This is required to be able to implement concurrent or
incremental garbage collection.
Change-Id: Ib3c5eee3779ca2ee08a57cd3961dbcb0537bbb54
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
* Only call ExecutionEngine::markObjects() on a full GC, it doesn't do
anything in the incrementall case anyway.
* Move the marking of child objects into it's own method for clarity
* Move collection of gray items down to happen directly before we drain
the mark stack
Change-Id: I41067e17d483067bd1c4d60da22c5628482dae78
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
It's only used during mark() calls, where we can be reasonably
safe that the assertion holds. But mark() needs to change and
become independent of the execution engine for concurrent collection,
so remove the assert as a preparation.
Change-Id: I52266af719afd76048d73efdcfcf96c0abc8843f
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Our two value encodings use different masks for the upper 4 bytes.
Depending on the target architecture we must use different values when
generating code that uses these masks.
This patch replaces the #ifdef'ed ValueTypeInternal_* enum values with
two C++11 scoped enums that allows for the co-existence of both
throughout the code base as well as selective use in the code
generators.
Change-Id: I380c8c28b84df2874cca521b78bfe7f9388ed228
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Move the Runtime function pointer array into EngineBase so that
we can eliminate the last use of qOffsetOf.
For improved cache locality the memory manager point is now also
located in the EngineBase.
Change-Id: I0b3cf44c726aa4fb8db1206cc414a56c2f522a84
Task-number: QTBUG-58666
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
The offsets of members encoded in JIT generated code differ between
32-bit and 64-bit architectures. This patch moves some of the
ExecutionEngine members into a separate standard-layout EngineBase class
(in line with the same class in commit
2a554434a5 and subject to merging). By
ensuring that the members are stored at pointer intervals, we can
translate from host pointer size to target when generating the code.
Task-number: QTBUG-58666
Change-Id: I1c38a7da059826848b80fd9972ed073214501386
Reviewed-by: Qt CI Bot <qt_ci_bot@qt-project.org>
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Add an incremental mode to the garbage collector, that will
get used for many collections. This should significantly
reduce average stop times for GC.
Make sure that manual calls to gc() still do a full collection,
to ensure consistency and keep tests that rely on gc() working.
Change-Id: I87b13529377b7639ce993dbd99e85ff0a555acd8
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Implement a Steel write barrier for our objects. The barrier is
interesting as it can also be used for incremental GC runs by
simply turning the barrier on and leaving old objects marked as
black.
Change-Id: I0b273974d94a990dee3cd9298089b8b202c75bf2
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
The new set() method also taked an ExecutionEngine pointer. This makes
it trivial to now add a write barrier for those operations.
Change-Id: I321eccfe6fb279cc240b5c84910e6854f71759f6
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
SimpleCallData doesn't need any loca variables, so
move it into a separate CallData Heap object. This
also allows getting rid of the manual markObjects()
implementation for CallContext.
Change-Id: I9014eb2f815d3e2fe63a951a9d126c38e8aaa0a3
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Introduce a ValueArray class, that defines an array of
Values at the end of a Heap Object.
Change-Id: I00efbf6f5839a6687dd5bc5fc037ec8f06e0936e
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Create a PropertyList prototype, and add the push method to
QQmlListProperty that call the append function if it has been defined.
Added a unit test and updated the documentation.
Change-Id: I2647766e98b60bf0546f6d6ed1422a616e0d3a07
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Per the spec, these use the same FunctionObject as the global object.
Implements:
20.1.2.12
20.1.2.13
For es6 test/built-ins/Number, we go from:
=== Summary ===
- Ran 490 tests
- Passed 412 tests (84.1%)
- Failed 78 tests (15.9%)
to:
=== Summary ===
- Ran 490 tests
- Passed 416 tests (84.9%)
- Failed 74 tests (15.1%)
Change-Id: I7e6f84ef81a429a4bce4895663ad6dc10f389f86
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
That way qmldevtools can - in the future - instantiate that also for the
cross-compilation targets.
Change-Id: If15b195f24e54226dc4f2fd9f0ad6874f44cb5af
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Full definitions of types are needed in static_casts.
Change-Id: I028ffc40a76cdb16cb297a181b3d9dfe9d09c945
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
It's a hack we needed when we still had a conservative GC, but
it is not required anymore. The only thing we still need is the
protection against running the GC recursively.
Change-Id: I55cd51d4929c828db5b61b38e781467c5bf77314
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Now that the code paths are very similar, we can simply to the
check whether to do a fast or slow function call in
ScriptFunction::call/contruct. To make this fast, cache the
result of the required check in QV4::Function
Change-Id: I03085ca2beb83b1721b60b0d7b2ab4c9266d1e48
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Factor the common code out into separate methods that can
later on be reused by the QQmlJavaScriptExpression.
Also ensure a CallContext is safe to use with a 0
FunctionObject.
Change-Id: I1181a8e320b8c931d9df5b2c91bc143d8587fb60
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
And return a QV4::Function from now on. This simplifies code in
other places and provides all the info required for stack traces
and debugging.
Change-Id: I512a8ac3932268d8cfc60675e75c4661d1f16fd8
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
The class should get merged with the QV4::QmlContext class.
Simplify the cleanup by moving both classes into a common
file.
Change-Id: I0074da79701d5f41eb51681b70fcde85bfd45fc1
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Try to avoid calling both as objectValue() already
checks isObject().
Change-Id: I1d770d4d9dabed4ea4cc3e322b8fdc5a64f5bd2b
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
It's enough to just call stringValue(), as that already
does the isString() check.
Change-Id: I7be0e643a7975c0704b4c9c43b337deb8db9fce0
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
This removes the destructors of subclasses of Base, making them nearly
trivial.
Change-Id: Ia6f7d467e87899b5ad37b8709a8f633a51689d59
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
So now VariantObject is nearly a trivial struct.
Change-Id: Ifc54c04d9686c03e12066c5287823dd3b1315d2a
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
This change also adds a check to the d() calls for Managed, verifies
that the object has been initialized. This is only done for debug
builds.
To prevent other code from tripping the check, a number of other classes
are either marked as trivial, or do initialization in the constructors.
Because of template function changes in them memory manager (those now
call init() instead of in-place new), String has an extra parameter
to force it to temporarily use an old/unmodified template function.
Change-Id: I8c35161ce7680835d830638b6d21498c5129b02b
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
By specification, date conversion functions for dates before the epoch
are not DST corrected. We converted QTime to a QDateTime where we set
the date part to Jan. 1, 1970, and then convert that to msecs since the
epoch UTC. For places on Earth where they had DST on that day (e.g.
Hobart in Australia), strange things happen: conversion from a QTime to
DateObject will use DST (because it's after the epoch in local time),
but conversions from DateObject to QTime won't use the DST because it's
before the epoch (in UTC).
Now as everyone knows, a 24-hour clock time has no meaning without a
date, only "elapsed time" has. But users still expect to be able to pass
QTime to QML/JS. So, we do the conversion on day 0 of month 0 of year 0,
and all of it in local time. This gives a stable conversion in both
directions, and the values in both C++ and QML/JS are the same for any
timezone (with or without DST) on this planet.
Task-number: QTBUG-54378
Change-Id: I892e16a93f015e92d311c6cae3ae7768b7373f6a
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
This is a partial revert of 90b06e2773, as it had unwanted side
effects. The original intention was to make assignment from char to
string possible, or more specifically, we wanted a solution where a
QChar could be assigned to a QString, as a character and not a string
representation of its value. While this behavior is desirable for
QChar, we most likely want the opposite for the regular character types.
Task-number: QTBUG-49232
Change-Id: I82d5f72b900fe984c4db1478fd52a9eb69ad2ee6
Reviewed-by: Michael Brasser <michael.brasser@live.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
[ChangeLog][QtQml][Important Behavior Changes] A JS null value converted
to a QVariant now has type QMetaType::Nullptr rather than
QMetaType::VoidStar.
Change-Id: I91a64e444ada0f1884fe807f9973348ba1a878ff
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
It's debatable if the uglification is worth the overhead of a few
pointers.
Change-Id: I63b55b2043b8752a94d4b862c3892915135a72b7
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
This will allow us to #define them away on -no-qml-debug, saving two
pointers per engine.
Change-Id: I400cffd32cd7f55ff0e68565734b6002b9f901d5
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
... or with QL1C in such cases:
- if there is overloaded function
- in QStringBuilder expressions
Saves ~1.5 KB in text size.
Build config: ubuntu 16.04 x64, gcc 5.3
Change-Id: Icc0789f1c244ce20a3182494b0c7f35c9d77e41d
Reviewed-by: Shawn Rutledge <shawn.rutledge@qt.io>
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
The implementation of many (or all) runtime functions consist of first
creating a QV4::Scope, which saves and restores the JS stack pointer.
It also prevents tail-calls because of that restoring behavior. In many
cases it suffices to do that at the entry-point of the runtime.
The return value of a JS function call is now also stored in the scope.
Previously, all return values were stored in a ScopedValue, got loaded
on return, and immediately stored in another ScopedValue in the caller.
This resulted in a lot of stores, where now there is only one store
needed, and no extra ScopedValue for every function.
Change-Id: I13d80fc0ce72c5702ef1536d41d12f710c5914fa
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Part of 0e053528 was reverted in the merge, about lastTimestamp. It
will be applied later in separate commit.
qmltest::shadersource-dynamic-sourceobject::test_endresult() was
blacklisted on linux.
Conflicts:
.qmake.conf
tests/auto/qml/debugger/qqmlprofilerservice/tst_qqmlprofilerservice.cpp
tests/auto/qmltest/BLACKLIST
tests/auto/qmltest/qmltest.pro
Task-number: QTBUG-53590
Task-number: QTBUG-53971
Change-Id: I48af90b49a3c7b29de16f4178a04807f8bc05130
Allocating from the JS stack will zero out the memory, and thus indicate
to valgrind that memory is not uninitialized. By first marking the whole
stack as uninitialized and only then allocating 2 entries, the behavior
for those two entries will now match the allocation behavior. This
fixes a false positive when using valgrind.
Change-Id: Icdb5279e1cfbfe6b5c385cc42c556edf721fa74b
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
All other changes are just to be able to include qv8engine_p.h in
qv4engine_p.h
Change-Id: I71ba2ec42cdc20d5c7d97a74b99b4d1ada1a5ed8
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
This is a regression from commit 94e337fa95 where
we accidentally ended up not having a calling QML context set anymore when
initializing the properties on newly incubated objects as provided by the
caller. The QML context is necessary as for example when we set a URL property,
the URL can be relative and it will be resolved to the base url of the context
when written, such as in in QQmlPropertyPrivate::write.
Change-Id: I1d896381fc92f653a7d76f4d82174bca48828f5e
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Getting the native stack size can be really expensive. For example, on
Linux/x86_64 (Ubuntu 15.04), it is at least 200,000 instructions for a
single-threaded application. With more threads (like qmlscene) it typically
ends up around 1M(!) instructions. Worse, it is called twice in the
ExecutionEngine constructor.
So, now we limit the depth of JavaScript calls to a fixed number, 1234 by
default. This can be changed by setting the environment variable
QV4_MAX_CALL_DEPTH to the desired depth.
Change-Id: Ic13c8efb2769e64fbc73deee6f6fa39d7c0b7af5
Reviewed-by: Simon Hausmann <simon.hausmann@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>
This only works for Windows 10 and later.
Setting the "Code Generation" capability in the manifest, WinRT
applications can use JIT for Windows 10 Desktop. Update the Allocator
in regards to this and also update the VirtualProtectFromApp section.
When the engine gets initialized, check for marking an area as
executable. If it fails, fallback to interpreter mode.
This does not affect ARM platforms, as we do not have calling
conventions yet. It is implicitly disabled as V4_ENABLE_JIT is not
defined.
[ChangeLog][WinRT] Enable JIT for x86/x64 targets on Windows 10 and
later.
Change-Id: Ie05add5263f71387c5ce98456b9bec86c6c07ceb
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
The wiki will provide information on how to work around the issue.
Change-Id: I4409f5b4af0668ab28197adb8766898124e8cec8
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
All identifiers starting with an underscore and a capital letter are
reserved to the compiler and must never be used by the user code. Try to
find a better name or, in the worst case, move the underscore to the
last position in these identifiers.
See commit cf63c63d558227fdbef09699c261560e7474f5ea in qtbase for a case
of such an identifier causing a build breakage when the compiler began
treating it specially (it was _Nullable).
Change-Id: I1d0f78915b5942aab07cffff140f9f39c29f0fdf
Reviewed-by: Erik Verbruggen <erik.verbruggen@theqtcompany.com>
Where possible, use qEnvironmentVariableIsSet()/
qEnvironmentVariableIsEmpty() instead of checking on the
return value of qgetenv().
Where the value is required, add a check using one of
qEnvironmentVariableIsSet()/Empty().
Change-Id: Ia8b7534e6f5165bd8a6b4e63ccc139c42dd03056
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Otherwise different debug services could steal each other's debugers.
Change-Id: Ic0a50333d21c7d20a7124240ea598f8446400ae3
Reviewed-by: hjk <hjk@theqtcompany.com>
... into a pure interface and a QV4::Debugging::V4Debugger implementation.
This is in preparation of a second implementation of this interface
to be used with 'native mixed' debugging.
Change-Id: I3078dcfe4bdee392a2d13ef43a55ca993e7b88d8
Reviewed-by: Ulf Hermann <ulf.hermann@theqtcompany.com>
Move the connect/disconnect methods for signal handlers into
their prototype, so that we don't need to define them per instance
anymore.
Change-Id: Iac1e6d1dd7bce86730dbb6c51e2c3f79713641f7
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Gives around 10% speedup on the v8 splay benchmark.
Change-Id: I47f64e7b73bde59ac3bdd2c94fc199ecfbbf290e
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Move the name and message property into the prototype as
per JS spec. Only define the message property in the
object itself if the value used for construction is not
undefined.
In addition, clean up creation of the objects and centralize
it in a few template methods.
Change-Id: I014017b710575b30bf4e0b0228111878f5c73b9a
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Also disable the old way of constructing objects.
Change-Id: Ib4e69087cd563ae1481da116d6caf97876239798
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Declare the default prototype and internal class
in the class itself.
Change-Id: I08c2b42aa61a886580061336ae60cef8dedc0f03
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Gives around 10% speed improvement on the v8 regexp
benchmark.
Change-Id: Iad37bcbc79ccbfb92f65852b660364c919862a75
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Append the part of the objects property data that is
known ad instantiation time to the object itself and
by that avoid creating a separate MemberData. Saves
some memory and should speed up object creation.
Currently implemented only for Object and ArrayObject.
Change-Id: I7693bf2f3a28fb718522398ebb94ac115e021fa4
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Now that the other method is gone, let's use
the shorter currentContext
Change-Id: I2a6fb3b77f83a1ffdf314ad29081e303d17030ed
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
This saves one pointer per allocated execution context.
Now every execution context that is pushed, allocates two
Values on the js stack. One contains the context itself, the
other one the offset to the parent context.
Things are a bit tricky for with and catch scopes, as those
are called from the generated code, and can't open a Scope
anymore. In addition, all methods iterating over the js
stack frames need to work with ExecutionContext pointers,
not ScopedContext's.
Change-Id: I6f3013749d4e73d2fac37973b976ba6029686b82
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Avoid the implicit push inside the execution context constructor
and rather make this explicit in the code.
Change-Id: I1bb0fb523fddbb273fc666370d619f55f49cd40a
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
One more step towards removing the class alltogether.
Change-Id: Ic9f6794eb3c5c6605ee43ad23a6d432ebbf321a1
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Conflicts:
src/qml/debugger/qv4debugservice.cpp
src/qml/jsruntime/qv4value_inl_p.h
src/qml/jsruntime/qv4value_p.h
src/qml/memory/qv4mm.cpp
src/qml/memory/qv4mm_p.h
src/qml/qml/qqmlnotifier_p.h
src/qml/qml/qqmlproperty.cpp
src/quick/items/qquickflickable.cpp
src/quick/items/qquicktextedit.cpp
tests/auto/quick/qquickwindow/BLACKLIST
The extra changes in qqmlbinding.cpp are ported from changes to
qqmlproperty.cpp that occurred in parallel with writeBinding() being
moved to qqmlbinding.cpp.
Change-Id: I16d1920abf448c29a01822256f52153651a56356
This is required, so we can safely access the vtable even while
we're marking objects during GC.
Change-Id: I34f56b61b4bca0d0742faf607eb5ab8b2c30685e
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Add some runtime methods to access properties of the scope
object directly (using the QmlContext), and generate proper
code to call those.
Change-Id: I0b29357c9a3b9ad53ba568ec6cb763e8ecb10f21
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Our generated code (JIT and interpreter) should operate on the
QML context to retrieve QML related things. That's better than
operating on 4 different temps.
So this commit introduces the QML context as a temp in the
code we generate for QML. The next commits will move things over
to use that context with specialized runtime methods instead of
using generic subscript/get calls on the different subobjects.
Change-Id: Ia05cf339de9cdd23003f35cf78ede17d2590f8de
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
qtdeclarative\src\qml\jsruntime\qv4engine.cpp(179) : warning C4172: returning address of local variable or temporary: dummy
Disable warning as using the address is intended.
Change-Id: Ide894a8dc2fb94f11d0455723c46567c84d91f8d
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
... and do a GC run when it exceeds a threshold. The issue with Strings
is that they hold on to QString instances that store the real content.
However, the GC only sees the light-weight JS handle, and doesn't take
the size of the backing content into account. So it could happen that
big QStrings accumulate in the heap as long as the GC didn't reach its
threshold.
The newly introduced unmanaged heap threshold is upped by a factor of
two when exceeded, and lowered by a factor of 2 when the used heap space
falls below a quarter of the threshold. Also grow the threshold if there
is enough space after running the GC, but another GC run would be
triggered for the next allocation.
There is a special case for Heap::String::append, because this method
will copy the data from the left and right substrings into a new
QString. To track this, append notifies the memory manager directly of
the new length. The pointer to the memory manager is stored in
Heap::String, growing it from 40 bytes to 48 bytes (which makes it still
fit in the same bucket, so no extra memory is allocated).
Task-number: QTBUG-42002
Change-Id: I71313915e593a9908a2b227b0bc4d768e375ee17
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Create a specialized QmlContext instead of re-using
a call context with a QQmlContextWrapper as activation
object.
This saves some memory and opens up the route to getting
rid of the context wrapper in a future commit.
Change-Id: I1591c73932a08564fddf5137ac05bbc6f31dd4d5
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Add a const several places and pass values by pointer, not
reference. This allows us to also get rid of some Scope's.
Change-Id: I3b4c29c22511c947d024457e7b8fa2aaa6bec55d
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Robin Burchell
Lars Knoll
Tasuku Suzuki
Simon Hausmann
Liang Qi
Ionut Alexandrescu
Kimmo Ollila
Erik Verbruggen
Erik Verbruggen
Anton Kudryavtsev
Christian Strømme
Christian Kandeler
Ulf Hermann
Erik Verbruggen
Simon Hausmann
Andy Nichols
Nobuaki Sukegawa
Jani Heikkinen
Maurice Kalinowski
Mathias Hasselmann
Thiago Macieira
Sérgio Martins
Friedemann Kleint
Ulf Hermann
hjk
Lars Knoll