On platforms where char is signed, like x86, the following is an error
(narrowing conversion):
unsigned char x[] = { '\xc3' };
Change-Id: I495bc19409f348069f5bfffd15518f9ef4e43faf
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
It's more optimised and it vectorizes better, due to one level of
indirection fewer.
Change-Id: I495bc19409f348069f5bfffd1551e85092ed8dc2
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
...not just the debug flag. The information is saved outside of the CBOR
map for two reasons:
1) removing the hack that depended on how QCborStreamWriter and
TinyCBOR internally work, allowing for the extra parameter to be
written directly. We wouldn't be able to use that hack anyway and would
have needed a further, uglier hack to encode a byte whose value we
don't know.
2) outside the map, this information can be parsed more quickly and
then we can discard any plugins we shouldn't actually load.
Since we're doing this for a flag, I decided to move the Qt version
there too for reason #2.
Change-Id: I61ecce6b1324410bbab4fffd153d4e5fc696d19e
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
In preparation for Qt 6 deprecating the binary JSON format. Also reduces
the size of the metadata a little: for the xcb platform plugin, it went
down from 264 bytes to 138; for the jpeg image plugin, it went from 320
to 135.
I've had to change the signature so older versions of Qt won't try to
parse the CBOR data as Binary JSON. Unfortunately, before QJsonDocument
could get a chance to reject it, qJsonFromRawLibraryMetaData() needed to
allocate memory and that causes crashes with Qt < 5.11.2.
Change-Id: Ieb48f7c0dd0e4e0fb35efffd153bee34e16ce347
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Adds an enumName to QMetaEnum to carry the name of the enum since for
flags that doesn't match the name of the Qt type, but is needed if the
flag is scoped.
Change-Id: I1c0f77eb9e40e6fd1eb6a59bea77caf0f33fcf43
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Keep the original class name around for a little longer so we can
generate the correct scoped enum in the moc output.
Task-number: QTBUG-47652
Change-Id: Ib5934316fa786cc475335b03c86b8ec2dc239055
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
This follows on from a232251992 which
covered a similar instance of this. As with that change, we should not
abort the compilation, just ignore it.
Task-number: QTBUG-63772
Change-Id: Ide958080a90f43ed19edd8a320e7d45de1c96821
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The referenced static meta object for the superclass might be in a
different DLL. In this case, the whole QMetaObject can't be initialized
all via preinitialized data in the data section of the binary, but must
run code at runtime to fill in the value of the dllimported pointer.
In these cases, both GCC and MSVC initialize as much as possible statically,
while only filling in the dllimported values (QMetaObject::d::superdata)
at runtime. Clang, on the other side, initializes the whole struct
at runtime if some part of it needs runtime initialization, leaving
the struct completely uninitialized before constructors are run.
In C++, there are no guarantees for in what order constructors in
different translation units are executed. This in particular means
that there are no guarantees as to whether qRegisterWidgetsVariant()
in qwidgetsvariants.cpp runs before or after the runtime initialization
of QWidget::staticMetaObject.
With GCC and MSVC, this doesn't seem to have mattered since only the
superdata pointer of the staticMetaObject was uninitialized - everything
else was initialized, and the superdata pointer doesn't seem to be
accessed during qRegisterWidgetsVariant.
With clang, the whole staticMetaObject is uninitialized, unless the
staticMetaObject has been initialized before (and the initialization
order is undefined).
By setting a manual priority (which is a GCC extension that also
clang supports) for the staticMetaObjects, we can be sure that
these are initialized before the actual explicit constructor
invocations (without any explicit initialization priority) that
can access the staticMetaObjects.
Change-Id: I64a82f12d690528567509791bae088b6304e189b
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The C standard library functions cannot handle UTF-8 filenames. Instead,
we need to use the wide-character versions which accept UTF-16 input.
Task-number: QTBUG-65492
Change-Id: If4b3b4eeeec4f3bbb428b8f6b0311a65d01463b0
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
In Qt 5.7, it was possible to call moc "-bfoo.h" or "-b foo.h" and it had the
same effect. With the port to QCommandLineOption, we broke the -b option as it
was not annotated as a short option.
(Regression in a7e3c17e75)
Task-number: QTBUG-63706
Change-Id: I161d0f1a4e65d129063b5e8431802257677da19d
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
_a[1] was reinterpret_casted twice in a row, which triggers
clang's warning undefined-reinterpret-cast:
"dereference of type '_t *' (aka ...) that was reinterpret_cast
from type 'void **' has undefined behavior "
only the last reinterpret_cast is kept
Change-Id: I71d52c5ff08c674003aec29f8a907c90905c0d4c
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
qt_metacast is not const so there is no need to use const_cast.
This fixes a warning in generated code.
Task-number: QTBUG-63352
Change-Id: I0c37442ac268a654316bc0e7e04f77fb51cae019
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
'using' is recommended by C++ Core Guidelines for improving readability:
https://github.com/isocpp/CppCoreGuidelines/blob/master/CppCoreGuidelines.md#Rt-using
In the case of generated code it can be useful to leave
moc files unchanged when modernizing code with clang-tidy's
'modernize-use-using'.
Change-Id: Iabb4de2aa8d2f9396d8c8d4ee21f80fffff1dadc
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
This is shown in the file properties, and in crash dialogs.
Task-number: QTBUG-61970
Change-Id: Icf12fabb7fad30638c4e94c8ea8729c9106d01d9
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Present for example in the Windows's Atlbase.h header.
We should not abort the compilation, just ignore that construct
Task-number: QTBUG-56634
Change-Id: Id6e4c9f03cb1cef46e330f4fbcae80ce4f3730c6
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
Normaly, in C++ It's not valid to define a keyword, but it turns out that some
system header do, so we just silently accept it.
[ChangeLog][moc] moc no longer errors out if a C++ keyword is #define'ed
Task-number: QTBUG-61204
Change-Id: Ia4d3ff9c77b6ff261b6140c220cfb81bd13f1d6d
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
It seems this change is preventing us to integrate qt5.git in '5.9'
This reverts commit c3030d7163.
Task-number: QTBUG-61204
Change-Id: Id98afaa23be0a8dd6f2c54a899f46542c65436aa
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Because we obviously don't support Microsoft's extensions in C++. This
is required because some MS headers have code that isn't proper C++,
like iso646.h:
#if !defined(__cplusplus) || defined(_MSC_EXTENSIONS)
#define and &&
Do not pass /Za to MSVC to generate moc_predefs.h, because this option
is incompatible with compiler options like /fp:fast that may be
user-specified.
This reverts commit e1a70ce4 and re-fixes the issue similarly to commit
d72ac3f3.
Task-number: QTBUG-58391
Change-Id: I5c0143283afed09f98200806c87259c039c00ae1
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@qt.io>
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@qt.io>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Limitation is that the signal needs to be parameter-less
[ChangeLog][moc] moc now supports NOTIFY signals of parent classes in Q_PROPERTY
Change-Id: Iad64c96c3ec65d4be8ad9ff1a9f889938ab9bf45
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Brett Stottlemyer <bstottle@ford.com>
... to reduce text size by replacing QStringLiteral with QL1S.
Change-Id: I8166282d915f081ba816989039451466ec29e208
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
By adding std::move where it makes sense.
This is not only good for move-only types, but for any type which
can be moved as it saves copies of the return value in any case.
[ChangeLog][moc] Move-only types are now supported as return types
of signals and slots.
Change-Id: Idc9453af993e7574a6bddd4a87210eddd3da48a9
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Even if it is normaly not used, templated code might still try to access it
Task-number: QTBUG-59414
Change-Id: I9f7aadd714843059c8f89cdac48c60a3e2ca7294
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
[ChangeLog][moc] Added Support for C++17 nested namespaces
Change-Id: Ib83fc5bf48f66546fa97b49710582fbf9c984503
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
These files are unused.
They still contains Qt3 code (QCString for example)
Task-number: QTBUG-59302
Change-Id: I661f08db4092311df1d6c7d54b9780d86adab8aa
Reviewed-by: Simon Hausmann <simon.hausmann@qt.io>
This is unnecessary (as the fallback behavior is return _id anyway),
and it makes coverity unhappy.
Coverity-Id: 173293
Change-Id: I91c016f3ed363319c6413ab3c2688698faf4f10f
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
We already check methodList, and as it is comprised of signalList,
slotList and methodList from cdef, this is redundant.
Change-Id: I0d1791f821134060aa20a8ea9b57f049b2701bf9
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
nullptr can be used directly in the Qt code since Qt 5.7.
Use it in the generated code for consistency.
Change-Id: Ic2c37e2757c9cebb3bccb8eb0f2c808fc0e83e35
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@qt.io>
This re-fixes commit d72ac3f35f, which
simply removed the #define but did so at the wrong place. Instead of
forcing the macro to be removed, let's simply not have it defined in the
first place.
Change-Id: Ie6dbad9bbbd9488887e8fffd148dd67d9a31b32e
Reviewed-by: Jake Petroules <jake.petroules@qt.io>
Code generated by moc very often calls deprecated functions, like
deprecated slots, signals and property getters and setters. There's no
way around that unless the class in question is willing to break binary
compatibility, so those warnings are actually harmless.
Change-Id: Iaeecaffe26af4535b416fffd1489d1a98ef8b34a
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
This improves moc performance on Windows where file-stat'ing is slow
and where the number of project include paths to search is often high
because project third-party headers are installed in separate
directories rather than a shared include path such as /usr/include.
In a real project of non-trivial size it reduces the total
from-scratch build time of an optimized build using CMake+ninja with
32 cores by 11% from ~11m35s to ~10m15s.
Change-Id: Ieed59646927ed75c55ed9efa97600c328b2fed2c
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Nested namespaces are quite common, therefore moc should support them.
Task-number: QTBUG-55415
Change-Id: I756cab36d498eb4342b402d255836d5d30f07b30
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Regression introduced in c32ef0a725
The expansion vector can be empty, in that case it is not allowed to
call constLast()
Task-number: QTBUG-55853
Change-Id: I47aa8eb7507ee91662215df42b4a66eebaa32bb5
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
Parsing the other files will (re)generate the same metaobject info in two places
Change-Id: I8984ed30751a7587de870f55dd427f067d1b2495
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Replaces our mix of comments for annotating intended absence of break
in switches with the C++17 attribute [[fallthrough]], or its earlier
a clang extension counterpart.
Change-Id: I4b2d0b9b5e4425819c7f1bf01608093c536b6d14
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Q_NAMESPACE is useful to add Q_ENUM_NS/Q_ENUMS, Q_FLAG_NS/Q_FLAGS and
Q_CLASSINFO to a namespace.
[ChangeLog] Added Q_NAMESPACE which can be used to add Q_ENUM_NS/
Q_ENUMS, Q_FLAG_NS/Q_FLAGS and Q_CLASSINFO to a namespace
Task-number: QTBUG-54981
Change-Id: Ic61b972794063e77134681fb347d6c4acddcdb44
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Conflicts:
configure
5.7 now supports clang on android; but dev re-worked configure
src/gui/kernel/qevent.h
One side renamed a parameter of a constructor; the other added an
alternate constructor on the next line. Applied the rename to both
for consistency.
tests/auto/tools/moc/tst_moc.cpp
Each side added a new test at the end.
.qmake.conf
Ignored 5.7's change to MODULE_VERSION.
configure.json
No conflict noticed by git; but changes in 5.7 were needed for the
re-worked configure to accommodate 5.7's stricter handling of C++11.
Change-Id: I9cda53836a32d7bf83828212c7ea00b1de3e09d2
Conflicts:
src/corelib/io/qtemporarydir.cpp
One side encapsulated a repeated piece of #if-ery in a local define;
the other added to the #if-ery. Made its addition to the other's.
src/corelib/kernel/qeventdispatcher_unix_p.h
One side moved some members into a struct; this collided with a #undef
check that neither side now has. Discarded the #undef part.
src/gui/opengl/qopengltexturehelper_p.h
5.7 deleted a bunch of methods; not clear why merge got confused.
src/tools/moc/moc.cpp
One added a name to the copyright header; another changed its URL.
Change-Id: I9e9032b819f030d67f1915445acf2793e98713fa
We should not call prev() if we had already reched the end.
Task-number: QTBUG-54815
Change-Id: I56bc86880a0dbfdce57fc4a08e5950f2ff3a5958
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
C preprocessors augment their standard list of include paths from the
environment: Unix preprocessors use $C_INCLUDE_PATH (for C) and
$CPLUS_INCLUDE_PATH (for C++), plus CPATH for both, whereas MSVC uses
the an environment variable simply called "INCLUDE". Handling this for
MSVC is particularly important because the VCVARSALL.BAT script sets the
necessary #include paths in the environment for important things.
Without that being parsed, moc won't find some #defines, like
WINAPI_DESKTOP_FAMILY.
[ChangeLog][moc] qmake and moc now cooperate to use the Visual Studio
environment variables (set by the VCVARSALL.BAT script) to find system
include files. A possible consequence is that moc parses application
headers slightly differently, depending on #if conditions that depended
on macros that previous versions had not seen #define'd. Implementers of
other buildsystems are advised to pass the --compiler-flavor=msvc option
to moc.
Change-Id: I7e06274214d1939b0124e5b4bf169cceaef9ca46
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
In order for moc to properly parse #ifdefs and family, we've had
QMAKE_COMPILER_DEFINES as a list of pre-defined macros from the
compiler. That list is woefully incomplete.
Instead, let's simply ask the compiler for the list. With GCC and
family, we use the -dM flag while preprocessing. With ICC on Windows,
the flag gains an extra "Q" but is otherwise the same. For MSVC, it
requires using some undocumented switches and parsing environment
variables (I've tested MSVC 2012, 2013 and 2015).
The new moc option is called --include to be similar to GCC's -include
option. It does more than just parse a list of pre-defined macros and
can be used to insert any sort of code that moc needs to parse prior to
the main file.
Change-Id: I7de033f80b0e4431b7f1ffff13fca02dbb60a0a6
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Conflicts:
qmake/library/qmakebuiltins.cpp
qmake/library/qmakeevaluator.cpp
qmake/library/qmakeevaluator.h
qmake/project.h
QMakeEvaluator:
* evaluateConditional(): one side changed return type, the other
changed a parameter type.
* split_value_list(): one side changed a parameter adjacent to where ...
* expandVariableReferences(): ... the other killed one overload and
changed the survivor
src/corelib/io/qlockfile_unix.cpp
One side changed a #if condition, the other moved NETBSD's part of
what it controlled.
src/corelib/tools/qdatetime.cpp
One side fixed a reachable Q_UNREACHABLE in toMSecsSinceEpoch(), the
other moved it from the private class to the public one, in the midst
of the "short date-time" optimization, which confused diff entirely.
One side changed a QStringLiteral to QLatin1String, the other rewrote
adjoining code.
src/network/kernel/qauthenticator.cpp
Both rewrote a line, equivalently; kept the dev version.
src/platformsupport/fontdatabases/mac/qfontengine_coretext.mm
src/platformsupport/fontdatabases/mac/qfontengine_coretext_p.h
One side changed #if-ery that the other removed.
tools/configure/configureapp.cpp
One side added a check to -target parsing; the other killed -target.
tests/auto/testlib/selftests/expected_cmptest.lightxml
tests/auto/testlib/selftests/expected_cmptest.teamcity
tests/auto/testlib/selftests/expected_cmptest.txt
tests/auto/testlib/selftests/expected_cmptest.xml
tests/auto/testlib/selftests/expected_cmptest.xunitxml
Regenerated using generate_expected_output.py
I note that quite a few other expected_* come out changed, now.
There was no git-conflict in
src/widgets/kernel/qformlayout.cpp
but it didn't compile; one side removed some unused methods; the other
found uses for one of them. Put FixedColumnMatrix<>::removeRow(int)
back for its new user.
Change-Id: I8cc2a71add48c0a848e13cfc47b5a7754e8ca584
Conflicts:
qmake/library/qmakeevaluator.cpp
One side changed the iterator to use ranged-for, the other changed its
body; they only conflicted because the latter had to add braces around
the body, intruding on the for-line. Trivial resolution.
Change-Id: Ib487bc3bd6e3c5225db15f94b9a8f6caaa33456b
Do not increment 'data' past the buffer in case of invalid token.
Remove the left over qDebug so we can make a test.
Task-number: QTBUG-54609
Change-Id: I8f0dd3381fbdea3f07d3c05c9a44a16d92538117
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
"foo" ## "bar" doesn't make a lot of sense, but MSVC allows them
(although gcc errors out on them). Simply ignore the ## in this case
instead of aborting with an error.
Fixes parsing of the Windows winsock2.h header.
Task-number: QTBUG-54560
Change-Id: I84cd5fbb56a006cf379430708c955cf0da475cff
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Because we obviously don't support Microsoft's extensions in C++. This
is required because some MS headers have code that isn't proper C++,
like iso646.h:
#if !defined(__cplusplus) || defined(_MSC_EXTENSIONS)
#define and &&
Change-Id: Ib57b52598e2f452985e9fffd145a3d4ec28e773d
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
make the 'cleaned' more robust by making sure we do not read past the
buffer in some cases. We must also use resize and not reserve on the
outpt buffer because reseve is meant as a hint and we are not supposed
to write past the size of the QByteArray even if it is reserved.
[ChangeLog][moc] Fixed crash on file ending with \\\r
Task-number: QTBUG-53441
Change-Id: I901e6c0ffc7f8877de3d07fd08cf26495461d294
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
Reviewed-by: Robert Loehning <robert.loehning@qt.io>
Port the last four remaining Q_FOREACH users in qmake
and uic to C++11 range-for and mark all qtbase tools
(incl. qmake) as Q_FOREACH-free, using QT_NO_FOREACH.
Change-Id: Ief4e5877269e7a853e4cf05e58861a448e822d3d
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@theqtcompany.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
moc.cpp(385): error #187: use of "=" where "==" may have been intended
Change-Id: Ibb6ef27030762cbe4f7c7002581a0955f0f9086d
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
So the command lines is more compatible with compilers flags. And new options
like -isystem can be used.
Change-Id: I7ad0997e5dab7d0425f12cde415abae6e9034a5b
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: David Faure <david.faure@kdab.com>
Adding strings in a single expression gets mediated
efficiently by QStringBuilder, where using += repeatedly
forces premature consolidation.
Change-Id: I6c1abb9c9801a016ce6c151cffdf8c7ac9502f4e
Reviewed-by: Edward Welbourne <edward.welbourne@theqtcompany.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
For CoW types, prefer const methods to avoid needless detach()ing.
Change-Id: Iefc33552d826aa30320e52acd2d421c9bdae127e
Reviewed-by: Edward Welbourne <edward.welbourne@theqtcompany.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Iterate the keys directly off the QMap itself.
Change-Id: I7792414db250d8ae63a935513a16a01eb184ebde
Reviewed-by: Edward Welbourne <edward.welbourne@theqtcompany.com>
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
If moc is invoked with the @ argument and no options file is specified
or the options file cannot be read, do not try to parse the empty
arguments list. Otherwise QCommandLineParser will print an additional
error message that is of no value for the user.
Task-number: QTBUG-51847
Change-Id: I9aa1eb20a44097b553123be8bc6fded87473a03a
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
These types are larger than a void*, so holding them in QLists
is needlessly inefficient. Worse, the code could come to depend
on the fragile property of (inefficient) QLists that references
to elements therein never are invalidated.
Fix by marking them movable, and holding them in a QVector instead.
Change-Id: I42f494ed87854eadc33d78db4479203ff5e0370f
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
Reviewed-by: Milian Wolff <milian.wolff@kdab.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
... by replacing them with C++11 range-for, or, for loops
over .keys() or .uniqueKeys(), with explicit iterator loops.
Saves 2300b in text size on optimized GCC 5.3 Linux AMD64 builds.
Change-Id: I6e1d4f5e56895dfd74aba21a3d4e913b5825645c
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The old code searched for any Qt containers by looking at
- each class definition
- each property's type
- each function (signal, slot, method)
- each argument type
and matching each against "Container<", building
the pattern string each time through the loop.
It would then collect hits in a QSet to be converted
to a QList and sorted at the very end.
The new code pulls the iteration over the candidates
out of all other loops. By doing so, it can stop
looking at classes, properties, functions etc when
it finds the first hit, and it inserts every candidate
at most once.
By iterating over the statically-known list of
candidates, the result is not a sorted set of Qt
containers, as before, but it still has s fixed order
across runs, which was the purpose of the sorting in
the original code.
In the implementation, make liberal use of C++11
range-for, which is safe, as we're passing everything
around as const.
Change-Id: If76dd3f57aa1b544a9cf1de2dca94ca7999220f0
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
When resolving the signal/slot of connect(&Foo::bar, ...) we place a meta-call
to map the address to the method index. Once we have found the index, we don't
need to continue but can return the result right away.
Change-Id: I67bb22df394d7c22dc1731367c0961b958ed77b3
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
From Qt 5.7 -> tools & applications are lisenced under GPL v3 with some
exceptions, see
http://blog.qt.io/blog/2016/01/13/new-agreement-with-the-kde-free-qt-foundation/
Updated license headers to use new GPL-EXCEPT header instead of LGPL21 one
(in those files which will be under GPL 3 with exceptions)
Change-Id: I42a473ddc97101492a60b9287d90979d9eb35ae1
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
A QByteArrayList is almost as efficient as a QVector<QByteArray>.
More importantly, the QByteArrayList case can share code with
the myriad of other QByteArrayList users, in particular
because std::stack is but the thinnest of wrappers around
its underlying container.
For moc, saves almost 1KiB in text size on optimized GCC 4.9
Linux AMD64 builds.
For qdbuscpp2xml, which re-uses moc code, saves ~1.6KiB.
Change-Id: I861e92b3c79e47e0ce892ccf54c9041182aaf212
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Without this, the symbol list is frequently reallocated. The value
is a guestimate. It corresponds to a block of memory with the size
262144 * 3 * 8 byte = ca. 6.3 megabytes
on a 64 bit machine. Looking at the final size of the symbol list,
this seems to fit the common case nicely.
Change-Id: Ib546a1ea479f2c6d8ab57be783cdf630f9b54f77
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
This was never a documented feature, but happended to work before Qt 5.5.
It broke because the peoperty access went into the static function and are
now prefixed with '_t->'
So restore the behavior as it was by not including the parentheses in the
member name.
Task-number: QTBUG-47695
Change-Id: Ic3509ddea7ac9abc871e71f5bfbe81d04d08e9bc
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
If a macro is used with too few parameters, complaining about its
definition using '#' followed by something other than a macro
parameter name is apt to be confusing - reading the definition will
reveal that the name in fact is a macro parameter after all. The
reader needs attention directed to the invocation, not the definition.
Split the test in two: one to test the prior error message does in
fact get produced for an invalid macro definition, the other to test
the invalid invocation case.
Task-number: QTBUG-46210
Change-Id: Ie177a56d346e553bf9d67e2008a4352633afa1ae
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The value was found by looking at the common ratio between
input size and final size of the result list.
Change-Id: I5762c15156afad4a7b8c1538e886058b3b5b0673
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
In the common case, macroExpandIdentifier was called in the
"not a macro" case, and a temporary vector with a single item was
allocated. Now, we catch this common case at the caller site
and put the single item directly into the result set, bypassing
the temporary list.
Change-Id: I71d92afc486ccdaae5930405d028f53f48073b8c
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Previously, a temporary list was allocated and then fed into the
bigger list of results. Now, we push data into the final list
directly, removing the overhead of the temporary allocation.
Change-Id: I9bea0fd3c23b1434b4be2728c60ac22a66908efc
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Milian Wolff <milian.wolff@kdab.com>
Previously, this allocated temporary QByteArrays in the frequent
QVector<Symbols>::indexOf calls (macro.arguments.indexOf(s)).
Now, the we reuse the fast SubArray::operator== implementation.
Change-Id: Idbc6e3cf5fd353e8e36d51ba88d31e4c3bfd121d
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
When invoking a macro with less argument than it expect, we would
crash trying to access the vector of arguments from the invocation
as we are trying to substitute an argument.
(Note that we do not show an error in case of argument mismatch
because ithat might happen parsing valid code as moc's c++ parser
is not 100% accurate (that was QTBUG-29331))
Task-number: QTBUG-46210
Change-Id: I3f08d7f5049e593a5bdc02a594ea63cadf66e7a4
Reviewed-by: Gabriel de Dietrich <gabriel.dedietrich@theqtcompany.com>
Prior to this, moc would not generate the function unless the
gadget class had a property or a non-constructor invokable.
Change-Id: Ic020ea5f8f59702f5e9e194a46e26850e53e5cfe
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
An error similar to the one below would be emitted
by the compiler on the moc generated file:
error: assigning to 'QObject *' from incompatible type 'Gadget *'
if (_a[0]) *reinterpret_cast<QObject**>(_a[0]) = _r; } break;
Change-Id: I75ae7bd6c46d20db2d47a80eaa08aae302d7d6c8
Reviewed-by: Simon Hausmann <simon.hausmann@theqtcompany.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Change 8836a4d5f0 already added Q_UNUSED in one place,
but that was not enough.
Change-Id: Ib98ba3fe47c0903cc01c56fd3eb8461c8938c75e
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This makes it easier to debug QJsonDocument binary bugs.
Note that the last character is never printed as printable, but by way
of construction it's always binary data anyway (the offsets table is at
the end after parsing JSON sources).
Change-Id: I8a7a116f51864cecb52fffff13bc24ad01ad8a49
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
...where passing them by value was not intentional.
Change-Id: Ifd5036d57b41fddeeacfbd3f5890881605b80647
Reviewed-by: Shawn Rutledge <shawn.rutledge@digia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.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.
Outdated header.LGPL removed (use header.LGPL21 instead)
Old header.LGPL3 renamed to header.LGPL3-COMM to match actual licensing
combination. New header.LGPL-COMM taken in the use file which were
using old header.LGPL3 (src/plugins/platforms/android/extract.cpp)
Added new header.LGPL3 containing Commercial + LGPLv3 + GPLv2 license
combination
Change-Id: I6f49b819a8a20cc4f88b794a8f6726d975e8ffbe
Reviewed-by: Matti Paaso <matti.paaso@theqtcompany.com>
In replacement for Q_ENUMS and Q_FLAGS.
Q_ENUM(Foo) has to be put after the declaration of Foo in an object.
It will tell moc to include the enum in the meta object (just like
Q_ENUMS) and will allow templated code to get the metaobject for
that enum.
Will be used by QDebug and QMetaType
Change-Id: Iefaf8ae07dc0359828102bf384809346629b3e23
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
Contrarily to Q_OBJECTs, Q_GADGETs are not guaranteed to
descend from a Q_GADGET. Here, we ensure that if the first
superclass is a Q_GADGET, then the derived class will be
treated as one. This allows gaps in the Q_GADGET hierarchy
while preventing from trying to link to the inexistent
staticMetaObject if there's no such ancestor.
Change-Id: If10fb952e23655102a425bb18fe8babaf447a47f
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@theqtcompany.com>
MSVC prints warnings if all slots are actually static methods:
moc_xxx.cpp:71: warning: C4189: '_t' : local variable is initialized but not referenced
This can be easily avoided by using Q_UNUSED.
Change-Id: Ice134a8a8d3e6aef1df1e7a89e9a2867281580fd
Reviewed-by: Friedemann Kleint <Friedemann.Kleint@theqtcompany.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
C++ standard advise to place 64k char limit for string literals, this
patch improves moc output so it is not affected anymore.
Task-number: QTBUG-36500
Change-Id: Iece630faaef45baebe8c7afe4fc51e0362c713de
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
When tokenizing, after macro expansion, moc needs to concatenate
subsequent string literals, because parser do not check for such
expressions.
Change-Id: Icc4f01395a5a7b67368eb8341a45ee74ade7d7f5
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
This commit changes several instances where moc was generating code that
used 0 as a null pointer constant. The Q_NULLPTR define is the more
idiomatic way to do this, and additionally this silences warnings
generated by e.g. GCC's -Wzero-as-null-pointer-constant.
[ChangeLog][Tools][moc] Fixed "zero as null pointer constant" warnings
in moc's generated code.
Change-Id: Ibe382b7bdbdddaf20cb4bdfd075fcdd1f363f9d3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
The change reduces heap allocations by using file mapping instead of
reading a whole file into memory just to create a slightly modified
copy of it.
For this small test case:
moc <<EOF
class X : public QObject { Q_OBJECT Q_PROPERTY(int x) };
EOF
massif shows improvement from:
peak cost: "26,8 KB" heap "2,1 KB" heap extra "0 B" stacks
to:
peak cost: "11,3 KB" heap "2,2 KB" heap extra "0 B" stacks
In general, depending on source file high peak memory usage is reduced
from few to few hundreds KB, especially that the allocation used to
happen for each include file too.
Change-Id: I9c1c848be848444156af25a991b67161fb9d8b29
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
We do not need to emit a diagnostic at definition time.
The diagnostic will be emit at expansion time.
Fix error when parsing boost header:
/usr/include/boost/fusion/container/vector/vector.hpp:25: Error: '#' is not followed by a macro parameter
Task-number: QTBUG-42233
Change-Id: I27deab362341f17ca3b0160615bb1b0934c3d5c3
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>
That way we will be able to access property of objects that are not QObject
(Q_GADGET)
Change-Id: Ib8ef6e52fc621e0b0d6530f82b1aa205f1ed5fd9
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This provides a way to invoke Q_INVOKABLE method
Change-Id: I301cec2a21c2889fb9dde0b1eb6e92d0bd561bd7
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
The byte array literals are only used to append them to another
QByteArray, so they offer only the static size calculation as a
benefit.
However, there are several drawbacks:
- QByteArrayLiteral data cannot be shared the way string literals
can be, not even within a single TU, and they add a few ints
for the QByteArrayData header which cannot reside in BSS, but
need to be stored in DATA.
- QByteArrayLiteral *does* allocate when the compiler doesn't
support C++11 lambdas.
- QByteArrayLiteral, when not using RVO, litters the code with
QByteArray dtor calls, which are not inline, and thus can't
be optimized away.
In particular, when used like this, they do not prevent any
memory allocation (in fact, they might add some, absent lambdas).
So, just append (C) string literals.
Change-Id: Iee5dba8dd970c5cc6df116afc1f8709a62356b06
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
The moc preprocessor is not necessarily fully compatible with the native
compiler preprocessor, which can lead to annoying warnings.
This fixes a problem particularly with the boost headers that rely on
MSVC only preprocessor features (to work around other MSVC preprocessor
deficiencies).
Task-number: QTBUG-29331
Change-Id: If884452969b512a746c81e235d31636b39c45b27
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
At the moment, it's possible to have 2 properties with the same name,
which doesn't make much sense. Notify the user about that so she can
react on it.
Change-Id: I4865b71730921b79ce9dd8abb0cc760b3f1dbfd8
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Frederik Gladhorn <frederik.gladhorn@digia.com>
Conflicts:
mkspecs/qnx-x86-qcc/qplatformdefs.h
src/corelib/global/qglobal.h
src/network/socket/qnativesocketengine_winrt.cpp
src/plugins/platforms/android/androidjniaccessibility.cpp
src/plugins/platforms/windows/qwindowswindow.cpp
Manually adjusted:
mkspecs/qnx-armle-v7-qcc/qplatformdefs.h
to include 9ce697f2d5
Thanks goes to Sergio for the qnx mkspecs adjustments.
Change-Id: I53b1fd6bc5bc884e5ee2c2b84975f58171a1cb8e
When compiling with the -Wswitch-default flag, the compiler
spits warnings for the moc generated files.
In certain development environments (e.g. where MISRA rules
have to be followed) every switch statement needs to have a
default case. Unfortunately the moc generated files contain
switch statements without default case. This patch adds a
default statement to all switch statements that are generated
by the moc compiler.
Change-Id: I8f3d81e3463fce4d3abf2ea5d0fa8727a7d9ba2e
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The comment shows to which string a QT_MOC_LITERAL is pointing.
Change-Id: Ia389d750b1b1c21e2242bad6beceea4f9298ff8e
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
moc would skip the 'operator' keyword as unknown and try to parse a type again
but as it sees the '<' it looks for the corresponding '>' which does not exist
types can't start with '<' anyway, so return an invalid type and continue
parsing as usual
Task-number: QTBUG-36834
Change-Id: If3d27076ef9947abf8c57c594713eece9334d0b0
Reviewed-by: Simon Hausmann <simon.hausmann@digia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>
At first, my goal was just to fix Moc::until() to parse properly
template arguments containing expressions containing > or >>
such as Foo<(8>>2)>
But with the test, I realized that normalizeType also requires change
not to split the > > too much.
And QMetaObjectPrivate::decodeMethodSignature should not interpret
the ) within the template parameter as the end of the function.
Change-Id: Ia9d3a2a786368aeda1edcf66280d70f64cf05070
Reviewed-by: Lars Knoll <lars.knoll@digia.com>
Better late than never :)
Task-number: QTBUG-33749
Change-Id: I5035255e66a56754b609441f5b81ab119565a7cb
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Make it easier to diagnose quoting errors in scripts.
Change-Id: I17894a426faa5cdf50f5ace4ed422ab2bd202558
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Moc should check full scope of any related objects or
gadget when it constructs extra data.
Change-Id: Ibd1b607a389cd4e788c0916984464cd9103d9c59
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
There is no point in saving \0 twice. Any code that could relay on it
is broken anyway, because moc saves \0\0 for an empty string.
Change-Id: I28fa4f78aae8c883c088df43ec89d608a99b3bdd
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Since now in Qt5 the moc does full macro substitution, it needs to handle
the defines passed is command argument, even if they span over multiple
tokens, or if they do not have any token.
Example:
moc '-DCOMPLEX=QVector<int>' '-DEMPTY=' foo.h
[ChangeLog][moc] Fixed passing -D of a macro defined to something more
complex than a single identifier.
Task-number: QTBUG-33668
Change-Id: Ie8131de215f1659a24af4778d52ee40cda19759f
Reviewed-by: Oswald Buddenhagen <oswald.buddenhagen@digia.com>
Reviewed-by: Jędrzej Nowacki <jedrzej.nowacki@digia.com>
There is the warning it fixes:
moc_foo.cpp:28:9: warning: conversion to ‘qptrdiff {aka long long int}’ from ‘long unsigned int’ may change the sign of the result [-Wsign-conversion]
- idx * sizeof(QByteArrayData) \
^
qarraydata.h:282:49: note: in definition of macro ‘Q_STATIC_ARRAY_DATA_HEADER_INITIALIZER_WITH_OFFSET’
{ Q_REFCOUNT_INITIALIZE_STATIC, size, 0, 0, offset } \
^
moc_foo.cpp:26:5: note: in expansion of macro ‘Q_STATIC_BYTE_ARRAY_DATA_HEADER_INITIALIZER_WITH_OFFSET’
Q_STATIC_BYTE_ARRAY_DATA_HEADER_INITIALIZER_WITH_OFFSET(len, \
^
moc_foo.cpp:33:1: note: in expansion of macro ‘QT_MOC_LITERAL’
QT_MOC_LITERAL(1, 23, 5),
^
[ChangeLog][moc][QTBUG-36128] Fixed sign conversion warning in generated file.
Task-number: QTBUG-36128
Change-Id: Ibf00974dbfb419bb53a3109d93e4cee598bb0588
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
When declaring a Q_PROPERTY(SomeType::SomeEnum foo ...) and SomeType is not a
QObject but a gadget, then we must still include SomeType's meta object in the
list of related meta objects.
Task-number: QTBUG-35657
Change-Id: I46195140cb5d180c4f03bb1fe06a876e3fe11267
Reviewed-by: Olivier Goffart <ogoffart@woboq.com>
