| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
| |
The StaticData flag needs to be saved to disk, but removed again
afterwards so that we can free() the malloc'd data. This also allows us
to avoid copying all the data into a byte array before saving.
Change-Id: I96513f8d98acf0ea0b4514d96376b487e8444917
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
| |
We need a CompilationUnit that only holds the data needed for
compilation and another one that is executable by the runtime.
Change-Id: I704d859ba028576a18460f5e3a59f210f64535d3
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|\
| |
| |
| |
| |
| |
| | |
Conflicts:
src/qml/compiler/qqmlirbuilder.cpp
Change-Id: I2cfda470515e2df778ad3c89105c07344af07c6d
|
| |\
| | |
| | |
| | | |
Change-Id: I3eb5d1affe64b6ae709d1154cc37de91db3816b6
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Some compilers seem to miscompile this construction. Furthermore, it
doesn't really add to the readability of the code. Inline the code in
question at the only place it's used and avoid most of the const_cast by
adding a non-const accessor to CompiledData::Unit.
Fixes: QTBUG-75392
Change-Id: I015317f28a92817d08d616cc35956745758d7847
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
The dependencies are only hashed if a dependencyHasher is given. This is
generally not the case when compiling ahead of time. There is also no
need to hide the declaration of DependentTypesHasher from the bootstrap
code.
Change-Id: I0ea74c3079656ce1fe353956999820916c8ff626
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We don't really need the dynamic strings at compile time, but having
them defined simplifies some code.
Change-Id: Ibcfaae7834f8aa63918da6787d222fe71657c4ee
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
We don't need it to build the IR and we can drop a few checks for
V4_BOOTSTRAP this way.
Change-Id: I9464e65528c70c42ebc8ddad576eaab001dc9d2f
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|/ /
| |
| |
| |
| |
| |
| |
| | |
The "early" compilation doesn't use it and we can get rid of a few
V4_BOOTSTRAP checks this way.
Change-Id: I1c4845aba445b105ddace0b6810e0e5c28a25b29
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|\|
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
.qmake.conf
src/qmltest/quicktest.cpp
tests/auto/qml/qmlcachegen/qmlcachegen.pro
Change-Id: I70e96e8817d59647f876b8b77b30cdeede8f0662
|
| |
| |
| |
| |
| |
| |
| |
| | |
Otherwise the resulting files differ subtly.
Fixes: QTBUG-74532
Change-Id: I12b4f1ba6dda781d63ad50cce87861ba24582bf7
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|\|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Conflicts:
src/qml/compiler/qqmltypecompiler.cpp
src/qml/compiler/qv4bytecodehandler.cpp
src/qml/compiler/qv4codegen.cpp
src/qml/compiler/qv4compileddata_p.h
src/qml/compiler/qv4compiler.cpp
src/qml/compiler/qv4instr_moth.cpp
src/qml/compiler/qv4instr_moth_p.h
src/qml/jit/qv4baselinejit.cpp
src/qml/jit/qv4baselinejit_p.h
src/qml/jsruntime/qv4function.cpp
src/qml/jsruntime/qv4vme_moth.cpp
Change-Id: I8fb4d6f19677bcec0a4593b250f2eda5ae85e3d2
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
After enabling lookups in QML files, we can remove all the code that
tries to deal with (type) compile time detection of access to id objects
and properties of the scope/context object. This also allows removing
quite a bit of run-time code paths and even byte code instructions.
Task-number: QTBUG-69898
Change-Id: I7b26d7983393594a3ef56466d3e633f1822b76f4
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When resolving names in the context of QML bindings, we now direct
runtime access to QQmlContextWrapper::resolveQmlPropertyLookupGetter. At the
moment this does basically the same as Runtime::method_loadName, which
we called earlier. However this now provides the opportunity to optimize
lookups in the QML context in a central place.
When performing a call on a scope or context object property, we also
did not use a CallName() instruction - which would have gotten the
thisObject wrong - but instead we use a dedicated
CallScopeObjectProperty and CallContextObjectProperty instruction. These
rely on identifying these properties at compile time, which goes away
with lookups (and also doesn't work when using ahead-of-time
compilation). Therefore the qml context property lookup is using a
getPropertyAndBase style signature and
Runtime::method_callQmlContextPropertyLookup uses that.
For the tests to pass, some error expectations need adjusting. In
particular the compile-time detection of write attempts to id objects is
now delayed to the run-time.
The old code path is still there and will be removed separately in the
next commit (as it is massive).
Task-number: QTBUG-69898
Change-Id: Iad1ff93d3758c4db984a7c2d003beee21ed2275c
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
|
|\|
| |
| |
| |
| |
| |
| | |
Conflicts:
src/qml/compiler/qv4codegen.cpp
Change-Id: I66b7db42bf208855889094ace0267326595ce03c
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Result objects are rather large, 96 bytes here. In a recursive algorithm
such as our parser, we should not keep too many of them on the stack.
Also, the size of Reference can be reduced by employing a bit field
rather than a number of booleans.
Also, try to convince the compiler to inline the accept() functions. The
extra stack frames those create are unnecessary.
Task-number: QTBUG-74087
Change-Id: I5c064491172366bb0abef99ffe9314080401a7d1
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| | |
So now the FixedPoolArray can be re-used in other places.
Change-Id: I0e0504892944722a0c18f207dc6400d5b314f6ae
Reviewed-by: Ulf Hermann <ulf.hermann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This gives us the opportunity to map the JavaScript null to QVariant's
concept of isNull().
[ChangeLog][QML] Assigning JavaScript null to incompatibly typed
properties generates a compile error now, rather than a runtime error.
Fixes: QTBUG-72098
Change-Id: I72fd1c30d84128c774230eaaea10455b2a0e064c
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|\|
| |
| |
| |
| |
| |
| | |
Conflicts:
tests/auto/qml/debugger/qqmldebugjs/tst_qqmldebugjs.cpp
Change-Id: Ic1dace832ad4b29023d24808b8617b5dcc915eb5
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Since change 9333ea8649838d7e0400b0e94c8cbd4fa5d216b0, we lookup
properties in the QML context object before the global object to
have proper scoping rules for QML.
Unfortunately this lead to a performance regression when using global
properties such as Math in imported script files, as the lookup would
always go through the qml context first.
This can be fixed, as we know that the global object is frozen in qml
mode, and the standard names of properties in the global object are
illegal to use in QML. So simply check for those names in the code
generator and create lookups into the global object for those.
Change-Id: I4b2089178c9e5f9440abdfd834cf7d92c3c0e2c3
Fixes: QTBUG-71591
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
|
|/
|
|
|
|
|
|
|
|
|
| |
Fixes -Wclazy-unused-non-trivial warnings, in preparation for using
gcc/clang's warn_unused attribute in all containers.
Maintainers please check if some unused variable isn't hidding an
actual bug.
Change-Id: I83c5a11aa2328db3dce4c6c402295d86ef297c83
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Commit 939014cb9cad2f3357f47b28a4580397c17b913c improved performance of
property lookups beyond the scope object, with the unfortunate
side-effect that the previously polymorphic lookup of methods broke.
Fix this by moving the handling to the caller side and falling back to
the string lookup for functions.
Fixes: QTBUG-71204
Change-Id: I2d9924034a9c14e7d161fa49d51b1f876ab5bc0f
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
Don't return a name reference if we fail to lookup something in
the scope object, but rather continue trying in the context
object.
Task-number: QTBUG-70315
Change-Id: I73f8aa7b648320434ef0ef37a4c12dca1eb7b209
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
| |
Doing this in different ways could lead to broken object layout or other
bad things later.
Change-Id: Ifd7b4e535d6fad4a867a6ea0c594df18247ee943
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This is a straight-forward hook into the module implementation in
QV4::ExecutionEngine. Modules are pre-compiled in the QML type loader
thread. That thread keeps track of all pending loading scripts through
the type loader's m_scriptCache. Once a module is compiled, it's
thread-safely registered with the execution engine.
Script instantiation and evaluation is done solely in the QQmlEngine's
thread.
ES Modules are identified in imports as well as qmldir files by the .mjs
extension.
Change-Id: Ie9c59785118afcb49f43a1e176a9f7db00f09428
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
| |
We only support 'pragma Singleton' currently, so there is no need
to parse the right hand side of pragma as a member expression.
Change-Id: Ic0dcbedb52cb58db2fd1cc099f14fd399b7162e3
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Separate the qml data (objects/imports) from the general compilation
unit data. It's only the former that needs to be re-generated as part of
the type re-compilation and by separating it we can allocate memory just
for that and keep using the mmap'ed general unit data for everything
else (including byte code).
Another upside of this change is that it allows eliminating the recently
introduced concept of a backing unit again.
Saves ~149K RAM with the QQC1 gallery.
Task-number: QTBUG-69588
Change-Id: Ie88a4286feb7e2f472f58a28fa5dd6ff0a91c4b6
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
| |
This will make it easier to move the data out of CompiledData::Unit and
into a separate data structure.
Change-Id: I32e6233a66f2279b44cc06ef7c3505db4a565f98
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
| |
Change-Id: Ib9028ffa8c7d014598da60bfe44e004031bf6017
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
For AOT loaded cache files, the CompiledData::Unit we produce is
ephemeral by design. That same applies to the strings, where we don't
really need to write out a new string table but we might as well just
keep the QStringList around. That avoids copying all the string data and
saves another 21K RAM with the gallery. It might even save a little more
if some of the new strings are truly shared via implicit QString
sharing. Also their "extraction" via stringAt(index) becomes as fast as
for the strings that come from mmap'ed memory - with no copying
involved.
Task-number: QTBUG-69588
Change-Id: I06cf0f083e3b8d3c5dbabb22beb0711f88fc8692
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
| |
Since the callers always end up assigning the returned unit pointer back
to the compilation unit's data member, we might as well do that inside
the function and return void instead.
Change-Id: I7a7f3e7a0c89ffe2f9474149fcf61736609b363d
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
| |
Consolidate the old "createUnitData" code that used to be a virtual
function into the one call site and avoid a malloc that way.
By the way of that, the string table gets a guard to prevent accidental
registerString() calls at a later point.
Change-Id: Ia3f8d3e874b7733822e6942e411c940ec60d95a5
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
When loading a pre-compiled cache file, the strings contained therein
will remain available in memory since commit
7dcada48d2435e8ceb0cc8a6771f79b76979e11f. While for aot built cache
files we may have to add new strings (for example for signal handler
parameters), we can re-use the existing strings by omitting them from
the intermediately created string table.
This saves ~283K RAM with qtquickcontrols1 gallery.
Task-number: QTBUG-69588
Change-Id: I8ea807f6dea4cc35d8b7e5f7329809ed1cd12880
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
| |
Replace use of CompiledData::Unit::stringAt with
CompilationUnit::stringAt as central place, in preparation for allowing
derived compilation units to retrieve strings from base units.
Change-Id: Ifafe6c18a9fd8d1e2259a60c17200124869b4c2d
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Move the translation data out into a separately indexed table, which
allows to shrunk the value union down to 4 bytes, together with the
previous commit.
Saves ~4k with examples/quickcontrols/extras/flat/Content.qml and ~37K
RAM with the gallery.
Task-number: QTBUG-69588
Change-Id: Ia5016b072320ebb6b8fcfbb4dad128d53c901c74
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
|
|
| |
Store doubles in bindings in the constant table of the compilation unit
instead of each binding. This removes one of the two 8 byte members of
the value union and also allows for sharing of constants throughout a
.qml file.
Change-Id: I1d7daafdb7f24e34c14cd160d2dcb2c5aaac1c50
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
|
| |
This allows updating the constants table when the unit data is set /
changes and removes the tie to the engine.
Change-Id: Ice553650390589e30e18421c4e55422a55d0df89
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|
|
|
|
|
|
| |
This is useful for profiling the sizes of cache files.
Change-Id: I831d0a20c4eda3d707221a61fa7290857605c7fb
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
|\
| |
| |
| | |
Change-Id: I0127f2c16de1d930bdd8cbccd42ec6785f31ab96
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This code path was previously missed, leading to some AOT-compiled types
generating errors of the form:
Unable to assign [undefined] to int
[ChangeLog][QtQml] Fix QML declared enums with CONFIG+=qtquickcompiler.
Change-Id: Ib46a2b2505aa3863f091a6ccdebf8425e62fc38f
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
We used to set the name later on when calling
Codegen::defineFunction(), but this is too late in some
cases, especially when using the name to determine if
this function could be a QML signal handler.
Change-Id: Ie620a65ac8f17906cd9eba338cbdd3563004375d
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
|\|
| |
| |
| |
| |
| |
| | |
Conflicts:
src/qml/qml/qqmltypeloader.cpp
Change-Id: If9874faf23e813b05ee4cf042133a528d73ba72a
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
When loading a .js file without QQC, we scan the sources and use the
ScriptDirectivesCollector to extract things like .pragma library or
.import ahead of time. That information is passed on to the compilation
unit generator for serialization. When compiling .js files ahead of
time, we also used the same collector, but we forgot to save the data
into the right location before serialization, so we essentially lost the
imports. This patch fixes that by centralizing this code into the
ScriptDirectivesCollector itself.
[ChangeLog][QtQml] Fix regression with .import in .js files not working
when using CONFIG+=qtquickcompiler.
Change-Id: I5413c14b1b8bd3114a997011534fe55cdb7634aa
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This is still to some extend work in progress as
lexically scoped for loops won't yet do the right
thing.
let and const variables are still accessible before
they are declared, and the global scope doesn't yet
have a proper context for lexically declared variables.
Change-Id: Ie39f74a8fccdaead437fbf07f9fc228a444c26ed
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
So far, when instantiating QML bindings, the node parameter
could be the same or a child of the body. This will break
badly when we introduce lexical scopeing as that node
could be an AST::Block that opens it's own context.
Changing this requires some smaller adjustments to our autotests,
as error locations will now be slightly different (pointing to
the beginning of the binding, not the beginning of the RHS of
the binding).
Change-Id: I2c536a4fe6d8b549a138cc7967ef034eb2523f3b
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
And make it an enum class. The new name fits better, as it's mainly
used to determine the type of the context when parsing. Also already
added the 'Block' value that will be needed.
Change-Id: I70d963b6a0b22db1a3c607cce6bdd2054b29e000
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Commit da5fffbd34d8be68f8ee4c649881dbb673c9c0a5 introduced deferencing
of the formals parameter list that can be a null pointer if the declared
function has no parameters.
Change-Id: Id7dce0f78b16266e672f0ae430ee4f979de5734d
Reviewed-by: Erik Verbruggen <erik.verbruggen@qt.io>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This basically updates all grammar rules in the
qqmljs.g file to be in line with the ES7 specification.
Some special handling for the lookahead rules appearing
in the spec was needed and is implemented through empty
lookahead rules in the grammar, that might push an
additional token into the token stream.
Renamed some classes in the AST to be in line with
the names used in ES7, and removed some other ones
(SourceElements) that are no longer used.
The ES7 grammar rules contain lots of variations of
the base rules (with In/Return/Yield/Default suffixes).
With the exception of the In and Default rules, these
are implemented through state tracking in the parser
and lexer.
Change-Id: I4017d97cd050ed816c1dad11833e882cba30801a
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Destructuring objects works, but arrays are not
yet supported.
Change-Id: I61e917e1964e3c719f71b8f11d194e09dfe288c2
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|
| |
| |
| |
| |
| |
| |
| |
| | |
It's not required and makes the refactoring of the function
parameter handling required for ES6 a lot harder.
Change-Id: I0273449219ea9f3324353d6cf9166c1b58eb914b
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
|