AppKit expects rendering to happen on the main thread, or at least any
interaction with AppKit UI classes such as NSView. Our OpenGL helpers,
such as QOpenGLContext, do not enforce this, and we may end up calling
into AppKit UI classes on the render thread, deadlocking the application.
Until this can be investigated and new APIs possibly introduced that allow
a more fine grained control in our own classes, we disable threaded GL
as a capability of the platform, which will inform clients such as
QtQuick to use the basic render loop.
[ChangeLog][macOS] Threaded OpenGL usage has been disabled when building
using Xcode 10/SDK 10.14 and later. Qt Quick defaults to the 'basic' render
loop now on macOS.
Task-number: QTBUG-71731
Change-Id: I6fc3295e833ecd48ad49382b8275c762fa7978a6
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
We document being able to parse more than 8-bit per color, but were
ignoring everything after the first 8 bits.
Change-Id: Ic85ab04b0836e6979a623e294eebd5084c1a9478
Fixes: QTBUG-71373
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
When passing \ to Gradle it will strip these out, so by using / it
ensures that it is able to find the Android SDK directory fine.
Change-Id: I053f087438ade6c30d015abe00e9958beb90a947
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
We had a global static, a setter for it (that's nowhere used within Qt
code) and a getter for it whose only use was commented out. Neither
was declared in any header; the getter's commented-out client had a
local extern declaration at the point of (non-)use.
Found while reviewing a change to the next few lines of code after the
commented-out use of the getter.
Change-Id: I393d56219cb7dd7cf836ca80e1bdd605a2914003
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
QImageWriter's plugin can open devices while operating and only free
them on destruction which means that if one wants to act on the file
written if must first destroy the writer's instance. This patch adds
the suggestion to use a scope to avoid that problem.
Change-Id: I239157ea86c4a93faab237fe1860312a3c7ac7a2
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Foreword:
- During a file or directory move the inotify id for an
entity is not changed.
- QFileSystemWatcher implementation uses a QMultiHash for
mapping an id to its path.
Suppose this filesystem hypothetical directory structure
- A
|--> B
and user watches both A and B directories.
Suppose that the B directory gets moved by calling "mv B B1".
The user receives a directoryChanged event for parent directory
A and scan filesystem for changes. During this scan the
user notices:
- a new directory B1
- a deleted directory B
The user simply invoke QFileSystemWatcher::addPath(B1) and
QFileSystemWatcher::removePath(B).
With the actual implementation the second operation could fail:
- The call QFileSystemWatcher::addPath(B1) insert a duplicated
records in the QFileSystemWatcher::idToPath
multihash ( {0, "A"}, {1, "A/B"} {1, "A/B1"}
- The call QFileSystemWatcher::removePath(B) fails because
- it first retrieves the the id for path B ---> pathToId("A/B") <-- return 1
- Then it calls idToPath.take with the id obtain in the
previous step <--- idToPath.take(1)
This last operation could take the record {1, "A/B1"} instead
of the {1, "A/B"} (because both have the same key) meaning that
the next check "x != path" evaluates to true (making the
removePath function fail).
This patch fixes the problem in the following way:
- Use idToPath.equal_range in order to obtain all paths with
the given id
- Remove the correct record that match (id, path)
- Prevent the removal of the inotify watch if another record
with the same id is still present in the multihash (like a
simple reference counting).
Change-Id: I9c8480b2a869d91e500af5c4aded596b9aa53b46
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The initial mouse press should create an implicit mouse grab so
that all subsequent mouse events are delivered to the widget that
was pressed first. After commit a4f7bb8733, every mouse press
would reset the implicit grab. This change checks the previous
button state, and does not reset the mouse grab if other buttons
are already pressed.
Fixes: QTBUG-70816
Change-Id: Icdd215c2f4aaee3c3f34607d68c1d8878155ec17
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Numbers look completely out of thin air, and while I trust it was not
an error back then it was introduced, today we end up with QComboBox
vertically translated and thus misaligned.
Task-number: QTBUG-69908
Change-Id: I784e06f00e4c92c4af67e9bd885b86648183f2e0
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Avoid having the reset in waitForDone interfere with other uses of the
thread-pool by locking the mutex higher, and maintaining the state
so the queues doesn't have threads not in allThreads.
Task-number: QTBUG-62865
Change-Id: I17ee95d5f0e138ec15e785c6d61bb0fe064d3659
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Document how to do an update, fix the bit-rot that had crept into
main.cpp since last it was compiled, correct the qt_attribution.json
to use the actual version number of UCD (its Revision number) instead
of the (admittedly correlated) Unicode release number. Updated to
Release 22 (which came with Unicode 11.0.0) in the process; but this
doesn't change our actual qunicodetables.cpp (so is incidental).
Task-number: QTBUG-71281
Change-Id: Ieb7a6e1a4d49f639993f76ff82c8f12a572db3c3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
"the GNU LGPL version 3" -> "GNU (L)GPL" because some parts of Qt are
GPL v3 (Qt Charts, Qt VirtualKeyboard, etc.).
Fixes: QTBUG-57697
Change-Id: Iceb88244e28b6900c5282b070468fb65b2bf52d2
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
When the on-screen keyboard completes a word via text prediction, the
message contains VK_PACKET as identifier for a character sequence.
While each character is send, the code only contains the first character
of the sequence.
Hence, resolve the actual code manually in case of a sequence.
This does not modify the virtual key, so that users are able to
distinguish between manual and predictive input.
Fixes: QTBUG-71210
Change-Id: I787f1f2d83acaf124dfbbab6c4614a1bfe7bb2eb
Reviewed-by: Andre de la Rocha <andre.rocha@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
The rgbSwap produces an image that fits the X-server, but not one that
fits our internal image definitions, so instead return our internal
image.
Task-number: QTBUG-56806
Change-Id: I25aedf7279bcd86792213b11dbd07a77b49538de
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
This is a whitespace-only change; removing a couple of newlines that
broke qdoc formatting of the enum values documentation.
Change-Id: Id371a4519922c71d79a11f3cda131e6683812696
Reviewed-by: Venugopal Shivashankar <Venugopal.Shivashankar@qt.io>
When a TCP connection timed out a QAbstractSocket::NetworkError was set.
To enable a more precise error handling for timeouts
QAbstractSocket::SocketTimeoutError is now set instead.
Separated ETIMEDOUT from other errors in nativeRead() and take over
responsibility for setting the error, which was previously handled by read().
Change-Id: Iccd45bdbb3d944cd160ae50c257d3256e05b1ae5
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: David Faure <david.faure@kdab.com>
Foreign windows do not have Qt backingstore content,
and are also not capable of accepting Qt content.
Change-Id: I959c7cdc32e6f4322497e132a436ce7d610a4106
Fixes: QTBUG-71183
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Commit 17b73b0d2b introduced a regression where the
grab rect position was added to the size when bounding
to the display size. This is incorrect.
Change-Id: I11d7ba7f53b96badfdead190ef9ddb525ed4ba99
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
This matches the behavior of QScreen::grabWindow(),
and gives the caller direct access to the scale factor.
Change-Id: Ia3ed165a62eaa0f386f8b508ea6b1128ba6be604
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The devicePixelRatio on the returned pixmap should
be the product of the Qt and platform scale factors.
This handles the corner case of setting QT_SCALE_FACTOR
on macOS with a high-dpi display.
Change-Id: I3600165d47c03c4e043bcc5e375932cc3fc0c544
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Allows a qt build to be configured to target arm64 desktop apps cross
platform and build them with nmake.
Change-Id: I99fed12047b45a504a1644201bcc19b18c69f3e6
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Output URL string and file system name, too.
Task-number: QTBUG-67932
Change-Id: Ic5d1927d70d98f7c081bee06af85b9f3a2a09812
Reviewed-by: Andre de la Rocha <andre.rocha@qt.io>
Assert on receiving double clicks which are currently not implemented.
Task-number: QTBUG-71263
Task-number: QTBUG-70999
Change-Id: I85cd21665ecaf118584053de63745044728d8f5b
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Makes clang-tidy not trip over generated code while running the
modernize-use-auto checker. In theory clang-tidy just shouldn't look at
generated code of course; but in this case just modernizing the
generated code is easy, so let's do it.
Example:
.../moc_kastentoolviewwidget.cpp:78:9: warning: use auto when initializing with
a cast to avoid duplicating the type name [modernize-use-auto]
KastenToolViewWidget *_t = static_cast<KastenToolViewWidget *>(_o);
^~~~~~~~~~~~~~~~~~~~
auto
Change-Id: I10c287320e1d5b5b8e66da3e0a22d517d0275dd1
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The input validation did not check for unreasonable use of the group
separator character.
Fixes: QTBUG-65024
Change-Id: If9d70d990fc6d5b298f3bde5b1604bf7e16dce24
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
The function's documentation needlessly repeated parts of its first
line. The BindFlag enum it takes as a parameter confused readers by
saying an option is ignored on Windows, failing to make clear that it
is so because that option is what Windows does by default.
Tidied some phrasing and typos in the process.
Fixes: QTBUG-52364
Change-Id: Ia6510caff7ec80216eefccf41fb009b1357e4b2e
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Instead of explicitly enabling layer-backing for Qt 5.12 on all macOS
version, we follow the macOS default to enable it for 10.14 if the
application binary was built against the 10.14 SDK.
Aligning ourselves with Apple's switch to forced layer-backing means
we have an easier story when it comes to supporting different runtime
configurations.
Fixes: QTBUG-71499
Change-Id: I34ee49b3daeb6ed8df444a3759d3573ebc9ea30f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The coordinates of the WM_CONTEXT message may be out of any screen in
PROCESS_DPI_UNAWARE mode since hi-res coordinates are delivered in this case
(Windows issue). Default to primary screen with check to prevent a crash.
Fixes: QTBUG-67966
Change-Id: I1950360520e93cbf3509611b3057635769f6543a
Reviewed-by: Andre de la Rocha <andre.rocha@qt.io>
When calling setDocument (directly or through the constructor) a delayed
rehighlight is initiated. Previously, if any text was changed before
this rehighlight could run it would cancel the rehighlight, even if the
changed text only caused a new block of text to be highlighted.
Fixes: QTBUG-71307
Change-Id: Ib09b664d90906f5b4427105f0e45469806f3a779
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
%p just prints a pointer. The operator<<(QObject*) member will print the
class type and object name.
Change-Id: Iba4b5c183776497d8ee1fffd1564951da0c6bebc
Reviewed-by: Sérgio Martins <sergio.martins@kdab.com>
This check, which was only done once, was wrong:
const int bytes_per_line = ((width * depth + 31) >> 5) << 2;
// sanity check for potential overflows
if (std::numeric_limits<int>::max()/depth < width
If width*height overflows, then it's already UB and checking afterwards
with a division is pointless and slow.
The other instances weren't properly guarding against overflows.
Change-Id: I343f2beed55440a7ac0bfffd1563350d4cfa639c
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Instead of doing byte comparisons, let the compiler do 16- and 32-bit
comparisons on its own.
Change-Id: If7e743cf8476463880ccfffd155f8629991b0b87
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Tobias Hunger <tobias.hunger@qt.io>
It can detect the standard UTF codecs, but not non-standard like
UTF-7[1], UTF-9 or UTF-18[2].
[1] https://tools.ietf.org/html/rfc2152
[2] https://tools.ietf.org/html/rfc4042
Fixes: QTBUG-67188
Change-Id: If7e743cf8476463880ccfffd155f853dc947421a
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The number of instructions is the same. But if the CPU can issue
32-byte-wide loads, this will be faster. For CPUs that would do two
16-byte loads, this is no worse than current code.
Change-Id: I8f261579aad648fdb4f0fffd1553d060b4fc852f
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
GCC 4.8 is complaining about perfectly valid (and recommended) code but
we still support it, so...
qcborvalue.h:74:25: warning: missing initializer for member ‘QCborError::c’
Fixes: QTBUG-71222
Change-Id: If7e743cf8476463880ccfffd155e8775b6b95469
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The unit tests weren't running into this problem because the every
setDevice() was preceded by the object being initialized with the exact
same data, so there was never a previous error state. I've only changed
a couple of tests, left the other setDevice() unchanged so we test both
behaviors.
Fixes: QTBUG-71426
Change-Id: I1bd327aeaf73421a8ec5fffd1561a590e3933376
Reviewed-by: Nils Jeisecke <nils.jeisecke@saltation.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Needed in qHash(QCborValue).
Change-Id: If7e743cf8476463880ccfffd155eeca91369b356
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Amends a952fd7d5b.
The mentioned commit started to skip QObjectPrivate::isSignalConnected() call if
the connectionLists are dirty, which lead to tst_qqmllanguage::receivers() test
inside qtdeclarative breaking.
Declarative signals were not checked if that function was not called. It
previously also wasn't called for signals higher than 64. Fix that by checking
for declarative signals after the connectionLists search is unsuccessful.
Fixes: QTBUG-71550
Change-Id: Ifcb5fdd0dc9a6b14b9f448a016fd09356a55b985
Reviewed-by: Liang Qi <liang.qi@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The current state of (emulated) mouse buttons was being incorrectly
reported for touchpad events under some conditions. In the handling
of pointer messages, GetAsyncKeyState() was being used to retrieve the
mouse button state. However, it does not seem to work always with all
touchpads. Furthermore, its use is not necessary, since the button
state information comes as a set of flags with the pointer message
itself. This change makes the handler use these flags instead.
Fixes: QTBUG-71470
Change-Id: Ie2e35bd80778ef74db672604a0f2af659785efbf
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
On Linux the correspondence between cursor functions and names
of cursors has never been standardized. Projects have either
assembled their own cursor function-to-name lookup table or
borrowed the table from other projects. The origins of our table
is described in QTBUG-71423.
On Ubuntu the default theme is called Adwaita. Before
bd72950fbe, we would not find a
cursor for 'openhand' and would fall-back to QXcbCursor::
createNonStandardCursor(). Which was sub-optimal, because the
cursors created by the fall-back path don't look like the
themed ones. But the situation was worse after bd72950fb (hence
the regression) - the 'openhand' fall-back name 'fleur' is a
symbolic link to 'grabbing', so we would get into a situation
where Qt::OpenHandCursor displays the same as Qt::ClosedHandCursor.
This patch adds a correct fall-back name for 'openhand' on Adwaita,
which is 'grab'. 'grab' actually is a symbolic link to 'hand1', but
'hand1' with other theams is a pointing hand cursor, that is why we
use the symbolic link's name in this case.
The lookup table still appears to be incomplete when comparing e.g
with KWin. Eventually we need to revise the table and put in a common
place so it can be shared between X11 and Wayland, but is out-of-scope
for this patch (see QTBUG-71423).
Fixes: QTBUG-71296
Task-number: QTBUG-71423
Change-Id: I247ed4b346c2cd3fe1c7fd0440d3763e0033346b
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
The engine used to send a UDP datagram to the local address to check
the proxy setup, but the check fails in case of the proxy hosted in
WAN and the local address hidden behind a NAT. In other words the
check fails because a public proxy hosted somewhere in internet has
no access to local addresses such as 192.168.1.2.
Remove the check to fix the issue; we still have other means to
detect network errors.
Change-Id: Ib6df263c87ebd7d6e88a0b5e024e78a559995234
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
When a popup is active, we set the receiver of mouse events to
be the active popup widget. But when we send a mouse event to
the popup, the receiver might start a new QEventLoop (e.g by
executing a new dialog). And in the meantime, the popup will
be destroyed. This will cause a crash in the line after the
event delivery (where we sat "qt_last_mouse_receiver = receiver"),
since at that point, "receiver" would be a dangling pointer.
This patch will use a QPointer instead of a raw pointer to
store "receiver", to ensure that it's set to null for
such cases.
Fixes: QTBUG-71062
Change-Id: Ie017cfa97370513ecfdd62c056fcb0e6c991f9f6
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Also clean up QTextCodec usage in qmake build and some includes
of qtextcodec.h.
Change-Id: I0475b82690024054add4e85a8724c8ea3adcf62a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
1. Remove the conditional inclusion of DTLS versions, they made difficult
and unnecessary ugly adding new protocols (something like TlsV1_2OrLater + 4).
2. OpenSSL 1.1.1 first introduced TLS 1.3 support. OpenSSL 1.1 back-end is
compatible with OpenSSL 1.1.1, but would fail to extract/report protocol
versions and set versions like 'TLS 1.3 only' or 'TLS 1.3 or better' on a
new context. Given 1.1.1 is deployed/adapted fast by different distros,
and 5.12 is LTS, we fix this issue by introducing QSsl::Tls1_3 and
QSsl::Tls1_3OrLater.
SecureTransport, WinRT and OpenSSL below 1.1.1 will report an error in case
the application requests this protocol (SecureTransport in future will
probably enable TLS 1.3).
Saying all that, TLS 1.3 support is experimental in QSslSocket.
Done-by: Albert Astals Cid <albert.astals.cid@kdab.com>
Done-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Change-Id: I4a97cc789b62763763cf41c44157ef0a9fd6cbec
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
For this we can use whatever the current NSColor.gridColor contains.
While this is mostly needed by the 'Dark' appearance, it also affects
the 'Light' theme, since the color QCommonStyle returns is different.
Let's use whatever Apple suggests.
Task-number: QTBUG-71048
Change-Id: I084414bad546755e9e67792484fe4601826ed0fa
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The issue we had has been fixed for years, but was unfortunately in
libxcb which we can't check at runtime. Instead assume very recent
Mesa drivers works.
Change-Id: I5fdd726b480b77edbedc0f369ae82ab4acbb77c9
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Document the version in the process; and make the wordings of
copyright notices match those in DEJAVU-LICENSE.
Task-number: QTBUG-70008
Change-Id: I1c965e5d7afb18dc4dbdffed908512c5771ab717
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The existing copyright notice gave me a name; github found me a file;
but it doesn't match what we had before. Made some guesses at which
parts of the change relative to upstream to keep, documented the diff
as a patch and recorded details in qt_attribution.json
Task-number: QTBUG-70008
Change-Id: I423724435eaeeda7237f8b3df8691b436fed8652
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The implementation calls GetCurrentThreadId, not GetCurrentThread, so
the return value is not the pseudo-handle.
Task-number: QTBUG-67686
Change-Id: Ifde0cf603dcea01bc1c454a8bebe1e5c0f22617f
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Previously no documentation was generated for the global qScopeGuard()
function. Create a class documentation page and add the the function
as a related non-member using \relates.
Task-number: QTBUG-71502
Change-Id: Ida5d7044f4de962360dfee9321feb49005d4b299
Reviewed-by: Martin Smith <martin.smith@qt.io>
\variable must not include the variable type, QDoc will resolve that.
This commit resolves four documentation warnings.
Task-number: QTBUG-71502
Change-Id: I5e88cf66d3c3bb8f18495d5477e1271ac2cd9e74
Reviewed-by: Martin Smith <martin.smith@qt.io>
And fix a typo in Qt::ScrollEnd.
Task-number: QTBUG-71502
Change-Id: I3efdbd12415814e066edd1b2f102a792812d36d5
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Commit 0a7aebad inadvertently removed two ANGLE-related files for
license information. This caused the licensing documentation to
fail to generate for those components.
Task-number: QTBUG-71502
Change-Id: I33ee673267c43474304d577e78fc1a0c3bd8691f
Reviewed-by: Martin Smith <martin.smith@qt.io>
These include typos, marking functions as \internal, documenting
trivial things, and fixing the function signatures passed to the
\fn command.
Task-number: QTBUG-71502
Change-Id: I24a9e1f7e1cdb39e5c31b99202bdd593c6b789ff
Reviewed-by: Martin Smith <martin.smith@qt.io>
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
These are minor typos or documentation warnings that snuck in with
new features.
Task-number: QTBUG-71502
Change-Id: I03669cfecc3c3d80168ff7b1ca8bca7571e06d25
Reviewed-by: Martin Smith <martin.smith@qt.io>
When the environment variable QT_QPA_FB_HIDECURSOR is set to 0,
the two class members mCursorImage and mDeviceListener are nullptr
but this was not checked in the functions afterwards.
Task-number: QTBUG-64844
Change-Id: Ic0fd6a09851777643e59bedf2c006a6bb9a36801
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
This part was accidently left disabling after testing the
fallback still worked.
Change-Id: Ic2df939753641a9771e68bc8857c570d356cff44
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Configuration 'compile' is obsolete and has been replaced with
'implementation' and 'api'.
It will be removed at the end of 2018. For more information
see: http://d.android.com/r/tools/update-dependency-configurations.html
Task-number: QTBUG-71570
Change-Id: I6f498d8cb3ff01ad641aee697496e3dc56059a72
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Opt out of switching between the normal and OpenGL based flush paths.
Once a QOpenGLWidget or QQuickWidget becomes visible in a window, the
window contents will be composed using OpenGL from that point on, even
if said widgets become invisible afterwards. Now that Qt Creator does
not rely on QQuickWidget the issue is less burning anyways.
Task-number: QTBUG-68329
Change-Id: I177e6e6094ee06ea26d8d0343bd3d84aadfa5913
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
It appears we are missing some setups for SIMD optimization, and already
have a workaround for that for NEON, so do the same for VSX.
Fixes: QTBUG-66388
Change-Id: I1cc1d0fe9c5a9df97acb589d29dec4dceb8fc576
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Marking various as final because no upstream is known or available.
Listing versions of others, where I was able to discover them.
Updated a stale link (that helpfully redirected).
Task-number: QTBUG-70008
Change-Id: Id00f34827133c560735c68793b4f1353f2b2ca85
Reviewed-by: Tobias Koenig <tobias.koenig@kdab.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Jani Heikkinen <jani.heikkinen@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Due to name conflict with file from qtdeclarative
Task-number: QTBUG-65769
Change-Id: I9ebf237701ce76b424f528feacb24e4158f06c0d
Reviewed-by: Venugopal Shivashankar <Venugopal.Shivashankar@qt.io>
This only updates data on languages already present in 5.12; once it
has merged up to dev, the scripts need to be run again to pick up a
few more languages and possibly add any more new languages present in
v34. Change some tests to match changes in en_AU's abbreviated day
and month names.
[ChangeLog][ThirdParty][CLDR] Update locale data to CLDR v34.
Task-number: QTBUG-71144
Change-Id: I68402b5e7e9d3dba669b8ba31b9a8abd86675c6e
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
When QMimeProvider parses the shared mime database xml files,
it will read the <comment> element for mime comment and treat the
`xml:lang` attribute as locale language string. When no `xml:lang`
attr is provided, QMimeProvider will read the value and treat it as
a en_US locale string as the default key.
When we call QMimeType::comment(), it will try to get the locale
comment string with the default language (QLocale().name()), once
it can't find a matched result, it should return the default key
(which QMimeProvider set it as en_US locale before) as fallback.
Task-number: QTBUG-71314
Change-Id: I444f8159d6f19dfef6338cd79312f608d8f13394
Reviewed-by: David Faure <david.faure@kdab.com>
A hardcoded value of 16 was used in QColorDialogPrivate::_q_addCustom()
instead QColorDialogOptions::customColorCount()
Fixes: QTBUG-58425
Change-Id: I7ae9881abd5926e0c6b118d5c84c3f259c545d35
Reviewed-by: Samuel Gaist <samuel.gaist@idiap.ch>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Attempts to use QSslSocket and its OpenSSL 1.1 back-end with OpenSSL 1.1.1
in a very peculiar way (for some reason the reporter calls OPENSSL_no_config())
combined with a bug in OpenSSL 1.1.1 resulted in a QSslSocket dead-locking
in initialization. This was happening because supportsSsl() first reported
false (OpenSSL internally fails to initialize after OPENSSL_no_config()),
but we have s_libraryLoaded set to true too early, thus the first supportsSsl()
returns false, the second - true.
Move setting of s_libraryLoaded later so that we don't claim to support
OpenSSL when an earlier ensureLibraryLoaded() attempt failed.
Task-number: QTBUG-70956
Task-number: QTBUG-71446
Change-Id: I8ad8763d357c84fc38c62e2ce914366367c2b445
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
freetype is included depending on configuration by the top level .pro
file.
Fix build warning:
Makefile:1361: warning: ignoring old commands for target `.obj/qfontengine_ft.o'
Change-Id: I0a0f111a841b368196633bbc0f9c698197f64bb2
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Preparing the use of the option for Python.
Task-number: PYSIDE-797
Change-Id: Ia1267b227ceac7f9dcbcfde6ed7c1480ef790f2a
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Follow the pattern of char and float, and treat shorts as a more generic
type in QVariant::canConvert()
Task-number: QTBUG-60914
Change-Id: Ib1cc7941ee47cb0fc0098f22f98a03cd6f6b63fe
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Use function name overloading for all of _NET_WM_STATE setters for
a mapped window, instead of few similar sounding functions, where all
do the same thing. For an unmapped window rename updateNetWmStateBeforeMap()
-> setNetWmStateOnUnmappedWindow(). Merge setNetWmStates(NetWmStates) into
setNetWmStateOnUnmappedWindow() and add a comment explaining why it does
what it does, as it was not obvious.
Internally there are 2 variants:
a) When a window is already mapped, to change the window state we send
XCB_CLIENT_MESSAGE(_NET_WM_STATE) messages to the root window as per EWMH
spec. The Window Manager MUST keep this property updated to reflect the
current state of the window. If this variant of the overload is called
while a window is not mapped yet, it does nothing. WM would ignore this
message anyway. The state change is not lost, it will be set later and
that is where the second variant comes in.
b) On an unmapped window we can set _NET_WM_STATE by using setNetWmStateOnUnmappedWindow().
This function will read QWindow properties and set all of them via one
xcb_change_property(_NET_WM_STATE) call. We do this inside QXcbWindow::show(),
where we know that QWindow state properties won't change anymore until
it gets mapped. Once it is mapped, we use the other variant.
This patch doesn't change any functionality, just makes the code easier
to follow.
Change-Id: I4f4703a35de083fcfd398112eabd0a5aec358e51
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
If QT_QPA_EGLFS_HIDECURSOR was enabled, m_deviceListener was never
initialized, which caused segfault in the destructor.
Task-number: QTBUG-71507
Change-Id: Id8b17f5312073249cd12995317213fd746753521
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
... to QXcbConnection::handleXcbEvent(), which is where it belongs.
This patch amends bc6f5b3ff6 (Sep, 2013).
And some other design cleanups.
Change-Id: Iefa0793c58de16a59d2294f38311e1e8dfa3035b
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The bitmap cache for the first 64 signals being connected was only set when the
connection is added. It was never unset when the connection was removed.
Internal use of the connectedSignals bitmap is not hurt by it occasionally
saying a signal is connected even though it is not, since the purpose of those
checks is avoiding expensive operations that are not necessary if nothing is
connected to the signal.
However, the public API using this cache meant that it also never spotted
signals being disconnected. This was not documented. Fix the behavior by only
using the cache if it is up to date. If it is not, use a slower path that gives
the correct answer.
To avoid making disconnections and QObject destructions slower, the cache is
only updated to unset disconnected signals when new signal connections are
added. No extra work is done in the common case where signals are only
removed in the end of the QObject's lifetime.
Fixes: QTBUG-32340
Change-Id: Ieb6e498060157153cec60d9c8f1c33056993fda1
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
Terminating character is needed for gtk portal backend, which is not
able to contruct a string without it and thus not able to pre-select a
file or a directory in file dialog.
Change-Id: I5b40fb1ce584936fc92e93f38dc8559890852d99
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
UWP also supports SwitchToThread. The usage of Sleep(0) was a leftover
that was forgotten when porting to desktop Window's approach of handling
threads.
Change-Id: I5e3d6fb3eefe07407b910cc6a6b45781d320e151
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The opacity set on a QPainter was being ignored for drawImage()/drawPixmap()
calls when rendering to a PDF file through QPdfWriter/QPrinter. This change
fixes it for PDF files v1.4 (default version) or v1.6. For PDF/A-1b files,
opacity will remain ignored to ensure compliance with the specification.
Task-number: QTBUG-70752
Change-Id: I4280e898698e0367dfb4c6ac2cd14ca2bf98850e
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
When we try to gracefully destroy a QOpenGLVertexArrayObject it is not
possible to create an QOffscreenSurface from a thread that is not the
GUI thread. In this case we just need to bail out instead.
The side effect that was seen previously was that there would be a
warning and a deadlock on Windows when closing QQuickWindows that
contained a QQuickPaintedItem backed by a FrameBufferObject render
target (which would be using the OpenGL paint engine) when using the
threaded render loop.
Task-number: QTBUG-70148
Change-Id: I4a20d74d9af850bb90d243212ad9f65c3fc9e616
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
QPointer uses QWeakPointer / QSharedPointer internals in QObject and has
the code to make sure two threads won't stomp on each other if both try
to create a QPointer for the same QObject at the same time. The
threading code was fine, but had a mistake in the clean up code for the
loser thread: the QtSharedPointer::ExternalRefCountData destructor has a
Q_ASSERT for the state of the reference counts. So we need to set the
state correctly before calling the destructor.
But we don't want to do it in case the Q_ASSERT compiled to nothing. So
we use a hack that violates the Second Rule of Q_ASSERTs: don't do
something with side-effects. This way, we can insert code that will only
be compiled if Q_ASSERTs do something, without having to duplicate the
preprocessor conditions from qglobal.h.
Fixes: QTBUG-71412
Change-Id: I1bd327aeaf73421a8ec5fffd1560fdfc8b73b70c
Reviewed-by: Romain Pokrzywka <romain.pokrzywka@gmail.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
"xcb_randr_get_screen_info" can be slow and in some configurations can
cause short mouse cursor freezes (which will happen on Qt application
startup).
Initial screen rotation was used only to not handle possible redundant
screen change event.
Fixes: QTBUG-70760
Change-Id: I9f01325a045d2c82c4dd2fce91a18a34e54a4bcd
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
We used to have logic that explicitly marked a QWindow as "embedded",
but we now resolve this dynamically based on whether or not we have
a parent window owned by Qt or not.
As part of this refactoring the fix for QTBUG-63443 in ac35f9c44c
got lost. We restore the behavior by treating Qt::SubWindow as a
reason not to create a NSWindow for a view. This flag is set by
QMenuPrivate::moveWidgetToPlatformItem already.
Fixes: QTBUG-63443
Change-Id: I12bff546eefcb8f6c9ca43b1b879773d129a28f9
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
We need to to explicitly unhide popup windows that were previously
explicitly hidden by applicationWillHide, so that their visibility will
be effectively restored when the application is unhidden (otherwise the
windows are gone forever even though their internal visibility is set to
true).
Change-Id: I4dbd209b07f769cc815851b40c41db0739ca2dc9
Task-number: QTBUG-71014
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Since Qt 5.10, qTo/FromBig/LittleEndian<float/double> stopped working.
It may be confusing, but big endian floats do exist, so not to break old
code, we should support them.
Change-Id: I21cdbc7f48ec030ce3d82f1cd1aad212f0fe5dd0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Prevents no-longer used QNSWindows from staying around (closed and
invisible, but alive), due to being auto-released when there is no
pool in place (during main(), before exec).
Change-Id: I4eb63c7140290ffe6bded8a76732354c846ed579
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The SDK and deployment target versions are helpful to know when
diagnosing issues.
Change-Id: I85026bd9c1d706a923e8953837bd59bf9ed0266f
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Commit b4bd5f9df3 removed two members from
the mwm hints structure but told xcb_change_property that the structure
has still five members when it had been reduced to three. This lead to
xcb_change_property accessing memory out of bounds.
As identified by Gatis, the safest option to avoid the access is to add
the two members again. Other window managers may be expecting their
presence in the window property.
Change-Id: Id4f0c9536cd317c35f2c6ebd1ac9ccc6f72de6a5
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
Reviewed-by: Mikhail Svetkin <mikhail.svetkin@qt.io>
This patch amends d67214302f.
The issue was caused by mistakenly interchanging
m_hasXRender <-> m_hasXRandr. Also renamed
selectXRandrEvents() -> xrandrSelectEvents() to be more
consistent with xi2Select*() API. And moved the xrandrSelectEvents()
to QXcbConnection ctor for the same reason.
Fixes: QTBUG-71305
Change-Id: I26f9bac3ae1f997f53134eb97f3569fb6d3c13fe
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
It seems some touchpads only send legacy WM_MOUSEWHEEL/WM_MOUSEHWHEEL
messages for 2-finger scrolling, instead of WM_POINTERWHEEL/
WM_POINTERHWHEEL, even after EnableMouseInPointer(TRUE), in spite of
sending the expected pointer messages for normal pointer movement/taps.
This change adds a workaround to handle legacy wheel messages even when
in pointer mode.
Task-number: QTBUG-71257
Change-Id: Ib360051147c4521751a5b91d90fa7657496777fa
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Task-number: QTBUG-71112
Change-Id: If1b81589f295657597a9dfa3671128a927cbe488
Reviewed-by: Andre de la Rocha <andre.rocha@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
If a tooltip was shown on an application modal dialog then it would end
up blocking mouse events to the dialog while the tooltip was visible.
Since tooltips are special in this case, they should not cause mouse
events to be blocked.
Fixes: QTBUG-65599
Change-Id: Ibf1729ca4942f5854e4f9687c5586382e23c1c31
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
The AMD FirePro dual gpus on the Mac Pro have a problem with offline
renderers in Chromium. Therefore, Chromium and thus Qt WebEngine
disable this option via the pixel format attributes.
The Qt Cocoa plugin on the other hand enables it in the recent versions,
causing context creation in Qt WebEngine to fail when run on a Mac Pro
with dual AMD FirePro gpus due to incompatible context options.
This patch uses the environment variable QT_MAC_PRO_WEBENGINE_WORKAROUND
which is set by Qt WebEngine upon application startup if the application
is running on a late 2013 Mac Pro. It should typically not be set from
anywhere else.
[ChangeLog] Offline renderers will be disabled when the application is
using Qt WebEngine and running on one of the late 2013 Mac Pro models.
Fixes: QTBUG-70062
Change-Id: I0b0831efb6f4073ebd37672040aaed6370853fc0
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The expression "m_context.view = view" always evaluates to the view,
so the check was flawed.
Change-Id: Icef4ba5244975ccd78a151c7d40c53b2364c6148
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The Qt QPA does not handle native double clicks as such; change the plugin
not to send them.
For Ink, remove the doubleclick detection. For the old code path,
map them to Press.
Fixes: QTBUG-70999
Change-Id: I54b858f9e146bf325a861554d5ef74143db7d2b7
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
There is no sense in testing the 'm_array' pointer against null, as the
memory was allocated using the 'new' operator. The exception will be
generated in the case of memory allocation error.
Task-number: QTBUG-71156
Change-Id: Ib76d16d38d2b0b0c4c9fae3d8d5bdd86af0d08ff
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Qt's PDF engine previously supported only the PDF v1.4 standard, which
only allows pages of up to 200x200in (about 5x5m). This patch optionally
enables the generation of PDF v1.6-compliant files that allow the
redefinition of user space units, so that pages of up to 381x381km are
now possible. By default, generated files are compliant to v1.4 spec.
v1.6 compliance must be enabled by, e.g., calling QPrinter::setPdfVersion()
with QPrinter::PdfVersion_1_6. PDF v1.6-compliant files require Adobe
Reader 7.0 or newer (also worked with the built-in viewers in current
versions of Chrome, Firefox and Edge).
Task-number: QTBUG-69386
Change-Id: I21708e0d465d5d7d9e46ff06dd04acfe1dfb0858
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
The state was forgotten from the translation layer, this is important for tree views.
Fixes: QTBUG-71223
Change-Id: Ief4004fe455889f9d5a7eb018bf34d37c36a6bd9
Reviewed-by: Jan Arve Sæther <jan-arve.saether@qt.io>
There is no need to call a memset as long as class is initialized with
new and default-constructor will initialize values.
Change-Id: I02f9ec524e32cb72713f6b5e37b60f3dec72fb28
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
A part of conditional expression is always false: if "c > 'z'"
Task-number: QTBUG-71156
Change-Id: I6ee20c45d80e476d97e59167c481b157e4a233d0
Reviewed-by: Mikhail Svetkin <mikhail.svetkin@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
'QAbstractItemView::NoSelection' named constant with the value of 0 was
used in the bitwise operation
Task-number: QTBUG-71156
Change-Id: I2d5099f9ed03cc42061508cc78282412a09825cb
Reviewed-by: Mikhail Svetkin <mikhail.svetkin@qt.io>
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
Reviewed-by: Samuel Gaist <samuel.gaist@idiap.ch>
Some Linux distributions patch OpenSSL's soname, making builds
on such distributions not deployable elsewhere. The problem is that
the code loading OpenSSL symbols would attempt to use the soname
of the build machine, and therefore not finding the OpenSSL
libraries on the deploy system.
The binary builds of Qt for Linux are affected by this problem,
as they build under RHEL7.4 which changes to soname of OpenSSL to
a non-standard string. This makes the binary builds not pick up
OpenSSL 1.0 from the machine where the build gets installed on.
Given that in the pre-1.1 versions only the 1.0 series is supported,
bump the minimum requirement of Qt to that. The 1.0.x releases
(up to 1.0.2, at the time of this writing) have kept binary
compatibility, and advertise a soname of "1.0.0", which is used
by most distributions.
So, if loading of OpenSSL with the build-time soname fails,
try to load them with the "1.0.0" hardcoded soname.
[ChangeLog][QtNetwork][SSL] OpenSSL >= 1.0 is now required to build
Qt with OpenSSL support.
Task-number: QTBUG-68156
Change-Id: Ieff1561a3c1d278b511f09fef06580f034f188c6
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
This change is based on a patch started by Gabriel. It works around the
apparently noop (on macOS Mojave, light theme) -drawRect: call on NSBox,
replacing it with -displayRectIgnoringOpacity:inContext:. Unfortunately,
this turns out to be extremely slow with dark theme, so we have to resort
to a custom version of NSBox/-drawRect: which calls its super's -drawRect:
and look more or less correct.
Task-number: QTBUG-69650
Change-Id: I5a614d7cd8002b378c6c3804663b6559e227f628
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
square/flat button had an erroneous size adjustment, resulting in a
wrong highlighted area. This button is anyway nothing like a real
control on macOS - they don't have 'flat' button which highlights
when pressed. Anyway, now our flat button looks more like AppKit's
'square' button when pressed.
Task-number: QTBUG-69533
Change-Id: I9643c1d4a25329c3f3c2148cc1821922655d9a8b
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
QTableWidget::takeItem() emitted cellChanged with row and column set to
-1. The internal functions searched for item after it was reset to
nullptr and therefore it was not found.
Since the modified cell is known because it's passed to the takeItem
function, the correct row/column can be retrieved from there.
Task-number: QTBUG-70478
Change-Id: I5ff5991c49f3200efe95fde4c7d0d28e19be7ebf
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
There's been one minor change in a struct and some minor re-ordering
of other things within their files (reflected here to simplify future
checks); and qt_attribution.json didn't document enough details to
ensure reliable review.
Task-number: QTBUG-70011
Change-Id: Iccff9cfd899e58cb42837c4628acacd7877c5b01
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Commit 2af1277631 renamed out_depends to
_out_depends inside qt5_add_resources but the old (empty) variable was
used with add_custom_command.
Change-Id: I8005674992b4538bd82375a4f1f70484bc0f0ae5
Reviewed-by: André Klitzing <aklitzing@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Native applications do not use custom fonts in menus, so we should
avoid it - our own example app 'menus' demonstrates the problem.
To quote Gabriel: Although the Mac HIG don't say that you shouldn't, and there's
even a API to do it (which we're obviously not using properly), the truth
is that nobody does it. Simply put, it looks wrong on Mac (although it's
arguable whether it looks good or bad regardless of the platform).
Task-number: QTBUG-29654
Change-Id: Iffd08ad63d419164102b2e738cdf1ebda1967a05
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The event dispatcher can be null already but we may have outstanding
QWinEventNotifier objects (like in a QProcess).
Patch-By: Tamas Karpati
Task-number: QTBUG-70214
Change-Id: I5e432e273def425ea334fffd154f34abfd6cb11a
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
[ChangeLog][Third-Party Code] PCRE2 was updated to version 10.32.
Change-Id: Id3bf7df0003f626cd1135d0508a5a489ff02f1e5
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Commit 582d221b29 introduced a regression where
the default Base color is no longer white in light mode,
for some use cases like querying QApplication::palette()
and when using style sheets.
[NSColor windowBackgroundColor] is not the correct
color for QPalette:Base, [NSColor textBackgroundColor]
is, except for the Disabled state which uses windowBackgroundColor
to get a grayed-out look.
This also changes the dark mode Base color slightly, from
RGB(0.196, 0.196, 0.196) to RGB(0.118, 0.118, 0.118).
Task-number: QTBUG-70966
Change-Id: I1ff494a09980ff2d04b2e2c43ee0bcc8e26ba9e7
Fixes: QTBUG-70966
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The Drag&Drop functionality had stopped working with touchscreen/pen
after the WM_POINTER-based input handling was added. The Drag&Drop
functionality internally uses the DoDragDrop() WIN32 call which,
according to Microsoft docs, is not supported for invocation inside
handlers for touch/pen messages, and should be invoked in handlers
for mouse messages that are synthesized by the OS afterwards. The
result was that when DoDragDrop (which is a blocking function with
its own event loop) was called it would hang ignoring all touch/pen
messages until a mouse/touchpad message arrived. This change
implements a workaround for this issue by enqueuing Qt touch/pen
events that would be generated inside the pointer message handler,
and that could start a Drag&Drop operation, and only producing them
after the OS sends the associated mouse messages.
Task-number: QTBUG-70887
Change-Id: Id45e0ecc70358ba250de9b3268856781ed21c9dd
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
This patch is amended version of 67cc8fea10,
which was temporary reverted to simplify integration of conflicting
patches. What was amended:
- Dropped the factory interface. It is sufficiently clean to check for
QXcbConnection::isConnected().
Task-number: QTBUG-68859
Change-Id: I810897b3ea20e356fc4d62e6f01231fd287962dc
Reviewed-by: Johan Helsing <johan.helsing@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
A file with the correct QPicture magic bytes, but shorter than a full
QPicture file header, could cause the header decoder to access memory
out of bounds. Add a size check to avoid.
As a driveby, generally harden the parsing against malformed files.
[ChangeLog][QtGui][QPicture] Fix crash reading malformed picture file
Task-number: QTBUG-71208
Change-Id: I86eb1f915ca9b3a4b91c7433036d76ed6061e2f0
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Extends the code from qtbase/0d6612af65161f6e659af6985bb2c0b76accb856
to cover the case of windows without title bar, too.
Fixes: QTBUG-4362
Task-number: QTBUG-8361
Change-Id: I5cff8814174069922936f3fcfbb3aef154c7a7e7
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Commit b7887f9b4f removed this explicit
disabling because it shouldn't be needed anymore. Turns out it was, as
new Android SDK do include modern Linux headers and those define the
structs and constants needed for statx().
Repeat of 8eb3944dac.
Task-number: QTBUG-64490
Fixes: QTBUG-71200
Change-Id: If7e743cf8476463880ccfffd155e6d5c2b5a3da9
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: BogDan Vatra <bogdan@kdab.com>
Apparently it is a left-over from Qt 4.
Task-number: QTBUG-70240
Change-Id: I6b882728defef1ab78331b03e76459a7419cd386
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Use newly introduced SystemParametersInfoForDpi() API with some convenience
overloads.
Task-number: QTBUG-4362
Change-Id: I4c41c700007bf7cc4fd5868356e3145c136704c0
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
This is a better version of 23b139038a
which was reverted.
Task-number: QTBUG-68620
Task-number: QTBUG-53022
Task-number: QTBUG-57211
Change-Id: I0b37fc261945c542bbfb30cecfe4b0a97c655e3c
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
The existing code was difficult to follow, it contained some code
duplication, inconsistencies and legacy code (mwm/dtwm support).
Most of this code was copied over from Qt 4 as is, with some
unexplainable (accidental?) modifications.
Most of Motif code was never ported over to Qt5. In addition to the
properties and protocols described in ICCCM, Motif uses properties
and protocols of its own. In Qt4 we had an implementation of Motif
DnD protocol. This was never ported over and mentions of this support
was removed on Oct 2012, 4c41cb48d0.
Nobody has complained for all these years, so it is safe to remove
the remaining leftovers. Motif style support also was removed around
the same time 4c41cb48d0.
Keeping only those Motif hints that do not have any replacement in
modern window manager specifications - decorations hints.
MWM_INPUT_*MODEL* in modern specs was replaced by _NET_WM_STATE_MODAL
The existing code was setting _MOTIF_WM_HINTS from 2 places - from
QWindow setter (::setWindowFlags) and again before mapping the window
(::show). We don't need the logic from updateMotifWmHintsBeforeMap().
That function read the current value of _MOTIF_WM_HINTS and merged in
few additional hints, hints that are not relevant based on all the
reasons from above.
Change-Id: I9cb10bcad9bfac8762e3909895c2e9de613e622c
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Added the files libEGLd_mingw32.def and libGLESv2d_mingw32.def that were
deleted by mistake in the recent ANGLE update, causing failure when
building with mingw in debug.
Fixes: QTBUG-71189
Change-Id: I0bf446c3f2d189afb806bbfed2fdc1a3f70750c1
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Moved all screen handling method implementations into
qxcbconnection_screens, the same way we are doing with
xinput2 code in qxcbconnection_xi.cpp. The goal was to
reduce the size of qxcbconnection.h/cpp.
Change-Id: I9bad55ca4b0874171b7313d923b13c66034c3b3e
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
A basic base class that creates a connection and initializes
extensions. The goal was to reduce the size of qxcbconnection.h/cpp.
Made QXcbAtom into a class that handles atom initialization and
exposes the relevant APIs. Before this patch, all of that logic
was inside of qxcbconnection.h/cpp.
Change-Id: Ia893c3b31e2343dfbe62fe2aa6bfd0017abf46ea
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
Trigger documentSizeChanged when block visibility changes, since block
count remains constant in this case.
Task-number: QTBUG-69310
Change-Id: I5ec7a4f9008f26ea8602356bcbaefbda293e54a3
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
0.8.2 + subsequent commits up to 31f1f355700870c6615399fbfa7934934b3a9a57.
There were couple commits after 0.8.2. Mostly minor stuff, with 2 exceptions:
fix off-by-one error in xkb_file_type_to_string() and undefined behavior in src/x11/keymap.c
We currently don't use any of the features added in later releases of
libxkbcommon, therefore the minimal required version in src/gui/configure.json
remains the same.
[ChangeLog][Third-Party Code] updated bundled libxkbcommon 0.4.1 -> 0.8.2 (up to 31f1f355700870c6615399fbfa7934934b3a9a57)
Task-number: QTBUG-71109
Change-Id: Ia03c3dc31c5f39e78dcb5915a45e82797b065ccb
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The function, although implemented differently in each event dispatcher,
is not supposed to process newly posted events, only the events that
were queued at the time of the call.
This is tested by tst_QEventDispatcher::processEventsOnlySendsQueuedEvents,
which is not blacklisted on any platforms, so we know it's the behavior
in practice.
Change-Id: If9a874eeeb8ebcebe88ed119b065ae12fc545129
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
After qtbase/550d2a0a15c9403894448ab83863e71bbac2d349,
pixmap functions are used for icons as well which can cause compile
errors for old UI files that still use them.
Task-number: QTBUG-8563
Change-Id: I61bd4b9c1bf774e071a35c5806657054a77ff4d0
Reviewed-by: Jarek Kobus <jaroslaw.kobus@qt.io>
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
'description' is used only once in all configure.json files, so
I assume that was meant to be 'purpose'.
Change-Id: I66e9d9196c27d2f2131c2d57ea03895e8f5ce754
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
The script taken here is a QChar::Script, not a
QFontDatabase::WritingSystem. This means it was passing QChar::Unknown.
Change-Id: I919ae7187ba277346a7719116a94776dce24dd84
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
According to reports Intel on Windows gets confused when feeding
program binaries retrieved from x86 and x64 builds into each other.
Task-number: QTBUG-64697
Change-Id: Ia7748f532ad06942a92c6fbfc4c9d1ad16bc785a
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
Before commit 4d15f393a7 update requests
were handled by a timer on QWindow. Therefore they survived the closing
and re-opening of platform windows. Now, as the timer was moved to
QPlatformWindow, it gets reset when you close the QWindow, and any
pending update requests are lost. However, we do set the
updateRequestPending variable on QWindow when requesting an update.
Therefore, we can also restore the update timer on the platform window
when creating it.
Change-Id: I23b00f24a46706beac7d1455edd8a5623db46b22
Fixes: QTBUG-70957
Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
Reviewed-by: Thomas Hartmann <thomas.hartmann@qt.io>
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
This was a regression from Qt 4.
Before this patch, we supported filtering events only at QWindowSystemInterface
level, but to properly support filtering in QAbstractEventDispatcher::filterNativeEvent,
we have to filter the events earlier. Now it is possible to enable/disable this
feature for platforms that support native event filtering.
The mapping of which events are user input events were taken from
QWindowSystemInterfacePrivate::EventType.
Task-number: QTBUG-69687
Change-Id: I9a5fb9f999451c47abcdc83fdcc129b5eeb55447
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
This patch reworks 0b1ce5db9e.
Just by looking at the source code it was unclear why is this
signal-and-slot connection necessary. It doesn't do anything on
normal exit - at the time we call dtor of this class,
QCoreApplication::instance() already is nullptr, which means that
no further event processing happens and we never get this signal.
Without digging into git history it may appear that the goal was
to process the remaining events on application exit, which would
be a questionable code by itself.
Change-Id: I202d11584901e3b52254d8e3c347aece17844b72
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
The pdf engine uses a resource file, but Q_INIT_RESOURCE() was lacking.
Task-number: QTBUG-71070
Change-Id: I685961b3f2eea0ffe6b5313c72d504a8ad9a98e5
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Also, call it only if the state really changes. If we stay disconnected
the whole time, there is no point in trying to create the session over
and over.
Change-Id: Ic3a92dd0575bed1a23ae36a944cc51b9741fb64a
Fixes: QTBUG-49760
Reviewed-by: Jesus Fernandez <Jesus.Fernandez@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
[ChangeLog][Third-Party Code] Sqlite was updated to version 3.25.2
Change-Id: I2d74ee924745a5e1edd6fe511777965313a4b77a
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Added eglext_angle.h and gl2ext_angle.h to the list of header files
copied to include/QtANGLE during install. These two header files were
introduced by the recent ANGLE update but were not added to the install.
This was causing build failures when including the ANGLE headers (e.g.
in qtmultimedia).
Task-number: QTBUG-71158
Change-Id: If2f1a9ecfcdf509cccf2b3671adf575cc39892d4
Reviewed-by: Andre de la Rocha <andre.rocha@qt.io>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
- The only place where window flag Qt::WindowDoesNotAcceptFocus changes is inside
QXcbWindow::setWindowFlags and from there we call updateDoesNotAcceptFocus(). The
current code was redundantly calling xcb_wm_hints_set_input in various places.
- Similar as above: call xcb_wm_hints_set_normal/iconic only where it can change.
This hint depends on window state, so update it only from setWindowState().
Removed unnecessary call to setTransparentForMouseEvents(), which is already called
few lines above from setWindowFlags().
Change-Id: I8da919b135a4dfda0c8c1dad51d85d3e706153ab
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
QAbstractEventDispatcher::wakeUp is a thread-safe method, using
a queued connection to invoke it is wasteful. This type of connection
involves allocating temporary QMetaCallEvent on a heap and locking of
destination thread's post-event queue. In most use cases this is ok,
and really convenient when target method is not thread-safe. But in
this case the existing solution was suboptimal, especially because
the events we are reading can be high frequency.
The solution that we use here is lock-free. There can be only one
time when it might need to wait for the lock, which is upon exiting
the application. If we have entered the critical section in
QXcbEventReader::run(), then the registered post routine (qAddPostRoutine)
will block the QCoreApplication's dtor (this is where dispatcher is
set to 0) until we exit the critical section. We also know when not
to enter the critical section, in case dtor is already running.
With this approach we might need to compete for the lock at most
once, instead of whole application lifetime, which was the case
with the existing code.
Change-Id: If6737329c972347b0050d67658e28dbaa6f552e8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It is necessary for QTBUG-69687. The original code processes the xcb
event queue only when new events have arrived, but if we want to do an
event filtering that buffers some events and processes them later based
on set/unset flags (e.g. QEventLoop::ExcludeUserInputEvents), we need
to call processXcbEvents() on every event loop iteration, not only when
new events have arrived from X server.
The required functionality is implemented by having custom event dispatchers,
instead of using the generic ones from QtGenericUnixDispatcher::
createUnixEventDispatcher() / eventdispatcher_support-private. This also
enables for further customizations, as might be necessary by QTBUG-70095.
Task-number: QTBUG-69687
Change-Id: I1f8b2400d26cccf17279d57bb4b678e40c615f33
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
We need to override this snippet for the documentation
we generate for Qt for Python, and it is easier to have
it on a separate file.
Task-number: PYSIDE-801
Task-number: PYSIDE-691
Change-Id: Ideb5b6af25024279f167137d3b65660bb9c96a7e
Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
In some cases, a tool tip may be shown outside screen geometry, i.e. if:
- QToolTip::showText is invoked manually with a position outside.
- In tst_QToolTip::setPalette if there is no screen at (0, 0). This might
happen in a multi-monitor setups where one screen is taller than the other.
- On Wayland windows are (by design) not allowed to know their position on
the screen. This means that global positions can't be trusted.
This started crashing when QDesktopWidget::screenGeometry(pos) was replaced
with QGuiApplication::screenAt(pos)->geometry() because screenAt will return
null if no screen is found, while screenGeometry defaulted to the primary
screen.
This reverts to the old behavior of falling back to the primary screen.
This won't solve the issue completely for the Wayland case, but at least we
will stop crashing.
Change-Id: I42dd07cc21c2f9f0ea0d69f0c25bd46d8a2615a0
Reviewed-by: Filipe Azevedo <filipe.azevedo@kdab.com>
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
QImage's operator>>(QDataStream&) did not set an error mode on the
stream on read failures. That would break QDataStream transactions.
Since the current QImage serialization cannot differentiate between
truncated and corrupted data, we set the ReadPastEnd error as expected
by the transaction system.
Also specify the expected file format on decoding QImage from stream,
to avoid all the format handlers' canRead() being invoked. This is
necessary since some of them may call ungetChar(), which fails when
the stream is in a transaction.
Also add testing of this feature to the QDataStram transaction
autotest. That required a slight rewrite of the fake sequential
QIODevice subclass. The previous implementation had incorrect
behavior of peek(), which is required by QImage decoders.
Task-number: QTBUG-70875
Change-Id: If3f1ca7186ad1e6ca0e6e8ea81d2b2fbece6ea01
Reviewed-by: Alex Trotsenko <alex1973tr@gmail.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This fixes qopenglwindow shader in the triangle
Task-number: QTBUG-67338
Change-Id: I8552183bf9ca45e9b56760b340d014ddd34c21f4
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Our fallback logic for inexact matches was not very good at accepting
better suggestions.
Change-Id: I40fb78bf583171105725156148e4a2245fb81354
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
Wrong option was used to set zlib compression level for png
setCompression with negative value uses default compression
setCompression with value between 0-100 converts to zlib compression level 0-9
setCompression with positive value overrides Quality option
Change-Id: Ic4b048a1e30d6940019c2a00a6c24d0c11e3f821
Reviewed-by: Alexandra Cherdantseva <neluhus.vagus@gmail.com>
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
Handle both of them inside the QXcbConnection::compressEvent().
Change-Id: Ibe7184ba5c5b636013145e887c817dca701345ad
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
For details how this works refer to the documentation in the patch.
The follow-up patches will switch to calling processXcbEvents() on every
event loop iteration. With the existing code that would mean frequent
locking of shared data (event queue). Acquiring a lock is fast, but
lock contention isn't. To avoid potential problems, reimplement xcb event
processing to be lock-free. Besides theoretical performance benefits,
this definitally improves code readability in qxcbconnection.cpp. Thanks
to Mikhail Svetkin for questioning the design of the existing code.
Done-with: Mikhail Svetkin <mikhail.svetkin@qt.io>
Change-Id: I935f2b6ca802580f5c80205aef7b2f9afc172d26
Reviewed-by: Mikhail Svetkin <mikhail.svetkin@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This reverts commit 67cc8fea10.
I forgot about this patch and now it makes rebasing the local changes
too time-consuming. Besides, 67cc8fea10 broke a build for -no-xcb-xlib.
I will restore this patch, with adaptations to the new QXcb*Connection
hierarchy.
Task-number: QTBUG-68859
Change-Id: I938f32b5da22ce18f95d761f9b34e77fff923e24
Reviewed-by: Gatis Paeglis <gatis.paeglis@qt.io>
This is a better solution for fbb485d4f6.
The existing solution was working fine, but it was exposing logic that is
internal to QWindowSystemInterface and platform plugin interaction. Some
platform plugins do event filtering at native event level - those that
support QAbstractEventDispatcher::filterNativeEvent(). Other plugins rely
on QWindowSystemInterface to do the filtering. Dispatchers should not care
about this.
The new logic rely on the fact that QWindowSystemInterfacePrivate::handleWindowSystemEvent
calls QAbstractEventDispatcher::wakeUp(). The same way postEventSourcePrepare()
rely on QCoreApplication::postEvent() to call QAbstractEventDispatcher::wakeUp().
Event sources run in the order they are attached, postEventSourcePrepare runs
before userEventSourcePrepare(). We rely on that order to pass wakeUpCalled
value.
Change-Id: I0327f4f2398fd863fb2421b8033bb1df8d65f5af
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
While this requires from us calling a deprecated method, a (non-deprecated)
method we were using gives a wrong color which is too bright/saturated.
Task-number: QTBUG-70676
Change-Id: Icebeb53e351caa646c533595ca1a886e5eb6b5b8
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
that resulted in 'Cmd' reported as combo of Qt::Meta/Qt::Control and
Qt::KeypadModifier.
Task-number: QTBUG-71006
Change-Id: I3dddc56f4d404a1ceefb21d57ac120b6273456ec
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
When there were additional spaces between the function definition and
the first parameter, the parser failed to parse it when it contained
another function (e.g. 'qlineargradient(... rgb() ...)').
The reason for this was that ::until() needs the function at index-1 so
it can correctly count the opening parenthesis.
Fixes: QTBUG-61795
Change-Id: I992f556e7f8cd45550f83bc90aa8de2b4e905574
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Those system calls are present in glibc 2.28. Instead of using
syscall(3) to place the system calls directly, let's use only the glibc
functions. That also means we no longer accept ENOSYS from either
function, if they were detected in glibc.
Change-Id: I44e7d800c68141bdaae0fffd1555b4b8fe63786b
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Jüri Valdmann <juri.valdmann@qt.io>
QWinRTBackingStore::flush() was not considering the possibility
that the supplied region may fall partially outside the bounds
of the paint device. This resulted in possible accesses to invalid
memory addresses, causing a crash. This bug was exposed by an update
in ANGLE that was causing a crash when running tst_QTableView::bigMode
with a small screen size. With this fix the function will use the
intersection of the supplied region with the paint device bounds.
Change-Id: I2f0f0f7f5510688bfa1459320a0c146df6be65d1
Reviewed-by: Miguel Costa <miguel.costa@qt.io>
This reverts commit 341bfcd1ea.
As it turns out there might be use cases where we want to have proper
windowing system event integration with glib dispatcher via g_source_attach().
For example with gtk_dialog_run, where GTK blocks in a recursive main loop.
We want to continue dispatcing our windowing system events during this nested
event loop. Not having a proper glib integration can result in rendering issues,
e.g. when resizing parent window via mouse while GTK-based dialog is shown. Can
be seen on examples/widgets/richtext/textedit/ -> Format (from menu) -> "Color..."
The issue from 341bfcd1ea actually should be fixed inside XCB platform plugin,
by improving integration with event dispatcher. That is handled in follow-up patches.
Change-Id: Icabc6d841a554aefbdd460765a3165d22e65f651
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
xcb_poll_for_queued_event() was introduced in libxcb 1.8.
The minimal required libxcb version was bumped up to 1.9 in
1f5d791708. Before this version
bump we needed the runtime check to support older versions
of libxcb.
Updated connections in the event reader to use the new signal
and slot syntax. Removed threadedEventHandling() method because
now it is always 'true'.
Change-Id: I0bce61fd478a871d35e676239ee5280c4f40be8a
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The remaining diff to clean 1.6.35 is archived in the qtpatches.diff file.
[ChangeLog][Third-Party Code] libpng was updated to version 1.6.35
Change-Id: I925b3d95f6da20e12fc3274b4713e3fea05094f7
Reviewed-by: Kai Koehne <kai.koehne@qt.io>
Reviewed-by: Liang Qi <liang.qi@qt.io>
According to the Khronos documentation, gbm_surface_create() will give
a BAD_MATCH error if the format does not match the EGL_NATIVE_VISUAL_ID.
Newer drivers have started to enforce this.
Change-Id: I61360b0f52965ad8057e7de8f824ffca64fea904
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Dominik Holland <dominik.holland@pelagicore.com>
A few places we didn't check if detach() succeeded including
in reinterpretAsFormat(), where it can be undone.
Task-number: QTBUG-70785
Change-Id: Ibcc8e26e2961f6288eb7a045ae1cb28e59213a49
Reviewed-by: Eirik Aavitsland <eirik.aavitsland@qt.io>
If you, directly after connecting, call "ignoreSslErrors()" followed by
"resume()" then you will most likely crash.
It is very contrived and there's no reason to do this.
Change-Id: I949a303238f5012296d0e84eb76173764eb9de2e
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
and fix handling of incoming binary data
Change-Id: I31e97505ad4ff64cf8e380df5d0d6b70c3cd60b0
Reviewed-by: Ryan Chu <ryan.chu@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
It has had some changes and should be kept up to date.
Fixes: QTBUG-70386
Task-number: QTBUG-70852
Change-Id: I868a558811c34cf5a800c3087a0ca96e7fb49b1a
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
QList, QVector and QVarLengthArray check the validity of iterators
passed to member functions using isValidIterator(), which checks that
the underlying pointers are in the range [begin, end]. This check is
well-defined when the outcome is positive, ie. when the iterator is
valid. But if the iterator is not valid, and does not happen to point
into [end, begin + capacity], the comparison, which uses normal
operator<, invokes UB.
Fix by using std::less<T*>, which defines a total ordering.
Change-Id: I1e5757789b4b9779f5e3e298e7f2b2dd0b27576c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The version we're using is at least seven years old.
Let's try staying up to date ...
Change-Id: I450a2fc37c22740fe554a81b471bf0fad340498e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Use a template on the size of the char[], as suggested by Ville
Voutilainen. This resolves ambiguity about whether such look-ups
should be done via QString or QCborValue (not that it would have made
any difference).
When we come to add mutating indexing of QCborValue, chained
dereferences like map[i][j][k] need to stay in operator[] const
throughout, to avoid detaching intermediates to create references into
them due to using the mutating operator[] on the earlier dereference's
return. So const-qualify the QCborValue operator[] const variants at
the same time, to match those of QCborValue itself.
Change-Id: Ib1652ae9440fe3767a653afa2856b74040210e07
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Likewise have mutating operator[] insert an invalid entry at its
target index, if beyond the end of the array. This makes it possible
to fill an array from high index to low, for example.
Change-Id: If71699c20e2623142214ce2c11c4d6e4a120c989
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Applications with target SDK version lower then 28 running on a device
with version greater or equal to 28 will cause compatibility warnings,
so default to none when the extract_android_style value is set to
default.
Note that the new value "default" was introduced to allow this kind
of changes in the future, i.e., selecting the best solution based on
some simple heuristics. Adding a new value also keep compatibility and
allows the user to explicitly set a value when needed.
Task-number: QTBUG-69810
Change-Id: I68301716767870ce6de40e45742d9c5fc263ee25
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
Modal sheets are supposed to appear below the toolbar if a toolbar
exists, or below the title bar if a toolbar doesn't exist.
In the unified title and toolbar mode, sheets should to appear
directly below the toolbar, if the toolbar is positioned at the
top of the window.
Code-wise we achieve that by calling setUnifiedTitleAndToolBarOnMac on
a QMainWindow, which results in adjusting the top content border of
the NSWindow via [NSWindow setContentBorderThickness:forEdge], which
Cocoa uses to position a modal sheet (the sheet top edge is aligned
with the top edge of the content view + the Y border thickness
value).
The issue is that because NSWindow.titlebarAppearsTransparent is set
to YES, for sheet presentation purposes, Cocoa considers the content
view to begin at the position of the top left corner of the
frame (where the title bar is), and thus sheets appear somewhere
in the middle of the unified toolbar.
To fix that we need to account for the title bar height in the
border thickness value.
Compute the title bar height from the window frame height - window
content view height, and add it to the top border thickness value.
Amends 8ac9addd94
Change-Id: Icf85c513035cc3710b438e60eb14dc04c5bbaced
Fixes: QTBUG-65451
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
For winrt we cannot rely on the fact, that QThread::current will be called
from the correct thread for the first time. The application's main entry
point creates a suspended thread and starts it right afterwards. At that
moment, other functionality (QLoggingRegistry for example) might have
called QThread::current, which set the wrong thread as the main thread. In
order to avoid this situation, the main thread is explicitly set in
QCoreApplication's constructor.
Task-number: QTBUG-66418
Change-Id: I8b6347357a80eb395ae758bd3d420adef0826751
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
The height of a line edit depends on the height of its font,
which in turn depends on DPI. So use the DPI-scaled icon size
from a style instead of hard-coded values.
Task-number: QTBUG-65627
Change-Id: Ic1f5af61b0d6346cfbc828817c4a3a39296a41ba
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
There is no need to apply QMAKE_LFLAGS_NOUNDEF in a .pro file that loads
qt_module.prf or qt_plugin.prf, both of which do the same.
Change-Id: I1ec86cddb4d0991becc136f8bd50d6652c885a6b
Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Make the code consistent with the DomProperty::write() method.
Task-number: QTBUG-70613
Change-Id: I622b4a019a4473823584de97304f8324f2cf0c6b
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Check on the application state before requesting activation.
Fixes: QTBUG-70810
Change-Id: I550137dc00209b17f4b3c033287ceef1b871ff50
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
qabstractitemmodeltester.cpp:223:31: internal compiler error: in expand_expr_real_1, at expr.c:9170
Change-Id: I098c1bdf706512c91c649205f4675de0ca374227
Reviewed-by: David Faure <david.faure@kdab.com>
The documentation for key bindings was a little bit unclear for key
asterisk. Also sync the names of the left and right key was not
consistent (Left/RightArrow - Left/Right)
Change-Id: Icee0821880fcc3ebc2ade939cf80127ebf0976ad
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Luca Beldi <v.ronin@yahoo.it>
Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
In iOS, the system certificates cannot be accessed, so this function
will return an empty list for the default configuration.
Task-number: QTBUG-60407
Change-Id: I0d691a0dd5e6367594e71c7ebccfbdc866d4a3f0
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Andy Shaw <andy.shaw@qt.io>
Perform a @Nx image file lookup when loading pixmaps.
Make drawBackgroundImage() handle high-dpi pixmaps,
here the layout calculations needs to be in device-
independent pixels
Fixes: QTBUG-36825
Change-Id: I61e6f53c59f61f3bd88c34a036349e51e8c8ad92
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
When switching between different input sources, we have to update layouts.
Task-number: QTBUG-50865
Change-Id: I0c23c19b79a2102dcc533822b0f861c387582c6c
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Previously some of the members would have random initial values. Also,
on updateKeyboard() if we don't find usable uchrData, we should just
reset keyboard_layout_format and keyboard_mode, rather than keep the
previous values.
Task-number: QTBUG-50865
Change-Id: I1297fa55bb1593dd549d0bc122713d5d98f7b1fc
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Usually we focus in when we receive a click or equivalent.
QGraphicsScene by default also transfers the focus when you start a
touch on a trackpad or similar. Most of the time this also generates a
synthetic mouse click, so people don't necessary notice. However, at
least on macOS you can configure this behavior. With focusOnTouch
switched off, QGraphicsScene behaves as one would expect on macOS.
Fixes: QTBUG-59442
Change-Id: Ib87112640eef6b77892ad2490d80eedd055e6dce
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Venugopal Shivashankar <Venugopal.Shivashankar@qt.io>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
QAccessibleTable(Header)Cell::isValid() did not check if the view is
still valid as it is done in QAccessibleTable::isValid() which can lead
to a crash during destruction when e.g. operator<<(QAccessibleInterface)
is called.
Change-Id: I930bf4c22aa0aa6868f58a8e5d0dd1557f6b17af
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@qt.io>
Calling update has a cost, and should only be done when the drawable
object changes size or location. Instead of calling update each time
makeCurrent is called, we listen for the appropriate notifications,
limiting the number of update calls significantly.
We still call update on the thread owning the QOpenGLContext, which
is not ideal, as [NSOpenGLContext update] should only be called on
the main thread, but in practice this works. Getting out of this
situation is tricky, and setView has in theory the same problems.
Until those problems have been solved we keep the behavior as is.
Task-number: QTBUG-63572
Change-Id: Ibac9f8be7843f2aa006af6f7ee670bf027122440
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
When the font changes in NSFontPanel, it notifies NSFontManager via
-[NSFontManager modifyFontViaPanel:], which in turn sends the font
manager's action (by default changeFont:) to its target (nil, unless set).
Sending the action in -[NSApplication(NSResponder) sendAction:to:from:]
will sanitize the 'to' argument via _NSTargetForSendAction.
If the argument is non-nill (if we've set the NSFontManager target
explicitly), and we're running in an app-modal session (which we are),
the target is checked for worksWhenModal -- a property which is defined
on NSWindow, and only supposed to be set for subclasses of NSPanel.
Since our QNSFontPanelDelegate class doesn't implement this method, the
_NSTargetForSendAction function will return nil, and the action is never
sent.
If we don't set the NSFontManager target (leaving it as nil), the function
will skip the worksWhenModal check, and fall back to resolving the target
via the responder chain, which includes taking the NSPanel's delegate
into account:
#0 -[NSWindow delegate] ()
#1 -[NSWindow(NSEventRouting) supplementalTargetForAction:sender:] ()
#2 _objectFromResponderChainWhichRespondsToAction ()
#3 _NSTargetForSendAction ()
#4 -[NSApplication(NSResponder) sendAction:to:from:] ()
#5 -[NSFontManager sendAction] ()
...
Since we want to end up in the QNSFontPanelDelegate, we can rely on the
default logic to resolve the target based on the responder chain. But in
case _NSTargetForSendAction will at some point also check the resolved
target for worksWhenModal, we also implement the worksWhenModal method,
to be on the safe side.
Fixes: QTBUG-69878
Change-Id: Ie739d016fe0efd17b3d8a99cc1fb1ace81807aff
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Do not override a custom font with the style default.
Task-number: QTBUG-70276
Change-Id: I58bf43f791aee5409e29459d579725365703e4f3
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
This is a regression from commit 346c15102b, which creates a new QRegExp
in setFilterRegExp, losing previously set case sensitivity property
(i.e. when the code does proxy->setFilterCaseSensitivity(Qt::CaseInsensitive)
before setFilterRegExp).
Interestingly that commit ensured that setFilterFixedString would still
preserve CaseSensitivity, but not setFilterRegExp(QString).
Change-Id: I3d37d001ce6e86dd90e7e07431440a42607172f9
Reviewed-by: Samuel Gaist <samuel.gaist@idiap.ch>
Reviewed-by: Christian Ehrlicher <ch.ehrlicher@gmx.de>
Reviewed-by: Luca Beldi <v.ronin@yahoo.it>
