QNX posix_spawn() implementation actually allows for detecting whether a
non-existent process has failed to start.
Change-Id: Ic1bf8da0d4636f1d7d9b7b4cf6ad45376f6df0ed
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Due to a bug in the internal implementation of posix_spawnp on QNX, all file
descriptors with a value lower than the maximum file descriptor specified
in the file_actions structure are duplicated by default, ignoring the
FD_CLOEXEC flag. This includes all file descriptors that we are not working
with. So we add those file descriptors that have the FD_CLOEXEC flag to the
file_actions structure as close actions.
Change-Id: I316bc334addb46a4b84c199a69e9bd291ca706c5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Most posix_spawn implementations are done using fork(), so the only way
to report errors afer fork() is via a special exit code.
Reference: http://pubs.opengroup.org/onlinepubs/009695399/functions/posix_spawn.html
Change-Id: I3a37f81b0cb278bb31e5cb83c87e6b4c034dbc19
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Replace the existing code in QProcess that dealt with signaling of child
processes exiting with forkfd and spawnfd. The previous code was
convoluted and hard to maintain, having shown its age in the last
year. I've been running it for a year and a half and the new
implementation is definitely an improvement.
This change replaces support for the QNX Neutrino spawn() call with the
POSIX version. We lose the ability to do setsid(), but we gain quite a
few ioctls() that were done to fill in the file descriptor mapping
structure. That's also the only OS for which we have the ability to
thread-safely chdir() before the call to spawnfd().
Another advantage is that forkfd does not require a dedicated thread
running to handle child processes exiting.
Change-Id: I5eb76821dfdb6a8ed2989d7f53b3c31e515c3174
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
In certain environments, using fork() is not recommended due to the need
for an MMU. This commit adds support for those environments, by using
posix_spawn. Limitations of this environment are:
- we cannot reliably detect failure to exec (e.g. non-existing executable)
- we cannot do setsid(); we do setpgrp(0, 0) instead
- we cannot thread-safely chdir() to the requested dir
Because of the former limitation, the QProcess unit tests that rely on
failure-to-start error conditions are either skipped or marked as
expected failures. There's a non-reliable solution that is implemented
in a another commit.
This change also makes it easier to transition the QNX builds to using
fork(), which is supported from QNX Neutrino 6.6 and onwards.
Change-Id: I5cb46abf2ef8783941525d35cc991f00d2bf2d58
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
dtruss logs show that the signal handler does enter and is active, since
it does the first waitid() call, but then returns immediately:
waitid(0x0, 0x0, 0x7FFF62D7C468) = 0 0
sigreturn(0x7FFF62D7C9A0, 0x1E, 0x0) = 0 Err#-2
Since there was no error return, we conclude that si_pid was zero on
return. Source code for OS X 10.7 confirms that si_pid is set to zero
unconditionally, which is rather stupid:
http://fxr.watson.org/fxr/source/bsd/kern/kern_exit.c?v=xnu-1699.24.8#L1330
This is fixed for OS X 10.8:
http://fxr.watson.org/fxr/source/bsd/kern/kern_exit.c?v=xnu-2050.18.24#L1399
Without that information, we have to scan each child anyway, so
just disable the waitid() solution on OS X. This is a "hammer" solution
which will get forkfd working. We can later try and detect at runtime
whether waitid() is working.
Change-Id: Ic5d393bfd36e48a193fcffff13bb584927cdeafe
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
POSIX.1 does not guarantee the presence of the si_utime and si_stime
members. So instead of trying to set those members to zero, ask the
compiler to initialize everything for us.
This was found on OS X when HAVE_WAITTID was removed.
forkfd.c:192:11: error: no member named 'si_utime' in '__siginfo'
forkfd.c:193:11: error: no member named 'si_stime' in '__siginfo'
Change-Id: Ic5d393bfd36e48a193fcffff13b90aa6ccf592ae
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
The previous implementation required one syscall per child we're waiting
on to see which one exited. That means the algorithm was O(n).
This implementation uses WNOWAIT to find out which child exited and then
goes straight to that one. So it's O(1) on the number of children, but
runs 2 * number_of_children_that_exited + 1 syscalls, assuming there are
no race conditions with other threads. If there are or if a child not
started by forkfd exits, we'll still iterate over each child we're
managing to see which one exited.
It modifies the existing code so that it will do a waitid() with WNOWAIT
to check on the status of the child: if the child has exited, we'll try
to lock the entry so only one thread will do the final wait(). In the
case we read the PID, then the child exited, was reaped by another
thread, the PID got recycled and that child exited again, we'll fail to
lock the ProcessInfo entry so no harm comes. If by an absurd coincidence
this other child was started by forkfd() and its ProcessInfo is exactly
the one we are looking at, then we'll succeed in locking but that's a
benign race: we'll do what the other thread was trying to do and the
other thread will give up.
Future improvements to the algorithm are discussed in the Gerrit change.
Change-Id: Ie74836dbc388cd9b3fa375a41a8d944602a32df1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This also brings in the harness to forkfd that uses QBasicAtomic
instead of the generic GCC atomics.
Change-Id: Id5488bf192db0027bc684956ade0bf6c640c9512
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When the width of the decoration was floored then in some circumstances
(such as when latin and Hangul text was mixed) an underline of the text
would appear to be broken on HiDpi screens. Since the width is correct
then we should keep it as it is to ensure it meets up correctly.
Task-number: QTBUG-44062
Change-Id: I2cbf722a9cf9c7e15caa9aad061bf28d3bd0bb59
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
This fixes GCC warning on MinGW, when built with -Wattribute (the default)
Change-Id: I8e2b56a4a8fe9db0ec821f346a523b670df80f85
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
The most common use case for QTextStream::readLine() is reading
a file line by line in a loop. The existing readLine() method
allocates new memory for each line, that results in a loss of
speed. The introduced overload can use already allocated memory.
Besides it allows you to not think about filesystem specifics.
The current QFile documentation suggests a separate way to read
files from /proc filesystem. With this overload it's possible
to use the same idiom in all cases:
QTextStream in(&file);
QString line;
while (in.readLine(&line)) {
process_line(line);
}
The idea was inspired by the blog post of Ivan Čukić:
http://ivan.fomentgroup.org/blog/2014/10/03/api-design-and-impact-on-the-performance-qt-vs-stl-example/
Change-Id: I0c62b4a52681870589bc099905e83ed69e03dd40
Reviewed-by: Martin Smith <martin.smith@digia.com>
Reviewed-by: David Faure <david.faure@kdab.com>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Expose the same kind of TLS PSK client support we already have set
in place for QSslSocket.
[ChangeLog][QtNetwork][QNetworkAccessManager] It is now possible to use
TLS PSK ciphersuites when using HTTPS (or similar protocols working over
SSL).
Change-Id: I56a048e9f4f841f886758c781af2867d18538a3e
Reviewed-by: Richard J. Moore <rich@kde.org>
The call to QPlatformWindow::setVisible() will trigger flushing
of the window system events. Depending on the events, this could
lead to the window being hidden. Since platformScreen()->addWindow()
and removeWindow() was done *after* this flush, you could get the
remove for a window before it had been added, whereas the logic
in these functions were written under the assumption that there
is exactly one remove per add, and that add always precedes the
remove.
This has only been seen when running the QCompleter test so far.
This patch reorders the statements to make sure the events are flushed
after the window stack has been updated. In addition, it adds asserts
for the assumptions in the addWindow/removeWindow code to catch bugs
there earlier.
Change-Id: Ic67b03afbf7acbcb78be86bffa4c26360dc5832f
Task-number: QTBUG-43836
Reviewed-by: Christian Stromme <christian.stromme@theqtcompany.com>
rawMode only has an effect under X11 in pre-QPA times.
Change-Id: Iaff8fed8f4ae5af5dd0399bb3ebd9d590a39a758
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Endian conversion needs to be done on aligned data, but the data is
stored unaligned in the QByteArray. So the new qFromUnaligned()
function is needed.
Change-Id: I12f9e52cea81d06129b306709bb9d2cd004f04e1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
22px is not always large enough: KDE5 seems to like to make icons up to
28px, and they could be even larger on high-DPI screens.
Change-Id: Ifa8e0d49b310e4b4304207596f0f32c36a5db6a7
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
Sometimes KDE doesn't render non-square icons properly when they have
been received over the D-Bus protocol.
Change-Id: Icc6fa3d64a1598ea8f719192ae18d32f287d6a79
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
If StatusNotifier is working, then QSystemTrayIcon::showMessage() will
send notifications using the org.freedesktop.Notifications protocol.
https://developer.gnome.org/notification-spec/
Task-number: QTBUG-4011
Task-number: QTBUG-31762
Change-Id: Ia1925ec3dd81b1b7b8f3b490b6364aaf8f93f395
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
At startup or when a new UIWindow has been created but not associated
with a UIScreen yet, its geometry will be invalid, and cause the wrong
geometry for its corresponding QScreen when we use it to map geometry.
Instead we explicitly use the status bar orientation to map to the
correct geometry.
Change-Id: If37b3ab2ad5db65e20a7e3af5c3854b3e3ddff0d
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
We set the parent of the window as part of constructing it, which will
cause a layout of the QIOSDesktopManagerView's subviews, but in this
case we don't need to re-set the window state as that's taken care of
later on in the QIOSWindow constructor.
Change-Id: Ic197c9a50394908c8aa2155abdc97bc322937a85
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Change-Id: I78c47c6ccdb53045f3fa412b1489e08691d3e195
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Auxiliary screens are always in their primaryOrientation.
Change-Id: I078151ccbdb8a78eb095a05672f7804ab608ff24
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
[ChangeLog][QtNetwork][QSslSocket] It is now possible to use TLS PSK
ciphersuites in client sockets.
Task-number: QTBUG-39077
Change-Id: I5523a2be33d46230c6f4106c322fab8a5afa37b4
Reviewed-by: Richard J. Moore <rich@kde.org>
The "Fixed CE build of sqlite3" patch is preserved in this change,
which causes sqlite's own localtime implementation to be used for
wince. This is extended by #undef HAVE_LOCALTIME_S to override the
new assumption that localtime_s is available on wince and should be
used. Also, removed HAVE_LOCALTIME_S=0 since this must now be
undefined instead of 0.
Change-Id: I418e138ddc47d1bfbb80de0f4e4205a79c425f10
Reviewed-by: Mark Brand <mabrand@mabrand.nl>
On platforms like Windows (and presumably on mobile devices too)
the loss of the context (e.g. the underlying D3D device in case of
ANGLE) is an event that can happen randomly and needs sufficient
handling.
Enhance QOpenGLContext::isValid() with the purpose of indicating
context loss.
Currently only the Windows EGL backend (ANGLE) has support for it.
Other platforms may be added later.
Task-number: QTBUG-43263
Change-Id: I8177694c1ee7cebbd5d330e34757fd94c563e6d6
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
For windows that were shown on an external screen (not a virtual sibling
of the primary screen), eg. on iOS, it doesn't make sense to re-show the
window when moved back to the primary screen.
By moving the logic into the QScreen destructor, we ensure that the code
path is hit both for the old and unsupported style way of destroying
QPlatformScreen by deleting it directly, and the new and safe way
of using QPlatformIntegration::destroyScreen(), while still allowing
clients to manage the windows themselves by emitting screenRemoved()
before applying our fallback logic.
[ChangeLog][QtGui][Important Behavior Changes] QWindows will no longer
be re-shown automatically when moved from a destroyed QScreen, unless
that QScreen was a virtual sibling of the primary screen.
Change-Id: If1105bc5ef41a5392854bb97d121c998bffa3606
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Implementing org.kde.StatusNotifier DBus interface
http://www.freedesktop.org/wiki/Specifications/StatusNotifierItem/
as well as org.canonical.dbusmenu for the limited purpose of showing
the tray icon's context menu. If a desktop environment (such as
KDE or Unity) has a StatusNotifierWatcher listening, then tray icon
information is sent to be displayed by the tray implementation
instead of being rendered directly in an XEmbed window. This is
necessary because some modern tray implementations no longer provide
XEmbed "hosting".
[ChangeLog][QPA][Xcb] QSystemTrayIcon uses StatusNotifier D-Bus
protocol when the desktop environment supports it
Task-number: QTBUG-31762
Done-with: Marco Martin <mart@kde.org>
Change-Id: I3b1f744d621eefc7e9c61d1469460ebfcc77fc54
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
When run in single-exec mode, QDoc constructs only one instance
of HelpProjectWriter, even though it generates documentation for
multiple modules in one go.
This change adds a reset() function for the help project writer,
allowing new parameters to be passed to the existing writer
instance, thus ensuring that all .qhp files are correctly
generated.
Task-number: QTBUG-43815
Change-Id: I1d1c9f713eb5f574a6f8e56616cf5f61bb3e8ff8
Reviewed-by: Martin Smith <martin.smith@digia.com>
Equivalent of gluProject and gluUnproject.
[ChangeLog][QtCore][QVector3D] add convenience project and unproject methods
to use like gluProject and gluUnproject
Change-Id: I6e4e3e79ea6e34d1fb0c375e15185c950b699ef0
Reviewed-by: Sean Harmer <sean.harmer@kdab.com>
QmlClassNode is renamed to QmlTypeNode. This is done
in preparation for implementing qdoc support for
documenting javascript code. Next, QmlTypeNode will
be renamed to JsTypeNode, and a new QmlTypeNode will
be declared that will inherit JsTypeNode.
Change-Id: Ia5d0c367d06c26cb43f887927bbcb096afcb7301
Task-number: QTBUG-43715
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
At startup, iOS will report UIDeviceOrientationUnknown for the device
orientation, which toQtScreenOrientation() maps to portrait. We can be
smarter than that, by falling back to the orientation of the statusbar,
which in most cases match the physical orientation of the screen at
startup, unless the Info.plist file has been modified to limit the
possible orientations. See also:
https://gist.github.com/torarnv/40c1931205e33d2b1ed3
Change-Id: I5c78fbe5c670ed2909a51b478bd4814e2433554f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
Allows the helpers to be used at QPlatformScreen construction time, before
it has been associated with a QScreen.
Change-Id: Iab8f863ef5c9339ef6e88b3d844915c03cacda74
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
On embedded the mouse cursor will now appear and reappear regardless of
how the input handling code is loaded (via a generic plugin or compiled-in
to the platform plugin).
Instead of passing around QDeviceDiscovery instances that only works
when compiling-in the code into the platform plugin, introduce a new
internal central QInputDeviceManager. The single instance of this
provides a place to store any future input device related signals and
properties.
Also introduce mouse hotplugging support to linuxfb.
[ChangeLog][QtGui] The mouse cursor on Embedded Linux is now handling
hotplugging correctly with eglfs and linuxfb regardless of how the input
handling code is loaded (via a generic plugin or built in to the platform
plugin).
Change-Id: I147c1b04a193baf216598015264f2c06e1b20f84
Reviewed-by: Andy Nichols <andy.nichols@theqtcompany.com>
On Android the foreground activity can get intents
with onNewIntent. Those intents can not be received
in any other way. This is especially true in Android nfc.
This patch adds a way to receive those intents in Qt.
This patch heavily leans on the implementation of onActivityResult.
Change-Id: Ic4dca301f34afe9a528149c3653e545ed3265a3c
Reviewed-by: BogDan Vatra <bogdan@kde.org>
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
Some platforms have the concept of private fonts (e.g., OS X and
iOS) which are generally used as system UI fonts. Since 909d3f5c7,
the platform font database has a similar but private API.
[ChangeLog][QtGui][Fonts] Added QFontDatabase::isPrivateFamily()
Change-Id: Ibdce9de534fadbbc3965be8a942c8012edeed209
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@theqtcompany.com>
This update makes removes Qdoc's DITA XML generator.
Change-Id: Ibcfd013ace00e56a23268a2a5d850e6c9ea093d0
Task-number: QTBUG-43174
Reviewed-by: Martin Smith <martin.smith@digia.com>
This is like value(), but returns an iterator instead of the value().
[ChangeLog][QtCore][QAssociativeIterable] Added find().
Change-Id: I029fc8f91cef78f718d419587a2a50ffd2bf7632
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
QtCore now builds with no -Wnoexcept warnings
(that doesn't mean there aren't tons of functions
that should be marked noexcept, just that all
conditionally noexcept functions aren't noexcept(false)
just because of a forgotten noexcept elsewhere).
Change-Id: I10dacb6b9c9d41d3595fe2f306356d62d3d91c76
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Make it easier for the compiler to apply the
Named Return Value Optimization (NRVO) in the
shortName()/longName() functions by not returning
different objects in different return statements.
Change-Id: I1b6fa7e6121bc1c843378be33499728c56c97f92
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Rationale: the case of an invalid QSslEllipticCurve is already
dealt with before we'd call ensureInitialized(). But in order
to have a non-invalid QSslEllipticCurve, we must have called
one of the constructor functions first. There, we already call
ensureInitialized(), so we don't need to do it here again.
Change-Id: I96bdb5db63ec0165e6b8fac9469b5d81c6b2cdae
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
This leaves the decision on whether to do unaligned stores to the
compiler, as opposed to forcing it by ourselves.
Since we're now implementing them using two calls, this invalidates the
compiler bug that triggered the #ifdef for Sun Studio (whatever that bug
was).
Change-Id: I4e494ca860a15b9427b2a3000921bf5d92cbb5ef
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
