Requires adding member swap. Member-swap is also
necessary to fix the deleted move-assignment operator
of this class, on which QXmlStreamAttribute heavily
relies for its own move assignment operator.
That's why it's not proposed for dev, but 5.6.
Change-Id: Id7d0823d44fc6e55ada7c096ae95444f6bb50963
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Faster, and, thanks to the new reverse_iterators, just as convenient.
Change-Id: Ibc6c64051a8ede4a47428e9271ffbeaa921fc255
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: David Faure <david.faure@kdab.com>
Call begin/endResetModel() in QSqlQueryModel and all
derived classes.
Task-number: QTBUG-49404
Change-Id: I11492d6386efb4c945c246a6379aaa6ca4502a25
Reviewed-by: Mark Brand <mabrand@mabrand.nl>
statically linked qt does not compile, as versionKeyLiteral() is unused,
causing clang to complain:
error: unused function 'versionKeyLiteral' [-Werror,-Wunused-function]
Change-Id: I6a233081e7c58fce75ece82616f937f29a23a81b
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Since the iPhone 6(S) Plus devices have a PPI of 401, we change the
logic from storing the unscaled PPI to storing the scaled PPI, and
applying that to a scaled geometry when computing the physical size.
Task-number: QTBUG-49467
Change-Id: I1741ff075749a301d2434cd35f642fcc9ea4b581
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Otherwise, we have unexpected channel close initiated by reply
destructor.
Change-Id: I15ad076ff20546e78787e19155544a2e5f8047a1
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Markus Goetz (Woboq GmbH) <markus@woboq.com>
Low-level input handling uses the native coordinate
system and the (device independent) coordinates from
QScreen needs to be converted.
Change-Id: I501dc77f5e51be01a42e533cd0609e069b8d228b
Reviewed-by: J-P Nurmi <jpnurmi@theqtcompany.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
QPlatformScreen works in native pixels and the
(device independent) coordinates from QWindow
needs to be converted.
Change-Id: I85ff45d66dc501e8c3c1c7a18f87448c92d79b9d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: J-P Nurmi <jpnurmi@theqtcompany.com>
It's a QPlatform* class and works in the native (platform)
coordinate system.
Change-Id: Ibf36bc68d84ecd668b35e0c22e1fa35cfc456a5d
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: J-P Nurmi <jpnurmi@theqtcompany.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
Support use of Qt::AA_EnableHighDpiScaling and
QT_AUTO_SCREEN_SCALE_FACTOR.
This makes the high-dpi scaling system use the screen
configuration set with QT_QPA_EGLFS_PHYSICAL_WIDTH and
QT_QPA_EGLFS_PHYSICAL_HEIGHT.
Implement QEglFSScreen::pixelDensity() and make it return
the scale factor for the display. The scale factor is computed
as the ratio of the logical DPI to the base DPI.
Change-Id: I14ed5da058024128479cb5508e056c39bd2f7563
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
For some reason add_ConnectionReceived has to be
called on the Xaml thread now. Otherwise the callback
function won't be called and thus listening on a TCP
socket won't work at all.
Task-number: QTBUG-49121
Change-Id: I11ce2f72b0c1d3bb20e9579de5a2ce5150ca966a
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Samuel Nevala <samuel.nevala@intopalo.com>
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@theqtcompany.com>
Calling QWinRTScreen::onOrientationChanged in the class constructor
can lead to situation where QWinRTScreen is expected to be fully
constructed. Moving the onOrientationChanged call to
QWinRTScreen::initialize ensures that QWinRTScreen is fully
constructed when it is called.
Task-Id: QTBUG-49470
Change-Id: I52748f33a9011dec3a172c1a74023cad15aae38a
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
This brings MSVC 2015 support to almost C++11 feature completion.
Constexpr support still has one annoying bug, for which the fix is
promised for Update 2.
Change-Id: I92c2ee6638625c61e3a16690563434e32d853368
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The ffsll function is missing with that SDK, but Clang and GCC support
it as the builtin.
Change-Id: I9f4d900898814b569c1affff140e15fb59fb45d8
Reviewed-by: Samuel Gaist <samuel.gaist@edeltech.ch>
Reviewed-by: Jake Petroules <jake.petroules@theqtcompany.com>
When that happens, we need to detach (in-place conversion won't work),
so we recurse back into the same function, but the template version that
does detaching.
Task-number: QTBUG-49181
Change-Id: Idba8c29717f34c70a58fffff1412fea3acc95f98
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
that way other modules can use the headers without hacks.
this required making the base directory for paths in headers.pri
configurable in syncqt.
Change-Id: Id35cfe05bcf4c576d3f2d0d8d09590a5e23d21d3
Reviewed-by: Andy Nichols <andy.nichols@theqtcompany.com>
This commit allows to avoid superfluous calls to the X server.
We don't request _NET_WORKAREA values when the screen geometry changes,
and we avoid RandR calls on each _NET_WORKAREA change.
Besides, update the available geometry of all screens with the
same root window, rather than only that one which corresponds
to Qt::Desktop window.
Change-Id: I5ec624717f5f261c986cd9aaf2425f22985e11c0
Reviewed-by: Błażej Szczygieł <spaz16@wp.pl>
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Shawn Rutledge <shawn.rutledge@theqtcompany.com>
This makes the automatic center click feature of the mouse event methods
also work if the target window isn't positioned at 0,0.
Change-Id: I0d711e484620900ba2ffc4139f7e13346a7482d3
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
The native position is given in device-independent pixels, so multiply
that by the window scale factor. If no cursor is present (e.g. on
Windows Phone), the native API returns SHRT_MIN as the coordinate values.
Catch that and pass Infinity instead in order to enable the cursorless
fall-back behavior in QQuickMenu::popup().
Change-Id: Ibe8ebcdbe257e8df25eea9873e8f4fa647ae1ab2
Task-Id: QTBUG-48265
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
That is the new name for gnome-icon-theme for more than 1.5 years. See [1]
for details.
Still keep “gnome” as the fallback icon theme, so this commit should not
break anything.
[1]: https://mail.gnome.org/archives/desktop-devel-list/2014-April/msg00086.html
Change-Id: Ifc2178827ad2cf6fdbc4e5248c83789ebc5b797e
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Jens Bache-Wiig <jensbw@gmail.com>
QWinRTInputContext is created from the XAML Thread, which can cause
problems when handleVisibilityChange is invoked. Instead just query the
keyboardRect and skip the emit.
Task-number: QTBUG-49389
Change-Id: I158204a07b9e000adffdc308e68b0f1425ed7c62
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
QRect only deals with integers, and I had a window set on the corner and
without a size. When this screen was created, it was reported to be on the
wrong screen, thus created there wrong.
The reason for this is that ::center is not prepared for this use.
QRect(QPoint(500,500), QSize()).center() is QRect(499, 499)
Change-Id: I24e57182f84873306f180749e96368d6da1147a9
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Added that adding an item deletes an existing one
Task-number: QTBUG-42105
Change-Id: Ibcf08e6e2bf9c9239e7eca5aa48f2f30c8417fec
Reviewed-by: Martin Smith <martin.smith@digia.com>
Link to Using a Designer UI File
Task-number: QTBUG-49251
Change-Id: I9d2e5ad746ce5c3933f28b8159e381dcdb8e643a
Reviewed-by: Topi Reiniö <topi.reinio@theqtcompany.com>
Reviewed-by: Martin Smith <martin.smith@digia.com>
Wrongly used QSQLQueryModel previously
Task-number: QTBUG-35860
Change-Id: I059ba45e54ac67b75b68eb0b804f66642b87a29f
Reviewed-by: Martin Smith <martin.smith@digia.com>
these modules simply have no "proper" include paths (they are used only
within qtbase, with explicit INCLUDEPATH statements).
Change-Id: Iec6aa6ae0bc449eb8cac2fa1c86a791755eec06e
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Same as in QString, though simpler a bit due to fixed-length sequence.
Change-Id: Idd6670a7b6415f5d515987c30931d16b73dd280e
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
and implement TitlecaseTraits as well,
to be re-used in the following commit.
Change-Id: I5c0bff42fe3b9bc9c2454cc16916cc2be87f604f
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Avoids in-loop check for every pixel by costs of slightly bigger code.
Change-Id: I8321d0945ac6b677786431670b2316c60e4db492
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Use an optimized version of the intra-pixel filter found
in FreeType's _ft_lcd_filter_legacy().
Gives us somewhat better accuracy and better performance.
Change-Id: Id4200b67e6ff55f605037c3e5464fef8bb03fbde
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
In change 8f6b328410 we override
the fontDef of the fallback font with the fontDef of the
original font (with the family name replaced) to fix a
regression where boldness was not inherited by the fallback
font.
This caused a bug, though, since the contents of the fontDef
would now misrepresent the actual font used. The side effect
of this was that isSmoothlyScalable() returned false because
the family and styleName combination we claim to have did
not exist in the font database. Result: We fell back to
native rendering for the font even though it's scalable.
Weight and style are the only parts that should be inherited,
since they are synthesized. So rather than overwriting
the fontDef completely, we copy the weight and style when
needed.
The bug in QTBUG-42963 is still fixed after this change.
[ChangeLog][Text] Fixed problem where fallback fonts for text
with certain styles would be reported as unscalable.
Change-Id: I95ef67f818852aea5a6ae8df789a52364ecb59a6
Task-number: QTBUG-47547
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
676edc006e introduced a regression by not filtering the active
popup window for tooltips, which is what the corresponding mouse
handler in QNSView does.
Change-Id: I5091dba0dd43be2e859ea360dfd812c0f081fd32
Reviewed-by: Eike Ziller <eike.ziller@theqtcompany.com>
Reviewed-by: Jake Petroules <jake.petroules@theqtcompany.com>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
This makes it possible to reuse it for the minimalegl QPA plugin.
Change-Id: I1c3dbaf67f32294a5d0e03cc1eb8557049b810a5
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
Wrap related code in QNativeSocketEngine and the tuiotouch
plugin in conditionals.
Change-Id: Ic6861b1c6a9e041fa8a50f96149f7280473a9fba
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
CoreFoundation event dispatcher on OS X can help users working with
QtBluetooth from either non-GUI threads or non-GUI applications.
Change-Id: Ie0793e7f49074a8ea9059251cd7c9bdf7953c206
Task-number: QTBUG-48758
Task-number: QTBUG-46625
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Call QWindowSystemInterface::handleWindowScreenChanged() in
QWindowsIntegration::createPlatformWindow() should the effective
screen be different from the initial screen due to the geometry.
In order to avoid code duplication in the Direct2D plugin, fold the
helper method createWindowData() into
QWindowsIntegration::createPlatformWindow() and introduce a new
virtual helper to create the actual platform window for
QWindowsDirect2DIntegration to hook in.
Task-number: QTBUG-48379
Change-Id: I49c7a50a12245691414ab49bffe802239a87d2d6
Reviewed-by: J-P Nurmi <jpnurmi@theqtcompany.com>
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
Precalculate everything we can and use faster loop.
Make inlined as it is used just in a single place.
Change-Id: If3c33d60739eb4ce896020321442ae81edd1c13d
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
These tricks has no meaning to other font engines (i.e. to FreeType).
Change-Id: I0410693d4d159367d760433062506691dc7c4610
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
This fixes-up bf013c9e74 by specifying
graphite2 shaper in a "safe" shaper list, one used for non-CoreText
font engines (currently, FreeType font engine only).
Make the the comment and the code a bit cleaner about the fallback
shaper list.
Change-Id: Ib0cedc0b593023c9ee771e8e98c314d9c8f13037
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Use static_cast instead of C-style cast;
ushort -> unsigned short;
Avoid a lvalue in qGetProp();
Minor styling fixes.
Change-Id: Ic6412e28ac9316a587d32d0e674d21f672967fd6
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Just like any other x_dependency.pri, it shall be included instead of
linking with png lib directly.
Change-Id: I698c8a8137f937d2e2c5f5dc9b9188b780c9e49a
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
QTextImageHandler needs it.
And QQuickImageBase will also need it.
Change-Id: Ica5d17519c4c6c06e25bab7d2b1d3733f4058425
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
* -style option is also used in other cases than widget world
Change-Id: I8555d309a7b9df0d26ad7a7b930411260537180e
Reviewed-by: J-P Nurmi <jpnurmi@theqtcompany.com>
json_p.h(283) : warning C4800: 'uint64_t' : forcing value to bool 'true' or 'false' (performance warning)
json_p.h(546) : warning C4800: 'uint32_t' : forcing value to bool 'true' or 'false' (performance warning)
qjsonvalue.cpp(176) : warning C4244: '=' : conversion from 'int64_t' to 'double', possible loss of data
qjsonvalue.cpp(518) : warning C4244: 'return' : conversion from 'const double' to 'int', possible loss of data
Change-Id: I2a24f90f7615aeb47f747ecbe3b580f23773ebda
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: hjk <hjk@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
On Unix we get a detailed error message when a process fails to start,
but later on we overwrite it with a generic "Process fails to start".
Fix this by keeping the original error message (if one is available).
This fixes a regression introduced in commit 5147f73ac3.
Task-number: QTBUG-49286
Change-Id: Idd0f0fed9773d39f2947fc3e532b51e670952caf
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The first two are the "proper" terms; the last is a colloquialism.
Also amended "daylight savings" (which summons to mind a hybrid of
"daylight robbery" and "bargain-basket savings").
Improved related wording in many of the places amended.
Task-number: QTBUG-49308
Change-Id: I726f18a344b2fe37f765a14684d1447c8b7ab00c
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
Status bar visibility can be controlled from dialog and
dialog size is screen size.
Change-Id: Ia8e932a9e1e1549a17d12532639391335760f9e0
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
WinRT/Windows Phone "coding guidelines" are now used for the native
socket engine as well.
- Whenever an operation is expected to succeed
Q_ASSERT_SUCCEEDED is used.
- QWinRTFunctions::await is used for waiting for async operations
- Improved error handling
Change-Id: I6c8d64731da5c94b911a5190231c7c8f68d9c261
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
The authorization dialog that grants the application access to the
assets will only show after returning back from
applicationDidFinishLaunching. Therefore, trying to load an asset from
main before qApp->exec() would normally fail. To remedy that, we added
a path that starts a QEventLoop for a split second to force the
application init process to finish. But this can only work if we
start it from the main thread.
A bug with this is seen in QML FileDialog, since it (clumsy enough)
starts to iterate, in a separate thread, all the files in its currently
set directory upon start-up construction (which should be fixed as
well). The result is that the application gets dead-locked on start-up.
Change-Id: I0047272d53314752903960b33f88b33dc0d7e78d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
The file picker must be used inside the Xaml thread, otherwise execution
will fail without giving any warning or error.
Task-number: QTBUG-48389
Change-Id: I917e88e95993da04be3d7cd34344ffd68ee71cdd
Reviewed-by: Samuel Nevala <samuel.nevala@intopalo.com>
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
GetConnectionProfiles does not list the current profile, so
"allInterfaces" also has to use GetInternetConnectionProfile.
Task-number: QTBUG-49121
Change-Id: I80a3ea1bfddfe502c84376ca90a7e1fe5d3020c3
Reviewed-by: Samuel Nevala <samuel.nevala@intopalo.com>
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
This is follow-up for QTabBar fix ea47d152b3.
In native OS X applications using mouse wheel on combo boxes have absolutely no
effect. We should bring the same behavior to Qt based OS X apps too, as users
are complaining of unexpected behavior, eg. randomly switching Qt Creator
sidebar mode when scrolling file list and moving mouse pointer little bit
above. Moreover inertial mouse behavior on OS X makes combo box usually move
several indexes, rather than single one on slight finger slide.
This also applies to iOS apps so the change affects all Apple platforms.
Task-number: QTBUG-10707
Change-Id: I6582265039198707ad8c2f54de96ee2a0b0e0b47
Reviewed-by: Adam Strzelecki <ono@java.pl>
Reviewed-by: Jake Petroules <jake.petroules@theqtcompany.com>
This allows requests to proceed without needing bearer plugins.
Task-number: QTBUG-49267
Change-Id: Ie5ce188ddefebd14d666bb5846e8f93ee2925ed1
Reviewed-by: Markus Goetz (Woboq GmbH) <markus@woboq.com>
This speeds up tests/auto/corelib/tools/qchar which spends a lot
of time in QTestTablePrivate::elementAt(), QTestTablePrivate::dataAt()
from 6s to 500ms on an average Linux machine.
Task-number: QTBUG-38890
Change-Id: I16b75819f88ca7974a581affd8409f7b4038a712
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
By setting them always inline we ensure they will never generate symbols.
Change-Id: I2c815c4fe2cb498187656e0b25ef4499258885ff
Reviewed-by: Erik Verbruggen <erik.verbruggen@theqtcompany.com>
Fix MSVC warning:
sql\drivers\psql\qsql_psql.cpp(452): warning C4267: 'argument': conversion from 'size_t' to 'int', possible loss of data
Change-Id: I44a902a9169efa568dcd82bdc08c5b97e661d65a
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@theqtcompany.com>
Given that % and / have a defined consistency relation, neither is any
better defined for -ve operands than the other. The code is in fact
avoiding using -ve operands, not preferring "well-defined" / over
(allegedly less so) %, as it claimed.
Change-Id: I666690872f078c0f21f9af7a227c2bbf291704e8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When inheritedMask is 0, the font inherits just everything; in this case
`inheritedFont.resolve(baseFont)` could be replaced with `baseFont`.
Change-Id: Ic3ed8ef174493544ada32037e7bdded46eb4bd43
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Allow for using other EGL implementations besides ANGLE.
Change-Id: I64ceca477912150425e4a9328001cc911bd351bf
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
The "previous" value is always 1 when the compare-and-swap succeeded,
instead of the previous value.
Instead of fixing this, let's just remove this file a bit earlier than
the rest. All of them will be removed in Qt 5.7 anyway, so let's leave
MIPS atomics to the compiler.
Task-number: QTBUG-49168
Change-Id: Idba8c29717f34c70a58fffff14133304595165f5
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lisandro Damián Nicanor Pérez Meyer <perezmeyer@gmail.com>
It's used from a different TU, too, so we can't rely on the
compiler to inspect it to draw this conclusion itself.
This function has technically not a wide contract, because
the output buffer needs to be large enough to hold the result.
However, this precondition cannot be checked from within the
function, therefore no assertion can ever be added and the
nothrow marker becomes acceptable (even desireable).
Change-Id: I2dc6c4f3d9d8147c6483865c5c4bbc8e9af291b7
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
'need' was never anything but zero, so drop it.
Change-Id: I4b52107afc7ed47c19ae1942cef0c92cbd0e1a36
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
A QAbstractSocket can be close()'d at any time, independently of its
current connection state. being closed means that we cannot use it to
read or write data, but internally it might still have some data to
send or receive, for example to an http server. We can even get a
connected() signal after close()'ing the socket.
We need to catch this condition and mark any pending data not yet
written to the socket for resending.
(cherry picked from commit 0df5d07929)
Task-number: QTBUG-48326
Change-Id: I67d9ad36f7288c9c6bef51aa6253d7b187737601
Reviewed-by: Ulf Hermann <ulf.hermann@theqtcompany.com>
Reviewed-by: Markus Goetz (Woboq GmbH) <markus@woboq.com>
* Georgian lari currency symbol
* A large collection of CJK unified ideographs
* Emoji symbols and symbol modifiers
* Letters to support the Ik language in Uganda, Kulango in
the Côte d’Ivoire, and other languages of Africa
* A set of lowercase Cherokee syllables, forming case pairs
with the existing Cherokee characters
* The Ahom script for support of the Tai Ahom language in India
* Arabic letters to support Arwi—the Tamil language written in the Arabic script
For more details, see http://www.unicode.org/versions/Unicode8.0.0/
[ChangeLog][QtCore] Unicode data updated to v.8.0
Change-Id: If255f95c9c45655b721369a116299da3cabbba0a
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Looks like it's not a good idea to send constant 'fake' mouse move events while
mouse is actually moving (mouseDragged/rightMouseDragged) + we're receiving
-dragginUpdated:.
Change-Id: Ibed5def3d8f06b764dea6c2cd196e37ca19ce967
Task-number: QTBUG-49204
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
Reviewed-by: Filipe Azevedo <filipe.azevedo@kdab.com>
reserve() is a good thing, and key to getting top performance out
of any data structure that provides them, yet the existing docs
scared people away by claiming that "you will rarely ever need to
call this function".
Change-Id: I88e30d96960443c0ed759a79c6fa9ee6af0d1e07
Reviewed-by: Mitch Curtis <mitch.curtis@theqtcompany.com>
This is in preparation of adding -Wzero-as-null-pointer-constant (or similar)
to the headers check.
Since QtConcurrent is basically all-templates, not all uses of 0 as nullptr
might have been detected by the headersclean check.
Task-number: QTBUG-45291
Change-Id: Iffcabdbab26b56597700346cd039d0c32c4fddca
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
High DPI should be disabled when the user chooses small fonts on a
High DPI monitor, resulting in lower logical DPI. The auto-algorithm
should then be equivalent to that of the device pixel ratio scaling
of Qt 5.5.
Task-number: QTBUG-49195
Change-Id: I756770ec7251b5b9b901253ded82e829aa50d38b
Reviewed-by: Alessandro Portale <alessandro.portale@theqtcompany.com>
A dialog embedded into QGraphicsView has Qt::WA_DontShowOnScreen set
(similar to a native dialog). It must not trigger the modal handling
though as not to lock up.
Task-number: QTBUG-49124
Change-Id: I22ce3f18d01df017b9317666770686bd4491387f
Reviewed-by: Andreas Aardal Hanssen <andreas@hanssen.name>
Fixes possible deadlock that occurs when synchronous WS event handling
(introduced in ee767c8) is used across threads.
Task-Id: QTBUG-49051
Change-Id: Iae973c2d4f4619b9eeb6e9393330b166ec608d27
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
This way we ensure that creation always works and we can access the
input pane statics.
Task-number: QTBUG-49034
Change-Id: I5d0340a7f6304b717f17f2106134fa9d083a7fde
Reviewed-by: Samuel Nevala <samuel.nevala@intopalo.com>
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Commit d020e0781c added a workaround to
the QtCore build due to a qmake bug in handling the extra target on a
Windows host. The workaround removed the tagging symbols from QtCore.
This commit removes the using of those symbols from everywhere else.
Task-number: QTBUG-49208
Change-Id: Idba8c29717f34c70a58fffff14133399f9f0b7f2
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Blocking the XAML thread can lead to a deadlock, so switch to the
non-blocking version of handleExtendendKeyEvent.
Task-Id: QTBUG-49051
Change-Id: I65a348af1f77b6afcd7d0fb9a80c70d60fc94c27
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
This ensures any non-standard fields (e.g. glibc's tm_gmtoff) are
clear, as well as the other fields mktime overtly promises to ignore.
Change-Id: I45d69eff7b5cb37ab910bf3d4d2d1481fd93aedb
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Make it possible to enable and disable both at the
environment variable level and at the source code
level. This applies to scaling done by Qt using display
density information provided by the operating system.
Disabling is done with a 'veto' system: both the environment
and source code my prevent the other for enabling
scaling. This covers use cases of 'my system does not
provide correct display metrics' and 'my application
needs access to display pixels', respectively.
On the environment, scaling is now enabled with
QT_AUTO_SCREEN_SCALE_FACTOR=1 and disabled with
QT_AUTO_SCREEN_SCALE_FACTOR=0. In source code the
corresponding application attributes are AA_EnableHighDpiScaling
and AA_DisapleHighDpiScaling. Not setting any of these
indicates 'no preference'.
The global scale factor set by QT_SCALE_FACTOR is
not affected by any if the disablers.
Task-number: QTBUG-46615
Change-Id: If18607d4b56ace1914a710e5aa60b2f0968e0010
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
Allow setting tryShortcutOverride for handleExtendedKeyEvent
with no timestamps.
Change-Id: I469b144cfcaf063861debe86195e1f96ac94cc37
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
When rebuilding the tab bar after hiding several dock widgets,
the index gets offset.
Task-number: QTBUG-49045
Change-Id: I05f6a976ca1d8c6f7cdf6532f1a728483398eabc
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
There is already a "wl_display" resource used by the client-side
Wayland plugin. This is confusing, and would cause horrible bugs
if ever a Wayland server was run as a client (the nested compositor
usecase). Therefore, use "server_wl_display" for the name of the
resource to be used on the server side.
Change-Id: I30455177a154e89b98ad3620b6a14626b821e54c
Reviewed-by: Giulio Camuffo <giulio.camuffo@jollamobile.com>
glslang_tab.cpp was not created in the right subdirectory.
This amends commit 69167189.
Change-Id: I031499baa53b72a1923883c58eb5ddfc2a02789a
Reviewed-by: Liang Qi <liang.qi@theqtcompany.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Since hinted text layouts are not scalable, we should disable
hinting by default when the high-dpi scaling is active.
This is the Windows version of 0f7bc885aa,
which solved the same issue for the fontconfig database.
Change-Id: I48b5da5b045dec195fd257743175017f39cf9620
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
If a native widget has a width or height of 0 we don't have to
invalidate its backing store as that is done by the window
system. Certain applications rely on ... interesting ... behavior
of certain window systems in this case.
Task-number: QTBUG-48321
Change-Id: I78ef29975181ee22429c9bd4b11d96d9e68b7a9c
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Alexander Volkov <a.volkov@rusbitech.ru>
Reviewed-by: Lisandro Damián Nicanor Pérez Meyer <perezmeyer@gmail.com>
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
Outside qmlscene, a window may be created before its size has been
set. So check if the size is valid before using it.
Also, allow an app to bypass the adjustment to the requested window
size that is normally done in order to leave room for the system's
status panel. Such adjustment does not make sense in the context of
autotests, and would often cause false test failures.
Change-Id: I1627a2e4c37c68ac61c4976be5b73045eb2bb989
Reviewed-by: Paul Olav Tvete <paul.tvete@theqtcompany.com>
They were using QStyleOption*V<N>, which are all obsolete since
Qt 5.0.
Change-Id: I6a8a666cf18d52fcac8f305f38c8ef0843ec7bfd
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
After canceling the asynchronous read operation, the
notified() slot receives ERROR_OPERATION_ABORTED.
We must not handle this situation as an error.
This amends commit 5ce567c5.
Task-number: QTBUG-48336
Change-Id: Iff948ceb3ad1f805a9de8c188fbc39ed4c76ba82
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
A wayland compositor on i.MX6 needs to create the wl_display
before creating the EGL display. This wl_display then needs to be
exposed so that QWaylandCompositor can use it.
Change-Id: Id60f6dd2fbba05140ca0671da6f17dbc2ecce3a3
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Duplicate identifiers were introduced for QString code snippets,
resulting in the wrong snippets to be quoted.
Change-Id: I75e9c99cbc73bf39262079f7f53342a6ac90c2a7
Task-number: QTBUG-49111
Reviewed-by: Venugopal Shivashankar <venugopal.shivashankar@digia.com>
When calling handleMouseEvent and similar, there are two choices when it
comes to the global and local position: by specifying the window it is the
caller's responsibility to provide a valid local position. When the window
is null, QGuiApplication calculates the local position. The winrt plugin
chose the former and therefore passing the global position as local is
wrong.
Task-number: QTBUG-48208
Change-Id: I3e1137cdb5d023296c4d73899da016641303c7df
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Also involves adding support for sharing contexts.
Task-number: QTBUG-48663
Change-Id: I0b18846ae70b63a0a21132f820a12ea744c0e936
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@theqtcompany.com>
Mention that it is still used by Active Qt.
Task-number: QTBUG-49148
Change-Id: Ic31b2f60b25886cd55c5fa516082d79311ab84cc
Reviewed-by: Fredrik Orderud <forderud@gmail.com>
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
The compiler-generated ones would break class
invariants.
This internal class so far isn't copied. This patch makes sure
that it stays that way.
Change-Id: I8bf75058edaf2adf148ae9edff4bf4e9a3b3368d
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Pass a QWindow pointer to QIcon::pixmap(), which
allows it to select a pixmap based on the target
device pixel ratio.
Change-Id: Ifb8e49c0eff0a42233a90eee3dfb995c2441bfd6
Task-number: QTBUG-38100
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
Pass a QWindow pointer to QIcon::pixmap(), which
allows it to select a pixmap based on the target
device pixel ratio.
Change-Id: Ib592de2a25060658b70b4e4ab7ff5fbaa9b60260
Task-number: QTBUG-44424
Task-number: QTBUG-38100
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
Running for instance the widget auto tests reveals a lot of egl warnings
due to items being created and not shown. Hence no surface was created
but tried to be destroyed when window was deleted.
Change-Id: I5c99eeb94a8fc2cfeb98f85445e013de61ff9ca9
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
After returning from main() it is not sufficient to exit the main
thread, it also needs to be ensured that the Xaml::IApplication object
gets notified by invoking Exit.
Task-number: QTBUG-49141
Change-Id: I8ca434be5f17ddddd465ede2a79585c28c51b3ef
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
The QWinRTEGLContext must not store the display, as it might get
destroyed while other objects still need it, for instance QWinRTWindow
to delete the surface. Rather create a global static for creating the
display once and delete it when application lifecycle ends.
Change-Id: Id176b6934e1d1327f5bb70ad0d258de91f675041
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
The minimum feature level is now detected before requesting a display.
If it is less than 9_3, use the WARP device instead.
Task-Id: QTBUG-44495
Task-Id: QTBUG-44694
Change-Id: I9f81f4f92269fab73c291f7373aa07236c7e5f98
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@theqtcompany.com>
Brings down gui/painting/qcolor from 7s to 3s, reducing the calls
to printf() helpers from 16E6 to 10E6.
Task-number: QTBUG-38890
Change-Id: I34065e6f09fc9a14920b06aa6936908229f3f9c4
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@theqtcompany.com>
A recent change uses an accessibility function even if accessibility
is disabled.
Change-Id: Ibf9afbb90120772065743b2f9dd3615e4e6a2f37
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
QAbstractSocketPrivate::readFromSocket() had a qDebug() that
references its q->... so it also needs its Q_Q() declaration.
Only relevant when QABSTRACTSOCKET_DEBUG is enabled (which it
normally isn't).
Change-Id: Ib82fd032fb2c4143a0987b9162377d0d7e968e95
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Since QSet<>::intersect() modifies the original
set, exitSetT1 has wrong content for next iterations.
Use intersects() instead.
Change-Id: I09e0961ec6dfb34ade88d48d1e009529aeab82b4
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
1. _HAS_DINKUM_CLIB is defined whenever a C header is included, even when
Dinkum is disabled.
2. _HAS_* macros are always defined, as either 0 or 1.
Change-Id: I727b854a6a733e2028e6facc327e264d0c4c9e90
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Somehow qmake doesn't add the correct rules for the Android makefiles,
so the build fails when cross-compiling from Windows. The reason for
that is unknown (could be related to that "qt_android_deps" config, but
that isn't used anywhere in qmake or the buildsystem).
This isn't likely to be a problem, since there are no global installs of
Qt on Android.
Change-Id: I1d0f78915b5942aab07cffff140f95ce32324030
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The QGuiApplicationPrivate object is already destroyed by the time
the plugins are cleaned up during the application destruction, causing
a segmentation fault in updateInputDeviceCount().
There's no point in calling updateInputDeviceCount() in the destructor
anyway as the whole process is on its way out that stage, and we
don't support unloading plugins during the application lifetime
otherwise, so the call can just be removed from the destructor.
Change-Id: Id819d73cb8234ccedb6ea7c3e39950589ee680a1
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
The Windows code was always returning AnyScript, which in turn made
QLocale::textDirection() for the default and system locales always
return LTR, even if the Windows UI was in RTL mode.
[ChangeLog][QtCore][QLocale] Fixed a bug that caused
QLocale::textDirection() to always return Qt::LeftToRight and
QLocale::script() to return QLocale::AnyScript on for the Windows system
locale.
Task-number: QTBUG-49031
Change-Id: I7e6338336dd6468ead24ffff14112c8d348eedba
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Move touch device initialization code to QWindowsContext and
replace the assert on the touch device in QWindowsMouseHandler
by a call to the initTouch().
Task-number: QTBUG-48849
Change-Id: If8573b8283ef94e7fd015f6edc626e3c8cc0b139
Reviewed-by: Joni Poikelin <joni.poikelin@theqtcompany.com>
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
Reviewed-by: Andreas Holzammer <andreas.holzammer@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
The latter is much faster as it doesn't have to deal with
time zones.
This change is safe, because the QDateTimes are only used for
comparison with other QDateTimes, which, from a quick glance around,
seem to be mostly, if not exclusively, in UTC. Comparsions work
across time zones, but the comparison between UTC date-times is
fastest.
Credits to Milian Wolff, from whose QtWS15 talk this advice is
taken.
Change-Id: I6859d886d8dc8e0a52fbe394fbb7b93a87b4739a
Reviewed-by: Milian Wolff <milian.wolff@kdab.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Skip the correct number of bytes in the read buffer when expanding '\n'
into "\r\n" upon writing.
Change-Id: I5b01fc47c330dee5c83001abf0acd7d63d790b96
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Since there are circumstances where the session is not yet created after
start() is called then we should check if the session is valid before
connecting to it.
Change-Id: I94236f76e4be2433a8c96eb91ce2d4b4d42f2fd9
Reviewed-by: Richard J. Moore <rich@kde.org>
- On NetBSD, the defines were not properly defined;
- On all other BSD systems, we use statfs which does not have f_frsize
member, revert to using f_bsize there.
Task-number: QTBUG-48267
Change-Id: Ia1ed484ac61a615fcbb5b45affb516b5e86a64b0
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Lisandro Damián Nicanor Pérez Meyer <perezmeyer@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The flex and bison rules generated the same files for debug and release.
This conflicts in parallel builds, e.g. when using jom.
Task-number: QTBUG-49003
Change-Id: I6f64760d83292e8bfad6fd0b4f8fbdd386e2213b
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Enclose all debug operators within QT_NO_DEBUG_STREAM, declare
all public ones and make the others static. Add operators for
POINT and WINDOWPLACEMENT.
Task-number: QTBUG-48449
Change-Id: I33f2dba2bf486a8f5cb11f11f4e2b37cce086def
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
This may result in incrementing an invalid iterator after the
iterator's container has changed. Also, for this to happen,
the view needs to have an active editor.
The reentrant call happens as follows in QTreeView, after the
model adds new rows to the view:
QTreeView::rowsInserted()
QAbstractItemView::rowsInserted()
QAbstractItemView::updateEditorGeometries()
QTreeView::visualRect()
QAbstractItemViewPrivate::executePostedLayout()
QTreeView::doItemsLayout()
QAbstractItemView::doItemsLayout()
QTreeView::updateGeometries()
QAbstractItemView::updateGeometries()
QAbstractItemView::updateEditorGeometries()
Other concrete item view classes may be prone to the same issue.
The fix consists in relayouting the items if needed, which should
trigger calling updateEditorGeometries() again. This doesn't
invalidate previous optimizations regarding item relayouting since
we only force relayouting when it'll be done by visualRect().
Change-Id: Id31507fdc8d9a84d50265298191d690d1a06792b
Task-number: QTBUG-48968
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The printf-style version of QDebug expands to a lot less code than the
std::ostream-style version. Of course, you pay in type safety (but
compilers warn about it these days), you cannot stream complex Qt
types and streaming QStrings is awkward, but in many cases you
actually improve on readability.
But the main reason is that something that's not supposed to be
executed under normal operation has no business bloating executable
code size.
This is not an attempt at converting all qWarnings() to printf-style,
only the low-hanging fruit.
In this first part, replace
qWarning() << "...";
with
qWarning("...");
Change-Id: I1f6869b024103c98262f1deae014e0d7337a541c
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The latter is much faster as it doesn't have to deal with
time zones.
This change is safe, because the QDateTime is only used to
compare against a copy of itself, adjusted by some seconds,
which doen't care which TZ the date-time is in, to get a
time_t, which is much faster done from a UTC time, and
to calculate the seconds to expirationTime, which, from a
quick glance around, seem to be mostly, if not exclusively,
in UTC. secsTo() works across time zones, but the comparison
between UTC date-times is fastest.
Credits to Milian Wolff, from whose QtWS15 talk this advice is
taken.
Change-Id: I22aa5350b29493f01fb503ef5ec68a964ca95cf3
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Back in the old days, we would pad the output from the font engines
to work around problems with the GL1 paint engine such as issues
with linear sampling, etc. This is no longer needed. Padding is
moved into the glyph cache, and in addition to reducing performance,
the extra padding is also making alphaMapForGlyph() and
alphaMapBoundingBox() fall out of sync when you fall back to
QPainterPath drawing the glyph. The result of this was that,
when prepared, the glyph cache was sometimes not made large enough
to hold what alphaMapForGlyph() actually produced, depending on
the size and order of glyphs, and glyphs ending up at the end of rows
would sometimes be missing from the output.
[ChangeLog][Text] Fixed some instances of missing glyphs when
drawing large fonts.
Change-Id: Ia5982392fe1637f6ebc740db9f226fbb91f75166
Task-number: QTBUG-47547
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
This commit removes the legacy ptrsize check, which was deficient
because it did not work for multiarch systems (when we supported fat
OS X binaries) and did not work for bootstrap builds because the size
might be different when cross-compiling.
Instead, let's rely on the predefined preprocessor macros to detect
correctly. As a nice side-effect, this fixes 64-bit Android builds
cross-compiled from Windows.
Task-number: QTBUG-48932
Change-Id: I1d0f78915b5942aab07cffff140f9a52b9342f23
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Erik Verbruggen <erik.verbruggen@theqtcompany.com>
Windows 10 uses one enum for all target platforms. Use
ApiInformationStatics to identify whether enum values are accessible at
runtime.
Change-Id: Ib77c9d2a2b5cf1655fbe7d937d0c83cc4cdd9ee9
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
[ChangeLog][QtCore][Android]
Fixed the opening of a local file using QDesktopServices::openUrl().
For a local file to be viewed with
QDesktopServices::openUrl(), Android needs to be given
the MIME type otherwise it does not start an Intent to view the file.
Task-number: QTBUG-45585
Change-Id: Ifcfce4bff35011f205cfadbdb2b37a1780dac87d
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Do not emit _q_notified when we're in a wait function. Otherwise,
the queued signals could pile up in the event queue.
Task-number: QTBUG-48653
Change-Id: I071863e2356e17c7004e3b7ca359967cb115e343
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
That is the case on Hurd, where the code currently breaks because Hurd
does not have WEXITED or WNOWAIT defined.
Change-Id: I4b13633612b1168d36c949d9e8b35bc05bca7d5c
Reviewed-by: Lisandro Damián Nicanor Pérez Meyer <perezmeyer@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Handle various cases where we have null QScreen or QPlatformScreen
pointers. With this change, I can run Qt Creator for several days.
Before, it would crash multiple times per day with a two-monitor
setup.
Change-Id: I0923d886ae2a4199ac37edd711ddd4f6f99df93d
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@theqtcompany.com>
Opt-in by setting android.app.auto_screen_scale_factor
to true in AndroidManifest.xml. This will enable
devicePixelRatio scaling in QtGui using a scale
factor provided by Android DisplayMetrics.
Note that the Android style is not currently supported:
it already accounts for different display densities
which results in incorrect visual sizes when enabling
devicePixelRatio scaling.
Implementation: Bring DisplayMetrics::density through
to setDisplayMetrics in androidjnimain.cpp, similar
to what is done for "scaledDensity". Override
QPlatformScreen::pixelDensity(), which forwards the
scale factor to QtGui.
[The difference between "density" and "scaledDensity"
is that the former is a physical display density factor
which corresponds closely to devicePixelRatio in Qt,
while the latter also includes the Android global font
scale factor.]
Scale the global font pixel size in qandroidplatformtheme.cpp
to keep the visual font size constant.
Based on an initial patch from Daiwei Li <daiweili@suitabletech.com>
Task-number: QTBUG-46615
Change-Id: Ia51f99bf6dda485a57413949246c7b32cb47e8dd
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Call QHighDpiScaling::updateHighDpiScaling() in init_plaform(),
after the platform integration has been created and most platforms
have populated the screen list. Keep the existing udpate call for
the platforms that don't, but guard against calling it twice.
Task-number: QTBUG-47947
Change-Id: Ib73bea7c4ab42e7acf6532f3a3100e1fc29acc2c
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
The latter is much faster as it doesn't have to deal with
time zones.
This change is safe, because the timestamp member is only
ever handled inside, and the calculation of the time difference
does not depend on any particular time zone.
Credits to Milian Wolff, from whose QtWS15 talk this advice is
taken.
Change-Id: I6c9190a4253ce5972871ab1f12870f8ae9891966
Reviewed-by: Milian Wolff <milian.wolff@kdab.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Truetype fonts should be rasterized with a winding fill
as documented in e.g. Microsoft's specs.
Failing to do this caused a bug when doing native
rendering in Qt Quick for fonts that were large enough
that the fallback path was taken when drawing the glyphs
into the cache. If the glyph had overlapping contours,
they would be subtracted from the shape.
[ChangeLog][Text] Fixed an uncommon rendering error with
fonts containing overlapping contours.
Task-number: QTBUG-41197
Change-Id: I0e4a4432ba3f902bc3ea59d8f4dbd12a295630b2
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Partially revert c7abf81786. Instead of using the ES3 entry point,
use ES2 for framebuffer blitting. This means that a small change is
required to ANGLE for the blit behave the same as ES3 (applied only for
Windows Store apps).
Task-Id: QTBUG-48266
Change-Id: Idc51f00a659c91f740876be071eb71bff69e0e38
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
_mm_load_si128 requires 16 byte alignment. This crashes on 32 bit Windows
builds.
Change-Id: Ib6c30eba726747bbab56467eada820521981a80c
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
It should be possible to use these constants simultaneously and to
handle them separately from each other.
Change-Id: I0c48a3c25456b487c9d6139b05105ada20f34be6
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This is true for ILP32 on x86-64, IA-64 and AArch64.
Change-Id: I1d0f78915b5942aab07cffff140f9d4c277bb5d4
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
The latter is much faster as it doesn't have to deal with
time zones.
This change handles the trivial ones: Either the call to
currentDateTime() is immediately followed by a call to toUTC()
or toTime_t(). The latter is much faster on UTC QDateTimes, too.
Credits to Milian Wolff, from whose QtWS15 talk this advice is
taken.
Change-Id: I872f5bbb26cbecedc1e5c0dbee4d5ac2c6eb67ee
Reviewed-by: Milian Wolff <milian.wolff@kdab.com>
The QSslSocket::addCaCertificates() variant that takes a path
argument uses QSslCertificate::fromPath() in its implementation.
Edit the description of the former to match that of the latter.
Fix minor issues in QSslCertificate::fromPath() documentation;
add a missing word, limit code snippet line width.
Task-number: QTBUG-47359
Change-Id: Ibead74c998503e60a67d0b8eb551536bd20feff8
Reviewed-by: Venugopal Shivashankar <venugopal.shivashankar@digia.com>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
Reviewed-by: Richard J. Moore <rich@kde.org>
We check the name of the window class the widget's QNSView
changes window and set a flag when the that window is a native
Cocoa menu window. Later, only those views not inside a native
menu can become first responder, ensuring Qt won't deactivate
the main window.
We're allowed to reject becoming the first responder mainly
because Cocoa itself doesn't support sending key event to menu
views and, therefore, it doesn't change what's already possible.
This patch also sets the widget action visible, which needs to
be done right after reparenting it to the container widget.
Besides that, it also contains a few small code cleaning changes
related to Cocoa's support of QWidgetAction.
Change-Id: Ia2170bdc5e1f40bfa2f1091c05e9e99397c47187
Task-number: QTBUG-44015
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@theqtcompany.com>
dragImage:at:offset:event:pasteboard:source:slideBack: on NSView is
deprecated since 10.7. Use the one on NSWindow instead.
Change-Id: Ia1c2ea367ae2ca5194b52ea57ab261461bf8b529
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
Instead of converting the QLatin1String to a QString at the first
opportunity, keep it around until it is appended to one of the
internal QStrings in write().
Avoids a memory allocation per QLatin1String / const char* streamed.
Change-Id: Id973a9b743e5a6696defbc4ef4ed2db1ef54e9cc
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In case the static destructor has already been run, make sure we don't
crash. This shouldn't happen, but could happen if the QApplication
destructor is run in a weird order (after the static destructors have
begun running).
That's not usually a case we'd fix (unsupported), but since this change
improves the code and also avoids creating the pixmap cache if it hadn't
been used up until this point, the change is a net benefit.
Task-number: QTBUG-48709
Change-Id: Ia505aece07bf4e13a1faffff140f3e119cfc773e
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
On 32 bit platforms the pointers may end up being 4 byte aligned.
Happens with MSVC on 32 bit Windows. _mm_store_si128 is documented to
require 16 byte alignment.
Change-Id: I80737fedf9e7f436a51a83924117cc0bc63017cc
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
This is in preparation of adding a putString(QLatin1String) overload.
To keep the change simple, I kept the construction of the padChar
chain per putString() call. It probably makes sense to have a
QString::resize(int, QChar) to perform the padding operation in-place.
Let's leave that for another changeset.
Change-Id: I9ef66b8df38117e1669fd2bece6ee180a2ce3369
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Compare against isVisibleTo() in QLineEditIconButton::actionEvent()
so that action events received before show() are handled correctly.
Fix a regression introduced by change
4dccb2ca67 for handling action
events causing side widgets to overlap when added before the widget was
shown. Use QAction::isVisible() to determine visibility.
Task-number: QTBUG-48806
Task-number: QTBUG-48899
Task-number: QTBUG-39660
Change-Id: I7a39a3b9a094f2c74cde09544f1158deb2b81cf2
Reviewed-by: David Faure <david.faure@kdab.com>
This was never a documented feature, but happended to work before Qt 5.5.
It broke because the peoperty access went into the static function and are
now prefixed with '_t->'
So restore the behavior as it was by not including the parentheses in the
member name.
Task-number: QTBUG-47695
Change-Id: Ic3509ddea7ac9abc871e71f5bfbe81d04d08e9bc
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
Windows Runtime side callback is run from XAML thread and receiver
is at UI thread thus sendEvent asserts. Use synchronous system
interface key event handler to deliver the event.
Task-Id: QTBUG-48105
Change-Id: I91a8ef6fd29c277edfb699b688b9e7895dadda8f
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Marc Mutz
Friedemann Kleint
Tim Blechmann
Oliver Wolff
Tor Arne Vestbø
Alex Trotsenko
Morten Johan Sørvig
Samuel Nevala
Thiago Macieira
Oswald Buddenhagen
Alexander Volkov
Volker Krause
Dmitry Shachnev
Liang Qi
Maurice Kalinowski
Aleix Pol
Nico Vertriest
Konstantin Ritt
Eskil Abrahamsen Blomfeldt
Timur Pocheptsov
Olivier Goffart
Orgad Shaneh
Kai Koehne
Edward Welbourne
Richard Moe Gustavsen
Adam Strzelecki
Lorn Potter
Allan Sandfeld Jensen
Ulf Hermann
Joni Poikelin
Paul Olav Tvete
Eirik Aavitsland
Joerg Bornemann
Topi Reinio
Laszlo Agocs
Takao Fujiwara
Morten Johan Sørvig
Andrew Knight
Dyami Caliri
Jarek Kobus
Rafael Roquetto
Romain Pokrzywka
Andy Shaw
Gabriel de Dietrich
Martin Koller
Maks Naumov
Shawn Rutledge
Frederik Gladhorn
Erik Verbruggen
Lars Knoll
Oswald Buddenhagen