I *did* try to check that such kind of an issue was not introduced
by the rL315874, but clearly i failed to finish verification.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@315876 91177308-0d34-0410-b5e6-96231b3b80d8
Previously this broke the builders, when D38101 was committed.
Silence the warning so that it can be re-landed.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@315874 91177308-0d34-0410-b5e6-96231b3b80d8
The vcruntime headers are hairy and clash with both libc++ headers
themselves and other libraries. libc++ normally deals with the clashes
by deferring to the vcruntime headers and silencing its own definitions,
but for clients which don't want to depend on vcruntime headers, it's
desirable to support the opposite, i.e. have libc++ provide its own
definitions.
Certain operator new/delete replacement scenarios are not currently
supported in this mode, which requires some tests to be marked XFAIL.
The added documentation has more details.
Differential Revision: https://reviews.llvm.org/D38522
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@315234 91177308-0d34-0410-b5e6-96231b3b80d8
Make it clear that these are intended only to force a specific ABI when
the autodetection would give the wrong result by renaming the cmake
options and adding separate forcing macros, as suggested by EricWF in
the post-commit review of r314949 and further discussed on IRC.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@314965 91177308-0d34-0410-b5e6-96231b3b80d8
libc++'s current heuristic for detecting Itanium vs. Microsoft ABI falls
short in some cases. For example, it will detect windows-itanium targets
as using the Microsoft ABI, since they set `_MSC_VER` (for compatibility
with Microsoft headers). Leave the current heuristic in place by default
but also allow users to explicitly specify the ABI if need be.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@314949 91177308-0d34-0410-b5e6-96231b3b80d8
Some ABI macros affect headers, so it's nice to have a site config
option for them. Add a LIBCXX_ABI_DEFINES cmake macro to allow
specifying a list of ABI macros to define in the site config.
The primary design constraint (as discussed with Eric on IRC a while
back) was to not have to repeat the ABI macro names in cmake, which only
leaves a free-form cmake list as an option. A somewhat unfortunate
consequence is that we can't verify that the ABI macros being defined
actually exist, though we can at least perform some basic sanity
checking, since all the ABI macros begin with _LIBCPP_ABI_.
Differential Revision: https://reviews.llvm.org/D36719
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@314946 91177308-0d34-0410-b5e6-96231b3b80d8
This patch adds the newly standardized deduction guides
for std::pair, allowing it to work class template deduction.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@314864 91177308-0d34-0410-b5e6-96231b3b80d8
Previously LIT would often fail while attempting to set up/configure
the test compiler; normally when attempting to dump the builtin macros.
This sort of failure provided no useful information about what went
wrong with the compiler, making the actual issues hard --- if not
impossible --- to debug easily.
This patch changes the LIT configuration to report the failure explicitly,
including the failed compile command and the stdout/stderr output.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@314735 91177308-0d34-0410-b5e6-96231b3b80d8
This warning "structure was padded due to alignment specifier" says
that the compiler is going to do exactly what you asked it to do.
It's triggered by the tests for over-aligned dynamic memory allocation.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@314257 91177308-0d34-0410-b5e6-96231b3b80d8
Despite a strong CMake warning that this is an unsupported
libcxx build configuration, some bots still rely on being
able to check out lit and libcxx independently with no
LLVM sources, and then run lit against libcxx.
A previous patch broke that workflow, so this is making it work
again. Unfortunately, it breaks generation of the llvm-lit
script for libcxx, but we will just have to live with that until
a solution is found that allows libcxx to make more use of
llvm build pieces. libcxx can still run tests by using the
ninja check target, or by running lit.py directly against the
build tree or source tree.
Differential Revision: https://reviews.llvm.org/D38057
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313763 91177308-0d34-0410-b5e6-96231b3b80d8
Summary:
This patch replaces __sync_* with __libcpp_atomic_* and adds a wrapper
function for __atomic_exchange to support _LIBCPP_HAS_NO_THREADS.
Reviewers: EricWF, jroelofs, mclow.lists, compnerd
Reviewed By: EricWF, compnerd
Subscribers: compnerd, efriedma, cfe-commits, joerg, llvm-commits
Differential Revision: https://reviews.llvm.org/D35235
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313694 91177308-0d34-0410-b5e6-96231b3b80d8
After speaking with the libcxx owners, they agreed that this is
a bug in the bot that needs to be fixed by the bot owners, and
the CMake changes are correct.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313643 91177308-0d34-0410-b5e6-96231b3b80d8
This reverts commit 4ad71811d45268d81b60f27e3b8b2bcbc23bd7b9.
There is a bot that is checking out libcxx and lit with nothing
else and then running lit.py against the test tree. Since there's
no LLVM source tree, there's no LLVM CMake. CMake actually
reports this as a warning saying unsupported libcxx configuration,
but I guess someone is depending on it anyway.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313607 91177308-0d34-0410-b5e6-96231b3b80d8
Patch from Eddie Elizondo. Reviewed as D37830 (https://reviews.llvm.org/D37830).
On MacOSX the following program:
struct S { virtual void f() = delete; };
int main() { new S; }
Fails with the following error:
Undefined symbols for architecture x86_64:
"___cxa_deleted_virtual"
This adds a fix to export the needed symbols.
Test:
> lit -sv test/libcxx/language.support/cxa_deleted_virtual.pass.cpp
> Testing Time: 0.21s
> Expected Passes : 1
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313500 91177308-0d34-0410-b5e6-96231b3b80d8
Clang recently changed the way it outputs static assert diagnostics.
This patch fixes libc++'s -verify tests so they tolerate both the old
and new message format.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313499 91177308-0d34-0410-b5e6-96231b3b80d8
Microsoft's CRT already provides the const overloads, and it defines the
`_CRT_CONST_CORRECT_OVERLOADS` macro to indicate their presence. Check
for this macro before attempting to define our own const-correct
overloads, to avoid compiler warnings about casts dropping const
qualifiers.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313377 91177308-0d34-0410-b5e6-96231b3b80d8
The RTTI structure is different on Windows when building under MS ABI.
Update the definition to reflect this. The structure itself contains an
area for caching the undecorated name (which is 0-initialized). The
decorated name has a bitfield followed by the linkage name. When
std::type_info::name is invoked for the first time, the runtime should
undecorate the name, cache it, and return the undecorated name. This
requires access to an implementation of __unDName. For now, return
the raw name.
This uses the fnv-1a hash to hash the name of the RTTI. We could use an
alternate hash (murmur? city?), but, this was the quickest to throw
together.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313344 91177308-0d34-0410-b5e6-96231b3b80d8
There were a number of cases where __double_underscore functions,
for example __has_construct_test, were called without being qualified,
causing ADL to occur. This patch qualifies those calls to avoid this
problem.
Thanks to David L. Jones for point out the issue initially.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313324 91177308-0d34-0410-b5e6-96231b3b80d8
If we define cmake macros that require a site config, and then undefine
all such macros, a stale site config header will be left behind.
Explicitly delete any generate site config if we don't need one to avoid
this.
Differential Revision: https://reviews.llvm.org/D36720
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313284 91177308-0d34-0410-b5e6-96231b3b80d8
Using the system default 'ar' might not be the right choice when
cross compiling.
Don't prepend the ar options by a dash, not all ar implementations
support that (llvm-ar doesn't).
Also pass the 's' option when creating the merged library, to create
an index.
Differential Revision: https://reviews.llvm.org/D37134
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@313122 91177308-0d34-0410-b5e6-96231b3b80d8
This patch fixes llvm.org/PR34298. Previously libc++ incorrectly evaluated
the __invokable trait via the converting constructor `function(Tp)` [with Tp = std::function]
whenever the copy constructor or copy assignment operator
was required. This patch further constrains that constructor to short
circut before evaluating the troublesome SFINAE when `Tp` matches
std::function.
The original patch is from Alex Lorenz.
git-svn-id: https://llvm.org/svn/llvm-project/libcxx/trunk@312892 91177308-0d34-0410-b5e6-96231b3b80d8
