Even for non-complex types, it makes no sense to use the assignment
operator instead of placement new when constructing new objects.
Pick-to: 6.0 5.15
Change-Id: I5f15fe4b3397cf52d1d35e6c4dcc513b94b3cf14
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Types that throw in their destructors are strongly discouraged in C++,
and even the STL doesn't define what happens if such types are stored
in their containers.
Make this more explicit for Qt and disallow storing those types in our
containers. This will hopefully preempty any potential future bug
reports about us not handling such a case. It also helps simplify
some code in QList and other cases and makes it possible to explicitly
mark more methods as noexcept.
Some care needs to be taken where to add the static asserts, so that
we don't disallow forward declarations of types stored in containers.
Place the static assert into the destructor of the container where
possible or otherwise into the templated d-pointer.
Change-Id: If3aa40888f668d0f1b6c6b3ad4862b169d31280e
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This allows to use list-initialization when calling these functions,
for instance list.removeAll({}).
Change-Id: I2828d900a44bd0bc5aea5fba4777304b09190bc9
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Just like QList.
[ChangeLog][QtCore][QVarLengthArray] The indexOf, lastIndexOf and contains
methods now take an object of any datatype -- and not just the
array's own value type. This allows for heterogenous lookup in
QVarLengthArray objects.
Change-Id: Ibc55191a140612a4e9be46b4d18457415ea3717f
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This will enable run-time debugging on Windows, using
_CRTDBG_MAP_MALLOC, which uses #define to override the
standard library memory management functions, including
realloc.
Fixes: QTBUG-86395
Change-Id: I51975dd74cab0ae8309436c86d17a59074c561e1
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Andrei Golubev <andrei.golubev@qt.io>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
Remove QTypeInfo::isStatic, as that's not used anymore in Qt 6.
Also remove sizeOf, it's unused, and we have QMetaType for that if
required.
Remove all typeinfo declaractions for trivial types, as the default
template covers them correctly nowadays.
Finally set up a better default for isPointer, and do some smaller
cleanups all over the place.
Change-Id: I6758ed37dfc701feaaf0ff105cc95e32da9f9c33
Reviewed-by: Fabian Kosmale <fabian.kosmale@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Remove the hand-rolled union used only to achieve alignment.
Change-Id: I8e635fa0d09042f30fd67d589db9bc384c0ec8ef
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
This had already been in very few places, where we ran into issues with
this before. More generic constraints here will significantly reduce the
amount of error messages a user has to parse in case he tries to instantiate
an operator by accident (or with a lacking comparison operator for one of
it's template arguments).
Change-Id: I1521d19c55d99732d9742402bd534c390a8e4242
Reviewed-by: Paul Wicking <paul.wicking@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
A QVLA is copyable, so it should be movable, too.
Added a helper function a la P1144's uninitialized_relocate_n to deal
with the QTypeInfoQuery stuff. This way, the code is re-usable
everywhere it's needed. The same cannot be said for QArrayDataOps,
which only a parent can love...
[ChangeLog][QtCore][QVarLengthArray] Added missing move constructor
and move-assignment operator.
Task-number: QTBUG-39111
Change-Id: If0dc2aa78eb29062d73dcd3dc4647ba345ae39e6
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
There is no reason for keep using our macro now that we have C++17.
The macro itself is left in for the moment being, as well as its
detection logic, because it's needed for C code (not everything
supports C11 yet). A few more cleanups will arrive in the next few
patches.
Note that this is a mere search/replace; some places were using
double braces to work around the presence of commas in a macro, no
attempt has been done to fix those.
tst_qglobal had just some minor changes to keep testing the macro.
Change-Id: I1c1c397d9f3e63db3338842bf350c9069ea57639
Reviewed-by: Lars Knoll <lars.knoll@qt.io>
This is required, so that QHash and QSet can hold more
than 2^32 items on 64 bit platforms.
The actual hashing functions for strings are still 32bit, this will
be changed in a follow-up commit.
Change-Id: I4372125252486075ff3a0b45ecfa818359fe103b
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
Amends 2e1763d83a.
The new range ctors need deduction guides, since the compiler can't
deduce the value_type from a pair of iterators.
Change-Id: I3ec1e5f91305b317c443b6a70246be416b55bad9
Reviewed-by: Volker Hilsheimer <volker.hilsheimer@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Don't call realloc() with all its machinery when we know exactly what
to do: destroy the last element and decrease the size by one.
Extend the test, removing the unused Foo class for a new Tracker one.
Change-Id: I568eef4f6335669689fb16fd23af92cb4d6464bd
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Otherwise "QVarLengthArray<Foo> x = {};" gives a warning. Also, some
compilers get confused about "QVarLengthArray()" this way.
Task-number: QTBUG-76199
Change-Id: I4296586c0181d3e6e82ca8b7b79aeb9a21645d1f
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
QList, QVector and QVarLengthArray check the validity of iterators
passed to member functions using isValidIterator(), which checks that
the underlying pointers are in the range [begin, end]. This check is
well-defined when the outcome is positive, ie. when the iterator is
valid. But if the iterator is not valid, and does not happen to point
into [end, begin + capacity], the comparison, which uses normal
operator<, invokes UB.
Fix by using std::less<T*>, which defines a total ordering.
Change-Id: I1e5757789b4b9779f5e3e298e7f2b2dd0b27576c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
With g++ 8.2.0, I get warnings when a QVarLengthArray<QString> calls
remove() or prepend(), as some tests in tst_QVarLengthArray do, as
they call memmove() "writing to an object of type ‘class QString’ with
no trivial copy-assignment; use copy-assignment or copy-initialization
instead"; which may indeed be a good argument for not using
QVarLengthArray<QString>, but its own tests do.
Change-Id: I4f8a64948b32a54e67a285df4ec7788f60739ffb
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
They were forgotten when the overloads for append()/push_back() were
added in Qt 5.6
[ChangeLog][QtCore][QVarLengthArray] Added missing rvalue overload of
operator+=() and operator<<()
[ChangeLog][QtCore][QVector] Added missing rvalue overload of
operator+=() and operator<<()
Change-Id: I20fedfba2bf282773bd1f9cf2c8ec06f05896a7d
Reviewed-by: Martin Smith <martin.smith@qt.io>
Reviewed-by: Mårten Nordheim <marten.nordheim@qt.io>
This is similar to commit 342bb5b03a.
From GCC 8:
qarraydataops.h:84:17: error: ‘void* memcpy(void*, const void*, size_t)’ writing to an object of type ‘class QStringRef’ with no trivial copy-assignment; use copy-assignment or copy-initialization instead [-Werror=class-memaccess]
[etc.]
Change-Id: I41d006aac5bc48529845fffd150e817e64973bec
Reviewed-by: Ville Voutilainen <ville.voutilainen@qt.io>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Improves performance and STL compatibility by adding rvalue versions
of prepend and insert.
[ChangeLog][QtCore][QVarLengthArray] Added rvalue overloads of
prepend and insert.
[ChangeLog][QtCore][QVector] Added rvalue overloads of prepend
and insert.
[ChangeLog][QtCore][QVarLengthArray] Can now contain movable but
non-copyable types, such as std::unique_ptr.
Change-Id: I6c946acc5b67502c91c52ac5dea67cedb1af93a5
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Something nice we'd like to detect for array-backed containers
is if the iterator passed is a Contiguous one; if the type is also
trivially copyable / Q_PRIMITIVE_TYPE, we could memcpy() the whole
range.
However, there's no trait in the Standard to detect contiguous
iterators (the best approximation would be detecting if the iterator
is actually a pointer). Also, it's probably not smart to do the work
now for QVector since QVector needs refactoring anyhow, and this work
will be lost.
QString and QByteArray are left in another commit.
[ChangeLog][QtCore][QVector] Added range constructor.
[ChangeLog][QtCore][QVarLengthArray] Added range constructor.
[ChangeLog][QtCore][QList] Added range constructor.
[ChangeLog][QtCore][QStringList] Added range constructor.
[ChangeLog][QtCore][QLinkedList] Added range constructor.
[ChangeLog][QtCore][QSet] Added range constructor.
Change-Id: I220edb796053c9c4d31a6dbdc7efc5fc0f6678f9
Reviewed-by: Milian Wolff <milian.wolff@kdab.com>
Side note: QHash has squeeze(), but there's no shrink_to_fit on
std::unordered_map.
[ChangeLog][QtCore][QByteArray] Added shrink_to_fit().
[ChangeLog][QtCore][QString] Added shrink_to_fit().
[ChangeLog][QtCore][QVarLengthArray] Added shrink_to_fit().
[ChangeLog][QtCore][QVector] Added shrink_to_fit().
Change-Id: Ifd7d28c9bed70727be6308f0191a188201784f61
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
This fixes a warning-turned-Werror in qdistancefield.cpp:
In member function ‘void QVarLengthArray<T, Prealloc>::realloc(int, int) [with T = bool; int Prealloc = 256]’,
inlined from ‘void makeDistanceField(QDistanceFieldData*, const QPainterPath&, int, int)’ at ../../include/QtCore/../../../../qt5/qtbase/src/corelib/tools/qvarlengtharray.h:275:10:
../../include/QtCore/../../../../qt5/qtbase/src/corelib/tools/qvarlengtharray.h:390:19: error: ‘void* memcpy(void*, const void*, size_t)’: specified size between 18446744071562067968 and 18446744073709551615 exceeds maximum object size 9223372036854775807 [-Werror=stringop-overflow=]
memcpy(ptr, oldPtr, copySize * sizeof(T));
~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Apparently GCC cannot rule out that copySize may be negative in the
call to memcpy. Put GCC on the right track by adding a Q_ASSUME.
Change-Id: I63e3801e52ebe2a7f77e3a97ef03ec3869319c8c
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Konstantin Ritt <ritt.ks@gmail.com>
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
It's private API, the last in-tree user was removed in acbd7999, but
failed to remove the class. Do it now.
Change-Id: I26294b535d80b419a2f545a4783014b493a5dc93
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
Improves performance when appending temporaries, esp. since
the aliasing fix in the lvalue overload in 0f730ef made that
overload correct, but a bit slower across reallocs.
The unit tests already also pass rvalues, so the function is
covered in the existing tests.
[ChangeLog][QtCore][QVarLengthArray] Added rvalue overloads
of append() and push_back().
Change-Id: If3a6970f03a160cba5b42d33d32d3d18948f6ce3
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Reviewed-by: Edward Welbourne <edward.welbourne@qt.io>
... except in QList.
When dafa3618 introduced isRelocatable and QTypeInfoQuery, the
intention was to decouple the memory layout of QList from the
reallocation optimizations in QVector. This never happened, as
QVector (and QVarLengthArray) continue to use isStatic to this
day.
Fix by porting both QVector and QVLA to QTypeInfoQuery and
isRelocatable.
Change-Id: I6951f2cf21f0cbb24e2dbd38f80f1bd82007d394
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Like the lvalue QVector::append() overload, when we reallocate,
we need to take a copy of the function's argument because the
reference will get stale upon reallocation.
Add a test.
[ChangeLog][QtCore][QVarLengthArray] Fixed a bug involving
appending an item already in the container to the container
again.
Change-Id: I06eeed6cb383dd5924e47a302bb3d1666d04c8e8
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
warning C4267: 'argument': conversion from 'size_t' to 'int', possible loss of data
while compiling class template member function 'QVarLengthArray<T,Prealloc>::QVarLengthArray(std::initializer_list<T>)'
Change-Id: I36f5ef65ec1f511eac7f3ad1a4717d18f7dc9ce4
Reviewed-by: Timur Pocheptsov <timur.pocheptsov@qt.io>
The C++ standard says in [support.initlist.access]/1:
constexpr const E* begin() const noexcept;
Returns: A pointer to the beginning of the array. If size() == 0 the
values of begin() and end() are unspecified but they shall be
identical.
So we can't assume it's non-null. I didn't want to remove the Q_ASSERT,
so passing a non-null pointer to append() remains required. This patch
simply won't call append() if the initializer list is empty.
This was already tested, but the failure is with a compiler that is not
part of the Qt CI.
Task-number: QTBUG-57277
Change-Id: Iaeecaffe26af4535b416fffd1489806872b412ee
Reviewed-by: Olivier Goffart (Woboq GmbH) <ogoffart@woboq.com>
The MSVC STL warns when passing naked pointers as non-bounded
iterators to algorithms such as std::equal and std::copy, in
an attempt to inform users that the range specified by that
iterator has an implicit minimum size that the caller of the
algorithm must ensure is met:
warning C4996: 'std::_Equal1': Function call with parameters that may be unsafe - \
this call relies on the caller to check that the passed values are correct. To \
disable this warning, use -D_SCL_SECURE_NO_WARNINGS. See documentation on how to \
use Visual C++ 'Checked Iterators'
When building Qt, as well as when building user projects with
qmake (cf. 0a76b6bc7f), we
globally disable this warning (with -D_SCL_SECURE_NO_WARNINGS),
but since we started using STL algorithms in public headers (e.g.
in qvector.h), users get this warning in their own projects now,
unless they, too, define said macro. But such a requirement is
against the Qt policy to have headers that are warning-free as
much as possible.
The suggested way of fixing this warning is to wrap the naked
pointer in a stdext::unchecked_array_iterator before passing it
to the algorithm, cf. examples in
https://msdn.microsoft.com/en-us/library/ttcz0bys%28v=vs.120%29.aspx
or, together with the capacity-made-explicit, in a
stdext::checked_array_iterator.
To avoid ifdefs for platforms that don't have these extensions
(which, incidentally, for the unchecked case, includes MSVC 2012),
wrap the calls in macros.
The end game here is to drop -D_SCL_SECURE_NO_WARNINGS, at least
for public headers, even though this commit also adds the wrapper
to implementation and private header files.
An alternative to the wrapper would have been the version of
std::equal that takes four iterators. However, that is a C++14
library feature, while this version of Qt still needs to compile
with a C++98 compiler, and, more importantly, there isn't, and
never will be, a corresponding 4-iterator version of std::copy.
Task-number: QTBUG-47948
Done-with: Stephen Kelly <steveire@gmail.com>
Change-Id: I1bbab257fb5f1c5042939c382a412b596112ff26
Reviewed-by: Stephen Kelly <ske@ableton.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: I046ec3e47b1876cd7b4b0353a576b352e3a946d9
Reviewed-by: Lars Knoll <lars.knoll@theqtcompany.com>
std::vector has them, too.
[ChangeLog][QtCore][QVarLengthArray] Added relational operators <, <=, >, >= if the element
type supports operator<.
Change-Id: I69e16d361fd4738a56b292ebfa78316d28871eda
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
Add a dedicated operator=(std::initializer_list) that
first resizes the QCLA, and then replaces the elements
one by one.
This should be usually faster than creating a temporary
QCLA and then copying it, except for the case where the
new array does not fit into the allocated stack - but this
is IMO nothing to optimize for.
Task-number: QTBUG-45041
Change-Id: I147d6d01186b1ca3c635b2c8365d8f6e638ce6fe
GPush-Base: 08de3113051e1289f0de0651ec5647c9ee6feb27
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Reviewed-by: Thiago Macieira <thiago.macieira@intel.com>
The documentation claims that operator== is needed, not operator!=.
While at it, we can also replace the loop with std::equal, which
might even allow STL implementations to choose a hand-optimized
version of the algorithm for C++ builtin types ...
Change-Id: I988b326d6af3b767526952e303468e18ff6594f9
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Implement an initializer_list constructor, which was probably
just forgotten so far. Technically this is a SC incompatible change,
since
QVarLengthArray<int> array = {10};
will now create an array with one element 10, instead of an empty array
with a reserved size of 10. Anyhow, keeping the inconsistency with the
STL / other Qt containers here would certainly do more harm than good
in the long run.
Task-number: QTBUG-45047
Change-Id: I4675880f93e141181250939942fa32300916b0e3
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>
Giuseppe D'Angelo
Lars Knoll
Allan Sandfeld Jensen
Jesko von Monkiewitsch
Marc Mutz
Liang Qi
Laszlo Agocs
Ulf Hermann
Edward Welbourne
Eric Lemanissier
Christian Ehrlicher
Thiago Macieira
Jesus Fernandez
Jani Heikkinen
Liang Qi
Kai Koehne