The test keeps failing, perfectly when run singly btw. However the approach of
sending mouse move events is inherently fragile due to the use of
QCursor::setPos and the expectation that that produces the correct sequence of
mouse move events.
Change-Id: I12ba1afcfd3fab965b8f93d5def48b435fd2ff33
Reviewed-by: Andy Shaw <andy.shaw@digia.com>
This patch fixes several upload corruptions if the server closes the connection
while/before we send data into it. They happen inside multiple places in the HTTP
layer and are explained in the comments.
Corruptions are:
* The upload byte device has an in-flight signal with pending upload data, if
it gets reset (because server closes the connection) then the re-send of the
request was sometimes taking this stale in-flight pending upload data.
* Because some signals were DirectConnection and some were QueuedConnection, there
was a chance that a direct signal overtakes a queued signal. The state machine
then sent data down the socket which was buffered there (and sent later) although
it did not match the current state of the state machine when it was actually sent.
* A socket was seen as being able to have requests sent even though it was not
encrypted yet. This relates to the previous corruption where data is stored inside
the socket's buffer and then sent later.
The included auto test produces all fixed corruptions, I detected no regressions
via the other tests.
This code also adds a bit of sanity checking to protect from possible further
problems.
[ChangeLog][QtNetwork] Fix HTTP(s) upload corruption when server closes connection
Change-Id: I54c883925ec897050941498f139c4b523030432e
Reviewed-by: Peter Hartmann <peter-qt@hartmann.tk>
When the global monitor was used it did not get the right position for
the mouse event and as a result it caused context menus to appear and
disappear instantly when right clicking over a non active window. This
ensures that the events are sent correctly with the right position and
button information.
Task-number: QTBUG-45015
Change-Id: I9b17a725e656c716c4e22117b4513e64c357b266
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
In rare cases the I/O operation was still running after the destructor
was running, which then modified free'd memory and caused a malformed
heap. To prevent this, we ensure that QWindowsPipeReader::stop() cancels
a running I/O operation and sets the readSequenceStarted flag correctly.
Also, we prevent the start of a new read operation after we called stop().
Change-Id: If8a28bdf23a39a0e88c1770a6f66e2b24ea426bb
Task-number: QTBUG-45601
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
There's exactly one caller for this private method, and future code
will be a bit simpler after moving the code to the calling site.
Change-Id: Ibc65f91c770f9f29b317ceddb39a67d52106da33
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
The zero timeout singleshot timer emitReadyReadTimer was used to emit
the readyRead signal via the event loop in case of a synchronous read.
In that particular case, ReadFile would return successfully, and the
notified slot would not be called.
Now, that we use an I/O completion port, the notified slot is always
called, even in the synchronous case. The emitReadyReadTimer is not
needed anymore.
This is also supported by the fact that the timer is immediately
stopped in notified() after it was started in completeAsyncRead().
Change-Id: I93bcde5f067bf89a1d49005a3fddda4c8c8c95fc
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
According to the documentation we should always pass a zeroed
OVERLAPPED object to ReadFile.
Change-Id: I3f822af46a2c38e029e02461f706c4fd91c00c50
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Currently pressed touch point is added to the list of active touch
points in Gui module. It must be excluded from consideration when
we traverse the list.
Task-number: QTBUG-43255
Change-Id: Idddab093b1f6a79122cf18fad7f43bfc93ce7eea
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
[ChangeLog][QtCore][QLockFile] Fixed a deadlock when the lock file
is corrupted.
Task-number: QTBUG-44771
Change-Id: Ic490b09d70ff1cc1733b64949889a73720b2d0f3
Reviewed-by: David Faure <david.faure@kdab.com>
This change brings the online template up to date with recent
changes:
- Update links in header and footer.
- To improve page load time and reduce load on qt.io,
do not load images from the main site.
- Fix auto-scroll/highlight jQuery to work on anchors
containing full stop ('.') characters
- Responsiveness: Fix menu issues with narrow views
- Fix styling of subscript/superscript HTML tags
- Make script URIs protocol-agnostic
Change-Id: I219e8ce6ff2e70f9374ad5e037ceda8f710a6a10
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
Merged in the upstream version, which obsoleted many of the local
patches. The remaining diff to clean 1.6.17 is archived in the
qtpatches.diff file.
Change-Id: I5065435dc5a922d3f4a46eb37a23a4877dde2ee6
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
The internalformat parameter seems to be GLenum (unsigned int) on QNX
based on the error message in the linked bug report. The standard is
GLint, though. To keep everyone happy, add a cast.
Task-number: QTBUG-45346
Change-Id: I57fece7b381e0d02acc842a21b1edc5451ea0224
Reviewed-by: Sérgio Martins <sergio.martins@kdab.com>
Reviewed-by: Rafael Roquetto <rafael.roquetto@kdab.com>
Dependencies to all header files generated by dumpcpp are now added to
every object file. This fixes parallel builds of projects that use
TYPELIBS.
Change-Id: I3c0456c7b182a42296ec6999aa86d1293ffd2e42
Task-number: QTBUG-45118
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
On Windows, we perform an extra certificate lookup for root CAs that
are not in Windows' (minimal) root store. This check can take up to
15 seconds. The SSL context can already be gone once we return. Hence
we now check for a non-null SSL context on Windows before proceeding.
Change-Id: I1951569d9b17da33fa604f7c9d8b33255acf200d
Reviewed-by: Richard J. Moore <rich@kde.org>
We need it because otherwise code like
QNetworkConfigurationManager ncm;
qDebug() << "ONLINE" << ncm->isOnline();
may give the wrong value because the queued signals that have been just connected a few lines above
may not have been processed yet
Change-Id: I959db75ed17497ab91eeba2669ee2c8947244f00
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
Takes care about unencrypted data in the socket writeBuffer when
close() flushes the output.
Change-Id: I301f41ea709817e215ee4246a3951e3182d94fbd
Reviewed-by: Richard J. Moore <rich@kde.org>
Use the name "OS X" instead of "Mac OS X", "Mac OS" and "OSX",
and mention iOS. Replace "Carbon Preferences API" by
"CFPreferences API" in the QSettings documentation.
Change-Id: Ia7f9fb874276c7c445a1649df521b96ff43daa0c
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Venugopal Shivashankar <venugopal.shivashankar@digia.com>
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
When using QtTest inside the notification example for QtAndroidExtras
there as a problem with the compilation as it saw it as being an ambiguous
check. Therefore doing an explicit case to uint ensures it is not a
problem.
Change-Id: Ibc9ce4c64292bf5ae7c501c592d8c42a66934b8b
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
This solves a problem when using a Qt application over remote desktop as
if it failed with even 1 sample then it would fallback to GDI which causes
an error if the software OpenGL option is used.
Change-Id: Ib311a7a657f92aab15277461bc8e040bebbe4753
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
previously the whole response was cached in a NSMutableData which leads
to high memory usage on large responses (e.g. downloading a large file).
With this patch only the part of the answer that has not yet been read
by the caller is cached.
Task-number: QTBUG-41356
Change-Id: Ic2fe822552620d8835a2c81f8c76dd170fe6ec97
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
Until now we applied the geometry in QWidget::setVisible() by
calling QWidget::move() and QWidget::resize(). But these
methods are unreliable when the window is created but not
visible yet. For example, specifying the window position by
"-geometry +0+0" will take no effect.
Apply the geometry directly to QWindow in QWindow::setVisible().
QWidget will update its geometry after the response of the window
system. Besides it allows to specify the geometry for QML
applications.
Task-number: QTBUG-44713
Change-Id: I9a0e110e81e569c81da802729707fec104fef887
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
It may change after the window has been created.
Change-Id: Ib81a7ad7353b1909cc42684fc70d6b7d2556106f
Reviewed-by: Uli Schlachter <psychon@znc.in>
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
This change is related to 6a7ee92b39
which added handling for SelectionClientClose. Further testing showed
that with e.g. Qt 4 applications the SelectionClientClose is not
emitted, but the selection window seems to be destroyed before the
client is destroyed.
Fur a destroyed selection window the same applies: the clipboard
content is no longer valid and we should emit the changed signal.
Change-Id: Id3778a28b9f5601bf2c6e0106981316e0efa6e7c
Reviewed-by: Gatis Paeglis <gatis.paeglis@digia.com>
Reviewed-by: Uli Schlachter <psychon@znc.in>
While most (all?) touchpads send WM_MOUSEWHEEL for vertical scroll
the story is quite different for horizontal scroll. Some of them
send WM_HSCROLL instead of WM_MOUSEHWHEEL.
Some of them even send left/right key event but those are lost cases.
Task-number: QTBUG-45120
Change-Id: I3bf86e25a6f4f3ba03ac7e89a23f4b7bc432f2de
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
At least the dismissive action should be on the left, but we can take
the OS X rules as reasonable in cases where a lot of buttons are in use.
Task-number: QTBUG-42808
Change-Id: If45f991a068d47009e02d39fbb3886ff4b31c8e1
Reviewed-by: BogDan Vatra <bogdan@kde.org>
The text was accidentally removed in
1afe110b8f.
Fixes qdoc warning:
qtbase/src/corelib/io/qstandardpaths.cpp:60: warning: Undocumented enum item GenericDataLocation in QStandardPaths::StandardLocation
Change-Id: I7f236c01d85ebf76b67ab7af7e61b8cbedee4c36
Reviewed-by: David Faure <david.faure@kdab.com>
The resize event for QMdiSubWindow may occasionally have an invalid
oldSize(). This has been the case since 2007. Fixing this looks like a
high risk, low reward endeavour.
Task-number: QTBUG-32446
Change-Id: I80ce43987c7b6e346cd44f7ac4cef01b01e5472b
Reviewed-by: Venugopal Shivashankar <venugopal.shivashankar@digia.com>
Thumb instructions trigger a compiler bug in e.g. Qt Script
when used with armv5 in the recent versions of the toolchain.
We already disabled this for the 4.8 toolchain, but the bug
is also present in 4.9, so we need to disable it there as
well.
Change-Id: I6cadc7efd2b59b9a2ffec038559edd0e7782a4b9
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
0c7241cc contains the wrong identifier for device capabilities. This has
been updated in a previous commit, but documentation was not properly
updated.
Task-number: QTBUG-45098
Change-Id: I300a2ea19ebdf8a1aaed492a1bdf24b9af2bd60a
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@theqtcompany.com>
Reviewed-by: Andrew Knight <qt@panimo.net>
If a cosmetic polyline contained a 1-pixel segment, the next segment
would be drawn with wrong starting point.
The original fix for QTBUG-26156 had some unwanted side effects
(QTBUG-31579 and now QTBUG-42398). It tried to skip start-point update
if stroke() did not actually draw anything (because the segment was
too small). However, to determine that, it tested for a change in
lastPixel. But that was not failsafe; in some cases (1-pixel segment),
lastPixel could be unchanged even though the segment had been
drawn. With this change, we instead test directly whether stroke()
skipped the segment or not.
Task-number: QTBUG-42398
Change-Id: Id751db69a18cd1af4f45070db9d5698aa532d22a
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@theqtcompany.com>
When continuing a previous layout, we would in some cases add a wrong
margin because we would treat the current frame as preceding itself.
Task-number: QTBUG-42269
Change-Id: Idf00fadbed890ce34b073cf670cc9b8e1d4b0d5d
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
It would happen together with an error:
QPainter::begin: Paint device returned engine == 0
and would be caused by the size provided to QIcon::pixmap being empty,
itself caused by the availableSizes list being empty for the Selected
mode.
This bug was most often hidden by the fact that the Selected icon mode
was not triggered properly since we usually only set menuVisible after
calling updateIcon, and most of the time when it did, we would overwrite
it right after with a Normal mode icon.
Fix the issue by disabling the broken feature completely since the
default Selected icon is grayed out while tray icons are now usually
black (or white when selected). To support the dark menu bar mode on
10.10 we'll need to use NSImage's setTemplate anyway and that
knowing in advance if we can invert the colors ourselves would also
better solve the menuVisible usecase.
Task-number: QTBUG-42910
Change-Id: If9ec9659af28ecceb841bfc2f11721e6029fe891
Reviewed-by: Morten Johan Sørvig <morten.sorvig@theqtcompany.com>
When the ComboBox currentText() was changed by key
LineUp or LineDown or mouse click, the completer still
contained the last inserted characters.
If now all text was selected (by Ctrl+A or selectAll()),
the old item and index was restored on next Enter press.
Task-number: QTBUG-41288
Change-Id: I6916fd31c8b8fbacfb12e1a62c3e46823cf918b4
Reviewed-by: Timur Pocheptsov <Timur.Pocheptsov@digia.com>
If shmget() does not work (probably because the backingstore size is
bigger than SHMMAX), don't create a QImage with the previous buffer and
the new size. That does not end well when we try to draw to the image
later. Instead, fall back to the malloc path, like we do when the system
doesn't support shared memory.
[ChangeLog][X11] Don't crash when resizing windows to bigger than 3840x2160
Task-number: QTBUG-45071
Change-Id: I009de7c2179ffde28e252593067756877cad1b1c
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
On some platforms (notably wayland and xcb) the openUrl() call returns
as soon as the request is passed to the underlying system. The call may
still fail, but this status is not reported to the application.
Change-Id: Iea554e9d5e7e371a5ce11454aef6f5c66ba36ca0
Reviewed-by: Valerio Valerio <valerio.valerio@jollamobile.com>
Reviewed-by: Topi Reiniö <topi.reinio@digia.com>
In some combinations the fusion style colorizeImage method would
generate invalid colors causing odd colors in the end result.
Task-number: QTBUG-41515
Change-Id: Ib16049be1f3e93ae4976524df2827076eac7d94f
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
Reviewed-by: Jens Bache-Wiig <jensbw@gmail.com>
On some devices, there is a possibility of getting a state change
while we are locking the event loop for suspend. This is
probably related to change 8c0ef140b3, but we should in
any case use the existing deadlock protection mutex mechanism,
just to make sure that there are not other cases which could
trigger the same deadlock.
Task-number: QTBUG-44339
Change-Id: I3e0b5fa2ddf4ef86e6b29cb1d67c4cccedd8242e
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@theqtcompany.com>
On Android 5, the Chinese fallback fonts are .otf format,
and these were ignored when we populated the font database.
Note that this also requires an upgrade of FreeType, otherwise
the Simplified Chinese font will not load.
[ChangeLog][Android] Fixed rendering Chinese text on Android 5.
Change-Id: I972d0fbc0b16458650a7b82ea82a0ddd5bdfa641
Task-number: QTBUG-44648
Reviewed-by: Liang Qi <liang.qi@theqtcompany.com>
When there is no composing currently being done then we need to get the
real virtual key rather than the one it was sent as with ImmGetVirtualKey.
This ensures that any shortcut associated with the key will be fired right
away.
Change-Id: Id526b7030ca21eaacbd6c74774392b0707cf762a
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Fuzzing test revealed that for certain malformed bmp and ico files,
the handler would segfault.
Change-Id: I19d45145f31e7f808f7f6a1a1610270ea4159cbe
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
