... by implementing a fake ln in qmake.
symlinks are supported only since vista (we officially still support
xp), and even there are permission-restricted (MS being (rightfully)
afraid of symlink attacks). so we fake the links by copying the files
instead.
the previous hack was a bit naive, simply using cp/copy instead of ln.
this didn't work with relative paths, as real symlinks are resolved
against their parent directory, not the working directory of the "ln"
command. the new fake does this correctly.
Change-Id: Ia2f5d68a39d6ffcc8a4383f9d0fc63a9da0a05c3
Reviewed-by: Joerg Bornemann <joerg.bornemann@theqtcompany.com>
Probably not performance-relevant here, but easier to read and safer,
too (consider what happens if frect becomes QRectF).
Change-Id: Ia5e4369b550b3a5b899b41ac4cbbda91229f2eb4
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
According to NETBSD manual pages, there's no statfs structure; statvfs
should be used instead, change introduces defines for the stat(v)fs
struct/function.
Task-number: QTBUG-40785
Change-Id: I98599e4635e46f90ffcc99c768f4c250f09f914f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The Windows 10 CRT does not export __getmainargs(), so we need to move
to GetCommandLine like on WinCE and desktop Windows. As CommandLineToArgv
is not available, the command line is split according by spaces, allowing
for quotes (and quote escaping) around arguments.
Change-Id: Ibc969df94ca5423a3a71d8190bbacd201189ea19
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Reviewed-by: Maurice Kalinowski <maurice.kalinowski@theqtcompany.com>
When running a modal session with NSOpenPanel (QFileDialog), our menu delegate
should not touch qApp->focusObject, since it's not what actually has focus
inside NSOpenPanel (will be some native view). Return YES instead.
Task-number: QTBUG-17291
Change-Id: I94f3281237fb25495d317b02310bf9d77b21d2ba
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
De-duplicates vtables and enables RTTI on this class
hierarchy.
Export, as QDeclarativeTranslatePrivate (QtQuick1)
inherits it.
Change-Id: I32439c59f7d4bff466ecae969bb761a42c3837f1
Reported-by: Volker Krause <volker.krause@kdab.com>
Task-number: QTBUG-45582
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Switch to using the pointer events from XI2 when touch is available (i.e.
version is >= 2.2). This allows us to select and grab the button and motion
events together with the touch ones. This prevents the issue of not getting
touch events when grabbing via the plain xcb functions.
To prevent touch sequences from being replayed after ungrabbing (for example after
dismissing a popup that caused a grab), we try to accept touches via XIAllowTouchEvents.
Unfortunately this leads to a deadlock and therefore we can only do it when we know
we have a new enough libXi. This is a configure time check which is not ideal since
the system on which apps run can have a newer libXi than the machine that did the Qt
build, but seems like the best we can do.
The environment variable QT_XCB_NO_XI2_MOUSE can be set to 1 in order to prevent
processing mouse events through XInput. This restores the old behavior with broken
grabbing.
[ChangeLog][QtGui] Pointer event delivery on X11 is now done via XInput 2.2+ when available.
Done-with: Michal Klocek <michal.klocek@theqtcompany.com>
Done-with: Alexander Volkov <a.volkov@rusbitech.ru>
Task-number: QTBUG-43525
Task-number: QTBUG-45054
Task-number: QTBUG-30417
Change-Id: I7cb2002b31bef4cd527aa427549dcf2d5467968e
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
With Qt 5.5 we are changing the license of the Qt for the WinRT &
WinCE ports to LGPLv3 / GPLv2+ / commercial.
Change-Id: I221559c5c42b1dcda172eb85e6bfa53c91976b23
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
De-duplicates vtables and enables RTTI on this class
hierarchy.
Export, as QPlainTextDocumentLayoutPrivate (QtWidgets)
inherits it.
Change-Id: I8aee43b43898a72430d8bc1247ea76c7d4e9031d
Reported-by: Volker Krause <volker.krause@kdab.com>
Task-number: QTBUG-45582
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Change-Id: Ia825ef6d932fbd45ceaf8c54a3207f207fa40994
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Subpixel rendered text doesn't work in the old OpenGL paint engine
because it assumes the glyphs are returned in RGB32 format, when they
may be in ARGB32. This patch changes the test of the returned format
to just check for 32bit matching the logic in the new OpenGL paint
engine.
Change-Id: Ib5b784dedba51cf22f216e2f035361518610b96b
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Pass a temporary QVector instead of a named variable. Apart from less function
calls, it also makes the code C++11-ready (enables move semantics).
Change-Id: Ia7d3bed8aeae48453f96dc3c979f7ae394a5a26d
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
QTextLayout::additionalFormats setters and getters using QList<FormatRange> have
been deprecated; port to the QVector versions.
Moved op== definition for FormatRange needed in tst_qsyntaxhighlighter.cpp
to a friend declaration in FormatRange itself, because MSVC 2008 doesn't find
it otherwise.
Change-Id: Ibab6589df057f02377d895079b56395859e3401e
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
To avoid a touch release outside the combobox triggering the popup,
a check was added that the combobox was hit. This fails if the combox
itself is only used for the popup and associated behavior, but does
not exist as widget. This patch changes the check so that only touch
release checks for a hit, but a generic click still behaves as in 5.4
to avoid regressions.
This fixes comboboxes no longer working in QtWebKit, since QtWebKit
renders its own comboxes in webpages, and only uses the popup of
the QComboBox.
Task-number: QTBUG-46152
Change-Id: I74fd57b2e42e77aa4a269d812ca4a6689f254889
Reviewed-by: Florian Bruhin <qt-project.org@the-compiler.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
When invoking a macro with less argument than it expect, we would
crash trying to access the vector of arguments from the invocation
as we are trying to substitute an argument.
(Note that we do not show an error in case of argument mismatch
because ithat might happen parsing valid code as moc's c++ parser
is not 100% accurate (that was QTBUG-29331))
Task-number: QTBUG-46210
Change-Id: I3f08d7f5049e593a5bdc02a594ea63cadf66e7a4
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
The pattern "mySet.intersect(other).isEmpty()" has been spotted in
the wild and in Qt codebase. intersects() is much cheaper because it
bails out as soon as we find one common item and doesn't do any
allocations.
[ChangeLog][QtCore][QSet] Added intersects().
Change-Id: I44a350dc4cdb9deb835a23eee99fc99d6ca24c82
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Needed for QSet::intersects() for optimization purposes. No need
to calculate the hash when we already have it.
Change-Id: I247602bb0558ca8d1fb8333de9d5f339146c576d
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
This reverts commit c9aaa3e2cd.
This fix broke fullscreen mode on Android. A better solution is
in the works, but we'll probably push that to Qt 5.5.1 instead
since it comes with a risk and we don't want to delay the release
any further.
Change-Id: I3aae6d52ebb8425089cdb6f7fc4c8ce9ad4911df
Task-number: QTBUG-38700
Task-number: QTBUG-46234
Reviewed-by: BogDan Vatra <bogdan@kde.org>
The range sampling operates the same in both directions just with
different step-sizes, so the code can be unduplicated,
Change-Id: I47805a7e925d4058f62c558ef08e79485915e937
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
The global variable is a left over from when there was an MMX assembler
implementation, and is now just making the compiler's job harder.
Change-Id: I686704b64a2f8c68ec8ca83f2ac3e465ded773e0
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
OpenSSL should automatically seed the random number generator, and the
current fallback code does not offer strong guarantees. It should
never actually be used, so let's remove it entirely.
Change-Id: I9a0a70345fe8df5913871bcac644623b687aea01
Reviewed-by: Peter Hartmann <peter-qt@hartmann.tk>
Add CONFIG += testcase to the .pro file which was missing.
The test thus was never executed by make check and left to
rot.
Fix up code:
- Remove module includes.
- Introduce explicit constructors taking parent object
for helper classes, removing calls to setParent().
- Ensure test does not leak objects by converting pointers
to stack variables or introducing QScopedPointer, verify
by checking for an empty window list in cleanup().
- Simplify code by removing unneeded variables.
- Split up conditions in QVERIFY().
Fix tests:
- Show windows were required when events are sent to
QWidget::windowHandle().
- Invert the conditions checking whether touch events are
accepted by widgets since widgets no longer accept them
by defaults in Qt 5 after
e50416066c.
- XFAIL multiPointRawEventTranslationOnTouchPad()
which started to fail at some point in Qt 5.
- Mark as insignificant on OS X due to crash.
Task-number: QTBUG-46266
Change-Id: I6676d021afb015411a24d97d9b8f7c327d4d3c3f
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
The positioning test has been observed to fail on X11.
Change-Id: I58727126a8742de93ec203e9992a9ae1b454f731
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Commit e486d69133 changed the behavior of
QString::fromLocal8Bit(const QByteArray &) to preserve embedded null
characters. The embedded null character from qt_error_string()'s buffer is not
something we want to preserve, it is merely a safe guard. Therefore let's strip
it away.
Change-Id: Iceac91551f51a1036a942ff30d246baea7a6fd7c
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
The QTextLayout::FormatRange is larger than void* and thus should not be held in
QList. Use a QVector instead.
Other parts of Qt already hold FormatRanges in QVectors, so this also makes
handling FormatRanges more consistent.
To avoid ugly names for the getter which doesn't overload on return type alone),
rename the set of function to format (from additionalFormats).
[ChangeLog][QtGui][QTextLayout] Added QVector-based alternatives setFormat(),
format(), and clearFormat() to setAdditionalFormats(), additionalFormats(),
and clearAdditionalFormats(), resp.
Change-Id: Ie04a561b43c91c3b2befb3cac2981821f84d5f77
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
As pointed out in the bug, it also fixes API use when configured
with no bearermanagement.
Task-number: QTBUG-46239
Change-Id: Ief8df85ad6acf61e8d5bb3eed54e7d6ecb84c1a0
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
QVistaHelper::drawTitleBar() used the font returned by
QApplication::font("QMdiSubWindowTitleBar")
(typically "MS Shell Dlg 2",16) to calculate the
bounding rectangle of the title text. However, if the window
is a toplevel QVistaHelper::drawTitleText() uses the theme font
obtained for WIZ_TMT_CAPTIONFONT (typically "Segoe UI",11.25)
to draw the title (since it is a window title). This causes the
font to be cropped when changing the application font or spurious
black rectangles to occur.
Fix this by exposing QWindowsFontDatabase::LOGFONT_to_QFont() via
QWindowsNativeInterface, and creating a QFont from the LOGFONT
obtained for WIZ_TMT_CAPTIONFONT and using that for the
bounding rectangle in the case of toplevel windows.
Split up the HFONT QVistaHelper::getCaptionFont(HANDLE hTheme)
into static LOGFONT getCaptionLogFont(HANDLE hTheme) and use
that to obtain the HFONT in drawTitleText() or QFont in
static QFont getCaptionQFont(), respectively.
Task-number: QTBUG-46360
Change-Id: I9069b403f7f948b6738eec452cb7584be45b8a29
Reviewed-by: Oliver Wolff <oliver.wolff@theqtcompany.com>
The -fPIE option is now accepted when using GCC 4, which means it
is available for backward compatibility for clients using
CMake 2.8.11 or older which makes use of the
POSITION_INDEPENDENT_CODE feature.
Conditionally use that feature for old versions of cmake with
GCC 4. Restore the tests for those versions, and clarify the
situation in the ChangeLog.
Change-Id: I5a06b155dda7db559d86841a2b34fd8ed95acbd0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Commit 3eca75de67 introduced the #error
nagging about use of -fPIE, but it makes the transition quite difficult
for people using other buildsystems. So let's give people a grace period
and enforce only for GCC >= 5.
Clang is affected, but differently. The problem only happens with -flto
-- that is, it happens when the linker detects that it's creating a
final executable. Maybe -Wl,-pie would fix it.
Change-Id: If4d5ac8db0ed4a84a3eaffff13e275edc29a72b7
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
We take some shortcuts still because we know that with ANGLE the header
with the extension constants is always available. A proper implementation
would not rely on the constants being available and would dynamically
check for the extension and would take care of defining the constants if
not available. However, just getting the extension list to check if the
functions needed to get the display are available is already a chicken-egg
problem so we won't go there.
Using eglGetProcAddress properly solves the issues with static builds too
since this always works.
Task-number: QTBUG-46284
Change-Id: Iff23669ebacaffa0c5f76fd2c928af689307874f
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Gunnar Roth
Reviewed-by: Andrew Knight <andrew.knight@intopalo.com>
Old implementation had complexity O(oldSectionCount); replace it with
O(hiddenSectionCount) algorithm. This boosts performance in case of the
vertical headers for models with big row count.
Change-Id: I7bb02f5579ce83fbdecf5f8c3aa7dcc0ac60dd40
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
This is a follow-up change to commit 3eca75de67
to mention changes required to CMakeLists.txt for users of Qt.
Change-Id: I1c9ed162427cdc620f998ccf266d59886901c28d
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Must not add -L without a directory path, so check that
QMAKE_LIBDIR_EGL actually has a directory to add.
Change-Id: I81920e3427f348739ced045a83f53265d72b261a
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
EGLFS backend does not use global sharing context. WebEngineWidgets was
failing to access textures created by WebEngineChromium and textures
were ending up as black rectangles.
This fix initialises window surface context with shared context and
fixes the bug.
Change-Id: I97189c06ee593ba55f353f44c23233175ebd3cba
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
The symbol is simply renamed in static builds, allowing wrappers of the
library to still call the function if needed.
Task-number: QTBUG-46209
Change-Id: I5d4ad2df59f206a3794b99364d122f9d0f12f8c6
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Tim Blechmann <tim@klingt.org>
Reviewed-by: Kai Koehne <kai.koehne@theqtcompany.com>
In order to ensure that the same text will be used in both QDateTimeParser
and QDateTimeEdit, use the QDateTimeParser context for the AM and PM
strings.
Task-number: QTBUG-251
Change-Id: I89b0809825251181440bf19cbe5828024a43acfb
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
After commit 3eca75de67 the use of
<Module>_EXECUTABLE_COMPILE_FLAGS becomes mandatory for 2.8.11 or older.
Therefore use it in test_use_modules - that's supposed to work with all
cmake versions it seems - and bump test_interface to require 2.8.12.
Change-Id: I7cfb6c6f1e8a97be916d372b9d9148490926693c
Reviewed-by: Lisandro Damián Nicanor Pérez Meyer <perezmeyer@gmail.com>
Reviewed-by: Stephen Kelly <steveire@gmail.com>
We have at least 5 different (but equal) implementations of a wrapper
in Qt, and some code uses explicit NSAutoreleasePools. Having a shared
implementation lets us clean up things a bit and makes it easier to
reason about which pools are actually needed.
Change-Id: I2fd8eefc3ae7308595ef9899b7820206268362a5
Reviewed-by: Tim Blechmann <tim@klingt.org>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Fixed regression from Qt4 that close button could't be disabled
from the tool window. With this patch buttons are hidden before
using button hints like it's on Windows.
Change-Id: I00f03ff9528ccae3b1136312404452b9415953b4
Task-number: QTBUG-45971
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
This involves deinlining some private class stuff to keep things
readable and maintainable.
Task-number: QTBUG-46002
Change-Id: Ie2888aa6c16a6f5182b61fbaa43288cfcc96cbc4
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
The connection to the deviceListChanged() signal may be queued. To make it
work our custom types have to be registered.
The problem is only visible with input backends like evdevtouch that live
on their own thread.
Task-number: QTBUG-46069
Change-Id: I4c03e8031e4337b5e711a3bd2cf405d15d6ce214
Reviewed-by: Gatis Paeglis <gatis.paeglis@digia.com>
