abc006123c
Task-number: QTBUG-50499
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
The previous version was good, just not optimal. Because the input was
an unsigned 64-bit number, compilers needed to generate extra code to
deal with HW instructions that only convert 64-bit signed input. And
that was useless because a double uniformly distributed from 0 to 1 can
only have 53 bits of randomness.
The previous implementation did exactly what the Microsoft libstdc++ and
libc++ implementations do. In my opinion, those implementations have an
imperfect distribution, which is corrected in this commit. In those, all
random input bigger than 0x20000000000000 has a different frequency
compared to input below that mark. For example, both 0x20000000000000
and 0x20000000000001 produce the same result (4.8828125e-4).
What's more, for the libc++ and MSVC implementations, input between
0xfffffffffffff001 and 0xffffffffffffffff results in 1.0 (probability 1
in 2⁵³), even though the Standard is very clear that the result should
be strictly less than 1. GCC 7's libstdc++ doesn't have this issue,
whereas the versions before would enter an infinite loop.
Change-Id: Ib17dde1a1dbb49a7bba8fffd14eced3c375dd2ec
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Let's make it happen even later: at the time of QtCore's unloading from
memory. This prevents issues with something using QRandomGenerator after
the global static destructor would have run.
Change-Id: Icaa86fc7b54d4b368c0efffd14eed56bbbb51cb6
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Mouse position is converted from native pixels later, so we must
provide native pixels for "QWindowSystemInterface::handleEnterEvent".
Amends 7091be1b79
Task-number: QTBUG-63865
Change-Id: I813c171f2fc1d321af702ac30eb5f2e4232e97c4
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Not properly initializing all members of the extended parameter struct
will cause an "invalid handle specified" exception on use.
Task-number: QTBUG-63883
Change-Id: Ic3a58df864c9e29ccbadc04bd71c18c8ef34374c
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@qt.io>
Both QNSMenu and QSystemTrayIconQMenu aren't referenced anywhere
else, including within qcocoasystemtrayicon.mm, since the QPA
backend was added.
Change-Id: I632c1b230226b2d08afce7f0f0019e9f7c030ba5
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
QBasicMutex and QMutex are the same in bootstrap mode.
Change-Id: Icaa86fc7b54d4b368c0efffd14eed63343ddb51b
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
A QVariant can only be converted to a QByteArray if it has user type
QMetaType::QByteArray or QMetaType::QString. The way it stood, we
always tried to convert the mime data to a QByteArray, and
then put the result into a QVariant. This would fail if the mime
data contained e.g a QPixmap.
This patch will inspect what kind of data the QMimeData contains, and
convert it to a QVariant using the expected API.
Backport of 6d3c483
Task-number: QTBUG-57428
Task-number: QTBUG-63660
Change-Id: I09b4a94aef7b52773e1a79c468ead71b36dfbfc5
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The path wouldn't match if the cookie's path was root ('/') and the
URLs path was empty.
Change-Id: I6dcd10f1fdf4f48f14e50f1b169cbdfda7005849
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Recent drivers no longer contain wintab32.dll, point out a version
that still has it.
Change-Id: I4125a0af3c11ab739f8006b91f58899aeed54458
Reviewed-by: Andre de la Rocha <andre.rocha@qt.io>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Before this patch a very accurate drop position below
or above an index was needed. Therefore it was not that
easy to do.
This patch increases the above/below area to be about
18% of the item (still leaving the most space for the item).
An average user will likely be 2-3x faster with dropping
below or above (while not losing much when dropping on items).
[ChangeLog][QtWidgets][ItemViews] Made it easier to
drop above and below items.
Change-Id: I47f0f80c76878c17ebf3f93d0a0cc82755971c2a
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
All systems must implement it by now. If there's any system still
without it, that means it has no IPv6 support, so they can disable
QtNetwork entirely.
[ChangeLog][Deprecation Notice] Starting with Qt 5.10, IPv6 support is
mandatory for all platforms. Systems without proper IPv6 support, such
as the getaddrinfo() function or the proper socket address structures,
will not be able to build QtNetwork anymore.
Change-Id: I3868166e5efc45538544fffd14d8c28046f9191b
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
All POST requests that were redirected would previously turn into GET
requests. This does not follow the standard for HTTP codes 307 and 308.
Task-number: QTBUG-63142
Change-Id: Ibd25a9566066e589670a9bc34e5dc5111f8139d5
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
In some cases when a session isn't needed (i.e. for localhost), the
session is not opened at all. If a program (e.g. our tests) redirects
from localhost to a different system (e.g. the qt network test
servers, or the internet) it will wait for a session forever. So, we
need to check if a session is needed for the redirect-target and then
open one. It is usually opened in
QNetworkReplyHttpImplPrivate::_q_startOperation
Change-Id: Id3b78182a3fb3f63f0235ecb1fb665df8bd0c4ca
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
We already cleared 'cookedHeaders', which is a QHash for 'known headers'
(enumerators as keys instead of strings), now do the same for 'rawHeaders'-
not to end up with some weird mix of headers from all possible redirect
responses and the final response.
Task-number: QTBUG-61300
Change-Id: Ifd6655c4167840bb00d29446d36ce65ba2d5491a
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
There's a common misconception that qDebug and friends are not
thread-safe, so let's explicitly state this.
Change-Id: I48d4ab8983017a9f2e7c9932a49ed573baa22929
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
I am not sure if this is going to help, but it is required that the
bridge checks that the interfaces it accesses are valid, since that
protects from accessing them when they are in the destructor.
This should be done, whether it fixes the issue or not.
Task-number: QTBUG-45855
Change-Id: I2b96999ca4043f8b33607c864d1d178695d03192
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
The semantics of QWindow::requestUpdate() is that it's used when the
window needs updates outside of the normal window invalidation callbacks
such as expose and resize events, e.g. when doing animations.
As a result, user code might not be prepared to handle window
invalidations in the update-request callback, assuming those
will still be delivered as normal, so that's what we do.
This was exposed by resizing one of the simpler Qt Quick examples,
where the resize's expose event was delivered as an update request,
but didn't trigger an unconditional draw of the window as it should,
as the scenegraph didn't change in response to the resize, which is
typical for an update request.
Change-Id: Ida8f85f1cf61c332aa9b199520e6854c48d3ab40
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Move the check for totalPacks below; it prevents leave notifications
from being handled.
Task-number: QTBUG-53628
Change-Id: I2436c51308803337e6d48ef958e03123283d4a1d
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
On macOS, absence of a QPA menu means that we should be using our
own internal logic since there's no entity on the QCocoaMenuDelegate
to take care of the shortcuts.
Change-Id: I35ed8f0b55445f61d0528709d4debb636a502002
Task-number: QTBUG-61039
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Otherwise we'll end up creating a NSPanel for the QMacNativeWidget which is
never closed, even if the backing NSView is moved to a new superview.
Ideally this would be based on [NSView viewDidMoveToSuperview] and
[NSView viewDidMoveToWindow], with retain/releases of the corresponding
NSWindow, but that needs more research, especially as AppKit on macOS
10.13 will always keep a strong reference to the NSWindow.
Task-number: QTBUG-63443
Change-Id: I9eec5ea871373d00dedf154600bf7005898cf37a
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
The modern approach to offscreen rendering on macOS is via FBOs, which
means there's no reason to allocate an NSView and corresponding NSWindow
just for that. In the offscreen case the NSOpenGLContext has a nil-view.
Change-Id: I2d1d407069af4d5283e6f56fba83db8eaf694ac6
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
We were using the default ones, provided by RFC7540. It appears they are way
too restrictive and conservative: when downloading something relatively big,
a stream keeps spending the whole session/its own 'recv' windows and thus
we have to constantly send WINDOW_UPDATE frames. This significantly slows
down our HTTP/2 implementation, making it orders of magnitude slower than
HTTP/1.1. To fix this:
- We send SETTINGS_INITIAL_WINDOW_SIZE in the first SETTINGS frame
to inform our peer that per-stream WINDOW is bigger than 64Kb
- We increase the session's receive window size.
Task-number: QTBUG-63722
Change-Id: I31312fcfd5f0fc0aee6aaa5d3562cc7d1b931adc
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Error out if it's missing or broken (Mersenne Twister not present).
This ensures that we never have a low-quality random generator in Qt.
Change-Id: I0a103569c81b4711a649fffd14ec80649df7087e
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
NSTimeInterval is a typedef for double, but the code stored its value in
an int, and only then multiplied by 1000.
Fix by only truncating NSTimeIntervals to int(64_t) *after* the
multiplication by 1e3 to get milliseconds.
While it's highly unlikely that a transition will have fractional seconds
length, don't assume if you can just calculate the more exact result.
Adapted-From: Marc Mutz <marc.mutz@kdab.com>
Change-Id: I0911b9c945a94ca24c3dfb23ed6a849141076326
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
We need it
1. to be sure, that thread is already attached to jvm
when we attach debugger to the process
2. to have a human-friendly name for main native thread
Change-Id: I1e572a0f09ec8af24a910835aaa6d302f6f2cb43
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Auto-reset events are automatically reset to non-signaled when we get
notified about the signaled state. This implies that we cannot check the
event state again in activateEventNotifiers. Instead, store the signaled
state in a QAtomicInt and test and decrement that.
This amends commit 85403d0a.
Task-number: QTBUG-63555
Change-Id: I0adee3d2929783f98d91060f9106c8b5266d72fa
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
The Qt documentation is using 'upper-left' everywhere except at
QRect(F) detailed description and Qt::CoordinateSystem enum description.
Therefore fix it in those four places to be consistent.
Task-number: QTBUG-59981
Change-Id: Ie652044d0207ea5a42888d9e1f1dc9a86b1e9410
Reviewed-by: Mats Honkamaa <mats.honkamaa@qt.io>
Reviewed-by: Tarja Sundqvist <tarja.sundqvist@qt.io>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
Use assignment. Fixes:
vulkan\qvulkaninstance.cpp(811) : error C2061: syntax error : identifier 'd_ptr'
vulkan\qvulkaninstance.cpp(812) : warning C4551: function call missing argument list
vulkan\qvulkaninstance.cpp(813) : error C2659: '=' : function as left operand
vulkan\qvulkaninstance.cpp(814) : error C2440: 'return' : cannot convert from 'QVulkanDeviceFunctions *&(__cdecl *)(void)' to 'QVulkanDeviceFunctions *'
There is no context in which this conversion is possible
vulkan\qvulkaninstance.cpp(832) : error C2061: syntax error : identifier 'd_ptr'
vulkan\qvulkaninstance.cpp(833) : error C2541: 'delete' : cannot delete objects that are not pointers
vulkan\qvulkaninstance.cpp(834) : error C2659: '=' : function as left operand
Change-Id: I859b141aa0cb24b1f85dc9f229262a4145651d7c
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Commit 282065d443 renamed the generator
functions but we didn't update all the docs.
Change-Id: I0a103569c81b4711a649fffd14ec877ffbfe710d
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
There are a couple of Q_ASSERT wrapped by the new noexcepts, but most of
those aren't validation of external parameters, only of internal
construction. The two exceptions are the checks for pointer alignment.
Change-Id: I0a103569c81b4711a649fffd14ec8523d741dfb6
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Incidentally, we can fix a typo.
[ChangeLog][QtNetwork][QNetworkInterface] Changed allAddresses() to not
include addresses found in inactive interfaces, matching the user
expectations of this function. If those addresses are needed for some
purpose, the application will need to call allInterfaces() and obtain
the addresses in each interface.
Task-number: QTBUG-51922
Change-Id: Iaf4157b7efa2416d898cfffd14d969c963ec0a2a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
I don't know why this needs to be a template specialization instead of
overload, like all the other toString overloads. With GCC 7, it was
calling QTest::toString<QHostAddress>, which returns nullptr.
Change-Id: I638cf58bfa7b4e5fb386fffd14ea7add0eef8f87
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
It must detach, otherwise it overwrites all QHostAddresses that were
sharing a given QHostAddressPrivate. This overload was introduced in
5.8 and probably as a result broke some pre-existing code, that previously
was using a conversion and the correctly working setAddress.
Conveniently, QHostAddress::clear() does: d.detach(); d->clear();,
exactly the first thing we do in other overloads of setAddress.
Task-number: QTBUG-63764
Change-Id: I63c36e877c9358d3ff26ba1f2e4adf35b771f426
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This code asks the platform theme to resolve SpellCheckUnderline to an
actual pen style (wave, solid, dash, etc.) but if there's no theme,
or if the default implementation in QPlatformTheme is used, the value is
still SpellCheckUnderline, which then casted to a PenStyle below in
qpainter.cpp:
pen.setStyle((Qt::PenStyle)(underlineStyle));
The value 7 is an invalid PenStyle, which leads to random behavior
when drawing the underline. Make it WaveUnderline if the platform theme
had no opinion on how to draw it.
Change-Id: I4f02f9b58f10582cee5aefce7a4d5cd300133140
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Unique begins with a "y" sound, hence a unique is correct.
Change-Id: I9eb6b4d4c9ddab45af931e97c041c24edf163eca
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
kernel/qopenglcontext_p.h included opengl/qopengl.h as "qopengl.h",
triggering:
In file included from ..\..\include\QtGui\5.11.0\QtGui\private/qopenglcontext_p.h:1:
..\..\include\QtGui\5.11.0\QtGui/private/../../../../../src/gui/kernel/qopenglcontext_p.h(58,10): warning: #include resolved using non-portable Microsoft search rules as: ..\..\include\QtGui/../../src/gui/opengl/qopengl.h [-Wmicrosoft-include]
Task-number: QTBUG-63512
Change-Id: I867a8f2ceeb8a9369dfe0163c179572feadc033e
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
The documentation of QKeyValueIterator was in the wrong place. This
patch moves it with the other iterators documentation.
Change-Id: I42ad529f72f9b19fdecfca7ddfeb6e6f4dcfebac
Reviewed-by: Sze Howe Koh <szehowe.koh@gmail.com>
That overload documentation was missing the "exact" qualifier regarding
the search result.
Change-Id: I27a08b60f9b88ae497f8bd9d9ba6a99f4a9ab4d6
Reviewed-by: Sze Howe Koh <szehowe.koh@gmail.com>
They are needed for older compilers that doesn't support
the __fp16 extension. Reverts under the assumptions other
compilers will optimize it away.
This reverts commit 6dc7e468df.
Task-number: QTBUG-63693
Change-Id: If780de001d8c12df0db12caaf62505f16e01b663
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
On high DPI (e.g Apple Retina) user resizes of
QDockWidgets (with custom titlebars) could fail.
There was a cursor position check in mouse
move event bailing out if the cursor pos was not within
the widget.
The problem was that we could be on the edge
(or maybe even cross it?). Furthermore there is (/was)
no similar check when setting the cursor
to be a resize cursor, so users will obviously expect
the resize to occur.
This solves a part of QTBUG-63526
[ChangeLog][QtWidgets][QDockWidget] Fixed an issue
in QDockWidgets where the widget would not resize
despite showing a resize cursor.
Task-number: QTBUG-63526
Change-Id: Ifa842a109071552506da3a82822d903dc252c8cd
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
308 Permanent Redirect was introduced after redirection support was
initially added to Qt.
[ChangeLog][QtNetwork][QNetworkAccessManager] Added support for HTTP status 308.
Task-number: QTBUG-63075
Change-Id: I1c6cda331d776237113ef8854de9abfe7e41ed3e
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Partially revert a72513cab7.
The value is too small for Chinese fonts.
Task-number: QTBUG-63654
Change-Id: If020bfc3044258b7abfd9d463bc9b292a9cc0839
Reviewed-by: Alessandro Portale <alessandro.portale@qt.io>
Documentation is missing the specialized event type names in the
non-client mouse event enum description.
Task-number: QTBUG-55018
Change-Id: Ica35994e13fc9a637a52eeca361898f8669fdbd1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
If the datagram reception failed, we forgot to set the buffer back to
empty. The returned QNetworkDatagram did report isValid() == false, but
it was possible to get the .data() and check its size, getting nonsense.
Tests in the next commit.
Change-Id: I638cf58bfa7b4e5fb386fffd14ea91adf2133d47
Reviewed-by: Jesus Fernandez <Jesus.Fernandez@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Amends c1cece8e54
For some reason, this was missed in the original change. This is now
outright prohibited in the watchOS 4 SDK and will cause a compilation
error on that platform.
Change-Id: Iaa2edf6256a54ca11dec9f1efd8a4d18ba7dc046
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Use the FreeType LCD rendering path with 2.8.1 even when it has lcd-
filtering disabled. This gives us proper subpixel rendering even with
a freetype build with the patented lcd-filtering code disabled.
The code is also simplified by removing the long pointless
ifdefs for 10+ year old versions of freetype.
Change-Id: I487e465317cb984b6e33c7bcc497f27cf29f9bcd
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
We should calculate the pen width based on the scaling similar to how
it is done for normal lines, otherwise we get the scaling applied
twice.
Task-number: QTBUG-61777
Change-Id: Iba71d55971a1d29537d2c9ff33749223a06160de
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
For gridlayout items the x position is rounded, but the QRectF changes
the width to keep the right edge. The width is yet calculated exactly
for the text so we need to preserve it.
Task-number: QTBUG-61244
Change-Id: I823ba742c9ab299740232b5d9b4ad5713e1782c4
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
This fixes a regression introduced in 8e70241dcc.
Task-number: QTBUG-63631
Change-Id: I5a2c23e06b790a482e1542ac2db3dcf25927caf2
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
If the native visual was BGR888, the XCB plugin would assert as soon as
it needed to find the corresponding QImage format. Fix by adding the
required mapping in qt_xcb_imageFormatForVisual().
Task-number: QTBUG-62840
Change-Id: Idd9eb01a60f605ad004d5b0c3025ded63ed64271
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
329a029c36 made ::raise and ::lower do nothing
for topmost and bottommost windows. This made it impossible to
e.g. raise one topmost window above another topmost window using
QWindow::raise.
Task-number: QTBUG-62021
Change-Id: I5f60816cbc48d69c0411e3bd68852d8bd8e300bb
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
<future> is needed by QThread::create. Instead of a fragile series
of preprocessor tests, move its detection to a configure test.
This dramatically simplifies the code, but on the other hand ties
the availability of QThread::create() to the system used to compile
Qt (rather the one used to compile an application).
Change-Id: If1b06363379bf29126cfa68f2a0651cbb78a67f7
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Commits aba336c2b4 (in Qt 5.2) and
aba336c2b4 (in 5.6) both tried to deal
with this problem, with different levels of success. This is the third
attempt (and hopefully the charm).
Instead of modifying the path that the user provides, go straight ahead
and declare it invalid. This is supported by RFC 3986, which declares
this expansion impossible:
relative-part = "//" authority path-abempty
/ path-absolute
/ path-noscheme
/ path-empty
path-abempty = *( "/" segment )
path-absolute = "/" [ segment-nz *( "/" segment ) ]
path-noscheme = segment-nz-nc *( "/" segment )
The "path-abempty" and "path-noscheme" cases are the two issues we
already handle. This commit adds the third one: path-absolute, which
requires that the first segment of the path be of non-zero length.
That is, it is now possible again to have http://example.com//path
constructed piece-wise, without it producing http://example.com/path.
Additionally, it catches the case of http://example.com//path parsed
from full URL then followed by setAuthority("").
Change-Id: I69f37f9304f24709a823fffd14e67a5e7212ddcd
Reviewed-by: David Faure <david.faure@kdab.com>
We don't support Windows versions prior to 7 nor Windows CE anymore.
Change-Id: I638cf58bfa7b4e5fb386fffd14ea930155d67689
Reviewed-by: Jesus Fernandez <Jesus.Fernandez@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
That was stupid for two reasons. First, applicationName() can be changed
by the user by calling QCoreApplication::setApplicationName(), so if
that happens, we won't be able to properly identify a lock belonging to a
given application because the name changed.
Second, because applicationName() is not what we compare to. Instead,
let's use processNameByPid() on both content creation and verification.
[ChangeLog][QtCore][QLockFile] Fixed a bug that would cause QLockFile
mis-identify valid lock files as stale if the application name was set
with QCoreApplication::setApplicationName().
Change-Id: I0b48fc8e90304e0dacc3fffd14e912a5c98c87e7
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
We duplicated both the creation of the contents and the check if the
file was stale. Centralize everything in qlockfile.cpp.
Change-Id: I0b48fc8e90304e0dacc3fffd14e91174af79841f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
F_SETLK is bad. Explanation in the comment. And flock(2) does work with
NFS on Linux, so let's just stick to that, which is simpler.
We only use the file locks when we attempt to delete an apparently stale
lock: that is, for a lock file that is at least staleLockTime old.
Change-Id: I0b48fc8e90304e0dacc3fffd14e908c8c4c9d59b
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: David Faure <david.faure@kdab.com>
It's confusing to keep it along with an unrelated class. Let's keep it
in its own file like for most other platform plugins.
Change-Id: I449ee061ff9fd5dc7ef06cadd633414d6b16358f
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
It's been for years that QMenu's rounded corners in qmacstyle_mac were
done via QWidget::setMask(QRegion). Unfortunately, QRegion mask does not
work well with retina displays and also does not support translucency.
That's why in this change we explicitly make QMenu's background
transparent and then draw a rectangle with rounded corners in
QMacStyle::drawPrimitive(PE_PanelMenu). This not only gives much better
result than the mask-based approach, but also de-HIThemes QMenu.
As a consequence, QComboBoxPrivateContainer doesn't get any mask from
QMacStyle anymore. Therefore, when the mask is empty, we need to paint
PE_PanelMenu before invoking QFrame's paint event handler.
Made-with: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
Change-Id: Ia9236176113f23b86e45507fa6ddf77236084ce3
Reviewed-by: Oleg Yadrov <oleg.yadrov@qt.io>
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@qt.io>
The qWaitFor functions themselves can not trigger a test failure, as that
will not result in the test function exiting early, so every single call
to qWaitFor needs to be wrapped in a QVERIFY.
Change-Id: Id15a1549f31d06cdbf788e1d84ea431c28636ec8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The resources allocated by QPlatformBackingStore are owned by the class,
and should be allocated in a context the class also owns. This removes
the asymmetry of having to pass in a context to composeAndFlush, while
having to make the same context current before destroying the platform
backingstore.
The context owned by QPlatformBackingStore is shared with the associated
window though a new QWindowPrivate::shareContext() API.
The result is that on e.g. iOS, the backingstore does not need to tie
the resource allocation of QPlatformBackingStore to the global share
context, but can instead tie them to the per-window context, and hence
clean them up after each window is closed.
Task-number: QTBUG-56653
Change-Id: Ic1bcae50dafeeafaa8d16a7febd83b840ec6367a
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
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>
When many runnables are executed, this improves the
performance by not resizing the queue for each runnable,
which was the case in the previous version, because of
many calls to QVector::takeFirst().
Also add a test that makes sure tryTake() is safe to
call and does not leave the queue in a bad state that
tries to use nullptr entries.
Change-Id: I608134ecfa9cfc03db4878dcbd6f9c1107e13e90
Reviewed-by: Lars Knoll <lars.knoll@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>
