The QQmlPropertyCacheVector in the CompilationUnit encapsulates the
property caches for the objects declared in the QML tree as well as the
bits indicating whether a VME meta-object is needed. The ref-counting
for the caches in that vector was done "manually" and thus error prone.
This patch replaces the vector with a wrapper container that has
explicit move semantics and takes care of the addref() and release()
calls upon insertion, replacement and destruction.
Change-Id: If805fe016f1a1c70e56f8a90909ab87b653ea026
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Replace more uses of manual reference counting with automatic counting.
Change-Id: Id3ce13814ac7f13d62562be21e4d59119d099944
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The unit data of QV4::CompiledData::CompilationUnit is read-only, which we
should "enforce" with a const, especially as in the future it may come from a
read-only mmap'ed chunk of memory.
Change-Id: I58dde9fe2494d3cfa52cdc9f612c7fa90b98aa26
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
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>
Change-Id: I1f3baa11e7c7860d6052182a1056f5b13d4ce593
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
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>
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>
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>
For each type of VME meta object we store an array of integers holding the
meta-type ids of the QML declared properties. We can replace that array with
access to the QV4::CompiledData::Property entry for each property, where the
type is also accessible.
This is a fairly straight-forward change, except for the bit in
QV4::CompilationUnit where we delay the release of the CompiledData::Unit and
friends until the destructor instead of releasing it at unlink time. That
should be a safe change and is necessary as there are a few tests around where
the VME meta object still needs access to this meta-data at a very late stage
in the life-cycle right before the deferred deletion is run.
Change-Id: I431de15d12766df837c0e0251192df16a5a76868
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Replace substring functions that return QString with
corresponding functions that return QStringRef where
it's possible.
Create QString from QStringRef only where necessary.
While touching the code, also port loops to C++11 style.
Change-Id: I04c99b24ea6afd3715e3edf9ea00bfab838fd53c
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Reviewed-by: Frank Meerkoetter <frank.meerkoetter@basyskom.com>
Reviewed-by: Shawn Rutledge <shawn.rutledge@theqtcompany.com>
Reviewed-by: Robin Burchell <robin.burchell@viroteck.net>
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>
This is a cleaner separation and further reduces include dependencies
in the definitions of our basic data structured.
Change-Id: I18aa86cdea0c0dfbc16075d4d617af97e638811e
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Avoid repeated string hashing and lookups in the property cache in order to
retrieve the property details when initializing literal and script bindings.
Instead we now cache the property data at type validation time, similar to how
the property data was encoded in the VME instructions in the old engine.
Change-Id: I3957c7c4c3e26dfa97c4880b23940a3755ee90e4
Reviewed-by: Lars Knoll <lars.knoll@digia.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>
All members are identifiers anyway, so this gets rid of a ### and
also simplifies some of the call sites by removing the need for a
scoped string.
Change-Id: Ic6b550cdb97afa5a4b0fa7e9b13e7768ed3f6bd8
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Changed runtimeStrings to be an array of Heap::String pointers instead of
indirect String pointers. Later that member along with other GC related members
will go into a managed subclass. Meanwhile the generated code no more loads
String pointers directly but just passes the index into the run-time strings to
the run-time functions, which in turn will load the heap string into a scoped
string.
Also replaced the template<T> Value::operator=(T *m) with a non-template
overload that takes a Managed *, in order to help the compiler choose the
non-template operator=(Heap::Base *) overload. This allows removing a bunch
of Value::fromHeapObject calls.
Change-Id: I20415c0549d33cca6813441a2495976b66d4c00e
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Avoid the use of Returned<String> for newString and changed the identifier
table to use Heap::String. This required moving some code back into
Heap::String, but that's code that doesn't call back into the GC, so
allocations and therefore future object moves aren't possible.
Change-Id: I1dca3e9c12a9c56f09419af8cc8cba39fe04f720
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
We don't need Returned<T> anymore with the QV4:: vs. Heap:: separation.
Eliminating Returned<T> simplifies also some code.
Change-Id: Ic2a9cd3c1a94f2ea37b539d3984d63997121c2b9
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
This is a step towards storing direct heap object pointers for the values
on the JS stack, to avoid the costly indirection for data access.
Change-Id: Ibb57ed6cf52a7088bbc95ee04ae3a4cb25b8c045
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
This makes following QML-code proper working:
ParentChange {
x: 0
Component.onCompleted: x = 10
}
where x is a QQmlScriptString.
Before this patch an error-message would be thrown that the
bool/int/string/etc cannot be converted to a QQmlScriptString.
With the patch primitive types including null and undefined are
proper converted to a QQmlScriptString. The patch ignores (as
in not implements) function/binding assignment.
Unfortunately since commit aa25ad8d5f its not possible any
longer to instanciate QQmlScriptString what means there is
otherwise no (easy) way to inject a QQmlScriptString from
within Javascript.
Change-Id: I18aac6a6e9a57f3b7d0a2d66cdab2be6c3c153c5
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
This is among other things needed to fix the qml import scanner to detect
dependencies from .js files correctly.
The patch also fixes the use of Q_QML_EXPORT towards Q_QML_PRIVATE_EXPORT
where appropriate and corrects the wrong include path for the double conversion
code to actually be relative to the file it is included from. This worked by
accident because of other include paths present in the build.
Change-Id: I338583dad2f76300819af8ab0dae8e5724c84430
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
* Encapsulate the string pooling for the V4 data generation into a StringGenerator class.
* Move type reference collection into the IR::Document, where it belongs (as it writes
into the typeReferences there)
* const fixes
* Remove unused methods and members
* Store unit and qml unit sizes right in the generated data structure. That makes copying
easier (like we do when generating the QML data based on the JS fields)
Change-Id: I053146ab0b00cc90ac7f72f867415962d1be121b
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Simple calls to qsTr and qsTrId are detected at type compile time and
reduced to a special Translation and TranslationById binding type, which
avoids allocating a QML binding at type instantiation type just to perform
a translation.
Change-Id: I61e4f2db2a8092b5e6870e174b832d9c20cd62b5
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
If QQmlCompiledData gets destroyed while somebody still has refcount on the
QV4::CompiledData::CompilationUnit, then unit's _data_ would be freed already
by ~QQmlCompiledData. Given that compilationUnit->data is pointing to the same
malloc'ed address as QQmlCompiledData::qmlUnit, we can just let the
CompilationUnit always own the data.
Fixes tst_qquickloader and makes it possible to run the qquickcomponent tests.
Change-Id: Ie3f3e5335139236d7c2524a327665bda0a9cc847
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Saves 12 bytes per Function
Change-Id: I9a495805f9201eb6162a520ff5c2defeb73dc37a
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Michael Brasser <michael.brasser@live.com>
Remove SafeValue, it was used to port over to an exact GC. Since
we now have that, we can now safely merge it with QV4::Value
again. Also rename SafeString to StringValue for better naming
consistency.
Change-Id: I8553d1bec5134c53996f6b0d758738a0ec8a2e4d
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
use this instead of the generic runtime method. This gives
around 10% speedup for array heavy Javascript such
as crypto.js.
Change-Id: Ic8f478c5b18893f2ef49e3f658b1ef24ae22e64f
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
This speeds up reading array data from objects significantly.
Change-Id: I5d17a7b3e7583a16dc76d1ee6cbc1d7134e4c2fa
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
These should later get resolved at compile time, like enum assignments.
Change-Id: I2f40c8d13330d2a101f79af12fe708f466eef225
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Don't use recursive function calls anymore. Instead, push marked
objects onto the JS stack, and then pop them off when their children
are being marked.
Should reduce stack memory usage, and improves performance by ~5%.
Change-Id: I2d37d97579144fcba87ec8e9fd545dd220c01fbb
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
This class is not required anymore to generate stack traces, as
we now store the required information in the JS context stack.
Change-Id: I3893c805ca89dda70efde07fdd120e7dfaf3639f
Reviewed-by: Simon Hausmann <simon.hausmann@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>
This can (and does crash) when a gc gets triggered during the
linking stage of a compilation unit.
Change-Id: I06f1299adab68ff8e0a4755d02e246870797e7f2
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Now that we store them per QML file, they live in the type cache and
can outlive the engine. Therefore the engine needs to free to unlink
any remaining units upon destruction. This needs to be done after the
"death" of the memory manager, which is likely to sweep away any
function objects that also hold a reference to the compilation units.
Change-Id: I3968d5995289e8d2bc1e3abbb1f8be88a0ab4e03
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Simon Hausmann
Simon Hausmann
Anton Kudryavtsev
Jani Heikkinen
Lars Knoll
Simon Hausmann
Jani Heikkinen
Sebastian Sauer
Lars Knoll
Erik Verbruggen