Robin Burchell [Mon, 12 Jan 2015 10:06:46 +0000 (11:06 +0100)]
QQuickParticleEmitter: Call particleCountChanged by hand on maximumEmittedChanged.
Analysis of QObject connections showed that a large number of connections were
set up to the particles system. As this particular case is cheap, we can do it
inline instead, and avoid the connection-per-instance.
Change-Id: Iec2049fff2c257dc39662ef2b5a8ecd75bbd870d
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
J-P Nurmi [Wed, 14 Jan 2015 10:28:06 +0000 (11:28 +0100)]
Fix Qt.inputMethod ownership
Change-Id: Ic51f9daf17742a421a810d95b2a510caf657e08f
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Robin Burchell [Sat, 10 Jan 2015 21:12:29 +0000 (22:12 +0100)]
Implement DefUses::Temps with a QVarLengthArray.
Avoiding heap allocations here shaves 100ms or so off the optimizer runtime.
Change-Id: If00c757532ffe90f2fa9c62dc999bb69e25bb71c
Task-number: QTBUG-43719
Reviewed-by: Erik Verbruggen <erik.verbruggen@theqtcompany.com>
Lars Knoll [Tue, 13 Jan 2015 16:22:08 +0000 (17:22 +0100)]
Don't require an ExecutionEngine member in String anymore
Change-Id: I7c1a5471507669871564d79dc17d1026c268b6d0
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Mon, 12 Jan 2015 20:55:51 +0000 (21:55 +0100)]
Refactor persistent values
Use a page wise allocation mechanism for persistent
values. This significantly reduces memory consumption
of persistent values and also improves their performance
a lot.
Change-Id: I8499d2ca5bdd871e029f643ae605a94544558bb5
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Robin Burchell [Sun, 11 Jan 2015 19:41:38 +0000 (20:41 +0100)]
QQuickItemParticle: delete delegates once they are completed.
This was one gigantic memory leak, if anyone tried to use this functionality, as
evidenced by the bug. The comment notes that deletion was only to be done if the
items were created by the ItemParticle, but I don't think this makes sense:
take()'s documentation doesn't imply that the ownership of the item stays with
the creator, and nor does this seem to be a particularly sensible API decision
to make, so I'm not going to implement that.
Change-Id: I307d77b70c2b4f7b34565865f8efcfe64817be2f
Task-number: QTBUG-37486
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Ulf Hermann [Mon, 12 Jan 2015 12:54:45 +0000 (13:54 +0100)]
Drop trailing vertices in GL_LINES and GL_TRIANGLES drawing modes.
This results in more correct behavior if you create nodes with a
number of vertices that doesn't match the drawing mode.
Change-Id: Ic0f59a4019a6b4087b527b7c9b38c35a9e02ece8
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Simon Hausmann [Fri, 16 Jan 2015 14:46:46 +0000 (15:46 +0100)]
Merge "Merge remote-tracking branch 'origin/5.4' into dev" into refs/staging/dev
Jocelyn Turcotte [Thu, 15 Jan 2015 15:05:26 +0000 (16:05 +0100)]
Avoid an incorrect warning when dynamically parenting a Window
"Created graphical object was not placed in the graphics scene."
QQuickWindow is the root of a graphics scene and doesn't need to be
inside another one.
It is already suggested in the Window documentation that Window can
be an inline child of a top-level QtObject. This patch fixer the
warning when dynamically creating a Window component.
Change-Id: Ie6d9d37b9e9ffdb61101aaaad6f4b722216ec759
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Simon Hausmann [Wed, 14 Jan 2015 11:00:56 +0000 (12:00 +0100)]
Merge remote-tracking branch 'origin/5.4' into dev
Conflicts:
.qmake.conf
src/qml/jsruntime/qv4context_p.h
src/qml/jsruntime/qv4debugging.cpp
src/qml/jsruntime/qv4engine.cpp
src/qml/jsruntime/qv4functionobject_p.h
src/qml/jsruntime/qv4qobjectwrapper.cpp
src/quick/scenegraph/shaders/visualization.frag
tests/auto/qml/qjsengine/tst_qjsengine.cpp
Change-Id: I492e8546c278f80a300a2129e9a29d861e144a30
Kai Koehne [Fri, 12 Dec 2014 10:10:45 +0000 (11:10 +0100)]
Redirect console.info to new QtInfoMsg
Change-Id: I6f4560a3baae16dd6db090d51a577060689825ef
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Holger Hans Peter Freyther [Sun, 11 Jan 2015 07:56:56 +0000 (08:56 +0100)]
Check for "c" being a nullptr in all places
So either c is never a nullptr or there is a potential crash in
the grab code. Replicate the check from above.
Change-Id: I2c93502c026cdd92dc5d20e3d6d59e6f6f92a6ce
Fixes: Coverity CID 10628
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Holger Hans Peter Freyther [Sun, 11 Jan 2015 07:53:58 +0000 (08:53 +0100)]
Add null check for target as well
I couldn't see any documentation for the implication that if
stackBefore is set then target must not be null. Coverity didn't
find that implication either.
Change-Id: Ifb93aa4c1e40f417033057b9a403d368dfdf0ba8
Fixes: Coverity CID 10627
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Holger Hans Peter Freyther [Sun, 11 Jan 2015 07:43:29 +0000 (08:43 +0100)]
Remove side-effect in assert
Fixes: Coverity CID 10587
Change-Id: I5101d209a54751532f48094f443afe35c6111e29
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Gunnar Sletta [Mon, 12 Jan 2015 13:35:24 +0000 (14:35 +0100)]
No need to include qsgbatchrenderer_p.h from qquickwindow_p.h
Change-Id: Iba97717ae8d485c217613475524a762949ea1dae
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Lars Knoll [Tue, 13 Jan 2015 08:01:29 +0000 (09:01 +0100)]
Make sure we always have an engine when assigning to a Persistent
This prepares things for a rewrite of the internals of Persistent.
Change-Id: Ib93ec5911984d1bfce87ffdc3f86bc75f6ecafe9
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Michael Brasser [Thu, 18 Dec 2014 16:20:37 +0000 (10:20 -0600)]
Fix flick() with ListView.SnapToItem at bounds.
Previously flick() would not always allow flicking to the very end of
a ListView if ListView.SnapToItem were set. Stop ListView from adding
overshoot, and then incorrectly correcting, if we are flicking exactly
to the end of a list.
Change-Id: Iad56e433bc1ba7d62f0553c4355469c9861df919
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Jocelyn Turcotte [Fri, 9 Jan 2015 17:52:56 +0000 (18:52 +0100)]
Use a list of lists for free chunks instead of merging in one list
This remove the tedious merging of free heap slots into one big list
after sweeping. It now lets each chunk keep its 0-terminated list of
free items and by then linking chunks together in a list for those
containing at least one free item.
The chunk metadata is moved at the beginning of the allocated VM
region to avoid moving it all around while sorting
MemoryManager::Data::heapChunks.
Change-Id: Ia93c2eb7feca0779a28aea5149dd066dacbfa494
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Jocelyn Turcotte [Fri, 9 Jan 2015 17:53:37 +0000 (18:53 +0100)]
Fix a harmless valgrind warning in RegExpObject
Only even values are initialized by the RegExp engine and the
compiler might reorder end to be tested before start, which would
let valgrind complain about branching on an uninitialized value.
Fix the issue by only checking if start is offsetNoMatch since it is
implied in the implementation that start == -1 => end == -1.
This matches the behavior of RegExpMatchesArray::reifyAllProperties in WebKit
Change-Id: Ifde1c5bc99da2a7929cd096bf477cae8799e4fed
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Gunnar Sletta [Sun, 11 Jan 2015 15:38:19 +0000 (16:38 +0100)]
doc: The correct enum name is 'Time', not 'LastModified'.
Change-Id: I4e75b97e9e70a440a2e2ec96af04d004323f09bb
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Simon Hausmann [Wed, 29 Oct 2014 08:58:03 +0000 (09:58 +0100)]
Improve startup of bindings using QQuickScreen
We can optimistically initialize the underlying QScreen to the primary
screen and this way ensure that in the common case the evaluation of
bindings using the screen's attached properties results in correct calculations
on startup. This way we can avoid re-evaluating the bindings again later.
Previously on startup all the returned values were zero and the bindings got
evaluated again once a window was assigned.
Change-Id: I98ba5905953f0b5054d924919239d178570250d3
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: J-P Nurmi <jpnurmi@theqtcompany.com>
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Pasi Petäjäjärvi [Wed, 17 Dec 2014 10:00:41 +0000 (12:00 +0200)]
Rename QML debug macro to Qt naming style
With this change the macro is in line with QT_NO_<feature> style
of naming macros. This way it will also be automatically added to
the file mkspecs/qmodule.pri in QT_NO_DEFINES value.
Change-Id: I96041438ef68bbcbc4e66301b0f5fe1212129c1a
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Ulf Hermann <ulf.hermann@theqtcompany.com>
Erik Verbruggen [Fri, 9 Jan 2015 13:17:31 +0000 (14:17 +0100)]
V4 IR: move the MemberResolver out of IR::Temp.
Temps are copied around a lot. This patch reduces the size by storing a
single pointer to the resolver.
Change-Id: I074b8b729fce310542cf4697ef42107085b304b3
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Simon Hausmann [Mon, 12 Jan 2015 12:07:38 +0000 (13:07 +0100)]
Add ability to retrieve QJSEngine for a given QObject
[ChangeLog][QtQml] Added qjsEngine(QObject*) getter similar to qmlEngine(object)
that allows retrieving the engine for an exposed object.
Change-Id: Id14adec58d45fe58133f65bab773b07d6bc18569
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Gunnar Sletta [Mon, 12 Jan 2015 13:50:36 +0000 (14:50 +0100)]
Clean up renderer visualization a bit.
viewport was currently unused and adreno 305 was confused by
the tweak.w > 0.0 when the input was 0.0f, so write it
in a slightly more straightforward manner.
Change-Id: I101ff71dcb04ca531e91cc5522876d71f368cdc8
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Lars Knoll [Mon, 12 Jan 2015 12:48:47 +0000 (13:48 +0100)]
QV4::String is not an Object
Remove all the code related to Object vtable's in here. This code
never got called anyway, as it would require casting to an Object
to call it (which would of course fail for Strings).
Change-Id: Ib62bb1b10999304cb65f6030d805698b9f60700d
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Sat, 10 Jan 2015 21:04:54 +0000 (22:04 +0100)]
Get rid of Value::engine()
This method is not guaranteed to return an engine. We're safer
checking for the value being an object first and then getting
the engine from there.
Change-Id: I5c95e675337e545f2421613bd31c42d1e58d6f9a
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Erik Verbruggen [Thu, 18 Dec 2014 13:59:49 +0000 (14:59 +0100)]
V4 JIT: ARM: move registers around.
In
6572d4e50d73ac60a8974d07de74c27a7f99ebef we moved the
addressTempRegister to r10, and in
d8b276a59402cbbe6d070ba38805350e7f3dd8a1 we made sure that the YarrJIT
saves it too. JSC solved this by moving it to r6, which is already
saved by the YarrJIT. To make a future update of the assembler easier,
we also move it to r6.
This requires that we move our scratch register too. But, because it is
used a lot, we don't want it above r7 for Thumb2 reasons. Therefore, we
move the engine to r10, and the scratch register to r5.
Change-Id: I35be539940d9fe80971973cfa7f3a8dab2196a1e
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Ulf Hermann [Fri, 12 Dec 2014 13:53:09 +0000 (14:53 +0100)]
Disallow QQmlDebugServer::enable for QT_QML_NO_DEBUGGER
This mirrors the behavior of the QQmlDebugServer constructor.
Unfortunately we have two #defines to disable the debug server,
QT_QML_NO_DEBUGGER and QQML_NO_DEBUG_PROTOCOL. We should honor both of
them as people might rely on either of them to effectively prevent the
debug server from working.
Change-Id: I0bc0d48632eaa05c471d0afed6fc48436116746a
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Ulf Hermann [Mon, 17 Nov 2014 13:43:27 +0000 (14:43 +0100)]
Make opengl context current in context2d's toImage if necessary
With the basic render loop the scene graph thread is the render thread
and thus we don't get a different opengl context for context2d in
immediate/fbo mode. We have to make sure the opengl context is current
when we call flush() from functions exposed to JS.
Change-Id: Idd91cf1ce9c299a7645cf3b78d4498652376b17e
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Lars Knoll [Sat, 10 Jan 2015 16:16:10 +0000 (17:16 +0100)]
Define vtable for CatchContext
Change-Id: I987e38f843881137b4e3241f68ecdf5516f95515
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 22:36:57 +0000 (23:36 +0100)]
Move extensible into the internal class
With this, we can now save one pointer per Heap
object.
Change-Id: I7f69193ff51c9fd9c5dbfba90aa1ebb3f93da2e6
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 21:02:40 +0000 (22:02 +0100)]
Fold the markBit into the internalClass pointer
Change-Id: I79f9286275f833dd01a3193042976d1f86ed5019
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 18:48:56 +0000 (19:48 +0100)]
Fold the inUse flag into the internalClass pointer
Revert the flag (ie. flag == 0 means the item is used), and fold it
into the second lowest bit of the internalClass/nextFree field.
Change-Id: I7b690fdce00d16aa538fa70a269a755511477c5c
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 14:10:52 +0000 (15:10 +0100)]
Move subType from Heap::Base into Heap::String
This is the only place it's still being used
Change-Id: I7341986278b1ae47d898cae3df605a23b34eebe2
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 13:32:57 +0000 (14:32 +0100)]
Remove subtype usage in ErrorObject
Change-Id: I77e23516270a2330ec22215aef6782a38d7fca09
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 12:56:47 +0000 (13:56 +0100)]
Remove the StringType_UInt subtype
It's not really used, and doesn't optimise anything really.
Additionally get rid of some code duplication.
Change-Id: I6502512e6df58db2c0264ea43d91a23c7585427c
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 12:28:40 +0000 (13:28 +0100)]
Get rid of subtype usage in Function objects
Change-Id: Ic84ddab292cb69e79dac0f2b8a87b96b096360d8
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 12:13:04 +0000 (13:13 +0100)]
Get rid of hasAccessorProperty in Heap::Base
This shouldn't affect performance as we can just as well
check for cases where we need to take the slow path
differently.
Change-Id: I4b9f69c39e9e64b437820ca3a6ea43e8877f2cf3
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 11:30:01 +0000 (12:30 +0100)]
Get rid of the needsActivation flag in Heap::Base
Change-Id: Idcec374cb6d16aabf919c23d7236b76d9d047527
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 11:11:09 +0000 (12:11 +0100)]
Get rid of the bindingKeyFlag
Instead use the vtable to identify that we have a
binding function.
Change-Id: I794aebb6fb83f648ba36f2f15cad94d2af3cae91
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 11:01:06 +0000 (12:01 +0100)]
Start cleaning up flags in Heap::Base
The goal is to get rid of the flags here and reduce
the size of each GC'ed object with a pointer.
Change-Id: I8c876810037756b85a646348cc5992a688588ea5
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Erik Verbruggen [Wed, 17 Dec 2014 12:25:22 +0000 (13:25 +0100)]
QML 3rdparty: revert unnecessary change.
It is the change to DataLog.h, which in turn forces the other changes in
order to match types.
Change-Id: Ie17e7efbd6a4d380a3b7383b0fd0243c6f68d0d5
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Laszlo Agocs [Thu, 8 Jan 2015 12:51:01 +0000 (13:51 +0100)]
Fix timing output from windows render loop
Print the frame time, not the total, and in milliseconds.
Change-Id: Ia7fe3bea8efafcce475c49517d7adab1b3841729
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Robin Burchell [Sun, 21 Dec 2014 23:00:26 +0000 (00:00 +0100)]
QV4: Use QVarLengthArray in cleanupBasicBlocks.
Avoids heap allocations, drops around ~60ms off my morbid testcase in doing so.
QBitArray is still allocating, meaning this function isn't "free", but it's now
at 0.9% of runtime vs the 1.3% it was before, so something for another day.
Change-Id: Ie0db8e0312bde5f67b37250d04b4d65e1f0b034d
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 9 Jan 2015 10:51:03 +0000 (11:51 +0100)]
Fix compiler warning about use of uninitialized variable
Change-Id: Ic9218d82c5f47d464a6536759cfe2c6dbfc9f985
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
J-P Nurmi [Mon, 5 Jan 2015 20:00:40 +0000 (21:00 +0100)]
Add Window.width and Window.height attached properties
[ChangeLog][QtQuick] Added Window.width and Window.height attached
properties
Change-Id: I3ef590a0d3e6fa660ed88992d5ae843deb09c7bc
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Albert Astals Cid [Fri, 9 Jan 2015 09:49:47 +0000 (10:49 +0100)]
Test: make mouse press, release, click and double click default to the center of the item
Most of the times you just want to act on the item not caring the position.
This helps for cleaner test code
Change-Id: I0e8a4919c503e3fd6dee60e1bd937ac57622279f
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Christopher Adams <chris.adams@jollamobile.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>
Kai Koehne [Mon, 5 Jan 2015 14:31:20 +0000 (15:31 +0100)]
Print exceptions in JS slots
Do print JS exceptions raised in slots to the console, and clear the
status afterwards. Otherwise exceptions raised might silently propagate
to later JS calls.
Change-Id: I122f3026aa18028415c5a4672cb3287e7d58ac4d
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Robin Burchell [Wed, 7 Jan 2015 21:00:00 +0000 (22:00 +0100)]
QQuickGridMesh: Avoid a signal connection per mesh instance by just calling the signal directly.
This can be instantiated a fair number of times, leading to a lot of signal
connections. Since the signal is so simple, just place a function call inline
and avoid unnecessary allocations.
Change-Id: I031245c46bf797898f22ed991d0e97d73728bdf5
Reviewed-by: Michael Brasser <michael.brasser@live.com>
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Robin Burchell [Mon, 22 Dec 2014 14:32:06 +0000 (15:32 +0100)]
V4 JIT: Hoist QVector allocation out of a loop.
Use erase() each loop iteration, but preserve the underlying allocation.
Change-Id: I673766f5567794215465daa597a4839a4900450b
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Robin Burchell [Sun, 21 Dec 2014 21:55:24 +0000 (22:55 +0100)]
QV4: Don't heap allocate blockNumbers.
Analysis of the allocations in BasicBlockSet show a huge amount of time spent in
allocating and freeing BasicBlockSet, which involves heap allocations. Most of
the time, blockNumbers is tiny (at least in my testing) - 0 to 1 in size at
deletion time. By inlining the (small) array, we save a nasty penalty.
This shaves around 400ms off the optimizer phase of my morbid QML testcase.
Change-Id: I46319173b5408a0d7a1b9663fdc516c9e5ca410e
Reviewed-by: Erik Verbruggen <erik.verbruggen@theqtcompany.com>
Erik Verbruggen [Fri, 9 Jan 2015 09:44:31 +0000 (10:44 +0100)]
QML: Remove unused field from struct.
Change-Id: I574dd5de67038ecb2e4d81ed69ed535e3be9ea05
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Erik Verbruggen [Fri, 9 Jan 2015 09:40:19 +0000 (10:40 +0100)]
QML: fix uninitialized variable use.
/Users/erik/dev/qt5-dev/qtdeclarative/src/qml/qml/qqmlpropertycache.cpp:1539:5: warning: variable 'offset' is used uninitialized whenever switch default is taken [-Wsometimes-uninitialized]
default:
^~~~~~~
/Users/erik/dev/qt5-dev/qtdeclarative/src/qml/qml/qqmlpropertycache.cpp:1544:15: note: uninitialized use occurs here
*index -= offset;
^~~~~~
Change-Id: I92538c8a3cf2a192be4c62e29b06b2c940b6def5
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Robin Burchell [Thu, 8 Jan 2015 00:28:58 +0000 (01:28 +0100)]
Replace InternalClass transitions hash with a sorted vector.
In a reasonable test application, there were some 1697 transition entries. Out
of these, 1663 of them had only a single item. The remainder, with the exception
of three, had <10 items. Only one of these had a count of >50 items (86).
As can be seen, most of the time, transitions is usually quite sparsely
populated, so using a hash is a large amount of overhead considering there's
just a few elements. For the times when it isn't, the vector being sorted
should help take care of that.
Since transitions are never removed, we can use a similar trick to
ba690fb73864915b4a35bbec5b7dc134ff1dafd0 and use a sorted vector to store them.
Compared to the hash approach, this saved ~412kb according to malloc_stats on a
reasonably comprehensive test application. Coincidentally, this also improved
v8bench for me by ~10%.
Note that this undoes
132cdfa69cae45d0c02ea715ce58722bbcd57e73, but the
expectation is that the fewer allocations done by using a vector will outweigh
the need to reserve any specific allocation initially.
Change-Id: Iec57a7db7e9a60347c9683b1cb1598f6d9c866f7
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Lars Knoll [Fri, 9 Jan 2015 09:53:12 +0000 (10:53 +0100)]
Stabilize the drag related autotests for higher dpi displays
Higher DPI displays might have a larger StartDragDistance (at
least on X11), breaking the autotest on these displays.
Change-Id: I9768a245a7654665dd9f718100289e9416d8c855
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Mitch Curtis [Thu, 8 Jan 2015 15:48:01 +0000 (16:48 +0100)]
Fix grammatical error in Text and TextEdit documentation.
Change-Id: I6fd68166108a060931cd6bcd64ccd865d251a4c3
Reviewed-by: Venugopal Shivashankar <venugopal.shivashankar@digia.com>
Erik Verbruggen [Thu, 18 Dec 2014 09:37:22 +0000 (10:37 +0100)]
Remove QQmlTrace. We have a real profiler.
Change-Id: I50d981b277187327c2c63f8372f64db1300ed9ef
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 2 Jan 2015 14:07:35 +0000 (15:07 +0100)]
Cleanup variant conversion code
Make public methods proper members of the ExecutionEngine, and
move private methods into the .cpp file only.
Change-Id: I3ca49e39bb1c4e559a2c63346e6ae6cfa446147d
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Fri, 2 Jan 2015 13:37:26 +0000 (14:37 +0100)]
Move the variant conversion methods from qv8engine to qv4::ExecutionEngine
Change-Id: Ibd529ae5cc3ba06f46152e9daa9119a4e7a2561c
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Thu, 1 Jan 2015 14:24:33 +0000 (15:24 +0100)]
This method is only used one place, so inline it there
Change-Id: Ia8b75965208c9687163ddaf982c4081c1ee504d4
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Thu, 1 Jan 2015 13:25:08 +0000 (14:25 +0100)]
Make the remaining variant conversion methods static
And make most private as well
Change-Id: If215aba4985a5150aa5a901f2e50e31bf0afa709
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Wed, 31 Dec 2014 15:15:42 +0000 (16:15 +0100)]
Remove most v8engine references in xhr
Change-Id: Ibbb38d57782f8566502852a2cb3cfd4a28f9dfc3
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Wed, 31 Dec 2014 15:03:34 +0000 (16:03 +0100)]
Remove v8engine from qmlcomponent and canvas2d
Change-Id: I575ba208de24ffe90bd02905aba67b0b08c5115e
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Wed, 31 Dec 2014 14:49:59 +0000 (15:49 +0100)]
Remove v8engine dependency in localstorage
Change-Id: I69616efe6629bcae154ff36f3e1aebd2dfa87aa4
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Wed, 31 Dec 2014 14:49:39 +0000 (15:49 +0100)]
Remove v8engine dependency in the builtin functions
Change-Id: If9d50709548a5295457a20907bdf5e1153f49071
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Wed, 31 Dec 2014 12:26:48 +0000 (13:26 +0100)]
Remove v8engine dependency from qqmllocale
Change-Id: Ib6f51915c5a049592c2958624017eb2f8ad962d8
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Wed, 31 Dec 2014 12:21:39 +0000 (13:21 +0100)]
Move extension mechanism over to use the v4 engine
Change-Id: Ib329fc7bcae3c78d962a116f53b2244a71f81228
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Tue, 30 Dec 2014 20:47:49 +0000 (21:47 +0100)]
Remove v8engine usage in the value type provider
Change-Id: I6b26a8249175a86c6e15d000b24ba3300e112715
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Simon Hausmann [Fri, 5 Dec 2014 15:32:56 +0000 (16:32 +0100)]
Clean up JS .import/.pragma directive scanning
There's a scanner in QQmlJS::Lexer::scanDirectives that can parse those, so
let's get rid of extra parser that operates on a string. Instead this way we
can do the scanning all in one shot, avoid detaching a copy of the source code
string and (most importantly) bring the parser closer to the copy in Qt
Creator, which uses the directives approach to extract imports and pragma.
Change-Id: Iff6eb8d91a45d8a70f383f953115692be48259de
Reviewed-by: Fawzi Mohamed <fawzi.mohamed@theqtcompany.com>
J-P Nurmi [Mon, 17 Nov 2014 11:36:30 +0000 (12:36 +0100)]
TextEdit: fix construction time text cache
QQuickTextEdit does not apply cached text to the underlying document
until component construction is complete. However, if text() was called
during the construction (eg. because of a binding), it would blindly
overwrite the cached text with (still) empty document content.
Task-number: QTBUG-41583
Change-Id: I6c5c1b6283d70d8a5ec2c8bd986095ee3c35a14c
Reviewed-by: Pierre Rossi <pierre.rossi@theqtcompany.com>
Lars Knoll [Tue, 30 Dec 2014 20:36:00 +0000 (21:36 +0100)]
Remove qv8engine usage in the contextwrapper
Change-Id: Iaf807add5d971e96cac57e38e13385e901f9c930
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Tue, 30 Dec 2014 17:29:21 +0000 (18:29 +0100)]
Get rid of QV8Engine::toString
Change-Id: Ib51fa09ae251c1b7b8878ecdf920016f8fcc0067
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Tue, 30 Dec 2014 16:06:02 +0000 (17:06 +0100)]
Remove v8engine pointer from QQmlV4Function
Change-Id: I68ddd79b8866ef9b8628a2eab1118d3ee3d72759
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Tue, 30 Dec 2014 15:38:20 +0000 (16:38 +0100)]
Get rid of more uses of QV8Engine
Change-Id: I763728b1685e8e68bcf815dda4dd8e5757f59dd2
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Tue, 30 Dec 2014 15:26:18 +0000 (16:26 +0100)]
Reduce v8engine dependency for the list model
Change-Id: Ib9e39a886c918a8f6ee1477465895739f90e1a0a
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Tue, 30 Dec 2014 15:04:32 +0000 (16:04 +0100)]
Further reduce v8engine dependencies
Change-Id: I9f50f5ed1928de0c389c9646b308b431b51dae37
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Tue, 30 Dec 2014 14:54:58 +0000 (15:54 +0100)]
The property() calls in the proeprty cache only require a QJSEngine now
Change-Id: Ic4a944b4f1dd02ad2f0284acad5f3135068f66df
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
J-P Nurmi [Wed, 7 Jan 2015 14:19:29 +0000 (15:19 +0100)]
Allow importing protected modules with different major versions
This allows QtQuick.Controls 1.x and 2.x imports to co-exist even
if they are two different plugins with the same module directive.
Change-Id: Idee302439e3c2fd6813ba2f41b69144fbae7902c
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Albert Astals Cid [Wed, 7 Jan 2015 15:17:52 +0000 (16:17 +0100)]
Make waitForRendering without item fail instead of crash
Change-Id: Icd276ed19087c2a1cde8c626347c696efb21ceb3
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Albert Astals Cid [Mon, 5 Jan 2015 13:44:36 +0000 (14:44 +0100)]
Parse dates as Qt::RFC2822Date too
[ChangeLog][QtQml] Fix JavaScript Date parsing to correctly accept RC2822 dates.
Task-number: QTBUG-38011
Change-Id: Ic7d3f5b8c624d6780d4c3595fd2efe22c427ffd3
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Robin Burchell [Sun, 21 Dec 2014 19:15:09 +0000 (20:15 +0100)]
QV4: Lower time spent in indirections and allocations.
Change data type for defsUntyped from QList -> QVector & reserve space when
creating the list. Drops ~250ms off a pretty morbid QML testcase, most, but not
all of which was spent in allocations.
Change-Id: I2ed8c62e7d41ab353a0194da268a2b430f079474
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Robin Burchell [Sun, 21 Dec 2014 18:44:41 +0000 (19:44 +0100)]
V4 IR: Change data type used in RemoveSharedExpressions pass.
Profiling the loading of a pretty morbidly large QML file consistently showed
that this was quite slow, around 1300ms in removeSharedExpressions, of which a
good >600-700ms (I didn't count exactly, but it was a very large amount) was
down to allocating and freeing QHash nodes.
As we don't require removals, leaving insertion and lookup as the only two
remaining options, a sorted vector becomes a viable alternative (credit to João
Abecasis for the idea).
An additional benefit of this change is that the two hash lookups are now
compressed into a single 'hash' lookup (via the lower_bound call) instead of
separately using contains() / insert().
Measuring the exact saving is difficult, but it looks like this saves between
700-1000ms off the runtime RemoveSharedExpressions. After this patch, malloc and
free are dominating the optimizer run, instead of any particular method.
Change-Id: I6c0bb8495eac4dd3613ba0274e8802d7bd609460
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Simon Hausmann [Wed, 7 Jan 2015 10:51:37 +0000 (11:51 +0100)]
Fix property access and method invocation on value types that use inheritance
For gadgets/value types we use moc's static_metacall, which doesn't call the
parent class implementation. Therefore before placing a static metacall we
must resolve the indicies and find the right meta-object.
Change-Id: I258e3d9ecfc704498c68772dc42b16134a3bfd83
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Alex Blasche <alexander.blasche@theqtcompany.com>
J-P Nurmi [Wed, 7 Jan 2015 12:32:33 +0000 (13:32 +0100)]
Unify QQmlDirParser version parsing and error reporting
Add a parseVersion() helper function to avoid duplicating
the same version parsing logic three times.
Change-Id: I4e5b6a8c86ee3a26c4eb91c660a81176ac0346cf
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Laszlo Agocs [Tue, 16 Dec 2014 11:49:08 +0000 (12:49 +0100)]
Prevent windows outside the gui thread in rendercontrol example
In the multithreaded case we attempted to create QWindows outside
the gui thread on platforms where QOffscreenSurface is backed by
the hidden QWindow in the absence of a better solution. This has
to be avoided. Therefore, pass a suitable surface from the gui
thread instead.
This will avoid "Attempting to create QWindow-based QOffscreenSurface
outside the gui thread." type of warnings when running on OS X with
--threaded.
Change-Id: Ie3ebeeeaa9e6bdf83e763e40213e2940fbfde667
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
Reviewed-by: Jørgen Lind <jorgen.lind@theqtcompany.com>
Robin Burchell [Tue, 6 Jan 2015 18:36:02 +0000 (19:36 +0100)]
qmltime: Correct module name in error message.
Change-Id: Icccad866bdbaf53bf2cb3f1aff30ecd9ff6d06a5
Reviewed-by: Michael Brasser <michael.brasser@live.com>
Jocelyn Turcotte [Mon, 5 Jan 2015 17:14:56 +0000 (18:14 +0100)]
Fix the vertex ordering of merged triangle strips
Culling and gl_FrontFacing depend on the vertex order matching the
value set through glFrontFace. To match the default counter-clockwise
order we need to preserve the parity of triangles relatively to the
start of the indices argument of glDrawElements, but we break this
parity because of the degenerate triangles between merged strips.
Fix the issue by skipping the first degenerate triangle, the parity
of following strips will be preserved since they are then separated
by two triangles.
This fixes a regression of
tests/manual/scenegraph_lancelot/data/shaders/culling/culling_1.qml
triggered by
38cab579a0c5398b7621221fd8609bc43cf1f3c5.
Change-Id: Iefa7eaab68112d31be9d8646bd288eb000528cd5
Reviewed-by: Gunnar Sletta <gunnar@sletta.org>
Gunnar Sletta [Tue, 6 Jan 2015 07:40:21 +0000 (08:40 +0100)]
Reuse the memory we use for vertex uploads.
Every batch would have its own buffer with its own memory
and the original idea was that it would stabilize so that
the same batch would be used over and over, so the memory
allocatd for each batch's buffer would be reused. In practice
batches get discarded and recycled in somewhat arbitrary
order so we always end up reallocating.
Instead, use a single buffer for all uploads. This saves us several
reallocations per frame, and also prevents pooling of several
potentially large allocated blocks inside the renderer.
The exception is when using visualization or for drivers
which have broken index buffer support (nouveau only currently),
where we keep the existing behavior as the buffer data is
needed during rendering and visualization.
Change-Id: Id29095e00128ec1ee6d4ac3caa8f2c17cc7bbd27
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
hjk [Tue, 6 Jan 2015 09:35:32 +0000 (10:35 +0100)]
Some minimal changes to make qmlpuppet compile again
Access to QJSEnginePrivate::createCache and the previous
signature of QQmlPropertyCache::invalidate is used.
Change-Id: Ie2578eee33ee4ebb2cdc4574d8d1cb7c8209868e
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Thomas Hartmann <Thomas.Hartmann@digia.com>
Gunnar Sletta [Mon, 5 Jan 2015 12:21:45 +0000 (13:21 +0100)]
Remove bogus assert.
It is perfectly valid to reparent nodes to a parent
that is part of an existing hierarchy.
Change-Id: Icf2dbe6030d399047e7ebe735e3096d1a8cd8c0e
Task-number: QTBUG-43278
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Reviewed-by: Laszlo Agocs <laszlo.agocs@theqtcompany.com>
Gunnar Sletta [Mon, 5 Jan 2015 11:38:18 +0000 (12:38 +0100)]
Traverse children correctly when reparenting.
Since the children are being removed from the parent, using
nextSibling on the child won't work. We need to extract the
first child from the parent until there are no more children
left.
Change-Id: Ifca5f2760e2b1c7ba56c198623c5dc9d82c7560f
Task-number: QTBUG-42530
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Albert Astals Cid [Fri, 19 Dec 2014 10:21:00 +0000 (11:21 +0100)]
Add propertyName to error
May help a bit to the reader, even if the line is already there
sometimes it can be confusing if it is a default property like data
Change-Id: I5cb8e8833c78a784ad4f5541d094840477e8f350
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Oleg Shparber [Wed, 31 Dec 2014 18:37:47 +0000 (10:37 -0800)]
Use QV4::ScopedObject typedef instead of actual type
Change-Id: I0b68c534ea513a7c230b12114f6b42b069f9864b
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Oleg Shparber [Wed, 31 Dec 2014 18:40:16 +0000 (10:40 -0800)]
Use QV4::ScopedArrayObject typedef instead of actual type
Change-Id: I975536745ac6c264aca074f84d223fbec7682d3d
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Oleg Shparber [Wed, 31 Dec 2014 18:43:46 +0000 (10:43 -0800)]
Use QV4::ScopedContext typedef instead of actual type
Change-Id: I71c6c9cf030e347fbc5e4073e9ca338a9ce95999
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Oleg Shparber [Wed, 31 Dec 2014 18:42:15 +0000 (10:42 -0800)]
Use QV4::ScopedFunctionObject typedef instead of actual type
Change-Id: I6b4effaa5bef992b4ae9402eea7fe655bc7b18f0
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Oleg Shparber [Wed, 31 Dec 2014 18:34:52 +0000 (10:34 -0800)]
Use QV4::ScopedString typedef instead of actual type
Change-Id: I64ecbf6cea463387a70e909ecc5f9165d22a7b0f
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Thiago Macieira [Fri, 19 Dec 2014 03:58:58 +0000 (19:58 -0800)]
Fix build: include <qdatastream.h> before using QDataStream
Change-Id: I779e0cd705f664368b47763d4367b9dc377c1770
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Lars Knoll [Tue, 30 Dec 2014 14:52:06 +0000 (15:52 +0100)]
Remove unused variable
Change-Id: I399e66822f52e35d122d3a20bb8de20206e5864d
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>