0
我準備了一個針對32位Windows的llvm-clang編譯器。我使用3.7.1並應用this patch。我得到的東西，幾乎的作品，除了我有提升錯誤。在Windows下Mingw + clang vs Boost：不同類型的錯誤
如果我編譯我的項目（即通常與MinGW的海灣合作委員會的工作）使用我平時提升1.53頭我有麻煩，包括boost/interprocess組件
In file included from [...]/Boost/include-WINDOWS/boost-1_53\boost/interprocess/sync/interprocess_semaphore.hpp:24:
In file included from [...]/Boost/include-WINDOWS/boost-1_53\boost/interprocess/exceptions.hpp:20:
In file included from [...]/Boost/include-WINDOWS/boost-1_53\boost/interprocess/errors.hpp:37:
[...]/Boost/include-WINDOWS/boost-1_53\boost/interprocess/detail/win32_api.hpp:878:48: error: conflicting types for 'VirtualProtect'
如果我再嘗試recompile boost有：
bootstrap.bat mingw
b2 toolset=clang --with-thread --with-filesystem --with-signals --with-test --with-system --with-regex --with-program_options --with-serialization variant=release link=shared threading=multi
取決於各種版本的提升，我得到很多警告，我可以嘗試忽略，但也有一些錯誤：
1.56 1.57只是這些：（並不多，但我不知道如何擺脫它）
In file included from libs\serialization\src\xml_wiarchive.cpp:35:
.\boost/archive/impl/archive_serializer_map.ipp:41:34: error: redeclaration of 'boost::archive::detail::archive_serializer_map::insert' cannot add 'dllexport' attribute
archive_serializer_map<Archive>::insert(const basic_serializer * bs){
^
.\boost/archive/detail/archive_serializer_map.hpp:42:17: note: previous declaration is here
static bool insert(const basic_serializer * bs);
^
In file included from libs\serialization\src\xml_wiarchive.cpp:35:
.\boost/archive/impl/archive_serializer_map.ipp:49:34: error: redeclaration of 'boost::archive::detail::archive_serializer_map::erase' cannot add 'dllexport' attribute
archive_serializer_map<Archive>::erase(const basic_serializer * bs){
^
.\boost/archive/detail/archive_serializer_map.hpp:43:17: note: previous declaration is here
static void erase(const basic_serializer * bs);
^
In file included from libs\serialization\src\xml_wiarchive.cpp:35:
.\boost/archive/impl/archive_serializer_map.ipp:61:34: error: redeclaration of 'boost::archive::detail::archive_serializer_map::find' cannot add 'dllexport' attribute
archive_serializer_map<Archive>::find(
^
.\boost/archive/detail/archive_serializer_map.hpp:44:37: note: previous declaration is here
static const basic_serializer * find(
^
爲1.59和1.60一堆MMX相關的錯誤，這就是，只要我所看到的，而不是功能準備我的LLVM-鐺時，我可以控制：
clang-linux.compile.c++.without-pth bin.v2\libs\thread\build\clang-linux-3.7.1\debug\threading-multi\win32\thread.obj
In file included from libs\thread\src\win32\thread.cpp:17:
In file included from .\boost/thread/thread_only.hpp:15:
In file included from .\boost/thread/win32/thread_data.hpp:11:
In file included from .\boost/thread/win32/thread_primitives.hpp:18:
In file included from .\boost/detail/interlocked.hpp:154:
In file included from C:\Programs\LLVM\bin\..\lib\clang\3.7.1\include\intrin.h:26:
In file included from C:\Programs\MinGW\i686-w64-mingw32\include\Intrin.h:154:
In file included from C:\Programs\LLVM\bin\..\lib\clang\3.7.1\include\mm3dnow.h:27:
C:\Programs\LLVM\bin\..\lib\clang\3.7.1\include\mmintrin.h:28:2: error: "MMX instruction set not enabled"
#error "MMX instruction set not enabled"
^
In file included from libs\thread\src\win32\thread.cpp:17:
In file included from .\boost/thread/thread_only.hpp:15:
In file included from .\boost/thread/win32/thread_data.hpp:11:
In file included from .\boost/thread/win32/thread_primitives.hpp:18:
In file included from .\boost/detail/interlocked.hpp:154:
In file included from C:\Programs\LLVM\bin\..\lib\clang\3.7.1\include\intrin.h:26:
In file included from C:\Programs\MinGW\i686-w64-mingw32\include\Intrin.h:154:
C:\Programs\LLVM\bin\..\lib\clang\3.7.1\include\mm3dnow.h:42:41: error: use of undeclared identifier '__v8qi'
return (__m64)__builtin_ia32_pavgusb((__v8qi)__m1, (__v8qi)__m2);
^
C:\Programs\LLVM\bin\..\lib\clang\3.7.1\include\mm3dnow.h:42:55: error: use of undeclared identifier '__v8qi'
return (__m64)__builtin_ia32_pavgusb((__v8qi)__m1, (__v8qi)__m2);
^
C:\Programs\LLVM\bin\..\lib\clang\3.7.1\include\mm3dnow.h:47:38: error: cannot convert '__m64' to '__v2sf' (vector of 2 'float' values) without a conversion operator
return (__m64)__builtin_ia32_pf2id((__v2sf)__m);
^~~~~~~~~~~
C:\Programs\LLVM\bin\..\lib\clang\3.7.1\include\mm3dnow.h:52:38: error: cannot convert '__m64' to '__v2sf' (vector of 2 'float' values) without a conversion operator
return (__m64)__builtin_ia32_pfacc((__v2sf)__m1, (__v2sf)__m2);
^~~~~~~~~~~~
C:\Programs\LLVM\bin\..\lib\clang\3.7.1\include\mm3dnow.h:52:52: error: cannot convert '__m64' to '__v2sf' (vector of 2 'float' values) without a conversion operator
return (__m64)__builtin_ia32_pfacc((__v2sf)__m1, (__v2sf)__m2);
^~~~~~~~~~~~
是否有一種方式來獲得提升鏗鏘的32位工作在Windows下的MinGW的設置的任何版本？
我也試過llvm 3.8.0，但我不能準備它，因爲我的MinGW工具鏈不支持C++ 11的某些部分（以及其他版本，我嘗試過，請參閱here）。
對於第二個問題，它可能與[該補丁（http://lists.llvm.org/pipermail/llvm-commits/Week -of-MON-20151109/312887.html） – Antonio