| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Freezing the global object using a script is pretty slow,
esp. given that the script needs to be compiled as well. Rather
do it programmatically.
The old code actually had a bug that would only cause the global
object to be frozen, not it's children. The new code fixes this,
but doesn't completely freeze the objects. Instead it makes all
the existing properties of the global object and it's children
readonly, but still allows extending existing objects with new
properties.
Change-Id: I0d7331cdc89a0ac717b8ed3b1a490b2a3742de02
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>
|
|
|
|
|
| |
Change-Id: I17f202cf58dc355524b89a456e6722d41e444ca1
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
| |
There's only one place where we need to resize our member
data, namely when we call setInternalClass() on an object.
In addition, encapsulate the access to the memberdata better
in preparation for inline property data later on.
Change-Id: Ia34d0253d5d1792f1d7c4981556d78375fa7a755
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Valgrind pointed out that Transition::id was used in a conditional jump
or move. Fixes regressions of the following tests on OSX:
ch15/15.2/15.2.3/15.2.3.12/15.2.3.12-2-1 in non-strict mode
ch15/15.2/15.2.3/15.2.3.12/15.2.3.12-2-2 in non-strict mode
ch15/15.2/15.2.3/15.2.3.12/15.2.3.12-3-28 in non-strict mode
Change-Id: Ia959ff6f9fdac8d4cb37f54f670fdff4c8ba9a67
Reviewed-by: Simon Hausmann <simon.hausmann@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>
|
|
|
|
|
| |
Change-Id: Ia52f0e6db325aab37477d455f163487b319dce29
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
We shouldn't have taken a copy here. I'm not sure why I did that. Also sprinkle
Q_ASSERT fairydust around to try ensure that this doesn't happen again.
This was exposed when we started trying to delete the transitions again, which
were leaked in 6421f275286b3238fe1a7a5e909225251f3e8dbf.
Change-Id: Id9272db7f1863d1ccc5b1f48b6382c68ae0da9da
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
Accidentally introduced in 6421f275286b3238fe1a7a5e909225251f3e8dbf.
Change-Id: I45d121a661ba403d1d4573f98fe0944f05319290
Done-by: Slava Monich <slava.monich@jolla.com>
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
With this, we can now save one pointer per Heap
object.
Change-Id: I7f69193ff51c9fd9c5dbfba90aa1ebb3f93da2e6
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
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>
|
|
|
|
|
|
|
| |
Profiling of allocations shows that this is one of the most frequent allocators.
Change-Id: Iae0350c61bfe37381577642b9a93064377f45c1d
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
This is the only way we can support a GC that moves
objects around in memory.
Change-Id: I1d168fae4aa9f575b730e469e762bc5b5549b886
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|\
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
src/qml/jsruntime/qv4arraydata.cpp
src/qml/jsruntime/qv4context_p.h
src/qml/jsruntime/qv4globalobject.cpp
src/qml/jsruntime/qv4internalclass.cpp
src/quick/items/qquicktext_p.h
src/quick/items/qquicktextedit_p.h
src/quick/items/qquicktextinput_p.h
Change-Id: If07e483e03197cb997ef47a9c647a479cdb09f4c
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
If there are deep object hierarchies connected to an InternalClass the
recursive nature of the destroy() method could lead to a stack
overflow. By changing the recursion into an iteration this is avoided.
Change-Id: I6667f268a366749c2dbc5f7147882388f192a9c5
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
| |
| |
| |
| |
| | |
Change-Id: I921cd8129acb47cffd58be9551a6925c4f296d41
Reviewed-by: Simon Hausmann <simon.hausmann@digia.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>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
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>
|
|/
|
|
|
|
|
| |
And do the same change for ArrayData.
Change-Id: Ia1ae56bd0ff586c9b987e15af7a53f395a37054a
Reviewed-by: Simon Hausmann <simon.hausmann@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>
|
|
|
|
|
|
|
|
| |
Remove the Ref classes, as they won't be required
anymore once Managed and Managed::Data are separated.
Change-Id: Ic6bec2d5b4ecf2595ce129dbb45bbf6a385138a5
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
| |
Change-Id: Ifa9aac63fdb270fb449f11832a1792caeb6c6724
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
| |
Change-Id: I1fcccd535ed356e736468337bd450270cd290044
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
| |
Change-Id: I95dcdda8c68e2a5c36244798c8c10dcfdd69d2c2
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
| |
This prepares for moving over to a d pointer scheme,
where Managed subclasses don't hold any data directly. This
is required to be able to move over to a modern GC.
Change-Id: I3f59633ac07a7da461bd2d4f0f9f3a8e3b0baf02
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
As per reported bug, we have to protect ourselves against potential loops
and can mark the internal classes much simpler by just walking through
the memory pool they were allocated in.
Task-number: QTBUG-38299
Change-Id: I3ae96e8082e76d06f4321c5aa6d2e9645d2830a0
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With a real prototype chain it can happen that an internal class' prototype's
class itself has a prototype. Therefore the first transition on the empty class
is a PrototypeChange one, but the class the transition leads to may have
PrototypeChange transitions itself, which weren't marked.
There are multiple solutions to this, but this patch is the minimal fix by
recursing fully through the internal class tree. That way it's easier to
back-port the fix also into 5.2.x based branches.
Task-number: QTBUG-37834
Change-Id: I901b13a2663fbad5844003ca5752f2f304de320c
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
|
|
|
|
|
|
|
|
| |
Move the allocated member data into the garbage collected
area, so that we can avoid using malloc/free for it.
Change-Id: I20625efa67ecd60238568742b74854b0c8cb2e3e
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
This cuts the memory required to store properties
in an object in half for the common case. Accessor
properties require two slots inside memberData,
but data properties only one.
Change-Id: I0bab1b88ca9ed5930abf065c77c89985b9ed5320
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
Reserve two slots in the internal class for accessor
properties. This opens up reducing the default storage
required per data property to one Value. In practice
this implies cutting the required memory in half.
Change-Id: Ifed897852bbdfd810018f0d6b049fca6690ead2c
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
| |
vtable changes need to happen when the internal class
is being constructed, not later on.
Change-Id: Ibb9515745c9fc3507a5a90b4cc50a33e2e0d3f99
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Encapsulate accesses to the current context, and rework
the way we push and pop this context from the context
stack.
Largely a cleanup, but simplifies the code in the long term
Change-Id: I409e378490d0ab027be6a4c01a4031b2ea35c51d
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
| |
Remove all the calls to setVTable that were in performance
critical parts of the code. This now brings performance
back to the level we had with the vtable inlined in the
Managed objects.
Change-Id: I76317cc5c53b5b700d1d3883b954407142a4c424
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
| |
This saves one pointer per object, and willmake other optimizations
easier in the future.
Change-Id: I1324cad31998896b5dc76af3c8a7ee9d86283bfe
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Due to a inefficiency in our InternalClass implementation,
we were not sharing the string and attribute table between
internal class instances.
This was extremely inefficient with the Qt object, as it created around
1000 internal classes with large string and property tables. With the
patch these tables are now shared.
Reduces memory consumption of a QQmlEngine instance from around 6.5M
to a couple of 100k.
Change-Id: Ib763f31deca0808c000ac2c30aa0b05e806bda40
Reviewed-by: Erik Verbruggen <erik.verbruggen@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>
|
|
|
|
|
| |
Change-Id: Ic0492fbe31a1e134674bc6c20381f735dd6d5b7a
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
| |
Don't have an implicit cast operator to Returned<T>
anymore, and return a T* from the operator->()
Change-Id: If4165071b986bfc84a157560d94d39c2dcfbc9e1
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
| |
Change-Id: I8e2dad0e9e34c5a549952bc0765cd57f6aa8aadf
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
| |
Rather use the correct internalClass directly when constructing
the objects.
Change-Id: I8e916f1ce8f83d291c08ca6332fe85b1f57b90b5
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The prototype is actually the same for most objects. By
moving it into the internal class, we can save 8 bytes
per object, as well as allowing for some future
optimizations.
Also fix a bug in the implementation of the Error
prototype objects.
Change-Id: I4d4b641055f644a9b088f27be34bfdb0446279b7
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
|
|
|
| |
Change-Id: Ifa97d3354a7a7afadf70f9ba540716bd5b1eef44
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|
|
Move the v4 engine classes from a subdir of qml/qml into
two subdirs (compiler and jsruntime) of the qml module
Remove an unsued qv4syntaxchecker class, and move
the moth code directly into compiler.
Change-Id: I6929bede1f25098e6cb2e68087e779fac16b0c68
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
|