Summary:
This patch fixes bugs.llvm.org/PR32979.
[util.smartptr.shared.const] says:
> In the constructor definitions below, enables shared_from_this with p, for a pointer p of type Y*, means
> that if Y has an unambiguous and accessible base class that is a specialization of enable_shared_from_-
> this.
This means that libc++ needs to respect the access specifier of the base class, and not attempt to construct
and enabled_shared_from_this base if it is private. However access specifiers don't affect overload resolution
so our current implementation will attempt to construct the private base.
This patch uses SFINAE to correctly detect if the shared_ptr input has an accessible enable_shared_from_this
base class.
Reviewers: mclow.lists
Reviewed By: mclow.lists
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D33033
llvm-svn: 302709
Libc++ is used as a system library on macOS and iOS (amongst others). In order
for users to be able to compile a binary that is intended to be deployed to an
older version of the platform, clang provides the
availability attribute <https://clang.llvm.org/docs/AttributeReference.html#availability>_
that can be placed on declarations to describe the lifecycle of a symbol in the
library.
See docs/DesignDocs/AvailabilityMarkup.rst for more information.
Differential Revision: https://reviews.llvm.org/D31739
llvm-svn: 302172
This patch overhauls both specializations of unique_ptr while implementing
the following LWG issues:
* LWG 2801 - This issue constrains unique_ptr's constructors when the deleter type
is not default constructible. Additionally it adds SFINAE conditions
to unique_ptr<T[]>::unique_ptr(Up).
* LWG 2905 - This issue reworks the unique_ptr(pointer, /* see below */ deleter)
constructors so that they correctly SFINAE when the deleter argument cannot
be used to construct the stored deleter.
* LWG 2520 - This issue fixes initializing unique_ptr<T[]> from nullptr.
Libc++ had previously implemented this issue, but the suggested resolution
still broke initialization from NULL. This patch re-works the
unique_ptr<T[]>(Up, deleter) overloads so that they accept NULL as well
as nullptr.
llvm-svn: 300406
Summary:
__compressed_pair takes and passes it's constructor arguments by value. This causes arguments to be moved 3 times instead of once. This patch addresses that issue and fixes `constexpr` on the constructors.
I would rather have this fix than D27564, and I'm fairly confident it's not ABI breaking but I'm not 100% sure.
I prefer this solution because it removes a lot of code and makes the implementation *much* smaller.
Reviewers: mclow.lists, K-ballo
Reviewed By: K-ballo
Subscribers: K-ballo, cfe-commits
Differential Revision: https://reviews.llvm.org/D27565
llvm-svn: 300140
For reference deleter types the const qualifier on the return type
of get_deleter() should be ignored, and a non-const deleter should
be returned.
This patch fixes a bug where "const deleter_type&" is incorrectly
formed.
llvm-svn: 300121
Summary:
Exactly what the title says.
This patch also adds a `std::hash<nullptr_t>` specialization in C++17, but it was not added by this paper and I can't find the actual paper that adds it.
See http://wg21.link/P0513R0 for more info.
If there are no comments in the next couple of days I'll commit this
Reviewers: mclow.lists, K-ballo, EricWF
Reviewed By: EricWF
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D28938
llvm-svn: 292684
In order to allow inlining of previously out-of-line functions without an ABI break
libc++ provides legacy definitions in the dylib that old programs can
continue to use. Unfortunatly Windows link.exe detects this hack and diagnoses the duplicate
definitions.
This patch disable the duplicate definitions on Windows by adding an ABI option
which disables all "legacy out-of-line symbols"
llvm-svn: 292190
Moves hot functions such as atomic add into the memory header file
so that they can be inlined, which brings performance benefits.
Patch by Kevin Hu, Aditya Kumar, Sebastian Pop
Differential Revision: https://reviews.llvm.org/D24991
llvm-svn: 292184
Summary:
Some parts of the FreeBSD tree are still compiled with C++98, and until
rL288554 this has always worked fine. After that, a complaint about the
newly introduced local _PairT is produced:
/usr/include/c++/v1/memory:3354:27: error: template argument uses local type '_PairT' [-Werror,-Wlocal-type-template-args]
typedef __scalar_hash<_PairT> _HashT;
^~~~~~
/usr/include/c++/v1/memory:3284:29: error: template argument uses local type '_PairT' [-Werror,-Wlocal-type-template-args]
: public unary_function<_Tp, size_t>
^~~
/usr/include/c++/v1/memory:3356:12: note: in instantiation of template class 'std::__1::__scalar_hash<_PairT, 2>' requested here
return _HashT()(__p);
^
As far as I can see, there should be no problem moving the _PairT
struct to just before the __hash_combine() function, which fixes this
particular warning.
Reviewers: mclow.lists, EricWF
Subscribers: cfe-commits, emaste
Differential Revision: https://reviews.llvm.org/D28472
llvm-svn: 291476
Summary:
On Windows the identifier `__deallocate` is defined as a macro by one of the Windows system headers. Previously libc++ worked around this by `#undef __deallocate` and generating a warning. However this causes the WIN32 version of `__threading_support` to always generate a warning on Windows. This is not OK.
This patch renames all usages of `__deallocate` internally as to not conflict with the macro.
Reviewers: mclow.lists, majnemer, rnk, rsmith, smeenai, compnerd
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D28426
llvm-svn: 291332
In ABI v1 libc++ implements std::pointer_safety as a class type instead
of an enumeration. However this class type does not provide
a default constructor as it should. This patch adds that default constructor.
llvm-svn: 291059
In the C++ standard `std::pointer_safety` is defined
as a C++11 strongly typed enum. However libc++ currently defines
it as a class type which simulates a C++11 enumeration. This
can be detected in valid C++ code.
This patch introduces an the _LIBCPP_ABI_POINTER_SAFETY_ENUM_TYPE ABI option.
When defined `std::pointer_safety` is implemented as an enum type.
Unfortunatly this also means it can no longer be provided as an extension
in C++03.
Additionally this patch moves the definition for `get_pointer_safety()`
out of the dylib, and into the headers. New usages of `get_pointer_safety()`
will now use the inline version instead of the dylib version. However in
order to keep the dylib ABI compatible the old definition is explicitly
compiled into it.
llvm-svn: 291046
The name _LIBCPP_TYPE_VIS_ONLY is no longer accurate because both
_LIBCPP_TYPE_VIS and _LIBCPP_TYPE_VIS_ONLY expand to
__attribute__((__type_visibility__)) with Clang. The only remaining difference
is that _LIBCPP_TYPE_VIS_ONLY can be applied to templates whereas
_LIBCPP_TYPE_VIS cannot (due to dllimport/dllexport not being allowed on
templates).
This patch renames _LIBCPP_TYPE_VIS_ONLY to _LIBCPP_TEMPLATE_VIS.
llvm-svn: 291035
libc++ no longer supports C++11 compilers that don't implement `= default`.
This patch removes all instances of the feature test macro
_LIBCPP_HAS_NO_DEFAULTED_FUNCTIONS as well as the potentially dead code it hides.
llvm-svn: 287321
Summary:
Libc++ still uses per-feature configuration macros when configuring for C++11. However libc++ requires a feature-complete C++11 compiler so there is no reason to check individual features. This patch starts the process of removing the feature specific macros and replacing their usage with `_LIBCPP_CXX03_LANG`.
This patch removes the __config macros:
* _LIBCPP_HAS_NO_TRAILING_RETURN
* _LIBCPP_HAS_NO_TEMPLATE_ALIASES
* _LIBCPP_HAS_NO_ADVANCED_SFINAE
* _LIBCPP_HAS_NO_DEFAULT_FUNCTION_TEMPLATE_ARGS
* _LIBCPP_HAS_NO_STATIC_ASSERT
As a drive I also changed our C++03 static_assert to use _Static_assert if available.
I plan to commit this without review if nobody voices an objection.
Reviewers: mclow.lists
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D24895
llvm-svn: 282347
call_once is using relaxed atomic load to perform double-checked locking, which contains a data race. The fast-path load has to be an acquire atomic load.
Differential Revision: https://reviews.llvm.org/D24028
llvm-svn: 280621
This patch adds the weak_type typedef in shared_ptr. It is available in
C++17 and newer.
This patch also updates the _LIBCPP_STD_VER and TEST_STD_VER macros to
have the value of 16, since 2016 is the current year.
llvm-svn: 273839
See https://llvm.org/bugs/show_bug.cgi?id=27115
The problem was that the conversion from
'const enable_shared_from_this<T>*' to 'const T*' didn't work if
T inherited enable_shared_from_this as a virtual base class. The fix
is to take the original pointer passed to shared_ptr's constructor in the
__enable_weak_this method and perform an upcast to 'const T*' instead of
performing a downcast from the enable_shared_from_this base.
llvm-svn: 273835
Summary:
Currently the implementation of [util.smartptr.shared.atomic] is provided only when using Clang, and not with GCC. This is a relic of not having a GCC implementation of <atomic>, even though <atomic> isn't actually used in the implementation. This patch enables support for atomic shared_ptr functions when using GCC.
Note that this is not a header only change. Previously only Clang builds of libc++.so would provide the required symbols. There is no reason for this restriction.
After this change both Clang and GCC builds should be binary compatible with each other WRT these symbols.
Reviewers: mclow.lists, rmaprath, EricWF
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D21407
llvm-svn: 273076
When you assign a shared_ptr, the deleter gets called and assigned. In this routine, the assignment happens inside a critical section, which could (potentially) lead to a deadlock, if the deleter did something wonky. Now we swap the old value with an (empty) temporary shared_ptr, and then let the temporary delete the old value when it goes out of scope (after the lock has been released). This should fix PR#27724. Thanks to Hans Boehm for the bug report and the suggested fix.
llvm-svn: 269965
These are the cases when an out-of-class definition of a method is
marked _LIBCPP_INLINE_VISIBILITY, but the in-class declaration is
not. This will start failing when (or if) we switch to
attribute((internal_linkage)).
llvm-svn: 255166
This change moves visibility attributes from out-of-class method
definitions to in-class declaration. This is needed for a switch to
attribute((internal_linkage)) (see http://reviews.llvm.org/D13925)
which can only appear on the first declaration.
This change does not touch istream/ostream/streambuf. They are
handled separately in http://reviews.llvm.org/D14409.
llvm-svn: 252385
Summary:
This patch properly constrains the converting assignment operator in C++03. It also fixes a bug where std::forward was given the wrong type.
The following two tests begin passing in C++03:
* `unique_ptr.single.asgn/move_convert.pass.cpp`
* `unique_ptr.single.asgn/move_convert13.fail.cpp`
Reviewers: mclow.lists
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D12173
llvm-svn: 246272
Currently we need an #ifdef branch every time we use pointer traits to rebind a pointer because
it is done differently in C++11 and C++03. This patch introduces the __rebind_pointer utility to
clean this up.
Also add a test that list and it's iterators can be instantiated with incomplete element types.
llvm-svn: 245806
Currently we need an #ifdef branch every time we use pointer traits to rebind a pointer because
it is done differently in C++11 and C++03. This patch introduces the __rebind_pointer utility to
clean this up.
llvm-svn: 245802
Summary:
After putting this question up on cfe-dev I have decided that it would be best to allow the use of `<atomic>` in C++03. Although static initialization is a concern the syntax required to get it is C++11 only. Meaning that C++11 constant static initialization cannot silently break in C++03, it will always cause a syntax error. Furthermore `ATOMIC_VAR_INIT` and `ATOMIC_FLAG_INIT` remain defined in C++03 even though they cannot be used because C++03 usages will cause better error messages.
The main change in this patch is to replace `__has_feature(cxx_atomic)`, which only returns true when C++ >= 11, to `__has_extension(c_atomic)` which returns true whenever clang supports the required atomic builtins.
This patch adds the following macros:
* `_LIBCPP_HAS_C_ATOMIC_IMP` - Defined on clang versions which provide the C `_Atomic` keyword.
* `_LIBCPP_HAS_GCC_ATOMIC_IMP` - Defined on GCC > 4.7. We must use the fallback atomic implementation.
* `_LIBCPP_HAS_NO_ATOMIC_HEADER` - Defined when it is not safe to include `<atomic>`.
`_LIBCPP_HAS_C_ATOMIC_IMP` and `_LIBCPP_HAS_GCC_ATOMIC_IMP` are mutually exclusive, only one should be defined. If neither is defined then `<atomic>` is not implemented and including `<atomic>` will issue an error.
Reviewers: chandlerc, jroelofs, mclow.lists
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D11555
llvm-svn: 245463
Although CMake adds warning flags, they are ignored in the libc++ headers
because the headers '#pragma system header' themselves.
This patch disables the system header pragma when building libc++ and fixes
the warnings that arose.
The warnings fixed were:
1. <memory> - anonymous structs are a GNU extension
2. <functional> - anonymous structs are a GNU extension.
3. <__hash_table> - Embedded preprocessor directives have undefined behavior.
4. <string> - Definition is missing noexcept from declaration.
5. <__std_stream> - Unused variable.
llvm-svn: 242623
Summary:
In some places in libc++ we need to use the `__atomic_*` builtins. This patch adds a header that provides access to those builtins in a uniform way from within the dylib source.
If the compiler building the dylib does not support these builtins then a warning is issued.
Only relaxed loads are needed within the headers. A singe function to do these relaxed loads has been added to `<memory>`.
This patch applies the new atomic builtins to `__shared_count` and `call_once`.
Reviewers: mclow.lists
Subscribers: majnemer, jroelofs, cfe-commits
Differential Revision: http://reviews.llvm.org/D10406
llvm-svn: 241532
Summary: Currently we only enable the use of __is_final(...) with Clang. GCC also provides __is_final(...) since 4.7 in all standard modes. This patch creates the macro _LIBCPP_HAS_IS_FINAL to note the availability of `__is_final`.
Reviewers: danalbert, mclow.lists
Reviewed By: mclow.lists
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D8795
llvm-svn: 239664
Summary:
In certain cases vector can use memcpy to construct a range of elements at the back of the vector. We currently don't do this resulting in terrible code gen in non-optimized mode and a
very large slowdown compared to libstdc++.
This patch adds a `__construct_forward_range(Allocator, Iter, Iter, _Ptr&)` and `__construct_forward_range(Allocator, Tp*, Tp*, Tp*&)` functions to `allocator_traits` which act similarly to the existing `__construct_forward(...)` functions.
This patch also changes vectors `__construct_at_end(Iter, Iter)` to be `__construct_at_end(Iter, Iter, SizeType)` where SizeType is the size of the range. `__construct_at_end(Iter, Iter, SizeType)` now calls `allocator_traits<Tp>::__construct_forward_range(...)`.
This patch is based off the design of `__swap_out_circular_buffer(...)` which uses `allocator_traits<Tp>::__construct_forward(...)`.
On my machine this code performs 4x better than the current implementation when tested against `std::vector<int>`.
Reviewers: howard.hinnant, titus, kcc, mclow.lists
Reviewed By: mclow.lists
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D8109
llvm-svn: 233711
Visual Studio's SAL extension uses a macro named __deallocate. This macro is
used pervasively, and gets included through various different ways. This
conflicts with the similarly named interfaces in libc++. Introduce a undef
header similar to __undef_min_max to handle this. This fixes a number of errors
due to the macro replacing the function name.
llvm-svn: 229162
Summary:
This patch add support for "fancy pointers/allocators" as well as fixing support for shared_pointer and "minimal" allocators.
Fancy pointers are class types that meet the NullablePointer requirements. In our case they are created by fancy allocators. `support/min_allocator.h` is an archetype for these types.
There are three types of changes made in this patch:
1. `_Alloc::template rebind<T>::other` -> `__allocator_traits_rebind<_Alloc, T>::type`. This change was made because allocators don't need a rebind template. `__allocator_traits_rebind` is used instead of `allocator_traits::rebind` because use of `allocator_traits::rebind` requires a workaround for when template aliases are unavailable.
2. `a.deallocate(this, 1)` -> `a.deallocate(pointer_traits<self>::pointer_to(*this), 1)`. This change change is made because fancy pointers aren't always constructible from raw pointers.
3. `p.get()` -> `addressof(*p.get())`. Fancy pointers aren't actually a pointer. When we need a "real" pointer we take the address of dereferencing the fancy pointer. This should give us the actual raw pointer.
Test Plan: Tests were added using `support/min_allocator.h` to each affected shared_ptr overload and creation function. These tests can only be executed in C++11 or greater since min_allocator is only available then. A extra test was added for the non-variadic versions of allocate_shared.
Reviewers: danalbert, mclow.lists
Reviewed By: mclow.lists
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D4859
llvm-svn: 220469
If you're crazy enough to want this sort of thing, then add
-D_LIBCPP_HAS_NO_THREADS to your CXXFLAGS and
--param=additiona_features=libcpp-has-no-threads to your lit commnad line.
http://reviews.llvm.org/D3969
llvm-svn: 217271
GCC will treat the default function template arguments as a
compilation error if C++0x is not enabled.
This commit workaround the compilation error by moving the
SFINAE check to function argument instead of the template
argument.
llvm-svn: 200523
pair, and a couple of pair-like implementation detail types. The
C++98/03 and 11 standards all specify that the copy constructor of
pair<int, int> is trivial. However as libc++ tracked the draft C++11
standard over the years, this copy constructor became non-trivial, and
then just recently was corrected back to trivial for C++11.
Unfortunately (for libc++1) the Itanium ABI specifies different calling
conventions for trivial and non-trivial copy constructors. Therefore
currently the C++03 libc++ copy constructor for pair<int, int> is ABI
incompatible with the C++11 libc++ copy constructor for pair<int, int>.
This is Bad(tm). This patch corrects the situation by making this copy
constructor trivial in C++03 mode as well.
Just in case it is needed for an incomplete C++11 compiler, libc++
retains the ability to support pair with rvalue references, but without
defaulted special members. However the pair needs non-trivial special
members to implement this special case, (as it did when clang was in
this place a couple of years ago).
During this work a bug was also found and fixed in
is_trivially_constructible.
And there is a minor drive-by fix in <__config> regarding
__type_visibility__.
A test is updated to ensure that the copy constructor of pair<int, int>
is trivial in both C++03 and C++11. This test will necessarily fail for
a compiler that implements rvalue references but not defaulted special
members.
llvm-svn: 194536
section in libc++. This requires a recompiled dylib. Failure to rebuild
the dylib will result in a link-time error if and only if the functions from
[util.smartptr.shared.atomic] are used.
The implementation is not lock free. After considerable thought, I know of no
way to make the implementation lock free. Ideas welcome along that front. But
changing the ABI of shared_ptr is not on the table at this point.
The mutex used to lock these function is encapsulated by std::__sp_mut. The
only thing the client knows about std::__sp_mut is that it has a void* data
member, can't be constructed, and has lock and unlock members. Within the
binary __sp_mut is currently implemented as a pointer to a std::mutex. That can
change in the future without disturbing the ABI (as long as sizeof(__sp_mut)
remains constant.
I specifically did not make __sp_mut a spin lock as I have a pathological
distrust of spin locks. Testing on OS X reveals that the use of std::mutex in
this role is not a large performance penalty as long as the contention for the
mutex is low (more likely to get the lock than to have to wait). In the future
we can still make __sp_mut a spin lock if that is what is desired (without ABI
damage).
The dylib contains 16 __sp_mut's to be chosen based on the hash of the address
of the shared_ptr. The constant 16 is a ball-park reasonable space/time
tradeoff.
std::hash<T*> was changed to call __murmur2_or_cityhash, instead of the identity
function. I had thought we had already done this, but I was mistaken.
All of this is under #if __has_feature(cxx_atomic) even though the
implementation is not lock free, because the signatures require access to
std::memory_order, which is currently available only under
__has_feature(cxx_atomic).
llvm-svn: 160940
Objective-C Automatic Reference Counting, where Objective-C object
pointers can have several different qualifiers (__strong, __weak,
__autoreleasing, __unsafe_unretained). These addressof() overloads are
only provided in ARC mode, and the __weak variant is conditionalized
on having weak-reference support in the ARC runtime.
For historical reasons, Clang provides these definitions itself, and
defines the macro _LIBCPP_PREDEFINED_OBJC_ARC_ADDRESSOF to note when
it as done so. The code belongs here, and this redundancy will be
eliminated in the future.
Addresses <rdar://problem/9658274>.
llvm-svn: 133656
Erik Pilkington