aboutsummaryrefslogtreecommitdiffstats
path: root/tests/auto/qml
Commit message (Collapse)AuthorAgeFilesLines
* Test whether the JIT is enabled on architectures/OSes as intendedMartin Storsjö2019-04-161-0/+22
| | | | | Change-Id: Ifdba6f08545e07b04cc7d9ace48ad0599c41229c Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
* QML: Allow fetchOrCreateTypeForUrl to report errors without qFatalErik Verbruggen2019-04-164-1/+6
| | | | | | | | Because qFatal will abort() the program. Task-number: QTBUG-71116 Change-Id: Ifd6be996cfbd6fff8e75ad2b26682c34f837ac88 Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
* QML: Check for (valid) compilation units while hashingErik Verbruggen2019-04-163-0/+6
| | | | | | | | | When an error occurred while creating a compilation unit, the pointer might be set to null. Subsequent use in hashing should check for this, and not use a nullptr. Change-Id: I62650917a740c9c1be29608285670153bed8703c Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
* Fix string replacement with invalid capturesUlf Hermann2019-04-151-0/+28
| | | | | | | | | | | | If we have a theoretically valid capture reference that just didn't capture anything in this match, we don't want to treat it as literal. Only capture references that clearly are outside the range of things we can possibly capture with this expression should be treated as literal strings. Change-Id: Iab0bf329d11a6b9e172aa662f11751d86cfc26a6 Fixes: QTBUG-75121 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Don't create value types for QImage and QPixmapUlf Hermann2019-04-121-0/+23
| | | | | | | | Those are "scarce" resources which need to be kept as QVariant. Fixes: QTBUG-74751 Change-Id: I28381e2a754ed4bbf4e409dc275f6288b64416cc Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Detect and reject cyclic aliasesUlf Hermann2019-04-093-0/+7
| | | | | | | | Previously those would result in infinite recursion. Fixes: QTBUG-74867 Change-Id: I6c0043b43e72fe7bc3a2a139ca600af2d5bca5ad Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Avoid INT_MIN % -1 and INT_MIN / -1Ulf Hermann2019-04-081-0/+17
| | | | | | | | | Those throw arithmetic exceptions as the result doesn't fit into an integer. Fixes: QTBUG-75030 Change-Id: Ibd978848f42cf1c9da1e4af2dc9d7da123ef8f5a Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Allow enums in QQmlType to be set up in two separate passesUlf Hermann2019-03-285-0/+110
| | | | | | | | | | | The baseMetaObject and the property cache can become available at different points in time. If we have initialized the enums before either of them is available we want to add the additional enums when the other one appears. Fixes: QTBUG-74677 Change-Id: I57276681a50b6c04181c6a29e736f2dc20632a0c Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Blacklist tst_qquickfolderlistmodel::nameFilters() on msvc2015/17Ulf Hermann2019-03-271-0/+3
| | | | | | Task-number: QTBUG-74739 Change-Id: Ib0f1d39cb5bc9864195b3c62d254fa38370faf0d Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
* Re-use existing IC when freezing propertiesUlf Hermann2019-03-271-0/+13
| | | | | | | | | | | | | There is no reason to start from the empty class in that case. Furthermore, if the properties are already frozen, starting from the empty class will walk the IC hierarchy to the current IC. However, if the garbage collector has removed the intermediate classes in the mean time, we end up at a new IC which is equivalent but not the same. Therefore, the freezing never terminates. Task-number: QTBUG-74190 Change-Id: Id544bd00d3b4b563fb06dfce0edd0385e1d32a6c Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Scope intermediate values when allocating objectsUlf Hermann2019-03-271-0/+13
| | | | | | | | | Otherwise the garbage collector might interfere and we might end up with dangling pointers in random places. Task-number: QTBUG-74190 Change-Id: I253c526ef930b8f0be14e96a42af8b66fda3b22d Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Trigger the garbage collector when allocating InternalClass objectsUlf Hermann2019-03-261-20/+34
| | | | | | | | | | | | | | | | | As we check the icAllocator's slots on shouldRunGC() we should also check shouldRunGC() when adding slots. Otherwise we might never run the GC when only allocating InternalClasses. In addition, account for the "unmanaged" size of the PropertyAttributes that are part of the InternalClass objects. Those can be large. In cases where an excessive number of large InternalClass objects is created the garbage collector is now invoked frequently, which costs a significant number of CPU cycles, but prevents the memory usage from growing indefinitely. Task-number: QTBUG-58559 Change-Id: Icf102cb6100f6dba212b8bffe1c178897880eda0 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Fix preview zoom problems on windowsTim Jenssen2019-03-221-6/+6
| | | | | | | | | | Apparently the window is moved to random places on multi-monitor setups on windows. Fixes: QDS-263 Change-Id: I21082c7031fefff3057074c147e82df7a88f4f78 Reviewed-by: Ulf Hermann <ulf.hermann@qt.io> Reviewed-by: Tim Jenssen <tim.jenssen@qt.io>
* Update tests/auto/qml/ecmascripttests/test262Jani Heikkinen2019-03-221-0/+0
| | | | | | | | Task-number: QTBUG-73454 Task-number: QTBUG-71209 Change-Id: I12925ce49cc18f4bb6908a5515fc476b32a222dc Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> (cherry picked from commit 896d49a4e1113e3eb4832b83920b0dfd76987259)
* Initialize TranslationData padding before writing cache filesUlf Hermann2019-03-2118-36/+103
| | | | | | | | Otherwise the resulting files differ subtly. Fixes: QTBUG-74532 Change-Id: I12b4f1ba6dda781d63ad50cce87861ba24582bf7 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* V4: Do not invert non-reflexive comparison binopsErik Verbruggen2019-03-211-0/+10
| | | | | | | | | This is only useful for the few (4) comparisons where we have specialized instructions, and it's very error-prone. Change-Id: I37efe94f54ba0adf393d9236df2d13aa6685eb46 Fixes: QTBUG-74476 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Add test for dynamic anchors to parents in PropertyChangesUlf Hermann2019-03-203-0/+41
| | | | | | Fixes: QTBUG-71042 Change-Id: I4d0927b0d6c9a4ccaaddfbfd6dd2a7841fbbc845 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* V4: Don't mark InternalClass::parent when garbage collectingUlf Hermann2019-03-201-0/+26
| | | | | | | | | | The parent pointer is only kept so that we can update the parent's transitions when removing a child. There is no need to keep the parents alive for the children. Fixes: QTBUG-58559 Change-Id: Ia28183966bde6d478ca030fe11195489925dfc13 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Implement dummy QML lookups for "global" variablesSimon Hausmann2019-03-204-5/+22
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | When resolving names in the context of QML bindings, we now direct runtime access to QQmlContextWrapper::resolveQmlPropertyLookupGetter. At the moment this does basically the same as Runtime::method_loadName, which we called earlier. However this now provides the opportunity to optimize lookups in the QML context in a central place. When performing a call on a scope or context object property, we also did not use a CallName() instruction - which would have gotten the thisObject wrong - but instead we use a dedicated CallScopeObjectProperty and CallContextObjectProperty instruction. These rely on identifying these properties at compile time, which goes away with lookups (and also doesn't work when using ahead-of-time compilation). Therefore the qml context property lookup is using a getPropertyAndBase style signature and Runtime::method_callQmlContextPropertyLookup uses that. For the tests to pass, some error expectations need adjusting. In particular the compile-time detection of write attempts to id objects is now delayed to the run-time. The old code path is still there and will be removed separately in the next commit (as it is massive). Task-number: QTBUG-69898 Change-Id: Iad1ff93d3758c4db984a7c2d003beee21ed2275c Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
* Raise timeout in QQmlPreview testUlf Hermann2019-03-201-1/+1
| | | | | | | Apparently on some platforms this can time out. Change-Id: I30cd274ff3127319c558b473b6ec7e7f5bc79c38 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Create import directory from intercepted URL, not orignial oneUlf Hermann2019-03-204-0/+42
| | | | | | | | | If the qmldir URL got intercepted, we should use the intercepted URL to get to the contents to be loaded. Fixes: QTBUG-73843 Change-Id: I51715575e767ed429a8237517f47196677409fe0 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Baseline JIT: Save accumulator in toInt32LhsAcc()Ulf Hermann2019-03-191-0/+12
| | | | | | | | | | toInt32LhsAcc convertes both the lhs and the accumulator to int32. If the accumulator is not saved, a GC run during the conversion of the lhs might trash its value. Fixes: QTBUG-74058 Change-Id: Ic42693061c7d483bb430d77bcc095de6ff9a6843 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Fix compilation for SLES 12Ulf Hermann2019-03-181-1/+1
| | | | | | | | Apparently only some compilers see through QPointer on connect() Fixes: QTBUG-74500 Change-Id: I4d0ca7c8ab8b7eae281deff14523a1628e669279 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Unify the JavaScript parsing recursion checksUlf Hermann2019-03-151-0/+5
| | | | | | | | | | | | | We only need to check in one central location and we can allow for more recursion. 4k recursions seem tolerable. A common default for stack sizes is 8MB. Each recursion step takes up to 1k stack space in debug mode. So, exhausting this would burn about half of the available stack size. We don't report the exact source location in this case as finding the source location may itself trigger a deep recursion. Fixes: QTBUG-74087 Change-Id: I43e6e20b322f6035c7136a6f381230ec285c30ae Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Don't keep raw pointers to SparseArrayNodeUlf Hermann2019-03-141-0/+22
| | | | | | | | | | | | The nodes are owned by the SparseArrayData and will be freed whenever an item is deleted from the array. Therefore, we have to look up the node for each iteration. This is slightly slower, but at least it doesn't crash. Fixes: QTBUG-74188 Change-Id: Id24324a8c83b00b3ad1212cdaabccabd6c8a999f Reviewed-by: Simon Hausmann <simon.hausmann@qt.io> Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Disable tail calls for function called with more arguments than formalsUlf Hermann2019-03-041-0/+13
| | | | | | | | | | | We cannot easily find the required stack space to store the extra arguments without adding a new stack frame. In principle it would be possible, but heavily recursing on such functions should be a rare problem. Change-Id: I1a53a6d29e37ce67aa7bd64acb7b1f41197e84c0 Fixes: QTBUG-72807 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* QAnimationGroupJob: Notify about removed children on clear()Ulf Hermann2019-02-261-2/+5
| | | | | | | | | | Derived classes, such as QSequentialAnimationGroupJob, might keep additional pointers to the children. Those need to be cleared, too. Therefore, use the regular removeAnimation() method for clearing. Fixes: QTBUG-73828 Change-Id: I64cc1fe4da59f10b29f27012b10f93b4289b6e5a Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* QQmlExtensionPlugin test: Don't fail when detecting old pluginUlf Hermann2019-02-261-1/+4
| | | | | | | | | | Some Qt modules still register plugins under the old IID. This makes the test fail if any of those have been compiled before. As the old IID is still recognized we don't really have to fail in this case. A warning should be enough. Change-Id: I409534d51c339ce361002866a50c47c721170b8b Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Unify and fix number to string conversion with radixUlf Hermann2019-02-251-0/+16
| | | | | | | | | | | Previously, the loop that generated the string could fail to terminate with certain numbers as input. Also, the algorithm was duplicated in two places. Change-Id: Ie2075148d931e7cfcedb5bcd23af61e2e8afc232 Fixes: QTBUG-73999 Reviewed-by: Robert Loehning <robert.loehning@qt.io> Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* V4: Rotate loop in ArrayPattern and eliminate "done" labelErik Verbruggen2019-02-251-0/+9
| | | | | | | | | | This prevents jumping over the resetting of the unwind handler when an exception occurs. (cherry-picked from commit 0282b89ec672e25a465a8e51bc74c7fd58a624b1) Fixes: QTBUG-73985 Change-Id: I4a4da815f54c13980d239e0492f9b013991cfbd5 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* Merge remote-tracking branch 'origin/5.12.2' into 5.12Qt Forward Merge Bot2019-02-222-30/+26
|\ | | | | | | Change-Id: If1c718780a48a3e818eb5a0c0907a31acb6c694d
| * Only warn about incompatible parameter types passed from QMLUlf Hermann2019-02-212-30/+26
| | | | | | | | | | | | | | | | | | Throwing a TypeError is too drastic. We need to properly deprecate this pattern before we actually enforce the compatibility. Fixes: QTBUG-73943 Change-Id: I00313ad7aed4021a7368fd014e2bfc6443b177e3 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* | QML: Pass type minor version when creating property dataUlf Hermann2019-02-213-0/+69
| | | | | | | | | | | | | | | | | | Depending on the type minor version recursive properties should be available or not. Check for that when resolving grouped properties. Fixes: QTBUG-33179 Change-Id: Id8f62befdc4a29d879710499e19d3d289bd18775 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* | QQmlPropertyCache: support setting an explicit metaObject revisionRichard Moe Gustavsen2019-02-211-0/+21
|/ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When creating a QQmlPropertyCache for a QMetaObject, there were currently no way to specify which revision to use. Normally this is not needed, since when creating property caches for types declared in QML, the correct revision would be filled in later, based on the import version found in the QML file. But sometimes we need to create a QQmlPropertyCache for a QMetaObject created in C++, that has no associated QML file and import version. And if that meta object has revisioned properties, we need to specify which revision of the meta object the cache should represent. Otherwise, the revision would just be 0, which means that revisoned properties would not be found by the V4 runtime later. As an example, QQmlAdaptorModel has a set of classes that wraps various models (QAIM, arrays, etc). When a new delegate item is created by a view, an instance of a model class will be created as well (from C++). This instance will be set as context object for the delegate item, enabling properties such as index, row, column and model roles. But since row and column should a revision (currently they don't), we need to be able to specify that the property cache should have a revision that matches the import version of the view. That way, we can ensure that they don't shadow any existing row and column properties that might exist in the application from before, and as such, cause regressions. This patch will add an extra argument to the constructor that lets you specify which revision of the QMetaObject to use. Task-number: QTBUG-70031 Change-Id: I1c245a0c8b6f071e35865966fedc97f2839cd2f3 Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
* Avoid unnecessary re-generation qml cache files in some circumstancesUlf Hermann2019-02-182-17/+46
| | | | | | | | | | | | The map of name IDs to resolved types so far is copied several times during compilation and different compile passes see different copies of it. Compile passes may add things to the map, and if they do that on copies that are inaccessible to other code, we get nondeterministic results. Furthermore all the copies and pointers are confusing and inefficient. Fixes: QTBUG-69340 Change-Id: I43ad3cbeeec34f90e05570eddc901fe8aa64c709 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* QML: Don't crash the parser on certain kinds of bad inputUlf Hermann2019-02-181-0/+10
| | | | | | Fixes: QTBUG-72734 Change-Id: I92146ec517c6a26a67e13830618e137f2a3021c2 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Check parameter types when invoking C++ functions from QMLUlf Hermann2019-02-172-26/+30
| | | | | | | | | | | | | | | We now check if the given parameters can be converted to the expected arguments of the function being invoked and throw a type error if not. Previously we would still invoke the method with random parameters. [ChangeLog][QtQml][Important Behavior Changes] The parameters passed to C++ functions from QML are now checked for compatibility with the expected arguments. If they cannot be converted, a type error is thrown in JavaScript and the function is not invoked. Fixes: QTBUG-73405 Change-Id: If16089510d314bb7cdb7d4db86478114c61281a8 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* PropertCache: Don't pass Q_GADGET value types as integersUlf Hermann2019-02-142-0/+51
| | | | | | | | We need to check for the IsGadget flag there. Fixes: QTBUG-73734 Change-Id: Ic4afd4215e6ed346bc40794d85397f0f262715e2 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Connect quit() and exit() signals with queued connectionsMichal Policht2019-02-137-13/+65
| | | | | | | | | | | | | | | | | | Class QQmlApplicationEngine connects QQmlApplicationEngine::quit() signal to QCoreApplication::quit() and QQmlApplicationEngine::exit() signal to QCoreApplication::exit(), but it does so with AutoConnection. This causes in some circumstances problems, which are described in Qt documentation (see QCoreApplication::exit()). This change modifies type of connections to queued connections. [ChangeLog][QtQml][QQmlApplicationEngine] QQmlApplicationEngine connects quit() and exit() signals with queued connections to avoid problems with AutoConnection, when connecting to QCoreApplication slots. Task-number: QTBUG-73649 Change-Id: Ib27738b5af2f879efee8862b1ca01613a2e8dc4e Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
* Don't optimize global lookups if fast QML lookups are disabledUlf Hermann2019-02-132-0/+38
| | | | | | | | | | | If fast QML lookups are disabled, we generally want to look up by string. If the name then happens to be a member of the global JavaScript object, we still don't want to directly access that, as the name could have been overridden in a deeper context. Fixes: QTBUG-73750 Change-Id: Id16110969123d91501064ba46bfad4c2a39e4650 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Improve error messageRainer Keller2019-02-042-2/+2
| | | | | | | Show more datails about what actually went wrong. Change-Id: I418a4d1f433bd4d440fc34e9a4932a9ea010b174 Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
* Correctly scope unwind handlers for try blocksLars Knoll2019-01-301-3/+44
| | | | | | | | | | | | | Make sure the unwind handler is always reset when leaving the try block. This exposes a couple of failures in the ECMAScript test suite that were before passing by pure luck. Task-number: QTBUG-72858 Change-Id: I014b1e37c2beff136ecd53a665a2f10933f7e12c Reviewed-by: Ulf Hermann <ulf.hermann@qt.io> Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
* Bring behavior of String.replace() in line with other enginesLars Knoll2019-01-251-0/+59
| | | | | | | | | | | | | | | | "x".replace("x", "$1") gives "$1" in both JSC and V8, as there are no captures that could be used as a replacement for $1. Implement the same behavior as it's the most logical thing to do (even though it's undefined according to the spec). Two digit captures ($nm) work in a way that they get applied if $nm captures exist. If there are less than nm but more than n captures available $n is replaced by the n'th capture and m is copied over verbatim. Change-Id: I8b5f576f2c42c8334859ab7854dcdf07104dd35b Fixes: QTBUG-73152 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* MemoryManager: Only clear weak values onceUlf Hermann2019-01-234-3/+90
| | | | | | | | | | | | We want to keep the weak values alive while the destruction callbacks are running, so that they can still access them. We set them to undefined later anyway because we expect the destruction callbacks to mess with the values. Therefore there is no point in also setting them in between. Fixes: QTBUG-72137 Change-Id: I83f70230f5b4ad2761c74770f975b14a5ca71f18 Reviewed-by: Lars Knoll <lars.knoll@qt.io>
* QQmlNotifier: Always keep the isNotifying flag when updating senderPtrUlf Hermann2019-01-232-0/+41
| | | | | | | | | | | When the sender gets deleted we still want to retain the flag that tells us that the notifier is currently active. Otherwise we can miss the error message about synchronously deleting objects while signal handlers are in progress. Task-number: QTBUG-73013 Change-Id: I8abba9b492327c15963d1875841c6822f345a89e Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Merge remote-tracking branch 'origin/5.12.1' into 5.12Qt Forward Merge Bot2019-01-213-0/+62
|\ | | | | | | Change-Id: I1e0bbc35807bd0f7f96694539d2fbec0f83ddb16
| * When matching RegExps catch JIT failuresUlf Hermann2019-01-093-0/+62
| | | | | | | | | | | | | | | | | | | | | | | | | | The Yarr JIT can generate code that fails to evaluate the RegExp at runtime. In that case we need to fall back to the interpreter. Also, don't needlessly cast the unsigned return value of RegExp::match to signed int before range-checking it. And fix some typos in the comments for the disassembler dumps. Fixes: QTBUG-72879 Change-Id: Ic8f80c076d6461d714816a9f66e1cac1d9b0c7a8 Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* | Annotate stack traces when frames are elided through tail callsErik Verbruggen2019-01-151-0/+16
| | | | | | | | | | | | Task-number: QTBUG-72407 Change-Id: I98b96852309fc783a945797185f666196513d24b Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
* | QQmlMetaType: Erase attached properties in dtorUlf Hermann2019-01-151-0/+39
|/ | | | | | | | | | Also, make it more obvious that the attachedPropertyIds are a static member of QQmlTypePrivate. Fixes: QTBUG-72972 Change-Id: If0a28e034dd46d7127993ed15aed11c7641d580e Reviewed-by: Harald Hvaal <harald.hvaal@gmail.com> Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Tests: Don't capture stack value by referenceUlf Hermann2019-01-031-7/+7
| | | | | | | | | We can actually not capture d at all there, as some of the items get deleted before the root object is deleted. Therefore, we need to iterate the children again on receiving the destroyed() signal. Change-Id: Iab7ebc3c731438a21b243284de7515530232828f Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>