summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* xcb: Flush scrolled region of backingstore to client when requesting imageTor Arne Vestbø2021-10-051-1/+6
| | | | | | | | | | | | We were doing this for composeAndFlush already, which accesses the image via graphicsBuffer(). Task-number: QTBUG-64504 Task-number: QTBUG-64414 Pick-to: 6.2 Change-Id: I51cc002db8a15991f8a8aa32e849c07318e6e74c Reviewed-by: Błażej Szczygieł <mumei6102@gmail.com> Reviewed-by: Laszlo Agocs <laszlo.agocs@qt.io>
* Add basic android multi-abi support for CMake projectsAlexey Edelev2021-10-053-15/+177
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Use CMake external project to crosscompile android libraries for mutliple android ABIs at the same time. The idea behind is to use pre-compiled Qt for each android ABI in external projects, compile libraries and then utilize results of compilation in common androiddeployqt call. By default multi-abi build uses the main ABI that qt toolchain file belongs to. The list of the autodetected Qt for Android ABIs is stored in QT_DEFAULT_ANDROID_ABIS cache variable. Users may change the set of the Android ABIs project-wide using QT_ANDROID_ABIS CMake variable or for the specific target by adding the ANDROID_ABIS argument when calling qt6_add_executable. To enable build for the autodetected ABIs user may set the QT_ANDROID_BUILD_ALL_ABIS option to TRUE. By default build procedure is looking for the respective android_<abi> folders to run per-abi build. If user's Qt for Android folder structure is different then one is created by Qt installer, path to the each Qt for Android might be overwritten using QT_PATH_ANDROID_ABI_<abi> CMake variable. Note: This only adds support for the multi-abi build of user's applications. TODO: This commit limits projects to not have in-tree library dependencies. That means that executable targets may have dependencies only from Qt directories or android sysroots. See QTBUG-94714 for details. [ChangeLog][Android][Platform Specific Changes] Added basic support for multi-abi builds of user projects. Task-number: QTBUG-88841 Change-Id: I3239ffe61e6b437e170c8decc5c36a9e774ed0fb Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
* QSFPM: remove redundant call to update_source_sort_column()David Faure2021-10-051-2/+1
| | | | | | | | | | | | | | | | | | | | | Everywhere else in this code, it says if (update_source_sort_column() && dynamic_sortfilter) so, doing the same here. This originally comes from commit 3ffcfc357dbc51bf62ffb but skelly isn't around anymore, and Milian (reviewer) had no objection to this change either. Looking into the implementation, calling it twice is a no-op anyway, create_mapping will early-return with "already done". On the other hand, now it has a chance of returning true, while before it would always return false in the second call ("return true if the column was changed"). I couldn't write a unittest for that case though, it would require that a source reset also changes the proxy->source column mapping somehow, but I don't see how that can be done (the latter being the QSFPM setup). Change-Id: Ie560facf1039a1afc02543e83bfdef0f8f2e09fc Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Remove qshadergraph filesJuan Casafranca2021-10-0517-2852/+0
| | | | | | | | | | | | - Those files were moved as part of Qt3D as its the sole user of these - Also removed associated unit tests Pick-to: 6.2 5.15 Change-Id: I302bc219218a58071c86d2447cb4449601fca32c Reviewed-by: Edward Welbourne <edward.welbourne@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Mike Krus <mike.krus@kdab.com>
* QNetworkInformation: refactor NetworkManager backend a littleMårten Nordheim2021-10-051-12/+10
| | | | | | | | | Using individual booleans to track that the value had changed was not necessary. Change-Id: Ieb4712776339aa78bdc75d7608fd74f96d19bbae Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io> Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
* Doc: Do not highlight DTLS client/serverKai Köhne2021-10-051-3/+1
| | | | | | | | | | | The client and server do in practice depend on each other. So setting up things is a tad more complicated, and the use case not big enough to warrant both to be highlighted. Task-number: QTBUG-96575 Pick-to: 6.2 Change-Id: I0ed69f7a20ec490cc977bde7f1b09162153d0bd2 Reviewed-by: Alex Blasche <alexander.blasche@qt.io>
* Doc: remove unnecessary newline in snippetKai Köhne2021-10-051-1/+0
| | | | | | Pick-to: 6.2 Change-Id: I6031fe8db971be4cbcaec8f1dc6fa7346edb601b Reviewed-by: Paul Wicking <paul.wicking@qt.io>
* Doc: Make it explicit when CMake API is definedKai Köhne2021-10-0520-0/+55
| | | | | | | | Task-number: QTBUG-96239 Pick-to: 6.2 Change-Id: Iaa99024f98f5422aa39f148a89594186b40344bf Reviewed-by: Edward Welbourne <edward.welbourne@qt.io> Reviewed-by: Paul Wicking <paul.wicking@qt.io>
* Remove QWeakPointer::operator->Fabian Kosmale2021-10-051-4/+0
| | | | | | | | | | | | | operator-> was only defined if QWEAKPOINTER_ENABLE_ARROW is defined. However, even in that case it would call QWeakPointer<T>::data, which does not actually exist. Take this as an indicator that nobody actually uses operator->, and remove the code completely. Note that the QWEAKPOINTER_ENABLE_ARROW was not documented, and neither was operator->. Change-Id: I2f4aa961a64281542c8c1b248a993e83471c059d Reviewed-by: Lars Knoll <lars.knoll@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
* QLocalSocket/Win: add missing check to waitForBytesWritten()Alex Trotsenko2021-10-051-1/+1
| | | | | | | | | | | | | | | Calling pipeWriter->checkForWrite() can indirectly close the socket even if the bytesWritten() signal has not been sent. So, we need to make sure the handle is valid before checking pipe state in the reader. There is no harm in calling PeekNamedPipe() with an invalid handle, but the wrong call should be avoided. This patch amends b2c3b3e8fe0d8bdc88051d0120aaa8d5cf8acce0. Change-Id: I5d2ecbbbe0af817aac68ad6f1173b0ed9b324e98 Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de>
* QPlugin: pass the direct header to qJsonFromRawLibraryMetaData()Thiago Macieira2021-10-045-7/+23
| | | | | | | | This is done in preparation for storing the metadata without the magic string in static plugins and in ELF notes. Change-Id: I3eb1bd30e0124f89a052fffd16a820454dd56d3e Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* QHash: suppress annoying GCC warning about allocation size too bigThiago Macieira2021-10-041-20/+31
| | | | | | | | | | | GCC is right, the maximum memory allocation is half the VM size, not the full VM size or multiple times that. QHashPrivate::Span is big. qhash.h:552:17: warning: argument 1 value ‘18446744073709551615’ exceeds maximum object size 9223372036854775807 [-Walloc-size-larger-than=] Change-Id: Ie72b0dd0fbe84d2caae0fffd16a071ffb5d0c70f Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* wasm: add simd supportLorn Potter2021-10-052-1/+8
| | | | | | | | | | | | | | | | | | Emscripten only supports SSE1, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2, and 128-bit AVX instruction sets at this time. https://emscripten.org/docs/porting/simd.html Browsers might need to enable simd support in the advanced configurations about: config or chrome:flags Enable by configuring Qt with -sse2 Pick-to: 6.2 Fixes: QTBUG-63924 Change-Id: Ifeafae20e199dee0d19689802ad20fd0bd424ca7 Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io>
* Fix __cpp_lib_hypot related error when building in macosAlexey Edelev2021-10-052-2/+2
| | | | | | | | | | | | | | | If __cpp_lib_hypot is undefined in macos you may observe the error: error: '__cpp_lib_hypot' is not defined, evaluates to 0 [-Werror,-Wundef] Adding the explicit check for definition suppresses the warning that is treated as an error. Change-Id: Ie4c185fefde2f5bab699d8fc79b6a170e64af393 Reviewed-by: Andrei Golubev <andrei.golubev@qt.io> Reviewed-by: Edward Welbourne <edward.welbourne@qt.io> Reviewed-by: Giuseppe D'Angelo <giuseppe.dangelo@kdab.com> Reviewed-by: Ievgenii Meshcheriakov <ievgenii.meshcheriakov@qt.io>
* QPlugin: add error messages to the qt_get_metadata()Thiago Macieira2021-10-041-6/+7
| | | | | | | This is the already-loaded counterpart to the file scanning code Change-Id: I3eb1bd30e0124f89a052fffd16a81f518fa95f0d Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* QPlugin: simplify handling of exact results from Q{Elf,Mach}ParserThiago Macieira2021-10-042-41/+27
| | | | | | | | | | Those two scanners always return exact results, if the metadata is present, so we don't need to re-scan (haven't needed since Qt 5.0). Especially since we scan from the end, we were spending cycles doing unnecessary work. Change-Id: I42eb903a916645db9900fffd16a4ccfdc7342278 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Q{Elf,Mach}Parser: harmonize the arguments passedThiago Macieira2021-10-045-69/+51
| | | | | | | | | | | | | | Both functions took a QString for the input file name, but while the ELF parser had an optional QLibrary pointer (which was never null) where to store the error string, the Mach-O parser received a pointer to a QString. So make both of them take a single in/out QString pointer, which has the file name on input and is cheap for us because of COW. Drive-by fix the name of the static function in qmachparser.cpp from "ns" (which stood for "not suitable") to "notfound". Change-Id: I3eb1bd30e0124f89a052fffd16a8182f4f8541c3 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* QPlugin: update the arch requirements to match the x86-64 ISA levelsThiago Macieira2021-10-043-7/+79
| | | | | | | | | | When we created the functionality, the levels were not yet standardized. Now they are and you can use -march=x86-64-v3 (for example) to get to them. We're making a split between "v1" and "baseline" here for the benefit of 32-bit, which is not included in the ISA levels. Change-Id: I2de1b4dfacd443148279fffd16a397a700b9a15a Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* QPlugin: use one global metadata magic stringThiago Macieira2021-10-042-34/+20
| | | | | | | And use QByteArrayView where applicable. Change-Id: I2de1b4dfacd443148279fffd16a39ec22908f75f Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* QPlugin: move some of the logic from moc's output to qplugin.hThiago Macieira2021-10-042-55/+129
| | | | | | | | | | | | | | This will allow us to make changes in QtCore itself, without having to worry about moc compatibility. The output uses an #ifdef so this version of moc can still be used to compile earlier versions of Qt (usually, in cross-compilation environments). See discussion in the mailing list[1]. [1] https://lists.qt-project.org/pipermail/development/2021-September/041732.html Change-Id: I2de1b4dfacd443148279fffd16a39784c80c5f3b Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* QPlugin: change the generic metadata scan to forwardsThiago Macieira2021-10-041-30/+10
| | | | | | | | | | | The reason it scanned backwards was valid, but unfortunately that doesn't work if the metadata magic appears in the debug info that comes after the metadata. Task-number: QTBUG-67461 Change-Id: I2bbf422288924c198645fffd16a9276767ca2589 Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* src/corelib/plugin: sort #includesThiago Macieira2021-10-047-44/+40
| | | | | Change-Id: I2de1b4dfacd443148279fffd16a3e4bdb1a49f88 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* QEventLoop: remove unnecessary bits of Qt 5 suggestionThiago Macieira2021-10-041-3/+1
| | | | | | Pick-to: 6.2 Change-Id: I2bbf422288924c198645fffd16a974803ab43b7f Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* CMake: Allow providing custom QT_ANDROID_SDK_BUILD_TOOLS_REVISIONAssam Boudjelthia2021-10-041-2/+5
| | | | | | | | | | | The current CMake configuration sets QT_ANDROID_SDK_BUILD_TOOLS_REVISION only internally, taking the latest build-tools found. A project might still want to manually set specific build-tools version. Pick-to: 6.2 Task-number: QTBUG-97002 Change-Id: I553563638160fdef7a27b9b334d1d9aae5d8b7e1 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Respect font stretch if set together with font styleAlbert Astals Cid2021-10-041-1/+1
| | | | | | Fixes: QTBUG-77854 Change-Id: I2bf9cea9d5ecd151a9d96bbe93e9477a9159ca1f Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
* Fix QPainterPath with QFont::SmallCapsAlbert Astals Cid2021-10-041-1/+1
| | | | | | | | | | | | | | Previous code was getting the QFontEngine from the font+script instead of from the QTextEngine. The font+script is not enough information to know if a given character is smallcaps or not, while the QTextEngine actually has access to the information needed and returns a properly a scaled fontengine if character is small caps Pick-to: 6.2 Fixes: QTBUG-13965 Change-Id: I9f95bd2f3c3bdff76c3acb94fa2edc99cdeb0a13 Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io>
* Remove an overly-complex assertionEdward Welbourne2021-10-041-9/+0
| | | | | | | | | | | | | | | | | | | | | | | | | It tried to sanity-check the correction being made to the time as a result of converting zone to UTC, but handling also the case where the correction avoided a gap. However, there are too many quirky cases to permit a simple condition. The latest to trip it up is a case of a zone leaving local solar mean time; a time in the gap that hits is shunted by the width of the gap, which could be any old thing, and won't show up on the data describing the zone before or after the transition. In any case, one branch of the check was doubling the zone's standard time offset, which makes no sense. As that was checking for double-DST, it would need to check standard offset plus twice the usual DST offset; but the data it's looking at probably has its DST offset set to the doubled one that needs to deal with. In general a change in DST without going via standard time will have problems here, and a change to standard time could also present problems. Stop trying to shore up the assert and just trust the offset we got. Change-Id: I4519f47e79dfb7595f4a12c259b80a338bc7033c Reviewed-by: Andreas Buhr <andreas.buhr@qt.io> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Doc: Add overview pages to CMake APIKai Köhne2021-10-0421-15/+100
| | | | | | | | | | | | | | So far the pages were only 'rooted' in the separate qtcmake documentation. Let them have a page per module too, so that it's easy to navigate. At the same time rename the cmake-macros-qtcore group to cmake-commands-qtcore. This will require a fixup in qtdoc repository. Pick-to: 6.2 Change-Id: Ifc09b27a6c220b5feab64686ca47c60d3342e71a Reviewed-by: Paul Wicking <paul.wicking@qt.io>
* Fix reference to quitOnLastWindowClosed in WA_QuitOnClose docsTor Arne Vestbø2021-10-041-1/+1
| | | | | | | | The property is nowadays on QGuiApplication. Pick-to: 6.2 Change-Id: I8759761b08b0cdcef786ca96a7fa6364331f157d Reviewed-by: Paul Wicking <paul.wicking@qt.io>
* Doc: Add summary to CMake commandsKai Köhne2021-10-0421-41/+41
| | | | | | | | | | | | For normal \page elements, \brief is just showing in the \generatelist or \annotatedlist commands. Make sure the description is also visible in the actual CMake command/variable/property page by defining a \summary macro. Pick-to: 6.2 Change-Id: I12bc854d547059a2f6309a5922bb0b2a36d4e41c Reviewed-by: Craig Scott <craig.scott@qt.io> Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
* androiddeployqt: only write enabled architectures to build.gradleAndreas Buhr2021-10-041-0/+2
| | | | | | | | | | androiddeployqt writes a list of architectures into the build.gradle file. This list should only contain the architectures which are enabled. This patch adapts the generation of this list accordingly. Amends 07cfab07a2a489e021b56226824a2d9b1211fad7. Change-Id: I5c7c1886c54cc3b9a7e42af6eefbaf2a067a67d5 Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
* QRingBuffer: port internals from int to qsizetypeAlex Trotsenko2021-10-032-15/+16
| | | | | | | | | | Since Qt6, QByteArray uses qsizetype as an integral type for offsets and sizes. In order to support large blocks, we have to migrate to that as well. Change-Id: I2c2983129d6a2e0a1e8078cc41d446a26e27288c Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@gmx.de> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Remove checks for features available in C++17Ievgenii Meshcheriakov2021-10-029-32/+2
| | | | | | | | | | This patch removes most of the checks that are made using C++20 __cpp_* macros for features available in C++17 and earlier. Library feature check macros (__cpp_lib_*) are unaffected. Change-Id: I557b2bd0d4ff09b13837555e9880eb28e0355f64 Reviewed-by: Edward Welbourne <edward.welbourne@qt.io> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Android: remove comments from the manifest fileAssam Boudjelthia2021-10-021-44/+6
| | | | | | | | | | | The comments are not needed, since we're adding a docs page for the manifest, which include the information from these comments and more. Keep the manifest clean. Pick-to: 6.2 Task-number: QTBUG-95285 Change-Id: I1fd8bc695f5aab1bcc69f549cb3ddeeaac168b98 Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
* Android: fix Gradle version in qt_attribution fileAssam Boudjelthia2021-10-021-1/+1
| | | | | | | | | | | | The Gradle version in qt_attribution.json doesn't mention the actual version that is currently used. This amends 67c3f3f4ffcadba32ea42522a36b3f6d88fc3482. Pick-to: 6.2 Fixes: QTBUG-71328 Change-Id: I50fbabd56beb44389f9453666822f9667bab7d61 Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
* Purge Q_NO_TEMPLATE_FRIENDS and platforms lacking support for themFabian Kosmale2021-10-023-48/+0
| | | | | | | | | | Given that we rely on C++17, it should be safe to mandate that level of language support. Change-Id: If07ccb36bea2a5113a8f5aacf635be7d2590baf7 Reviewed-by: Lars Knoll <lars.knoll@qt.io> Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org> Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
* Replace 0 with nullptrKarsten Heimrich2021-10-021-2/+2
| | | | | | Change-Id: I9ee25644a3e066391e881bd2f64274909eabfcbf Reviewed-by: Allan Sandfeld Jensen <allan.jensen@qt.io> Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
* Update PCRE2 to 10.38Giuseppe D'Angelo2021-10-0226-991/+1366
| | | | | | | | [ChangeLog][Third-Party Code] PCRE2 has been updated to version 10.38. Change-Id: Ib6ab544790747a94a00b8eb516314ff3c57e4c79 Pick-to: 6.2 5.15 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* QByteArrayMatcher: add QByteArrayView overloadsThiago Macieira2021-10-012-2/+33
| | | | | | | | | | | | | The modification to the existing constructor was necessary to avoid ambiguous overloads from plain strings. They'd previously only match QByteArray, but now they match QByteArrayView too. The QByteArray constructor must stay even in Qt 7 because the pattern is stored. The QByteArray indexIn also stays in because of ambiguous overloads. Unless we want to remove the const char* overloads. Change-Id: I2bbf422288924c198645fffd16a92859b67f3978 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Fix scrolling of QRasterBackingStore when operating on QRasterGLSurfaceTor Arne Vestbø2021-10-014-13/+15
| | | | | | | | | | | | We were to strict in what surface type we allowed scrolling for. The RasterGLSurface type is an odd one, used by widgets to compose GL and raster content, which means we still have a raster backingstore we can scroll. It's just the flush that's different. Pick-to: 6.2 Change-Id: Ia229c21c00ad38df9e87f4fc78e341e030ef228d Reviewed-by: Morten Johan Sørvig <morten.sorvig@qt.io> Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
* Use %.6g rather than %s and QByteArray::number()Edward Welbourne2021-10-011-3/+2
| | | | | | | | | | | | | The comment against the use of this claimed qvsnprintf() lacks 64-bit support, which is probably untrue by now but certainly irrelevant as the value being formatted is a qreal, not a 64-bit integral type. Since %g wants a double, make the value passed explicitly double by casting its denominator. Casting its numerator to qreal was superfluous as that's its type already. Change-Id: I5ff885fbeb9b638b2b0507061e0a19e3b8522143 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com> Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
* Clean up QXmlTestLogger's xmlQuote() and xmlCdata()Edward Welbourne2021-10-012-63/+63
| | | | | | | | | | | | | | | | | | Make the overloads taking a size private (they're only called by the implementation of the others) and have them take qsizetype for the size, rather than size_t. Make the still-public ones consistently return 0 on failure and move the documentation to them. Change some spurious checks (QTestCharBuffer::size() is necessarily positive) to assertions, use '\0' for the nul character to make it visually distinct from the integer 0. Use a do..while loop instead of break-ing out of the tail of a for (;;) loop. Remove spurious extra nul-termination - the function taking size alread did that, so the wrapper to size-expand needn't. Minor style cleanups. Change-Id: I7440f3d9be5e3b2dd88c2aad62bcd1af9fc02278 Reviewed-by: Thiago Macieira <thiago.macieira@intel.com> Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
* Tidy up QTest::qt_asprintf() and remove a spurious declarationEdward Welbourne2021-10-011-18/+11
| | | | | | | | | | | | | | | | The extern declaration for filter_unprintable() was nowhere used and no such function is anywhere defined. The for (;;) loop with lots of reasons to break; was better structured as a do {...} while loop. A comment on -1 as return from qvsprintf() was misplaced. The '\0'-termination after calling qvsprintf() was redundant as vsprintf() reliably '\0'-terminates anyway. Turned a static const into a constexpr. Assert size (a QTestCharBuffer necessarily has size >= 512). Change-Id: I5b7729b9bd66fea0ee7ce3e7cfdde6770f10b36c Reviewed-by: Thiago Macieira <thiago.macieira@intel.com> Reviewed-by: Tor Arne Vestbø <tor.arne.vestbo@qt.io>
* Disable the dnslookup feature for INTEGRITYJoerg Bornemann2021-10-012-1/+2
| | | | | | | | | | | | | | | | | | The QtNetwork project file excludes the dnslookup source files for Unix, despite the dnslookup feature is always on. This inconsistency leads to linker errors when building applications against QtNetwork: (error #412) unresolved symbols: 1 QDnsLookupRunnable::query(int, const QByteArray &, const QHostAddress &, QDnsLookupReply *) from libQt6Network.a(qdnslookup.cpp.o) Turn the feature off for INTEGRITY and remove the condition from the project file. Pick-to: 6.2 Change-Id: If086da51ca343f598eb273236e7b6338f5da1e97 Reviewed-by: Tatiana Borisova <tatiana.borisova@qt.io> Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Let QLocale::uiLanguages() use WinRT API when possibleIvan Solovev2021-10-012-2/+32
| | | | | | | | | | | | | | | | | | This patch introduces support for the WinRT UI languages API. We are using the Win32 API to get the list of preferred languages when the system locale is used. However, this API returns an incomplete list. As Qt 6 supports Windows 10 and above, we can make use of the WinRT API, if it's supported by the compiler. This API returns the full list, as reported by the Windows system itself. Note however, that this API can't be used with Clang and MinGW, so we still have to fall back to Win32 API for these compilers. We also do it if WinRT API returns an empty list of languages for some reason. Fixes: QTBUG-94341 Pick-to: 6.2 Change-Id: I1d23c68d2ec298ae7835d0d18718876ff041aede Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
* Don't export Qt6::ATSPI2_nolink target in Qt6GuiTargets.cmakeJoerg Bornemann2021-10-011-3/+11
| | | | | | | | | | | | | | We don't have any APSPI2 includes in our public headers and should therefore not export the target that provides ATSPI2 include paths for consumers. Link against PkgConfig::ATSPI2 instead of the _nolink target. The former will not be exported. Pick-to: 6.2 Fixes: QTBUG-97023 Change-Id: I4b12e0c2230917feeb963c02565e6db24f757bd3 Reviewed-by: Alexandru Croitor <alexandru.croitor@qt.io>
* Doc: Replace broken link to Roman Czyborra's personal pageLuca Di Sera2021-10-012-4/+4
| | | | | | | | | | | | | | | | | | | | The `QFont` documentation refers to Roman Czyborra's personal page with the intent of providing information about encodings. The page seems to have been down for many years and may contain information that is now deprecated such that it is now replaced with a link to the Unicode Technical Report 17 about the Unicode Character Encoding Model. The new link was chosen as a substitute as it seems to cover or provide the intuition for, standing by the last captured version of Roman Czyborra's personal page on the Wayback Machine, many of the same concepts. Task-number: QTBUG-96127 Pick-to: 6.2 Change-Id: Ibf901b5023688c14efb3d6f77a10609f7ba80e72 Reviewed-by: Paul Wicking <paul.wicking@qt.io>
* androiddeployqt: Allow to also provide multiple qt install directoriesRaphael Cotty2021-10-011-17/+91
| | | | | | | | | | | | | | | | | | The "qt" property can also be filled with a json object like this: "qt": { "x86":"/home/.../Qt/6.3.0/android_x86", "x86_64":"/home/.../Qt/6.3.0/android_x86_64", "arm64-v8a":"/home/.../Qt/6.3.0/android_arm64_v8a", "armeabi-v7a":"/home/.../Qt/6.3.0/android_armv7"} That allows to update the options.qtInstallDirectory when changing the current architecture. The "qtHostDir" property is also added to provide the path to the qt host installation (rcc, qmlimportscanner). Change-Id: I3f0c36cc7a7030d45adc1a4dccaaad61dd538aea Task-number: QTBUG-88841 Reviewed-by: Assam Boudjelthia <assam.boudjelthia@qt.io>
* wasm: fix runOnMainThread()Morten Johan Sørvig2021-10-011-6/+10
| | | | | | | | | | | Calling emscripten_async_run_in_main_runtime_thread_() with a pointer to a static lambda was too clever, use an anonymous function as callback instead. Pick-to: 6.2 Task-number: QTBUG-94344 Change-Id: I2d8a8b0ffc2dd1d02018aa5902550216d00f641d Reviewed-by: Lorn Potter <lorn.potter@gmail.com>
* QWinRtFunctions::await() - introduce early exit conditionIvan Solovev2021-10-011-8/+19
| | | | | | | | | | | | | | | | | | The await() method waits for the result of async operation in a non-blocking way (triggering processEvent periodically). It means that during this wait some other things might happen, and there would be no reason to wait for the end of the operation execution. This patch implements an additional parameter - a function that specifies a condition for an early return. When this function returns true, the await() method returns with E_ABORT, which makes it possible to distinguish it from timer expiration. Task-number: QTBUG-96057 Pick-to: 6.2 5.15 Change-Id: Ide73d768b7cbb3a35be7160ce7555aeb2dca5235 Reviewed-by: Oliver Wolff <oliver.wolff@qt.io> Reviewed-by: Juha Vuolle <juha.vuolle@insta.fi>
generated by cgit v1.2.3 (git 2.25.1) at 2024-05-27 10:22:49 +0000