aboutsummaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
...
* | Use qCpuHasFeature because the check can be optimized out of existenceThiago Macieira2013-12-101-1/+1
| | | | | | | | | | | | | | | | | | qCpuHasFeature also checks the compiler flags of this particular build. If the feature was enabled by the compiler in this .cpp, then no runtime check is performed and the test becomes an unconditional true. Change-Id: Ibb7d333e59a0f4ed06b7d0056547177fd69658a0 Reviewed-by: Simon Hausmann <[email protected]>
* | Mark item as dirty when nodes are cleanedMichael Brasser2013-12-102-0/+4
| | | | | | | | | | | | | | | | Otherwise they may not be recreated in the correct configuration. Change-Id: Ib103b1874d7f104680eae7b6f43f722b5fa93bf2 Done-with: Aaron Kennedy <[email protected]> Reviewed-by: Gunnar Sletta <[email protected]>
* | Fix broken Maroon game / regression in PropertyChanges {} elementSimon Hausmann2013-12-105-23/+52
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Commit 0aadcf8077840068eb182269e9ed9c31ad12f45e that pre-compiles the expressions in PropertyChanges {} introduced a regression in where the evaluation context was incorrect and thus bindings would not be able to access the correct properties. For example PropertyChanges { target: someObject y: height / 2 } Here height should be looked up in the context of "someObject", not of the PropertyChanges element. This patch introduces an auto-test that verifies that the lookup context is correct and fixes the bug by disabling accelerated compile time property lookups for binding expressions that are requested from a custom parser. Change-Id: I5cb607d07211b453ddfc9928ccbf5f9ecec85575 Reviewed-by: Lars Knoll <[email protected]>
* | QSGRenderer: sanity check attribute regs only when QSG_SANITY_CHECK=1J-P Nurmi2013-12-101-4/+2
| | | | | | | | | | | | | | | | | | | | The sanity check that attribute registers are disabled that was always done in debug mode took roughly 16ms on my machine and caused a trivial QML animation to eat 100% CPU. Task-number: QTBUG-35443 Change-Id: I8ba2a80db341d17e08216b3cad1678dd59b9b1a5 Reviewed-by: Gunnar Sletta <[email protected]>
* | V4: change variable renumbering algorithm from recursive to iterative.Erik Verbruggen2013-12-101-89/+237
| | | | | | | | | | | | | | | | | | Replace the recursive calls and subsequent clean-ups by pushing actions on a to-do stack, and processing that stack in a loop. Change-Id: I83536e88d400592b6e9f5fda3d795e41711a131a Reviewed-by: Simon Hausmann <[email protected]> Reviewed-by: Fawzi Mohamed <[email protected]>
* | Windows: Use Shell API for checking file case correctness.Friedemann Kleint2013-12-102-8/+27
| | | | | | | | | | | | | | | | | | The old method of converting to short 8.3 name and back does not work for drives where this is disabled. Change-Id: Ia0a46331a31eeb61578c31ba063a80665d5fc25c Reviewed-by: Michael Brasser <[email protected]> Reviewed-by: Alan Alpert <[email protected]>
* | Correctly profile the size of loaded pixmapsUlf Hermann2013-12-092-9/+5
| | | | | | | | | | | | | | | | | | | | Only set the sizes when they're known, prefer request size to implicit size (as the request size is what ends up in the cache), and don't set the size twice. Task-number: QTBUG-35337 Change-Id: Ie516a1cae2d9050f61362ee99cf8a6a9dd8ea3bb Reviewed-by: Gunnar Sletta <[email protected]>
* | Remove unused function isUnicodeNonCharacterThiago Macieira2013-12-081-14/+0
| | | | | | | | | | | | | | | | Found by ICC 14.0: qhashedstring.cpp(199): warning #177: function "isUnicodeNonCharacter" was declared but never referenced Change-Id: I62b113e41197dac12f73db8347e22c825e404627 Reviewed-by: Simon Hausmann <[email protected]>
* | Fix crash in QQuickLoader when source component is garbage collectedSimon Hausmann2013-12-062-0/+6
| | | | | | | | | | | | | | | | | | | | It may happen that the QQuickLoader is the last entity left in the system holding a reference to the QQmlComponent *sourceComponent. We have to let the garbage collector know about that by keeping a persistent value for it. Task-number: QTBUG-35334 Change-Id: I715864440378fd9dd4f2d5ef8ff2f171c81ed7ef Reviewed-by: Lars Knoll <[email protected]>
* | Profile a pixmap start event also when loading is asynchronous.Ulf Hermann2013-12-061-0/+3
| | | | | | | | | | | | | | | | | | | | | | If those events aren't profiled we get unmatched events for finished loading and errors. Task-number: QTBUG-35337 Change-Id: I132638edf5625566e37ee06453484f80b2adb74a Reviewed-by: Ulf Hermann <[email protected]> Reviewed-by: Gunnar Sletta <[email protected]>
* | Fix: QSGGeometry lineWidth in qsgbatchrendererTorgeir Lilleskog2013-12-061-0/+3
| | | | | | | | | | | | | | | | | | lineWidth was ignored in the new scenegraph batchrenderer. Regression from 5.1.1 Task-number: QTBUG-35346 Change-Id: I80eacc165f70b5f39d4a01cf458ab1a0e49cbd2d Reviewed-by: Gunnar Sletta <[email protected]>
* | Propegate filtering state when we extract texture from atlas.Gunnar Sletta2013-12-061-0/+1
| | | | | | | | | | | | | | Task-number: QTBUG-35395 Change-Id: I5ac08ade7763d14bca855b8300156aee0e6a6920 Reviewed-by: Laszlo Agocs <[email protected]>
* | Avoid warnings on cast from char* to QStringThiago Macieira2013-12-061-1/+1
| | | | | | | | | | | | | | | | | | Why isn't QtQml being built with those functions disabled, so they produce errors? Change-Id: I2c0cf945523142800982f100a61a6c10e1929632 Reviewed-by: Simon Hausmann <[email protected]> Reviewed-by: Kurt Pattyn <[email protected]>
* | Only emit focusObjectChanged when an actual change occurs.Michael Brasser2013-12-061-2/+6
| | | | | | | | | | | | | | Change-Id: If18b460a8773e5cac597c02c51836b79711c20f4 Done-with: Matthew Vogt <[email protected]> Reviewed-by: Simon Hausmann <[email protected]> Reviewed-by: Alan Alpert <[email protected]>
* | Clean up property dependency data structuresSimon Hausmann2013-12-057-141/+87
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | As a follow-up to the previous commit, this patch cleans up the data structures used to track dependencies of QML binding expressions and functions to context and scope properties, determined at compile time. Instead of "collecting" these depending properties upfront (codegen time), we propagate the information that a property is a context or scope property into the IR at codegen time and later in the isel collect these properties and their notify signal index in a hash in the IR functions. The CompileData structure generator then can read these hashes directly when writing out the dependency information. Change-Id: I32134706e2d24bf63d1b1abad0259ab072460173 Reviewed-by: Lars Knoll <[email protected]>
* | Fix dependency calculation for context and scope propertiesSimon Hausmann2013-12-053-4/+17
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | We were incorrectly calculating writing to a context or scope property as a dependency for an expression. We don't know whether a property is being written only or also being read from at lookup time, but we can make that decision in the isel then when generating the move instructions. So initially context and scope properties end up in a candidate set first and get promoted to real dependencies when they're being used in reading moves. Task-number: QTBUG-35210 Change-Id: Ia67057abafc2d611e1e6605327b4965ebe91cbed Reviewed-by: Lars Knoll <[email protected]>
* | Fix data layout of Managed classLars Knoll2013-12-051-4/+4
| | | | | | | | | | | | | | | | The internal class should really be the first member. The flags should go away over time. Change-Id: Id2373a438e2af63a55704819c9dd5569c3895080 Reviewed-by: Simon Hausmann <[email protected]>
* | Fix dynamic properties in QQmlPropertyMap not always being visible in QMLSimon Hausmann2013-12-053-0/+53
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | QQmlPropertyMap is a fully dynamic class that can add properties at any point in time. In order for these properties to be visible inside QML, we must disable the property cache (instead of trying to unsuccessfully re-fresh it). What happened in this particular case is that the QQmlPropertyMap derived type was instantiated and the VME instruction for creating it would also assign the property cache the compiler determined. There's no way for QQmlPropertyMap itself to access this property cache instance (stored in output->types[id].typePropertyCache) or invalidate it, so instead don't use the compiler's property cache when instantiating the type. This patch also disallows the adding properties to QQmlPropertyMap when it is used as base type for a new QML type, as we cannot provide the derived type to the QQmlPropertyMap constructor - this is only possible in C++. Task-number: QTBUG-35233 Change-Id: I7fa9e4a2224ccfdd7ccb3fd9f73919ecd46058a8 Reviewed-by: Alberto Mardegan <[email protected]> Reviewed-by: Alan Alpert (Personal) <[email protected]> Reviewed-by: Lars Knoll <[email protected]>
* | Fix Flickable StopAtBounds drag over, back, over behavior.Martin Jones2013-12-051-28/+46
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | A Flickable with StopAtBounds failed when: 1. position on a boundary. Without lifting your finger: 2. attempt to drag beyond the boundary -> doesn't drag 3. drag back to initiate dragging 4. attempt to quickly drag beyond the boundary. After 4, the view should be back on the boundary, but it could get stuck a little short of the boundary. Change-Id: I9bfbb4293f4d464bddb97c5c37e9bb91ed7d48e4 Reviewed-by: Robin Burchell <[email protected]>
* | V4 IR: reverse propagate int32 truncation.Erik Verbruggen2013-12-046-139/+393
| | | | | | | | | | Change-Id: I5cb0c7798d0e530f3137710bf0e723bd7b64dc89 Reviewed-by: Simon Hausmann <[email protected]>
* | QQmlMemoryProfiler: Fix dangling pointer use.Robin Burchell2013-12-041-2/+1
| | | | | | | | | | | | | | toUtf8 returns a temporary, accessing constData isn't safe Change-Id: I3c4d077f24cee0eaf1df230c4d8079619967b51a Reviewed-by: Simon Hausmann <[email protected]>
* | Remove unused memberLars Knoll2013-12-044-6/+4
| | | | | | | | | | Change-Id: I9926f1ab10ea04387f17794944dcc11f4a2a9054 Reviewed-by: Simon Hausmann <[email protected]>
* | Fix some header guardsLars Knoll2013-12-043-4/+4
| | | | | | | | | | Change-Id: Ib696c79754cc238e3f4fa356c461a07d9f06d6e7 Reviewed-by: Simon Hausmann <[email protected]>
* | Encapsulate the current context and fix it's usageLars Knoll2013-12-0445-222/+217
| | | | | | | | | | | | | | | | | | | | | | Encapsulate accesses to the current context, and rework the way we push and pop this context from the context stack. Largely a cleanup, but simplifies the code in the long term Change-Id: I409e378490d0ab027be6a4c01a4031b2ea35c51d Reviewed-by: Simon Hausmann <[email protected]>
* | Remove setVTable calls in performance critical areasLars Knoll2013-12-0421-75/+97
| | | | | | | | | | | | | | | | | | | | Remove all the calls to setVTable that were in performance critical parts of the code. This now brings performance back to the level we had with the vtable inlined in the Managed objects. Change-Id: I76317cc5c53b5b700d1d3883b954407142a4c424 Reviewed-by: Simon Hausmann <[email protected]>
* | Move the vtable pointer from the object to the internal classLars Knoll2013-12-0452-254/+309
| | | | | | | | | | | | | | | | This saves one pointer per object, and willmake other optimizations easier in the future. Change-Id: I1324cad31998896b5dc76af3c8a7ee9d86283bfe Reviewed-by: Simon Hausmann <[email protected]>
* | Fix a bug in the code for evalLars Knoll2013-12-041-5/+5
| | | | | | | | | | | | | | | | | | When eval was being used as an indirect call, the code didn't reset the current context properly before returning from the eval call. Change-Id: Id5c7e9a897101d25593ef0f3b9945adaf19360b3 Reviewed-by: Simon Hausmann <[email protected]>
* | Add a static toArrayIndex() method to QV4::StringLars Knoll2013-12-043-4/+12
| | | | | | | | | | | | | | | | This avoids a hack in QV4::Codegen where we created a V4::String on the stack to convert to an array index. Change-Id: I9a88d45817bbcde52a4037a52fbae299b8c9cb1a Reviewed-by: Simon Hausmann <[email protected]>
* | Fixup the implementation mess for QJSValue(QString)Lars Knoll2013-12-043-10/+111
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Until now we were using a QV4::String without engine to represent this case. But this leads to lots of quirks, where we ended up trying to access the engine (or the internalclass/vtable) of this string anyway. Now just represent it by using an QString in QJSValuePrivate, and use an empty value to represent it. This adds a little bit of code to QJSValue and QJSEngine, but is more stable and maintainable in the longer term. Change-Id: I3358165ee64e788274225743a95dfb13346225cc Reviewed-by: Simon Hausmann <[email protected]>
* | Optimize ExecutionContext::setPropertyLars Knoll2013-12-041-3/+12
| | | | | | | | | | | | | | | | | | | | No need to call hasProperty followed by put. Instead get the property descriptor and call putValue for it. Improves v8-bench by 5%. Change-Id: Ied047126c651c033f7ad4c27deaeec08e5fee7f9 Reviewed-by: Simon Hausmann <[email protected]>
* | Use an internalClass to represent formals and locals in CallContextsLars Knoll2013-12-047-67/+65
| | | | | | | | | | | | | | | | | | | | formals and locals in a CallContext where so far accessed through a linear search in ExecutionContext::getProperty. Fix this by introducing an internalClass for the Function used by the call context. Change-Id: I1141efa12b19d6de4a354bfd6e769c5ffcb8898b Reviewed-by: Simon Hausmann <[email protected]>
* | Remove unused inline wrapperLars Knoll2013-12-045-12/+9
| | | | | | | | | | | | | | | | | | The wrapper method for Function::code() was still there from the times we used C++ exceptions. It's not needed any more, so get rid of it. Change-Id: I2ec25fbca71eeef9d7a94a38b5adfa42e4de3a84 Reviewed-by: Simon Hausmann <[email protected]>
* | Speedup JS instanceof operatorLars Knoll2013-12-047-58/+50
| | | | | | | | | | | | | | | | Cache the prototype of the functionobject, and inline hasInstance. This removes a vtbl method and speeds things up quite a bit. Change-Id: Ic68f301f7e09763d445a98bffa2cd201303f902e Reviewed-by: Simon Hausmann <[email protected]>
* | Initialize variableLars Knoll2013-12-041-0/+1
| | | | | | | | | | Change-Id: I37b14a406ebb9cf87fef20426a94725a2441b7cd Reviewed-by: Simon Hausmann <[email protected]>
* | Fix qsTr() in .js contextJ-P Nurmi2013-12-041-1/+3
| | | | | | | | | | | | | | | | | | Don't assume a four characters long file name suffix (.qml) Task-number: QTBUG-32850 Change-Id: I522c06b71bf1b38f32f2947a6c06017f83eb50be Reviewed-by: Michael Brasser <[email protected]> Reviewed-by: Alan Alpert <[email protected]>
* | QQmlImport: Don't try use a dangling pointer.Robin Burchell2013-12-031-2/+1
| | | | | | | | | | | | | | | | | | | | | | toUtf8 would return a temporary, and constData would hold a pointer inside that temporary. This isn't even remotely safe. Move the pointer use down to the initializeEngine call so it is kept around long enough for us to do our stuff. Task-number: QTBUG-35355 Task-number: QTBUG-35343 Change-Id: Ie816d0d1a37e42607f26d9ad02cf999f3d459cd9 Reviewed-by: Simon Hausmann <[email protected]>
* | Merge "Merge remote-tracking branch 'origin/release' into stable" into ↵Frederik Gladhorn2013-12-039-93/+172
|\ \ | | | | | | | | | refs/staging/stable
| * | Merge remote-tracking branch 'origin/release' into stableFrederik Gladhorn2013-12-039-93/+172
| |\| | | | | | | | | | Change-Id: I894ad7b4888744833f487b481950c087fa960a5c
| | * Fix style animations to stop when the item is hiddenJ-P Nurmi2013-12-031-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | QStyleAnimation automatically stopped for hidden QWidgets, but didn't know anything about QQuickItems and kept animating regardless of their visibility. QStyleAnimation was changed so that it will keep animating only as long as the animation target accepts animation updates. This change ensures that the style animation updates are accepted only when the item is visible. Task-number: QTBUG-35319 Change-Id: I3c93a653316b8abbbc32940cd7499b660828eff8 Reviewed-by: Gunnar Sletta <[email protected]> Reviewed-by: Jens Bache-Wiig <[email protected]>
| | * Safely abort when we don't succeed in creating a GL context.Gunnar Sletta2013-12-032-3/+11
| | | | | | | | | | | | | | | | | | | | | | | | Task-number: QTBUG-33363 (cherry-picked from commit 12eab9162781) Change-Id: Ia2b0c329157786cb4ec703989f12d2fdb1ce6bc8 Reviewed-by: Laszlo Agocs <[email protected]>
| | * Doc: Change the order of QQuickText function definitionsTopi Reinio2013-11-301-49/+49
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | For Text QML type, some property and method documentation is missing from the output because in the source file, they appear before the '\qmltype Text' command. This change reorders the functions so that qdoc will see all documentation for Text QML type. Task-number: QTBUG-35018 Change-Id: Icd995f66679d5105912ee12a7aeffd510921a54d Reviewed-by: Martin Smith <[email protected]> Reviewed-by: J-P Nurmi <[email protected]>
| | * Doc: Update examples used in Writing QML Extensions tutorialTopi Reinio2013-11-302-16/+33
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Update the examples to use Qt's resource file system to store the qml files in, enabling them to run unmodified in any platform/build configuration. Also, - Chapter 6: Move the plugin code to a subproject and create a C++ application to use the plugin - Update documentation related to above chapter - Add thumbnails for the examples so they will appear in Qt Creators example list. Task-number: QTBUG-35001 Change-Id: I29122af11bb11c7e5e17993438e5fc18c7f96f89 Reviewed-by: Jerome Pasion <[email protected]> Reviewed-by: Alan Alpert <[email protected]> Reviewed-by: Kai Koehne <[email protected]>
| | * Handle boundingboxes with NaN in them.Gunnar Sletta2013-11-302-8/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | NaN does not compare well with other floats. The result is that the bounding box is left at its initial values, FLT_MAX for top-left and FLT_MIN for bottom-right. If so, treat geometry as invalid, aka infinite. Task-number: QTBUG-35192 Change-Id: I1df6987d56a0ce1f500b0eba344a5dcbc55f80a4 Reviewed-by: Mitch Curtis <[email protected]> Reviewed-by: Laszlo Agocs <[email protected]>
| | * Ensure that QML Windows respect the default platform window stateTor Arne Vestbø2013-11-301-1/+79
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The 'visible' property of a Window would be set on the baseclass QWindow like any other property during QML component creation, which would cause create() to be called and the platform window would be created. This left the 'visibility' of the QML window as Windowed, not respecting the platform defaults for how windows should be shown. The user would have to explicitly set "visibility: Window.AutomaticVisibility" for this default to apply, which doesn't make sense -- it should be the default. We solve this by deferring setVisible and setVisibility on the window until the component is complete and we have a full picture of its state. We then ask the platform for the default state based on the window flags (ensuring that eg "flags: Qt.Popup" will not result in maximized windows on iOS and Android), and apply the deferred visibility. The deferred visibility may still be 'false', but setting the window state makes sense anyways, so that a later "visible = true" will apply the default window state. Deferring platform window creation until the geometry has been potentially set from user code also has the benefit that the platform window can check the geometry and apply a default geometry if it's null. This was not possible when the 'visible' property was a regular property, as you could not know if the user's geometry changes would come after platform window creation. Task-number: QTBUG-35174 Change-Id: Icf3236187992048a85b2196c059f9b54699041a4 Reviewed-by: Simon Hausmann <[email protected]> Reviewed-by: Robin Burchell <[email protected]> Reviewed-by: Gunnar Sletta <[email protected]>
| | * Revert 99480d5420c0beea6771be582c039b550a4461f5Gunnar Sletta2013-11-272-18/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The Mac OS X platform plugin has been fixed so that this hack is no longer needed. Not to mention that it breaks on XCB. We keep the warning about bad exposes from the plugin in debug mode. These are still useful for tracking down future bugs. Task-number: QTBUG-35143 Change-Id: I5125f7ae2b7fd77c55e9a29b10aa5434598a9ea9 Reviewed-by: Ulf Hermann <[email protected]> Reviewed-by: Gabriel de Dietrich <[email protected]>
* | | Improve the Canvas threading modelGunnar Sletta2013-12-039-176/+251
|/ / | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The canvas classes were mixing scene graph resources and GL content across threads. This led to a number of potential crashes in addition to that the FBO based rendering had significant potential for stalling. QQuickContext2DTexture is no longer a QSGTexture with ambiguous ownership. Instead we use textureForNextFrame which is called on the render thread while the GUI is locked to synchronize state from the Context2D's "texture" into the actual QSGTexture. This means that cleanup of the QQuickContext2DTexture and the QSGTexture used for display is no longer in conflict. QQuickPixmap no longer contains a QSGTexture either as these are strictly for use on the scene graph thread. The Images are anyway loaded explicitly as QImage files in QQuickContext2DContext and uploaded again for every Canvas, so relying on the GL paint engine to do the caching will give us the same with less code. I also changed the default strategy to Immediate as that one supports the full API (cooperative does not support readback) and because cooperative is pretty bad for performance since the rendering happens in the sync() step. Task-number: QTBUG-34268 Task-number: QTBUG-31052 Task-number: QTBUG-21935 Task-number: QTBUG-30689 Task-number: QTBUG-29007 Change-Id: Ic540b22d5faa1188e21e56a3beee24191d13f423 Reviewed-by: Mitch Curtis <[email protected]> Reviewed-by: Laszlo Agocs <[email protected]>
* | Drop the "missing provider" warning from ShaderEffect.Gunnar Sletta2013-12-031-1/+0
| | | | | | | | | | | | Task-number: QTBUG-34676 Change-Id: I5f1c2f9ebe6048da5d5c1d1ea5e4799eacea3e8b Reviewed-by: Laszlo Agocs <[email protected]>
* | Fix typo in debug defineGunnar Sletta2013-12-031-1/+1
| | | | | | | | | | Change-Id: I4d024aeb4618228cad3000ddfda32e5c8aba5742 Reviewed-by: Gunnar Sletta <[email protected]>
* | Fix build in release mode with forced assertsKonstantin Ritt2013-12-031-1/+1
| | | | | | | | | | Change-Id: I0e35533af7f65200a8bc3c4024c29344fa6f4b7a Reviewed-by: Lars Knoll <[email protected]>
* | V4 IR: remove common toInt32 casts.Erik Verbruggen2013-12-031-2/+29
| | | | | | | | | | | | | | | | | | | | | | | | | | E.g.: a | 0 b & 0xffffffff These operations force the operands to be converted to int32 without changing their value. At this point we already added convert calls to the IR, so we can safely get rid of these operations. Change-Id: Ic4d3b989e13439eccd2c878fa7bf5030acae7630 Reviewed-by: Lars Knoll <[email protected]>
generated by cgit v1.2.3 (git 2.25.1) at 2025-05-23 04:07:54 +0000