| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
QQmlCompiledData used to contain the binary data for instantiating QML types in
the QML VME. Nowadays the QML type compiler as well as the JavaScript compiler
create a QV4::CompiledData::CompilationUnit.
Change-Id: I155f62a5ecfb55a3fe230520231b6d8fd5b28ac9
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When using CompilationUnit with a QML engine, ensure that they are deleted in
the same thread as the QML engine.
Only the QML engine has a secondary thread (not a plain QJSEngine) and there it
may happen that the last refcount drops within the loader thread. For example
when the trimCache() is called within the loader. The destruction of the
CompilationUnit however is not safe to perform in a secondary thread.
Change-Id: Ia0105a8885ec97b0b2159e32e637adbd4e99f016
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
| |
The clear() re-implementation from QQmlCleanUp was empty, so this added no
functionality. Commit 19c6f620dd35916466c36234231d798f79732ab0 removed the last
usage of it.
Change-Id: I499a6daeb1f74cc8bad1cacc5c367fde1e6eee75
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I808d0a36094e873b69cf24a5b0113e741ff2a25d
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I99bb37bf4d4aa4aedd8e02a0fb4afb4a908573a6
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
|
|
|
|
|
| |
Change-Id: I85e8267ce4cd26ae83fe567357e1368658fdb43d
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
| |
These allow pre-allocating some arrays when instantiating types.
Change-Id: I2ca4ba4a69429918f03a5ba4c501c763e7ffa8dc
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
| |
Change-Id: I77892919678cb01ba1e697a44122760679a72045
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
| |
This makes particularly sense as the binding property data per object that
allows us to avoid a by-name property lookup when instantiating types is also
stored there.
Change-Id: I4d9275c1d8fde252df83eb11a9dfea71e5e9583a
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
| |
Similary to the other hash tables we can store the actual information about
whether a binding is covered by a custom parser or not straight in the
CompiledData::Binding.
Change-Id: Iab9044af57338cec935d3ef38764d7dc1aa507e8
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Ultimately the decision which bindings to initialize in a deferred way depends
on the data in the meta-object (deferred property names entry). The hash in
QQmlCompiledData is just caching this information. We are better off storing
this single bit right in the binding itself instead of in a parallel data
structure.
Change-Id: Ib66d3550210af1f882b98b0ba9089391813d69ad
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
By storing the object indices of named objects in the CompiledData::Object of a component,
we can achieve two things:
(1) We can eliminate the hash of vectors in QQmlCompiledData for the object-to-id mapping
(2) We can store the mapping from object name to integer object id in the CompilationUnit and
share it across different QQmlContextData instances (as long as it is not modified).
Also added a new test that verifies the functionality of a .qml file starting
with Component{} itself with object names, something that was previously only
implicitly tested through some of the examples (corkboards.qml for example).
Change-Id: I28c70217222dc0e5252bf5247b7e3fc4def47446
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
| |
By storing the calculated integer id for an id-named object in
CompiledData::Object we can simplify the code and replace a hash table with a
plain vector.
Change-Id: I4a84cdd00e98766d603d152e5a6574b232771a02
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
This allows simplifying some code and reducing the usage of the objectIndexToId
hash maps.
Change-Id: I1f08d4b224c4f9fa498d90471fa545ae4e4f2af4
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
| |
This would not seem like a member variable that is hot enough to keep there and
it is two loads away.
Change-Id: Id7088771bd33545a2846cc3497e5904dd8ac4f5d
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
| |
It is unused now and we can remove it as well as its QByteArray based storage.
The non-emptyness of the meta-data QByteArray was also used to indicate whether
it is necessary to create a VME meta-object when instantiating an object. This
bit is now folded into the flag of the QFlagPointer storing the property
caches.
Change-Id: I3c3604c61ff16a4e76912e68b1c19afdb0f2bd9d
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
| |
Remove two dummy forwarding functions
Change-Id: I053f9aecc761772c932550a0ddbea390621049c1
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
| |
From Qt 5.7 -> LGPL v2.1 isn't an option anymore, see
http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-kde-free-qt-foundation/
Updated license headers to use new LGPL header instead of LGPL21 one
(in those files which will be under LGPL v3)
Change-Id: Ic36f1a0a1436fe6ac6eeca8c2375a79857e9cb12
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
|
|
|
|
|
|
|
|
|
| |
Qt copyrights are now in The Qt Company, so we could update the source
code headers accordingly. In the same go we should also fix the links to
point to qt.io.
Change-Id: I61120571787870c0ed17066afb31779b1e6e30e9
Reviewed-by: Iikka Eklund <iikka.eklund@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Removes URL parsing and allocation/deallocation overhead as a bottleneck from
cached instantiation. Takes total runtime percentage of beginCreate from 6208ms
to 4987ms out of a 15 second trace of the instantiation_cached librarymetrics
benchmark with the empty Item data.
Task-number: QTBUG-43096
Change-Id: Ie4ec8e83461b4926e502dd78a7178cc8e8131e70
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In the QQmlScriptString we store the binding id and it is an index into the
runtimeFunctions array of the compilation unit. However we don't store the
compilation unit and instead in QQmlBinding and QQmlExpression try to retrieve
it from the cache via the context url (we have the context after all). That
turns out to be not a reliable way, as sometimes the URL might slightly differ
from the originally compiled cache (qrc:/// turning to qrc:/ maybe).
Consequently the type is (unnecessarily) compiled again and unfortunately not
_linked_, therefore the runtime functions array is empty. Another option is
that when the component was created from a QByteArray, then no entry exists in
the cache in the first place.
This patch addresses the problem by storing a reference to the compilation unit
in the QQmlContextData. That we can safely retrieve and it'll make sure the
compilation unit also stays alive.
In the process of that the manual reference counting was switched over to
QQmlRefCount and QQmlRefPointer for QV4::CompilationUnit.
Task-number: QTBUG-41193
Change-Id: I9111f9a3b65618e453954abcd789c039e65a94f7
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
| |
- Renamed LICENSE.LGPL to LICENSE.LGPLv21
- Added LICENSE.LGPLv3 & LICENSE.GPLv2
- Removed LICENSE.GPL
Change-Id: I84a565e2e0caa3b76bf291a7d188a57a4b00e1b0
Reviewed-by: Jani Heikkinen <jani.heikkinen@digia.com>
|
|
|
|
|
|
|
|
|
| |
Merge QV4::CompiledData::QmlUnit into QV4::CompiledData::Unit. For pure JS
units it means a slight increase of memory usage by a few bytes, but overall it
makes the code a lot simpler.
Change-Id: Ib48927749720b056f004aac0fe22cb8ec729e3f6
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
This is part of the effor of moving members from QQmlCompiledData into
QV4::CompilationUnit in order to eliminate the former in the long run.
Change-Id: Icce7fe0ee9a49cb3a7677fd7020008fc55ecdcf6
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The custom parser design used to be so that the custom parser operates on the "AST",
creates its own binary representation of the data it needs, stores it in a QByteArray
and gets that at object instantiation time. That meant serializing everything necessary.
With the introduction of the "binary" QML data structure, that process of serialization
becomes obsolete and would require extra work in the custom parsers for example for QQuickStates
to store the translation parameters.
The clean solution is to eliminate this unnecessary serialization process and
instead let the custom parsers do a verification pass at type compile time and
then simply operate directly on the QV4::CompiledData::Bindings at object
instantiation time. That simplifies the code, and allows for support of
translations throughout all list model properties.
Additionally this speeds up the creation of state objects and reduces memory
consumption. Previously a text: qsTr("foo") binding in states would result in
an actual java script binding. After this patch it is merely stored as a string
and translated at object instantiation time.
Change-Id: I7550274513f54abb09a0ab4de51c4c0bcdb23cae
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
Connection objects
We can re-use the expression we've compiled at QML type compilation time, as
long as we "inject" the signal parameters in the dynamic qml lookup chain.
Change-Id: Icc417531c41dea06ff5d033011179af49b03f542
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
| |
* Get rid of members in QQmlCompiledData that were used by the VME
and are now unused
* Get rid of QQmlVME friend declarations that are not needed anymore
Change-Id: I11b4b6f0b4b0b60edf92a1256be3d0d44d76bbc9
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
What remains is the code for removing .pragma from script source code (and
replacing it with white-space to preserve line/column numbers). The previous
code even returned the value of the pragmas, but for the remaining caller sites
that value isn't used, so we can just return void.
Change-Id: I16db15da236970660b817d6c4493005365a7a1af
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
* QQmlCodeGenerator -> QQmlIR::IRBuilder (it doesn't generate code, it
generates the Object/Property/Signal/etc. IR of the .qml file, that's
going to get transformed to QV4::CompiledData::*)
* ParsedQML -> QQmlIR::Document
Change-Id: I329e858487b66e1ae528d44316761f5dd34b79f4
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
This removes the bulk of the code. A few smaller cleanups remain, to be
done in smaller changes as they move code around.
Additionally the "optimize" option of qqmlbundle was removed. It called QQmlScript::Parser::preparseData,
which however was not implemented and always returned an empty QByteArray. Therefore "optimize" would not
do anything and the class is gone now :)
Change-Id: I0c265e756704cb53c5250be1f69e4a3e1b6e64d5
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
| |
Change-Id: I592518444ef353cfcf153df0e6afa2fbac613560
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We need to track the objects created and pass them over to the VME guard used in the
incubator. The incremental build nature of the incubator requires that to avoid crashes.
For nested incubators we need to set the activeVMEData field in the root QQmlContext, to allow
child incubators to locate the parent.
Lastly we need can emulate most of the VME behavior in terms of build states when running with
QQmlInstantiationInterrupt by presenting four build steps: The initial state, two build steps,
a finalization step and the state when we're done.
Change-Id: I16cd7f71744decb9d4735ec77e9d944fad18e88d
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Make sure to pass onFooChanged handlers to QQmlConnection's custom parser by
not relying on the signal handler converter to set the
IsSignalHandlerExpression flag. That should only be set for real signal
handlers, the custom parser gets the raw bindings.
Also don't try to initialize bindings at creation time the custom parser
covers.
Change-Id: Iae22bc886c312843136f073959e59da440f4184c
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
QQmlJS::MASM -> QV4::JIT
QQmlJS::V4IR -> QV4::IR
Change-Id: I707e8990459114a699c200fe3c22cec3c8df1afc
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Pre-calculate the amount of space we need for binding and parser status
callbacks at compile time and therefore use a much simpler data structure
(vector) to store the points to the bindings and callbacks. They need to be
stored because during object construction and binding enabling phase, it may
happen that they get destroyed and thus their m_mePtr pointing into the array
gets deleted.
The contiguous vector will also make it possible to interrupt the completion
phase.
Change-Id: Ic7c985bb8325ab355112e30e9d33d6ae4e7476d1
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
| |
remove trailing spaces and expand tabs
Change-Id: Ieacb9d096b612c45d1a64700044c114d1f7522bc
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
List model definitions make heavy use of custom parsers, which requires AST
access as well as a general port to the new QQmlCustomParser API.
Additional fixes in the custom parser support were needed to pass all tests:
* Fix support for AcceptsSignalHandlers and AcceptsAttachedProperties
* Don't call setCustomData unless the compiler generated data earlier
Change-Id: Ic42f8a890391267c94f63d35f055b60fdbf3c83d
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
The TypeReference is not copy-safe, as it holds refcounted property cache
pointers. For the new compiler code path, don't copy them but keep pointers to
TypeReference objects around. Also make sure to ref the root property cache
correctly and avoid the unnecessary addref for the property cache when creating
new vme meta objects (initial refcount is 1).
Change-Id: I0c4b952c8300c2167d926d9c35b8579fd505d596
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
| |
This is in preparation for listmodel support, share the code for resolving
types and enums between the old and the new compiler, as all it needs is the
imports.
Change-Id: I4908d71eee50c769108e0e2b68b03496722fa49d
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
Move the outter loop into the builder class itself, use a vector instead
of a list (we know that it's a fixed size and we only do indexed access)
Change-Id: I933f0496ea47b3bc7c2bebde7f1a14b4f603b4c3
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
Enough to support the Connections {} element. What's missing are pre-compiled
bindings signal handlers.
Change-Id: I3ad1413fa636434d899ae8fb380249aaf40363dc
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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 <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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 <mardy@users.sourceforge.net>
Reviewed-by: Alan Alpert (Personal) <416365416c@gmail.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For example the x property in
PropertyChanges {
target: foo
x: someItem.x - other.width / 2
}
was compiled at run-time dynamically, which produces slower code (no type
information available) and slows down the type instantiation, because the
compilation happens every time at instantiation time (or later).
With this change, when the custom parser behind PropertyChanges requests a
binding ID for "x", the right hand side will be added to the bindings to
compile, then compiled and later at run-time the QQmlBinding constructor that
takes a QQmlBinding::Identifier can retrieve the correct compiled function from
the QV4::CompiledData::CompilationUnit.
Change-Id: I857fb2d39e82714b225bc9394b9904b795c6662b
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
GCC 4.7-4.9 are right that the "code" member is used uninitialized. In
fact, GCC 4.9 was quite assertive about it:
qqmlinstruction_p.h:538:102: error: ‘def.QQmlInstructionData<8>::<anonymous>.QQmlInstruction::instr_common::code’ is used uninitialized in this function [-Werror=uninitialized]
static void setData(QQmlInstruction &instr, const DataType &v) { memcpy(&instr.FMT, &v, Size); } \
^
(It says "is used uninitialized" for this particular case; the "may be
used uninitialized" appears in other places)
The analysis is as follows:
- variable declared on qqmlcompiler.cpp:1467:
Instruction::SetDefault def;
- type is POD, so no initialization is performed (def contains garbage)
- on qqmlcompiler.cpp:1468 we use the variable:
output->addInstruction(def);
- QQmlCompiledData::addInstruction is inlined and does:
QQmlInstructionMeta<Instr>::setData(genericInstr, data);
- which is the call above, doing a memcpy with a source (&v) equal to
the uninitialized "def" variable
- result: memcpy is copying uninitialized bytes
Valgrind doesn't report this because it doesn't care about copying
uninitialized data. It will only complain if a decision is made based
on it, which we don't since the first thing we do after the memcpy is
initialize the member.
The solution is simple to not copy the common part of the
instructions. This way, we save 8 bytes of unnecessary copying and we
still keep the warning if a member of an extended instruction isn't
set.
Change-Id: I940b40ea9aa61c7386e5cced4a7865be7bfddb5d
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
We store QQmlPropertyData pointers in our IR for Qt meta-object property
resolution at compile time. As it turns out however, it is possible that these
pointers change after retrieval from the QQmlPropertyCache, as the cache may
change later in the compilation process. Therefore we must do what also
QQmlCompiler does by storing a copy of the QQmlPropertyData. For the JS IR we
can do that conveniently through the IR memory pool.
A side-effect of this bug was that QQmlPropertyData pointers were re-used
and so the identity check in the isel later such as
_function->contextObjectDependencies.contains(m->property)
for dependency tracking failed. In the example given in the bug report it was
determined that the window.contentWidth property wouldn't need a property
capture, and therefore the binding was not re-evaluated as window.contentWidth
later in the binding evaluation phase received its correct value.
This patch also fixes the incorrect debug output names assigned to JS binding
expressions, where the index used to look up the name is per compiled object,
not per QML component.
Task-number: QTBUG-35063
Change-Id: I3e5bbfaac11e5c122a2ed15a3e486a93988e1b6e
Reviewed-by: J-P Nurmi <jpnurmi@digia.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
objects at QML compile time
This avoids having to do a string lookup for ids and in the import cache at
run-time, before we can do a string hash lookup in the property cache. Instead
we resolve final properties in the context and scope object at compile time and
look them up at run-time using their index instead. The dependencies to these
properties are also tracked separately and recorded in the compiled data.
This is merely the initial patch. There's a lot left to do, such as having
specialized getter and setters for specific property types. Setters are missing
altogether right now and will fall back to name lookup.
Change-Id: If3cb4e7c9454ef4850a615f0935b311c9395b165
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* Run the binding expressions, functions and signal handlers through
the V4 codegen _per_ component, and run the isel at the end for the
entire file. We need to do per-component codegen because we want to
set up the correct id and object scopes, which are different for the
root component and anonymous components.
* Changed V4IR::Module to allow for the concept of "qml modules" where
there is no root function defined. This is a logical consequence of
running v4 codegen multiple times with different input but the same
V4IR::Module.
Change-Id: Ib3a719f83507cbab7c2e4e145ccad5b663c795cf
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
| |
Handle them similar to function declarations, except that we need to synthesize
the expression into a function declaration that includes the signal parameter
names. This is done quite similar to the code path in the new compiler.
Change-Id: I751081f7f1052692da6e2ed60c7f5c017372d829
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
...instead of extracting the function body as a string and compiling it in the
GUI thread.
Change-Id: I3c3108f6e35464b5581a2d8b5799e7285858ce4d
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|