They are deemed good enough.
qt_deploy_translations is left in TP because with the current
signature, it's not clear yet how it's supposed to work for iOS
and other non-desktop targets.
[ChangeLog][CMake] The Core CMake deployment API
is out of Technical Preview status.
Task-number: QTBUG-108507
Change-Id: I384233c697b33644de3c9e1fb17d04f44ca16ea2
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Aside from the generation of the xml and junit xml files
for each test, also save the test output into a
tst_{name}-{timestamp}.txt file.
This will allow developers to inspect the output of passed tests,
by downloading the test results archive, even if we use ctest's
--output-on-failure.
Change-Id: I0a6c0ee04b4525d3ad9b207b28117d2182d29c28
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
It is deemed good enough.
[ChangeLog][CMake] The Core CMake QT_RESOURCE_ALIAS
property is out of Technical Preview status.
Change-Id: I50128a8ea4e82b3d15ff272eb713bc0f6a8b167d
Fixes: QTBUG-109077
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
The only usage of that function is to decide whether we should etch
disabled text. That should depend on the actual palette we are using,
not on a system setting (that might be ignored by an application or
style overwriting the palette).
Since styleHint might be called without option and widget, fall back to
the application default palette if needed.
Change-Id: Icf90eb4198890c613dccea6188733e74995962c5
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
The Vista style uses system assets for controls, and those are never
dark. Because of that we cannot support dark appearance with that style,
and applications using the Vista style should always use the light
system palette. Override QStyle::polish(QPalette &) in the vista style
to do that.
To make that palette available, move the code reading the light palette
into QWindowsApplication, and call that method from both the platform
theme (if the system is running in light mode) and from the Vista style
(only if the system is running in dark mode).
If the system is dark mode and another style is used (e.g. Fusion, which
works well with a dark palette), then the palette returned by the
platform theme gets used without any modifications.
This requires duplicating some small inline helper functions in
QWindowsTheme and QWindowsApplication. We can clean this up once the
implementation is complete for both Qt Widgets and Qt Quick.
Change-Id: Ia13f59a2d8414642603f9708926718daf9e8954d
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Santhosh Kumar <santhosh.kumar.selvaraj@qt.io>
This reverts commit 3ee7a9f85c.
Styles that don't support dark mode can overwrite the palette in
a QStyle::polish(QPalette &) override instead.
Change-Id: I8b84d822d91be5b3f67f76f791a1330bba7ea546
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Santhosh Kumar <santhosh.kumar.selvaraj@qt.io>
This will enable all qt internal tools to be able to access files
on paths longer than 255 characters.
Two examples that were not working before: moc and windeployqt.
Fixes: QTBUG-109207
Change-Id: I93f9770f1d3c4f3f2cca4655e4bed89c95b9786b
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
The test message sent to the dbus might never see a reply, e.g. in case
of a misconfigured dbus. Instead of waiting for a reply and stopping the
static top-level build, timeout after the CI had enough time to receive
the dbus message (5 seconds).
Change-Id: Ic6d60e7d1ddb2d864dc4a78d4edbd0ef7b6604c2
Reviewed-by: Alexey Edelev <alexey.edelev@qt.io>
Added Positioningquick, Sensorsquick, Webenginequick, and Webviewquick
modules to windeployqt list
Task-number: QTBUG-105135
Pick-to: 6.4
Change-Id: I26c0b5006453c8533ff94967b2e8f03ede88bf7e
Reviewed-by: Oliver Wolff <oliver.wolff@qt.io>
Most code uses the window/backing store DPR, except for this path that
uses the screen.
On wayland it's possible to get a different ratio for the screen and
window which causes some subtle rendering issues. For other platforms
behavior shouldn't change.
Change-Id: Ie390ba3bdfc183815df9e7e79e508d3e15d61f25
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
qtattributionsscanner expects file paths to be separated by a space.
Pick-to: 6.2 6.4
Change-Id: I4c9dfea0f086fc9631cb06f40e2d3cab0a32ca4e
Reviewed-by: Jörg Bornemann <joerg.bornemann@qt.io>
Amends a874087504, which tested whether
d->container is nullptr to decide whether to hide the popup, and then
dereferences d->container later without checking again. This raised a
correct static analyzer warning.
Simplify that logic. hidePopup() does nothing if there is no visible
container, and we don't want to accept() the cancel key if there isn't.
So the closeOnCancel logic isn't actually needed, we only need to accept
the ShortcutOverride to make sure that QComboBox sees the Cancel key
even if there is a shortcut registered, and then we can handle and
accept the cancel key to call hidePopup() only if the popup is visible.
Add test to verify that this interaction works as expected.
Pick-to: 6.4
Task-number: QTBUG-108908
Change-Id: I60d92b068f0f5139d629cf4a58e225512170df77
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Added some background about Serialization with the classes and used cases.
Fixes: QTBUG-103951
Pick-to: 6.4 6.3 6.2
Change-Id: I3ff179b814fc5d424f2ac2ffaf3237b90ddd7e2b
Reviewed-by: Vladimir Minenko <vladimir.minenko@qt.io>
Reviewed-by: Cristian Maureira-Fredes <cristian.maureira-fredes@qt.io>
Reviewed-by: Kai Köhne <kai.koehne@qt.io>
When the screen resolution or scale factor changes, then we have to
recreate the system tray icon so that it doesn't get blurry. Such
changes generate a WM_TASKBARCREATED message, which we already handle
through a self-registered message ID to re-add the icon when the task
bar/explorer crashed. So call updateIcon there to recreate the HICON
before adding it again.
For this to work, we have to actually store the QIcon in the already
present (but so far unused) m_icon member, and reset that member before
calling updateIcon, which would otherwise be a no-op.
Fixes: QTBUG-108641
Pick-to: 6.4 6.2
Change-Id: If3dd042416b7f61bbb3b43f7d563e52b406136a2
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
__ASSEMBLER__ and __cplusplus cannot both be defined at the same time,
so it suffices to check just one of them. Thanks to Thiago for
noticing.
Change-Id: I92e43d1a1c551ee48c51490b2878bb0c84bfb36f
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Partially reverts 50b05e3e2a.
The move of qVersion() from qglobal.h into qlibraryinfo.h, a header
not included by qglobal.h, was a conscious, documented, QUIP-0006-
allowed, but stetching the SiC Type A definition, and therefore
unwarranted, SiC.
Fix by moving qVersion() into its own header file, qtversion.h,
included from qglobal.h.
[ChangeLog][Editorial] Remove the entry for qVersion() being moved to
qlibraryinfo.h.
Fixes: QTBUG-108818
Change-Id: I3524802348635512a63fbbb55ddc28ad28f331e8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
A deleted mMimeData instance might be referenced if more that one file
item is present in the clipboadData. This fix uses a continuation
mechanism to only run the actual qWasmClipboardPaste when all the
data has been collected. The ownership of QMimeData is then transferred
to the global QClipboardData q_clipboardData.
Fixes: QTBUG-108841
Change-Id: I16def48d70ebbffc68462ed74ccd9ee8ee8053de
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
This was missed during original patch
Pick-to: 6.4
Fixes: QTBUG-108677
Change-Id: Icd8960f2f41468dc12c3269b12df777acd962753
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The algorithm is the same, differing only on how we load data onto
vector registers, so they can be merged, simplifying the code base.
For all strings over 16 characters in length, we loop and then we
perform a final overlapped comparison, if necessary. I've kept the 32-
byte-per-loop solution even for pre-AVX2, because that should pipeline
better. For any strings between 4 and 16 characters, we perform a pair
of maybe-overlapped comparisons, of either 4 characters or of 8, so we
consume the full string. That leaves a tail of at most 3 characters in
scalar code.
Change-Id: Ib42b3adc93bf4d43bd55fffd16c8c15d6d761af2
Reviewed-by: Lars Knoll <lars@knoll.priv.no>
Replace the regex that looks for 'We mean it.' disclaimer with simple
string lookup to allow the disclaimer to have a follow text or be in
the middle of the sentence.
Amends b89d63515b
Task-number: QTBUG-87480
Change-Id: I2a9d7306beb51de62d00b79980a12da6f170df93
Reviewed-by: Robert Griebl <robert.griebl@qt.io>
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
In CMake versions older than 3.20 add_dependencies have no effect
when adding interface libraries. So need to add the '_lib_pri'
targets to ALL to make sure that the related rules executed.
Amends 190e58e1f4
Pick-to: 6.2 6.4
Fixes: QTBUG-109240
Fixes: QTBUG-109239
Change-Id: I7c565ce3bc15e549569317454c2f7caac823cf66
Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
Coming work shall limit the feature to controlling these parts of the
class, while making the rest visible to all builds.
Change-Id: I602c604c8c66bd4f1816a9dabef19c72527831a5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In the process, split some long lines. The test relies on omitting the
hour (so as to get the default, 0) from both the format string and the
string parsed, so as to test that the parser correctly handles the
corner case where the zone skips the first hour of the day. This was
not entirely obvious when reading the row data, so make it explicit in
a comment.
Change-Id: I919b292b78bd399a8749806a0e913d43f5b414e1
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mate Barany <mate.barany@qt.io>
Test a few more cases are correctly handled.
Change-Id: I7f286ba93f59bf0168cac789cd30590f40e98cee
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The spec deseralized isn't a Qt::TimeSpec; handle it correctly instead
of taking for granted that QDateTimePrivate::Spec's values happen to
match.
Change-Id: I67f3c960f3a3b90cdad3c1eca673f7ec8fd10b82
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Mostly pedagogic checks in operator_insert_extract(), but also
sanity-checking, to confirm spec conversions produce results equal to
what each came from.
In daylightTransitions(), verify the spring forward goes from standard
time to daylight-saving time.
Change-Id: Ieb9c603ee2eadecea055da4e8889528161f4d999
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Use QCOMPARE(a, b) in place of QVERIFY(a == b), similar with
QCOMPARE_LT() and QCOMPARE_LE(); and use a scope-guard to emit a
message on failure instead of incurring the cost of building a string
for the message, even when the passing test doesn't need it.
Change-Id: I3884bc40e89a4b1ba881968b99faab27d4b1abc9
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Ivan Solovev <ivan.solovev@qt.io>
Principally to get Qt::LocalTime mentions out of the way ahead of the
QTimeZone work on Qt::TimeSpec, but also mop up trailing 0 parameters
to QTime for seconds and milliseconds.
Change-Id: I51041582faae100894a567c9e5ae96a60a3b2d8c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Mate Barany <mate.barany@qt.io>
Left over from long ago, making confused use of Qt::hex.
Change-Id: I7f411e4888ee1a637d2212fd6976dd003f8da9ce
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Several were overlong (or soon to get so by being made longer); others
were inconsistent with neighbors; one was inconsistent with itself.
Change-Id: I272680499605a757e4827d27021bf234a91cf77a
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
As a result, also make sure the test will fail if output to debug
stream doesn't produce the expected result.
Change-Id: I9914c9c41c8d8b79f32dfb8e0c735f12e2d59f5e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Put core includes all in a common form, sort alphabetically.
Remove a stray blank line.
Change-Id: I211c6b407f5e49d907cb065521883567f1dd30f4
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Make the tz variable more local (saving some #if-ery).
Recognize that the Qt 4 (and 5.1) format didn't store offset or zone
so we're treating offset from UTC as UTC and zone as local time.
Don't assign a variable the value it already has.
Omit parameters when they're the default value anyway.
Improved some comments.
Change-Id: Ie29a66b50ecb8ed8b0566a0ecbc6492c6ba63e61
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Comparing to true and false doesn't enrich the output.
Change-Id: Ie26a3f3d584f88310b8d4a31cad07be8dc8cb646
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Certain older SDKs don't declare the specific wait_for overload but
still uses it. So, we declare it.
Also add the factorycacheregistration include to wipe out any factory
that may be created and cached in the TU.
Change-Id: I778ab1b8724306b17bb6a2daa0b7364f268ded83
Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
Reviewed-by: Eskil Abrahamsen Blomfeldt <eskil.abrahamsen-blomfeldt@qt.io>
The QWasmScreen's top left coordinate does not precisely translate
to correct page coordinates, especially when fixed position is used
and page margins are set. Also, this is wrong in complicated setups
with e.g. multiple nested elements.
Therefore, to get the correct coordinates in pointer event handlers,
we have to assume the local coordinates of the screen, and translate
those to the (possibly incorrect) coordinates that QWasmScreen thinks
it has in page.
It is another problem to fix the wrong coordinates QWasmScreen thinks
it has in the page.
This has been checked with complicated setups with screens in scroll
containers, screens with fixed position, screens with relative position,
with and without body margins etc.
Change-Id: I749f2507fec7ae278b6f9d7d13ae288e65472dba
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
This got broken with the introduction of the compositor patch.
Change-Id: I7d85795eb537449855a4cce3c8b6b031095c3f7f
Reviewed-by: David Skoland <david.skoland@qt.io>
static Q_ALWAYS_INLINE __m128i mm_load8_zero_extend(const void *ptr)
is unused when compiled on Android.
This causes a compiler warning and blocks CI, if compiled with -Werror.
This patch marks the method [[maybe_unused]], in order to prevent the
warning.
Change-Id: Ife10c0a84ac34196405ce9c2356351d825751adb
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
The + operator is left-associative, so a + b + c is (a + b) + c.
Apply the same trick C++20 applied to std::string's op+ and overload
for rvalue left-hand sides. This means that a + b + c is now
equivalent to
[&] {
auto tmp = a;
tmp += b;
tmp += c;
return tmp;
}()
removing a ton of temporary buffers (not objects, because CoW makes it
impossible for the compiler to track the single conceptual object
passing through the chain) when not using QStringBuilder (which isn't
available for QList).
This is BC, because the operators are all inline free functions or at
least inline members of non-exported classes.
Use multi-\fn to document the new operators. No \since is needed, as
this doesn't change the set of supported operations, just makes some
of them faster.
[ChangeLog][QtCore][QList/QString/QByteArray] Chained additions (a + b
+ c) now produce only one temporary buffer for the whole expression
instead of one per addition. Using += or QStringBuilder is still
faster, though.
Change-Id: I87e837d8803e79dc29c9268f73e6df9fcc0b09a3
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
No need to involve the compiler, make the preprocessor error out.
Change-Id: I37d3f241bd0879ba123fe26342aaaaff74c2fb6b
Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
This structure was 216 bytes with 3 padding regions: 3 bytes between
continuationState and refCount, 4 bytes between m_progressValue and
m_progress, and 1 byte at the end after the booleans. The total of 8
bytes can be neatly reduced.
And since it's now a multiple of 16, we should get an effective 16 byte
reduction in malloc()ed size.
Change-Id: I69ecc04064514f939896fffd172e46d1d68cb28b
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
Move the assertions to qtypes.cpp, next to where one of them already
existed.
Change-Id: I51d12ccdc56c4ad2af07fffd172dae7c49d78273
Reviewed-by: Marc Mutz <marc.mutz@qt.io>
