They return a pointer to the actual header, skipping the magic string.
This is done in preparation for the header located in an ELF note, which
won't have the magic.
Change-Id: I3eb1bd30e0124f89a052fffd16a8229bec2ad588
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The .qtmetadata section is already suppressed for static plugins, but we
carried this unnecessary magic string. For library file formats where we
don't have a scanner (Windows COFF PE), a library that linked multiple
static plugins could end up with multiple metadata found.
We can't suppress the header because the version of moc could be
different to the version of Qt, so we need to read the header version
too. Right now, the version isn't output by moc (all the logic is in
qplugin.h), but this could change again in the future. In any case, 4
extra bytes are not a big deal, so the header stays.
Change-Id: I3eb1bd30e0124f89a052fffd16a82088d8303081
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The .pri files for public Qt modules were missing run_depends entries.
Those entries are used by qmake when linking plugins of a static Qt
build. In shared builds, run_depends entries are used to decide whether
to add the -rpath-link argument to the linker. In that matter,
run_depends entries serve a similar purpose as CMake's
IMPORTED_LINK_DEPENDENT_LIBRARIES.
In Qt5, run_depends entries consist of the qmake values QT_PRIVATE and
QT_FOR_PRIVATE.
QT_PRIVATE contains the Qt modules that are privately linked to the
public module, and QT_FOR_PRIVATE are the dependencies of the private
module. That loosely translates to LINK_LIBRARIES of the public module
and INTERFACE_LINK_LIBRARIES of the private module.
Private modules whose public counterparts are already public
dependencies don't have to be added to run_depends.
Pick-to: 6.2
Fixes: QTBUG-84310
Change-Id: If7025c98ea462a2d5a143202df3d98d88e4fee08
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Looks like it was copied from QUtf16::convertFromUnicode(), but for the
UTF-32 case that is not correct. UTF-16 to UTF-32 conversions can change
the length of the string due to surrogates.
There are unit test tests for creating and parsing UTF-32 headers, and
for detecting content as UTF-32, but there aren't any for UTF-32
conversions. I don't have time to add a full test for that.
Fixes: QTBUG-97122
Pick-to: 6.2 6.1
Change-Id: Ic17a33f599b844d8ab5dfffd16ab2c4cb6b0547d
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This allows Qt Quick Controls to detect if a dark theme is in use,
and if so, use a dark theme of the current style (if available).
Fixes: QTBUG-93955
Pick-to: 6.2
Change-Id: I15fc8a2271acf9ba27232359056d5a681d4cffca
Reviewed-by: Dmitry Shachnev <mitya57@gmail.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Leaving it set to 0 will cause mktime() to interpret other fields as
in standard time, potentially "correcting" them and tm_isdst to
represent the equivalent moment in daylight-saving time. Set it to -1
to tell mktime() to let the system work out whether the time is
standard or daylight-saving.
Change-Id: Id33d4cb0afdb14f236ca5ce04cf605610a30d712
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It's available on MinGW, so not limited to MSVC; and the MinGW
localtime() and localtime_r() fail in some cases where it works.
Use the generic name rather than _localtime64_s(), since localtime_s()
now just calls it; and, in any case, we were assuming time_t is
__time64_t when calling it.
Change-Id: I316cc5b1a3e19cd6725555042dfaba3124a25a03
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The skip-reason only cites Qt/X11. Convert the #ifdef into a platform
name check instead. The test also fails with the offscreen plugin, so
skip it for that as well.
Pick-to: 6.2
Change-Id: I49607b89f4b32359e81e1d9aadff2c3e03035c53
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
QDialog overrides setVisible to set focus on the default push button, or
(if there is no such button), make the first autoDefault push button the
default button. QDialog also explicitly sends a FocusIn event to the
focus widget in the dialog.
All this should not be done if the dialog does not become active after
getting shown, which will be prevented if the WA_ShowWithoutActivating
attribute is set.
Add a test case.
Fixes: QTBUG-8857
Pick-to: 6.2
Change-Id: If47021a4721a280ba45e95b43d0424cdacd9b4ea
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
The need for this union has disappeared at some point during refactoring
It does confuse MemorySanitizer, though, so it's good to remove it.
Change-Id: I5f5050adba302bf96db96d5f303d2a25630543ce
Fixes: QTBUG-97109
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
For a widget that implements height-for-width, the vertical scrollbar
becoming visible might be just enough to make the scrollbar unnecessary.
In that situation, the scrollbar flips on and off continuously.
To avoid that situation, make the width of the widget smaller until the
height fits without scrollbar, up to the point where we have space for
the scrollbar anyway.
The calcuation here is assumed to be cheap, but depends on the
heightForWidth implementation in the widget. Running the while-loop a
few dozen times should have no performance impact during resizing
and laying out the scroll area contents.
Add a test that confirms that within a brief period of time we only get
the one hide-event we expect.
Done-with: Zou Ya <zouya@uniontech.com>
Fixes: QTBUG-92958
Pick-to: 6.2 5.15
Change-Id: I0faeb5f9b1a226aada958c18333d9c2ac8203dd1
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
Previously, when deciding where the actual data is, Frame was calling
padding() to test if offset is needed. A curious case with a DATA frame
containing compressed body and having 'PADDED' flag set with a padding
equal to ... 0, ended in a decompression error (and assert in 6.2 code).
Pick-to: 6.2 5.15
Fixes: QTBUG-97179
Change-Id: I9341a4d68510aa4c26f4972afdcd09a530d5a367
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
When setCurrentIndex is called by removeTab, the old current index
might no longer be valid. Only update the lastTab value of the new
current tab if the old current index is still valid.
As a drive-by, use the validIndex helper function.
Fixes: QTBUG-94352
Pick-to: 6.2
Change-Id: I945e2093a90a1fccbba86d32b1113f83fedd41de
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
This simplifies the code. While at it, also expand default case
so changes in the QChar::Direction enum don't go unnoticed.
Change-Id: Id52140b0aaf7fd259c4b8079462811603eb3ff93
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Since commit f3c7d22dd0 we do not use
libraries from /usr/local and other non-system locations on macOS. But
our configure tests still did. This led to discrepancies between
find_package calls in configure tests and the Qt project itself.
Mentioned commit removed /usr/local and friends from
CMAKE_SYSTEM_PREFIX_PATH. But we can't pass this variable to the
configure tests, because CMake sets it up and overwrites our value.
Pass CMAKE_SYSTEM_PREFIX_PATH and CMAKE_SYSTEM_FRAMEWORK_PATH as
QT_CONFIGURE_TEST_CMAKE_SYSTEM_{PREFIX|FRAMEWORK}_PATH variables to
tests.
Tests with separate project files that call find_package() must add code
like this after the project() command:
if(DEFINED QT_CONFIGURE_TEST_CMAKE_SYSTEM_PREFIX_PATH)
set(CMAKE_SYSTEM_PREFIX_PATH
"${QT_CONFIGURE_TEST_CMAKE_SYSTEM_PREFIX_PATH}")
endif()
if(DEFINED QT_CONFIGURE_TEST_CMAKE_SYSTEM_FRAMEWORK_PATH)
set(CMAKE_SYSTEM_FRAMEWORK_PATH
"${QT_CONFIGURE_TEST_CMAKE_SYSTEM_FRAMEWORK_PATH}")
endif()
Adjust pro2cmake accordingly.
Task-number: QTBUG-97076
Change-Id: Iac1622768d1200e6ea63be569eef12b7eada6c76
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
No need to reinevent the wheel. Also, use nanoseconds
now that we have the precision available.
Change-Id: I287d06198edc3f5bdf7f85bd5f3e235b1fea95a3
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Unify cancel and finish in QPromise destructor in a single call. This
saves us one extra mutex lock and atomic state change.
Task-number: QTBUG-84977
Change-Id: Iac06302c39a2863008b27325fcf6792d4f58c8ae
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Change-Id: I5d66a67a5c16a479f62b9f43cb769d5ee5f66b60
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
When passing an absolute input file name to qlalr, the #line directives
in the output will contain absolute paths. For reproducible builds,
it's desirable to have relative paths.
Pass the input file as relative path to qlalr to produce relative #line
directives.
Fixes: QTBUG-96267
Change-Id: I4ef1e4d5be7cbaf25a54a34ab7c2dbec06411a1d
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Use '-G Ninja' instead of '-G"Ninja"'.
Pick-to: 6.2
Change-Id: I53a77ea3b2d1547bb516b3ab756f5ae0f24092b1
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
We prefer the shared cache path, because there is no reason to give up
benefitting from sharing the content between applications. If that path
is not QFileInfo().isWritable(), we fall back to the local cache path.
However, there are reportedly systems with security solutions such as
AppArmor where the writable check passes and yet attempting to create a
file still fails. Then there is no cache in effect, because nothing ever
gets written out.
Handle this better: if writing the file fails and we still use our first
choice, the global cache location, fall back to the secondary choice
(the app-local path) and try again.
Pick-to: 6.2 6.1 5.15
Fixes: QTBUG-96789
Change-Id: Ifea32e9af0cf85aa70f3069256ed3d6a7e2fbe90
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Dalton Durst <dalton@ubports.com>
Reviewed-by: Andy Nichols <andy.nichols@qt.io>
This was confusing for me, being unfamiliar with the API.
Pick-to: 6.2 5.15
Change-Id: I831c6d0aa30847e069a7c21c279f147a1b24e486
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Unfortunately we can't depend on the C++11, 14 or 17 Standard
Library, which is not properly and completely implemented
everywhere. All Library features above C++98 must be checked with
their corresponding __cpp_lib macro before use. This does not
apply to the C++17 Core Language.
qwaitcondition_p.h:144:20: error: 'mutex' in namespace 'std' does not name a type
144 | using mutex = std::mutex;
| ^~~~~
qwaitcondition_p.h:59:1: note: 'std::mutex' is defined in header '<mutex>'; did you forget to '#include <mutex>'?
58 | #include <condition_variable>
+++ |+#include <mutex>
59 | #include <mutex>
qwaitcondition_p.h:145:33: error: 'condition_variable' in namespace 'std' does not name a type
145 | using condition_variable = std::condition_variable;
| ^~~~~~~~~~~~~~~~~~
qwaitcondition_p.h:59:1: note: 'std::condition_variable' is defined in header '<condition_variable>'; did you forget to '#include <condition_variable>'?
58 | #include <condition_variable>
+++ |+#include <condition_variable>
59 | #include <mutex>
Pick-to: 6.2
Change-Id: I2bbf422288924c198645fffd16a9249e5330136a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
On a mobile device, selecting text in a line edit brings up the text
action popup for select/copy/cut. In a spinbox where the user changes the
value using the buttons, this can be very irritating, without providing
any usability - the user is unlikely to start typing, at least not
without first transferring focus into the lineedit first to bring up the
keyboard.
This style hint allows styles to override the default behavior of
QAbstractSpinBox. Implement the customization for the Android style, and
add a test case for QSpinBox.
[ChangeLog][QtWidgets][QStyle] A new style hint, SH_SpinBox_SelectOnStep,
specifies whether pressing the up/down buttons or keys in a spinbox will
automatically select the text.
Fixes: QTBUG-93366
Change-Id: If06365a7c62087a2213145e13119f56544ac33b5
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Richard Moe Gustavsen <richard.gustavsen@qt.io>
I staged the manual test a little too soon, forgetting it's not compiled
in CI
Change-Id: Iaae8b8caaf8433c45e66ff662bb9bb7b25a3b8bd
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
Moving the class out makes it easier to navigate the code
Change-Id: I8de67641512a7dbbf6446cdca9d8dea79a63c217
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
We were using the non-standard interface for PropertiesChanged, and
support for this was dropped in a Network Manager update earlier this
year[0].
Switch to using the standard interface.
[0] https://networkmanager.dev/blog/networkmanager-1-32/
Pick-to: 6.2
Change-Id: Ibdf2b635ae13a3150c4d2faa028c15c70c8af5ca
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
The first callback added for the android backend does not have the value
as a parameter which is something the other callbacks have. Change it so
it does.
And promote the lambda to a real function.
Change-Id: I06f10f7c79f33a5ea3154f2fc6d20d550cd7eca7
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
It's part of the capabilities which we are already using. It also lets
us work around a pre-existing edge-case where, if you have a VPN enabled
and enable Airplane mode it will continue to tell you it is Online even
when it is not. This happens because VPN is reported as a transport and
when Airplane mode is enabled it may be left enabled as the _only_
transport.
At the same time clear the default filters (if any), and filter out
suspended connections. May not necessarily make any difference.
And add a comment for why we cannot use a technically more suitable
type of callback.
Task-number: QTBUG-91023
Change-Id: Ic26c4d4e8da139ec8606a0b1bf5fb7157bd0beaf
Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
The new public API returns and notifies changes to the currently active
transport medium for the application. And there's a new private API to
report it, with backends to follow.
Task-number: QTBUG-91023
Change-Id: I527985f9dabcd7bc4a32f36597e21bc4ab664c4e
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
On macOS, when showing a window, we decide if it should be made
key and therefore active, if the app has no active modal session
or if the window's worksWhenModal returns true.
However, the window needs to be made key also when a modal window
is present, but not visible. Add this condition when checking if
the window needs to be made key.
This makes the behavior consistent with what happens when a modal
is minimized on macOS. The input focus is passed to the next window,
and the window appears active, even if it can not be interacted with.
Fixes: QTBUG-85574
Pick-to: 5.15 6.2
Change-Id: I204d4f912128f4a46840789fc2ee08e1b2716bfc
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
In QDoubleSpinBoxPrivate::validateAndInterpret() some code still
assumed checking one entry in a QString is enough (it isn't - the
grouping separator may be a surrogate pair) and nothing considered the
case of an empty grouping separator (as can arise if the user sets
that in their system configuration, as a brute-force way to suppress
digit-grouping). Failure to consider this case failed an assertion on
dereferencing the first character of the separator.
Handle the case of empty separator and suppress tests that try to
match a separator, when there is no separator.
Pick-to: 6.2
Fixes: QTBUG-96734
Change-Id: I63d3205ef8656a2cd2c0d855a25e712ad66e2429
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
For now it only has a trivial test (empty) and an all-surrogate test
(Chakma digits) - but at least now all conversions to and from UTF-16
are tested. In particular, there were previously no UTF-32 tests.
Pick-to: 6.2
Change-Id: I9317928a88b9990530126db80e4756b880a364df
Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
WM_GETDPISCALEDSIZE and WM_DPICHANGED.
Pick-to: 6.2
Change-Id: I0614b0f552e52fc77d026dbc7e8c13b05f7d0e9e
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The event flow for programatic window geometry change
(e.g. from MoveWindow()) differs from user-interactive
geometry change: We still get WM_DPICHANGED, but this
event is not preceded by WM_GETDPISCALEDSIZE, so we don’t
get to override the window size.
However, Qt has already scaled the window size for the
new DPI in this case (the scaled size is provided to
QWindowsWindow::setGeometry()), so we can omit making
second native set-geometry call.
Pick-to: 6.2
Change-Id: Ia7d42d7fee49adf757e7fe75d77f1731405ad519
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
By handling WM_GETDPISCALEDSIZE we can keep QWindow’s
device independent size constant across DPI changes.
This is done by scaling QPlatformWindow’s native size
such that the change of scale factor and change of
QPlatformWindow size cancels out.
Qt now handles DPI change using two events:
WM_GETDPISCALEDSIZE: Compute the new size for the window.
WM_DPICHANGED: Apply the new DPI and window geometry.
The reason for this complication is that Windows retains
control over the window position during the DPI change,
in order to e.g. accurately track the cursor position
during a screen change.
The default WM_GETDPISCALEDSIZE implementation (provided
by Windows) scales the win32 window size linearly with
the DPI change. We want to use linear scaling as well,
however the win32 window size includes the margins, which
do not change linearly as the DPI changes.
Instead, scale the QPlatformWindow size, and then add
the new margins.
Pick-to: 6.2
Change-Id: I4f225be8fad56b1fa77e9e3cfd6538a206589d73
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
We want to have as little code in the QWindowsContext
event switch as possible.
Pick-to: 6.2
Change-Id: I04d578aae81c4ee804310a70bd87ee60b2890b6a
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
Closing a window with a native child results in the native child's
QWidgetWindow being closed. That explicitly calls setVisible(false)
on the child, which will still have the ExplicitShowHide attribute
set from the initial (explicit) show. Even though we then reset
the ExplicitShowHide, the WState_Hidden attribute will still be
set, so Qt considers the window to have been hidden, and not show
it again when the parent becomes visible.
Add a test case.
Fixes: QTBUG-96286
Fixes: QTBUG-79012
Fixes: QTBUG-71519
Change-Id: I482e6d5236c959d82ce66798176b259a3176972c
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
The only place where we use it, we swiftly cast it to QMetaObject*
anyway. Generating an object with virtual methods, only to then cast to
its non-virtual base type is dangerous. Also, if we are required to
return a QAbstractDynamicMetaObject from a method of
QDynamicMetaObjectData, we cannot implement QDynamicMetaObjectData
without also implementing QAbstractDynamicMetaObject. This indirectly
forces us to copy metaobjects around as we cannot wrap an existing
metaobject into a QDynamicMetaObjectData subclass.
Copying metaobjects around makes QMetaObject::inherits() unusable
because inherits() assumes identity is determined by identity of the
pointer.
Change-Id: Icbe697efa6fd66649eef3c91260c95cda62e3c90
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Remove dead code
Make logic to position test widgets more consistent across test functions
Consistent cursor positioning logic
Get rid of unnecessary event processing
Reduce calls to qWait with hard coded values
Use qWaitFor to establish asynchronous preconditions
Ignore expected warning messages
Expect-fail (instead of blacklist or skip) some tests that always fail
Ignore unpredictable events in childEvents test
Split large "render" test in multiple test functions
Task-number: QTBUG-52974
Task-number: QTBUG-26424
Change-Id: I206e114c2c1d9801e08600d3341c2738dc881f17
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
The static analyzer can't see that the QPointer not being null implies
that the QWidget pointer is still valid. Since we reset the widget
pointer to nullptr earlier on if the QPointer is null, just test for the
widget pointer instead.
Amends f6befd3e63 and fixes report
020fbdde70065e8de370537304d13d21.
Change-Id: I715764956a02b1300766a3cea2d427f840f1cec8
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
This reverts commit 7fd9ed3201.
While trimming the font name worked for cases with application fonts,
it actually introduced an assert for system fonts that ended with a
space, because enumerating these failed. So the original assumption
that all Windows APIs also trimmed the family name was wrong.
The original bug was that the font with the trailing space could
not be selected, but when using setFamilies(), it can. So there is
a perfectly fine way around the original bug when using a font that
has this problem. Therefore, no additional fix is needed for that.
[ChangeLog][Windows] Fixed an assert that happened when the system
had a font with a trailing or leading space in its name.
Pick-to: 5.15 6.1 6.2
Fixes: QTBUG-93885
Task-number: QTBUG-79140
Change-Id: I6d9df31a4f2c6555d38d51da374f69b6fb0f1ecb
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
