aboutsummaryrefslogtreecommitdiffstats
path: root/src/qml
Commit message (Collapse)AuthorAgeFilesLines
* Doc: Use [] instead of new Array() in snippetPaul Wicking2018-08-011-1/+1
| | | | | | | | | | | This change removes warning about using new Array() as constructor with the snippet in Qt Creator. Task-number: QTBUG-65891 Change-Id: I858601c21d703e4226db4f4c4d9ef52ec6418ed8 Reviewed-by: Nico Vertriest <nico.vertriest@qt.io> Reviewed-by: Mitch Curtis <mitch.curtis@qt.io> (cherry picked from commit a30138cae86268d71415bc88cfba157977f4a442)
* Doc: Extend ItemSelectionModel QML documentationPaul Wicking2018-07-311-4/+59
| | | | | | | | | | This change provides a bare minimum documentation for the ItemSelectionModel QML type. Task-number: QTBUG-58090 Change-Id: I0e232f8e05e7629d6f573f8dce21154d0ec307e5 Reviewed-by: Topi Reiniö <topi.reinio@qt.io> (cherry picked from commit 9ee7b52f3234d12c7617ddf716f54e1923d3cded)
* Fix truncating of doubles to integersSimon Hausmann2018-06-213-0/+41
| | | | | | | | | | | | | | | | | When analyzing the data flow and determining that a double is or'ed with 0, we correctly assume a cast to an integer and insert a convert instruction in the IR instead of the actual bitwise or.. The JIT implements that, but the interpreter doesn't, resulting in incorrect values. This problem is specific to the SSA optimizer and the underlying isel requirements in 5.9 and does not apply to the design of the code generator in 5.11 and onwards. Task-number: QTBUG-68640 Change-Id: I954086b118b022bc6a8e00f17517c2d378825613 Reviewed-by: Lars Knoll <lars.knoll@qt.io> Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
* Fix out of bounds stack access with the byte code interpreterSimon Hausmann2018-06-211-1/+1
| | | | | | | | | | | | | | | | | | | | | | | When generating the byte code, we use code like this to initialize the specialized instructions: Instruction::Binop op; op.foo = 1; ... to conveniently initialize the fields and then use the addInstruction template specialization to copy the non-generic bits (for example binop specific parameters) across into the Instr union. We copy InstrMeta<InstrType>::Size bits, which includes the alignment padding applied via MOTH_INSTR_ALIGN_MASK. However the source data type (Instruction::Binop in the above example) does not include fields and therefore the memcpy will read more bytes off the stack than were allocated. Discovered via ASAN during the investigation of QTBUG-68640. The issue is specific to Qt 5.9, it does not apply to the code base of 5.11 or later, where the byte code is strictly integer organized. Change-Id: Ia12b75f6c1ec6c9d77ee6e7c4253c1e18c46291c Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* On network redirects, update finalUrl, not urlUlf Hermann2018-05-251-8/+11
| | | | | | | | | | | | | | | | We want all further imports to be relative to the redirected URL, not the base one. Note that this will incorporate any prior URL interceptions into the final URL if a redirect happens. We don't really want this to happen because the result of interception is not meant to be the base for further URL lookup. However, as interception occurs before redirection, this is unavoidable. Don't use URL interceptors on remote URLs. Task-number: QTBUG-67882 Change-Id: I6693d14c8af8212dda9954d0bd0293c3c85441ef (cherry picked from commit cda2680d801acce4e221b23e88d9b3c5504f86e8) Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix QML context leak with visual data model and list property modelsSimon Hausmann2018-05-073-6/+14
| | | | | | | | | | | | | | | | | | | | | | | When using the VDM or QML list properties as models, the delegate model injects an intermediate QQmlContext that provides access to the properties of the exposed QObject as context properties. Before commit e22b624d9ab1f36021adb9cdbfa9b37054282bb8, that context was marked to be owned by the parent QQmlContext. When the reference counting was introduced, that parent became referenced from the cacheItem (DelegateModelItem), but that intermediate QQmlContext became floating and was leaked. This can be observed by running the objectListModel test of tst_qquickvisualdatamodel with detect_leaks=1 in ASAN_OPTIONS. The leak is fixed by re-introducing the exceptional case of a parent holding a strong reference to the child, in just this one case. Change-Id: Iabc26990d39757b0abe0cddf69e76e88e40fba40 Reviewed-by: Lars Knoll <lars.knoll@qt.io> Reviewed-by: Michael Brasser <michael.brasser@live.com> (cherry picked from commit 01df9e5f46fd05a80f8f6fcaa91204e6184ded6f)
* Fix crash when calling QQmlEngine::clearComponentCache()Simon Hausmann2018-04-106-61/+71
| | | | | | | | | | | | | | | | | We must protect various resources in the type loader with our existing lock. The QQmlTypeLoaderQmldirContent is now value based, so that we can release the lock on the shared cache early. Copying it involves adjusting the refcount of the QHash and QString instances in the QQmlDirParser. The safety of this was verified with a TSAN build and the example supplied in the task. It crashed reliably with TASN errors first and with this patch it runs without errors. Task-number: QTBUG-41465 Change-Id: I616843c4b8bdfd65d1277d4faa8cb884d8e77df8 Reviewed-by: Lars Knoll <lars.knoll@qt.io> (cherry picked from commit a3ad52526f79c1528f170c8affe5af00b68ca61d)
* Fix out of bounds reads in Array.concatLars Knoll2018-04-031-1/+1
| | | | | | | | | | | In some cases, when our simple array data had an offset and data would wrap around, ArrayData::append would write out of bounds data into the new array, leading to crashes. Task-number: QTBUG-51581 Change-Id: I55172542ef0b94d263cfc9a17d7ca49ec6c3a565 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> (cherry picked from commit f495d4b660107536d0a67ba48e88550278f13893)
* Handle function expressions as signal handlersErik Verbruggen2018-03-288-5/+41
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | There are two ways to use function expressions on the right-hand side of bindings: property var somethingPressed somethingPressed: function() { /* ..press something else.. */ } signal buttonPressed onButtonPressed: function() { /* ..handle buttonPress.. */ } In the former case, it declares a property that holds a function. So on initialization, the right-hand side of the binding returns a closure that gets assigned to the property 'somethingPressed'. In the latter case, the signal handler is explicitly marked as a function for clarity. So, the handler should not be returning the closure, but the handler should *be* the closure. In general, it is not possible to detect if the left-hand side is a property or a signal handler when generating QML cache files ahead of time. So for this case, we mark the function as only returning a closure. Then when instantiating the object, we check if it is a signal handler, and if the handler is marked as only returning a closure. If so, we set that closure to be the signal handler. This patch also handles function expressions in SignalTransition and function expressions as signal handlers. Task-number: QTBUG-57043 Task-number: QTBUG-50328 Task-number: QTBUG-50328 (cherry picked from commit 22b13921f8067f8a93164875a4ad59bed85b0400) (cherry picked from commit dc4d6293f9473c0f03c570430d08867d2d01c6e2) (cherry picked from commit 21301c1dbb00f4a2cd991e520423ed039b297ffb) Change-Id: I3008ddd847e30b7d0adef07344a326f84d85f1ba Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix JITted code for jump strict-not-equal undefined on 32bitErik Verbruggen2018-03-151-10/+27
| | | | | | | | | | | | | | The generated code for jump-on-strict-not-equal-undefined used the same logic (but with inverted conditions) as the equal case. For equality, one can jump to else if the value parts are not the same. So, for not-equal, if the value parts are the same, it would jump to the else block if they are the same. Meaning, an encoded int value of 0 (which is strict-not-equal to undefined) would end up being evaluated as equal. Task-number: QTBUG-66832 Change-Id: I5c6b8e9b11be53ae21a7164e0a1e0cbfd204f401 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix issue with bindings to aliases that cannot yet be resolvedErik Verbruggen2018-03-155-60/+100
| | | | | | | | | | | | | | | When an alias points to a child object which has not yet been initialized, it's id won't have been registered yet, so setting up a binding to it will result in a crash. The fix is: when setting a binding target fails, and its target property is an alias, queue them until all bindings have been set up, and try again. Task-number: QTBUG-57041 Change-Id: I4dc5a6d25c0a32fed9fd952c955e2006c76be45a Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> (cherry picked from commit aa94c6c0469b0595f483f13ac88459f0035deef9)
* Fix issue with allocating huge objects in the memory managerLars Knoll2018-03-082-3/+24
| | | | | | | | | | | | | | | | We shouldn't allocate objects that are larger than the size of a standard memory segment through the chunk allocator, as this can lead to problems when freeing the segment and then re-using it again. Instead allocate a private MemorySegment for these objects, and free it when the object gets garbage collected. Task-number: QTBUG-66732 Change-Id: Ic24ff65d204977f313ab0adaf7a8132883e525f0 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> (cherry picked from commit c99abf1851bbbfcec11eb77173df620746940ab0) Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
* Allow setting values in value type group properties in "on" assignmentsSimon Hausmann2018-02-261-1/+7
| | | | | | | | | | | | | | Assigning to a group property inside a property value source or interceptor as part of an "on assignment" is perfectly valid. That is because while "color" is a value type property, the on assignment means we're actually setting easing.type (in the example and test) on the property value source, not the color, and that one is a QObject. The same goes for interceptors. Change-Id: I505a658977a578894d6dfb00bf5c65b41e42b12f Task-number: QTBUG-56600 Reviewed-by: Michael Brasser <michael.brasser@live.com> (cherry picked from commit 2659c308792967322564b5088e0e21bb371e0283)
* Fix ListModel.get(idx) == ListModel.get(idx)Simon Hausmann2018-02-232-9/+16
| | | | | | | | | | | | | | | | | | | This is a regression introduced with commit 4876ea6a18ccdfd72014582aa5d50ab9f6b6ec9e. Where we previously always returned the same JS object, we would afterwards return a new JS object for every invocation, which breaks reference comparison. As we store the JS wrapper for the list element in the QQmlData->jsWrapper we can avoid repeated allocations. In order for that wrapper to keep working after modifications (insertion, etc.) to the list model, we have to replace the static element index with a reference to the node model meta-object, which also has an element index that however is kept up-to-date by the list model itself. Change-Id: I4368de6b6d86687fe96fbf73bd60b80b69d7b058 Task-number: QTBUG-52017 Reviewed-by: Michael Brasser <michael.brasser@live.com> (cherry picked from commit 44a89492b49f23a975377795dbb7a48916cb5081) Reviewed-by: Mitch Curtis <mitch.curtis@qt.io>
* Remove superfluous assert when traversing IRErik Verbruggen2018-02-221-1/+0
| | | | | | | | | | | | When accessing/calling a property on an object, it is possible (and perfectly fine) for that object to be a constant value. I.e. Undefined. All code handling such a call do handle constants correctly. Note: this is a 5.9 specific change, because 5.11 got rid of this code. Task-number: QTBUG-66027 Change-Id: Ied9d0c9c8f8bf958f8634f7be196900b3ea64861 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix crash when changing from a simple to a sparse arrayLars Knoll2018-02-162-1/+5
| | | | | | | | | | | | | | After that change, if we ran out of slots in the freeList, the last entry would point to the first Value in the value array, not indicating that we ran out of free slots. Conflicts: src/qml/jsruntime/qv4sparsearray_p.h Task-number: QTBUG-65828 Change-Id: I3e57bb7a0c2dc29172a485a6ea957b6ab5ac962e (cherry picked from commit 16ca5eab9bdd31774dc8e657f217e044640eecff) Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Correctly set this object when calling scope/context functionsErik Verbruggen2018-02-151-0/+4
| | | | | | | | | | | | When a function is called that is in a QML scope or a QML context, set the 'this' object to the QML scope. Note: this patch is 5.9 specific. 5.11 has a similair issue, but the implementation is quite different, so that needs a separate fix. Task-number: QTBUG-59357 Change-Id: Ia78e012d413c40a094e957f4020502cd055ac286 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Clear the last value when removing propertiesErik Verbruggen2018-02-141-38/+11
| | | | | | | | | | | | | | | | | | | When we remove a property from an object, we shrink the used entries by one (or 2 when an accessor is defined) by moving subsequent entries "down" over the removed entry. We also have to set the last entry (or 2) to Undefined, otherwise any heap objects referenced there would be retained. This cherry-pick into 5.9 also fixes an issue where the MemberData was accessed out-of-bound: some entries stored in memory after the array were copied in, resulting in invalid pointers, leading to a crash whenever the garbage collector would run. Task-number: QTBUG-66090 Change-Id: I75905fafd0d88891820d894a869b9714bc9807e0 Reviewed-by: Lars Knoll <lars.knoll@qt.io> (cherry picked from commit 9e32901835de1c6729ac8bf228148c1e03c4c5a5) Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Use only cache path to cache .qmlc files on AndroidBogDan Vatra2018-02-131-0/+2
| | | | | | | Task-number: QTBUG-58223 Change-Id: Ibc599ac2e62aa60405af0022c7f5bab6eac3e3c4 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> (cherry picked from commit ff08272245c099cadd433c8b5d4f98301f5e585b)
* Fix memory leak with ListModel.getSimon Hausmann2018-02-131-3/+7
| | | | | | | | | | | | | | | | | | | This is a regression introduced with commit 3cc589c98390992e3ee8a7970dc2913ea857d623, which in turn fixed a leak with QV4::QObjectWrapper objects. Unfortunately the allocate() call into the persistent (weak) value storage in the list model introduced a leak of the weak value itself. This is fixed by replacing the free standing weak value allocation with the use of the existing jsWrapper weak value in the declarative data (QQmlData). That weak value is freed property in the destroy() method of the QV4::QObjectWRapper. The extra QQmlData allocation is hidden behind a unified allocation, similar to what we do in void QQmlType::create(QObject **, void **, size_t) const. Task-number: QTBUG-66189 Change-Id: I5351e3e484542709a6b210e84aa19b14d28e11ad Reviewed-by: Lars Knoll <lars.knoll@qt.io> (cherry picked from commit 22d43f74e264626d0c28654c42c91839f9de45b5) Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
* Prevent huge arrays to overflow the JS stack during GCErik Verbruggen2018-02-131-2/+23
| | | | | | | | | | | | | | The JS stack is used as a worklist while marking in order to prevent recursion overflowing the C stack. Now if all contents of an array are pushed onto the stack, it can easily cause an overflow. To prevent this, drain the stack periodically. This is fix that should not go into 5.11, as it's already fixed there by using a ValueArray that will have this exact behavior. Change-Id: Id5bd28879f6ef0265344d9a70c25f6c66b067309 Task-number: QTBUG-62087 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Revert "Restore a temporary QQmlInstanceModel::object() overload"J-P Nurmi2018-02-121-1/+0
| | | | | | | | | This reverts commit 823428d994b0ec0f6b85288d74554660a51b2406. No more needed after qt3d commit 78f77f80057b1e77c3f47d52de3e0b3f0c5d8d6e that took the new API into use. Change-Id: I4acb707cf363a625ae8e90c4560a0d5140cc4011 Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
* Fix memory leak with JS importsSimon Hausmann2018-02-093-1/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Strictly speaking this is a regression introduced with commit e22b624d9ab1f36021adb9cdbfa9b37054282bb8, making the QQmlContextData objects reference counted, especially from the V4 QML context wrapper objects. That change (correct as it is) introduced an accidental circular dependency in the simple scenario of importing a .js file in a .qml file: Each time the type in the .qml file is instantiated, we create a dedicated QQmlContextData for the .js file. If the .js file has no imports itself, that new context will get the same ctx->importedScripts JS array as the QML context of the .qml file. That is a strong reference via QV4::PersistentValue. That array in turn contains the QV4::QmlContextWrapper that belongs to the imported script, which in turn holds a strong reference (via refcount) to the script's context. This patch breaks the circular reference when we perform context invalidation, as the least intrusive measure. For the auto-test to work, we must also clear the qmlContext persistent of the QV4::Script that's used to evaluate the .js file. In subsequent imports that persistent will be initialized to new values, so it will only hold a strong reference to the last import, but strictly speaking that is still a leak - hence also part of this fix. Change-Id: I3e543c946e5e683425072dc3df7e49ca0e0c0215 Task-number: QTBUG-66189 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Document how to work with arrays using QJSValueMitch Curtis2018-02-081-5/+81
| | | | | | | | | | | | | | | - Mention (in the detailed description) that Array is indeed supported. - Provide examples for getting and setting individual array elements, and how to read the length of the array. - Properly document the property() and setProperty() overloads that take an index. - Link to the overloads where it makes sense. These changes make the intended workflow for using arrays much more obvious. Change-Id: I4657a7b1e2b4c2977120ee8e345ee9ae7d2bbc2d Reviewed-by: Topi Reiniö <topi.reinio@qt.io>
* Fix memory leak with deferred propertiesSimon Hausmann2018-02-061-1/+2
| | | | | | | | | | | | | | | | | | | | | | This is a regression introduced with commit 3b6eeee177b64eebe240d51be0c7bb5f031471d8 in the 5.9 branch. When constructing an object with deferred properties and not running qmlExecuteDeferred, then the deferred data would never get deleted because the bindings list remains non-empty and we would leak the deferred data as well as the entire compilation unit behind it. This happens for example when declaring when instantiating a QML file with states: states: [ State { ... }, State { ... }, ... } Unless every state is entered, its deferred changes property is never applied (via qmlExecuteDeferred) and thus the defer data is leaked. Task-number: QTBUG-66189 Change-Id: I1b2119c601d1e0ab4e37f53d4cf2f569586ee883 Reviewed-by: J-P Nurmi <jpnurmi@qt.io> Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Clean up manual refcounting of compilation unitsSimon Hausmann2018-02-064-13/+23
| | | | | | | | | | | | | | | | Replace manual use in QQmlData and QQmlData::DeferredData with QQmlRefPointer. Due to forward declaration trouble this required declaring a non-inline constructor/destructor for QQmlData and DeferedData and disabling copying, so that not every C++ compilation unit including qqmldata_p.h needs to instantiate the QQmlRefPointer destructor and thus know whether QV4::CompiledData::CompilationUnit has release(), etc. The out-of-line declarations however should not have any negative impact as the only call sites are within qqmlengine.cpp, too. Change-Id: I2e8295cb0d7f876a5d7d18765dbac285184e6c99 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Add back the optional heaptrack trace pointsAndras Mantia2018-02-061-0/+15
| | | | | | | | | | | This approach tracks object allocations only, when slots from already allocated memory segment are used. Change-Id: I514b974d7580c1236264bec96dc1abe594585e86 Reviewed-by: Milian Wolff <milian.wolff@kdab.com> Reviewed-by: Ulf Hermann <ulf.hermann@qt.io> Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Document that Q_GADGET classes cannot be used with newQMetaObject()Mitch Curtis2018-02-052-5/+39
| | | | | | Task-number: QTBUG-62007 Change-Id: I63d5a57163b36bc8629930e1cda8d5afa1e77d15 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* QML: Collapse all NaNs into one single (encoded) NaNErik Verbruggen2018-02-051-0/+4
| | | | | | | | | | The idea of NaN boxing is to use one single NaN as a "true" NaN, and all others as a boxed value. So when encoding some NaN, be sure to use that one "true" NaN. Otherwise, it will be interpreted as an encoded value. Task-number: QTBUG-65998 Change-Id: Ia6e4641be180f3d626c40a57b473f181358e04db Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Doc: add C++11 lambda examples for qmlRegisterSingletonType()J-P Nurmi2018-02-021-0/+25
| | | | | | Change-Id: I444137fd10041781df232447b8e2bf712582f079 Reviewed-by: Mitch Curtis <mitch.curtis@qt.io> Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
* Fix memory leak with QtQuick compiler generated filesSimon Hausmann2018-02-021-0/+3
| | | | | | | | | | | | | | | | When for the QQC code path we do QML type re-compilation, we allocate a new QV4::CompiledData::Unit. We must make sure that this dynamically allocated memory is released in QV4::CompiledData::CompilationUnit's destructor, by ensuring that the StaticData flag is not set. This isn't directly applicable to the ahead-of-time generated cache file unit data as they will always be re-generated (and thus the unsetting of StaticData at the end of createCompilationUnit::createUnitData()), but I've added a test-case nevertheless to ensure the correct engine behavior. Change-Id: I16973d7989567892bf8bf9dd6214bf293055d260 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Fix memory leak with value typesSimon Hausmann2018-02-021-0/+2
| | | | | | | | | | Commit 3b14e2ffdd8eb4b7f7f4508768b75f2acc399370 replaced the QQmlRefPointer<QQmlPropertyCache> with a raw QQmlPropertyCache pointer and added a V4_NEEDS_DESTROY tag. However unfortunately the destroy() method in the heap class does not decrease the reference count. Change-Id: I90a8c56cd638592b67aae7041fbb57c879c4146c Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Fix dead lock / race in QML type loader when importing pluginsSimon Hausmann2018-02-021-66/+84
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When importing modules - in the QML loader thread - with plugins we keep globally track of the Qt plugins that we have loaded that contain QML modules, to ensure that we don't call the engine-independent registerTypes() function on the plugin multiple times. After registerTypes() we may also call initializeEngine() on the plugin for the engine-specific initialization, which - as a QQmlEngine is provided as parameter - must happen in the gui thread. For that we issue a thread-blocking call that waits until the gui thread has woken up and processed the event/call. During that time the global plugin lock is held by that QML loader thread. If meanwhile the gui thread instantiates a second QQmlEngine and attempts to issue a synchronous type compilation (using QQmlComponent::CompilationMode::PreferSynchronous), then gui thread is blocking and waiting for its own QML loader thread to complete the type compilation, which may involve processing an import that requires loading a plugin. Now this second QML loader thread is blocked by trying to acquire the global plugin registry lock (qmlEnginePluginsWithRegisteredTypes()->mutex) in qqmlimports.cpp. Now the first QML loader thread is blocked because the gui thread is not processing the call events for the first engine. The gui thread is blocked waiting for the second QML loader thread, which in turn is stuck trying to acquire the lock held by the first QML loader thread. The provided test case triggers this scenario, although through a slightly different way. It's not possible to wait in the gui thread for the plugin lock to be held in a loader thread via the registerTypes callback, as that also acquires the QQmlMetaType lock that will interfere with the test-case. However the same plugin lock issue appears when the first QML engine is located in a different thread altogether. In that case the dispatch to the engine thread /works/, but it won't be the gui thread but instead the secondary helper thread of the test case that will sit in our initializeEngine() callback. This bug was spotted in production customer code with backtraces pointing into the three locations described above: One QML loader thread blocking on a call to the gui thread, the gui thread blocking on a second QML loader thread and that one blocking on acquisition of the plugin lock held by the first. Fortunately it is not necessary to hold on to the global plugin lock when doing the engine specific initialization. That allows the second QML loader thread to complete its work and finally resume the GUI thread's event loop. Change-Id: If757b3fc9b473f42b266427e55d7a1572b937515 Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
* Improve persistent/weak value allocation performanceSimon Hausmann2018-01-311-10/+24
| | | | | | | | | | | | | | | | | | | When the persistent storage becomes fragmented, we would find the page with a hole in it, but we wouldn't put it to the front of the page list. So upon the next allocation we would begin iterating through firstPage again. This wasn't quite visible in callgrind as overall not many instructions were executed, but in perf this function showed up as hotspot because the search for free pages ends up with a lot of cache misses. In delegates_item_states.qml this was about ~7% of measured cycles with perf. Change-Id: I2bfa337ea9be14d1321756963c72d31336790a0a Done-with: Erik Task-number: QTBUG-65708 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Disallow registration of types beginning with lowercase lettersMitch Curtis2018-01-313-0/+13
| | | | | | | | | | | | | | Allowing types with lowercase names causes ambiguity, as can be seen in QTBUG-43567 and the comment in IRBuilder::visit(), which explains that "the grammar can't distinguish between two different definitions" whose only difference is casing of the first letter. - Prevent registration (return -1 with e.g. qmlRegisterType()) when a type name doesn't begin with an uppercase letter. - Document the uppercase type name rule in more places. Change-Id: I4e522c65990f418eaafa45a256e3cb07a3e01ba4 Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
* Fix exposure of -1 as enum value in QML exposed C++ singletonsSimon Hausmann2018-01-311-10/+12
| | | | | | | | | | | | | | | | | When a C++ singleton has an enum with the value -1, we would expose that value correctly when taking the accelerated property access code path in the optimizer, but when going through the slower QQmlTypeWrapper we would return undefined. This turned out to be a silly logic error that assumed that -1 is not a valid value for an enum and instead indicates an enum value not present. [ChangeLog][Qml] Fix -1 as enum value in QML exposed C++ singletons showing up as undefined. Task-number: QTBUG-66067 Change-Id: Ib66dad7a4b59822b2c40ad6bd9af4b72469582e9 Reviewed-by: Lars Knoll <lars.knoll@qt.io> Reviewed-by: Michael Brasser <michael.brasser@live.com>
* Fix segfault when alias target refers to lowercase-named typeMitch Curtis2018-01-272-10/+23
| | | | | | | | | Create an error via QQmlCompileError and return it instead of asserting. Task-number: QTBUG-43567 Change-Id: I0c0741943d30516379eff5f44ed8618a0f0116a4 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix namespace builds with -no-qml-debugJoni Poikelin2018-01-231-0/+5
| | | | | | Task-number: QTBUG-65924 Change-Id: I47b3afbb8235900156c814874d5ae2250cf13da8 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Doc: Do make 'Licenses and Attributions' valid for multiple Qt versionsKai Koehne2018-01-221-2/+3
| | | | | | | | | | | The Qt documentation is supposed to be valid for older Qt versions too. Anyhow, the generated attributions are only valid for the exact version the documentation was generated from, so make this explicit. Also mention since when the libraries are under LGPL3/GPL3. Change-Id: Iec8f67e5e43be456cc77283ca6d2a7ebe142f501 Reviewed-by: Leena Miettinen <riitta-leena.miettinen@qt.io>
* Avoid repeated calls into thread local storage to get the animation timerLars Knoll2018-01-222-43/+41
| | | | | | | | | | Instead hold a direct pointer to the animation timer and make it's methods non static. Change-Id: I6382fd2a1c02464ddb573f0210a14c603fd932db Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> Reviewed-by: J-P Nurmi <jpnurmi@qt.io> Reviewed-by: Robin Burchell <robin.burchell@crimson.no>
* Fix QQmlDelegateModel::object documentationPaolo Angelelli2018-01-181-5/+5
| | | | | | | It is talking about item() but such a method doesn't exist anymore. Change-Id: I1935d8b9e88b27a9db1122545a2a82a42d827671 Reviewed-by: Michael Brasser <michael.brasser@live.com>
* Allow for currentQmlContext to be inlinedErik Verbruggen2018-01-182-6/+12
| | | | | Change-Id: Ic2a98a3a4b4362036222df05a92c0bed633c1d1c Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Optimize QQmlProperty constructor for the common property caseSimon Hausmann2018-01-171-66/+70
| | | | | | | | | | | | | The common case is that QQmlProperty is constructed on the property of an object, not a group property. Therefore we should do the QVector<QStringRef> split on the property name by '.' only if a dot exists, and can avoid the allocation and deallocation of the vector. Shaves off ~1.2% off delegates_item_states.qml. Task-number: QTBUG-65708 Change-Id: Iffbde176e616beec0ae0a47216360558adc793ee Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Speed up PropertyChange state applicationSimon Hausmann2018-01-172-0/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Every time we decode a potential binding of a PropertyChanges{} object, we call qmlContext(this) and we go through a full QQmlProperty construction (which involves property name decoding by dots and property lookups), just to determine if we're doing a binding on a property or a signal. QQmlProperty::isSignalProperty() will only return true if the property is valid and if it's a "function" type. The QQmlProperty constructor on the other hand only constructs a valid regular property if it's _not_ a function type and a signal property _has_ to start with "on" followed by an upper case character. We can copy this shortcut out into decodeBinding() to avoid the QQmlProperty construction in the common case of plain property bindings. This is also legit in the scope of group properties, as signal bindings on group properties are not supported (we always use the state's target object for signal lookup, never the group object). In addition, avoid creating a public QQmlContext for the PropertyChange object by allowing for the construction of the QQmlProperty object via the QQmlContextData, as that's the only data structure we really need. These two changes used to be separate, but they need to go together to keep the tests passing, as the property validation and warning issuing is now moved from decodeBinding() into ::actions() itself. Shaves off 1.5% off delegates_item_states.qml Task-number: QTBUG-65708 Change-Id: I32a17d815bd3495a907a51068a971eb7cb69c6ef Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Prevent a QVector detachErik Verbruggen2018-01-171-1/+1
| | | | | Change-Id: Ibda07de7a83cf9a1434532c485583b8b49b0a605 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Revert "Allow canceling incubation"J-P Nurmi2018-01-176-49/+0
| | | | | | | | | | | This reverts commit ca6b787a01ea289bd5c2a3e4ff3c7442a4ff58fc. This internal API was added as a workaround for Qt Quick Controls 2. It is no longer needed now that Qt Quick Controls 2 are using deferred execution. Task-number: QTBUG-50992 Change-Id: Iaddf22460f091743e1a68acd16813a28f3e82ecb Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Use a more optimized lookup for global propertiesLars Knoll2018-01-177-14/+23
| | | | | | | | | | | | | | Force the use of a global lookup if we know that the property can and will be found in the global object. This is possible, as the global object is frozen in QML mode and can't be overwritten. Shaves of .5% on the delegates_item_states benchmark, and will significantly speed up all accesses to e.g. the Math object. Change-Id: Ia1e248781a13ebaeb8bc43652e53a6fdde336d0d Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Have more bindingBits available inlineLars Knoll2018-01-172-43/+40
| | | | | | | | | | Many QML items have more than 32 properties, so we ended up malloc'ing the binding bit table on the heap quite often. Extending the inline data to be able to accommodate for up to 64 properties fixes that. Change-Id: I90a42d601a5406ffacf2506f1957b0c2080bbb7b Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Speed up V4 QML context creationSimon Hausmann2018-01-163-0/+6
| | | | | | | | | | | | After commit cae7975a036352ca4bbcf1381a445362f8e01367 the vtable became part of the internal class, which meant that for allocating the V4::QmlContext and the QV4::QmlContextWrapper we had one additional IC transition upon allocation. We can avoid that by providing a default IC for both types. Task-number: QTBUG-65708 Change-Id: I22815a24a8389b59103309806f6ac4ca382405f0 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Optimizations for Repeater::clear() and ~QQmlItem()Lars Knoll2018-01-151-1/+1
| | | | | | | | | | | QQmlRepeater::clear() had quadratic complexity in the number of items, because the items where removed from the back. Fix this by searching the cache from the back as well as searching for child items to remove from the back. Change-Id: I92e491a8abf47cee9d382ef15cd2471f722fa6dd Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
generated by cgit v1.2.3 (git 2.25.1) at 2024-06-07 17:28:08 +0000