[Mingw-w64-public] [PATCH] Merge libapi-ms-... into libruntimeobject

2016-04-20 Thread Hugo Beauzée-Luyssen
--- mingw-w64-crt/Makefile.am | 16 + mingw-w64-crt/Makefile.in | 39 -- mingw-w64-crt/lib32/Makefile.am| 2 +- .../lib32/api-ms-win-core-winrt-l1-1-0.def | 9 + .../lib32/api-ms-win-cor

Re: [Mingw-w64-public] [PATCH 3/5] Revert "Added libruntimeobject.a and use it instead of api-*.a files."

2016-04-20 Thread Kai Tietz
The issue is that multiple imports of different DLLs are poorly handled by dlltool. There happens some issues in that case. We might want to consider to provide our own tool for just generation import libraries Cheers, Kai 2016-04-20 15:15 GMT+02:00 Jacek Caban : > On 04/20/16 15:12, Hugo

Re: [Mingw-w64-public] [PATCH 3/5] Revert "Added libruntimeobject.a and use it instead of api-*.a files."

2016-04-20 Thread Jacek Caban
On 04/20/16 15:15, Jacek Caban wrote: > On 04/20/16 15:12, Hugo Beauzée-Luyssen wrote: >> This reverts commit f3f8cd3173d82fa6f4ae31ecf8189e5ff19b1ca9. >> > Why? Ah, I see that in the other patch. I'd just send one patch instead of a separated revert, but that's fine with me. Thanks, Jacek -

Re: [Mingw-w64-public] [PATCH 3/5] Revert "Added libruntimeobject.a and use it instead of api-*.a files."

2016-04-20 Thread Jacek Caban
On 04/20/16 15:12, Hugo Beauzée-Luyssen wrote: > This reverts commit f3f8cd3173d82fa6f4ae31ecf8189e5ff19b1ca9. > Why? Jacek -- Find and fix application performance issues faster with Applications Manager Applications Man

[Mingw-w64-public] [PATCH 3/5] Revert "Added libruntimeobject.a and use it instead of api-*.a files."

2016-04-20 Thread Hugo Beauzée-Luyssen
This reverts commit f3f8cd3173d82fa6f4ae31ecf8189e5ff19b1ca9. --- mingw-w64-crt/Makefile.am | 3 +- mingw-w64-crt/Makefile.in | 39 -- mingw-w64-crt/lib32/Makefile.am| 3 +- .../lib32/api-ms-win-core-winrt-

[Mingw-w64-public] [PATCH 5/5] Merge libapi-ms-... into libruntimeobject

2016-04-20 Thread Hugo Beauzée-Luyssen
Using a single .def file seems to make the WACK confused --- mingw-w64-crt/Makefile.am | 19 +++ mingw-w64-crt/lib32/Makefile.am| 3 +-- mingw-w64-crt/lib32/runtimeobject.mri | 5 + mingw-w64-crt/lib64/runtimeobject.mri | 5 + mingw-w64-crt/winrt/l

[Mingw-w64-public] [PATCH 1/5] QueueUserAPC is allowed on Windows 10

2016-04-20 Thread Hugo Beauzée-Luyssen
--- mingw-w64-headers/include/processthreadsapi.h | 3 +++ 1 file changed, 3 insertions(+) diff --git a/mingw-w64-headers/include/processthreadsapi.h b/mingw-w64-headers/include/processthreadsapi.h index c04e79c..f8c1480 100755 --- a/mingw-w64-headers/include/processthreadsapi.h +++ b/mingw-w64-

[Mingw-w64-public] [PATCH 2/5] _begin/endthread(ex) are not allowed on winrt

2016-04-20 Thread Hugo Beauzée-Luyssen
Provide a redirection from _beginthreadex -> CreateThread _endthread -> ExitThread, as the documentation recommends. --- mingw-w64-libraries/winstorecompat/Makefile.am | 1 + .../winstorecompat/src/beginthread.c | 54 ++ 2 files changed, 55 insertions(+) cre

[Mingw-w64-public] [PATCH 4/5] processthreadsapi: Also define {TLS|FLS}_OUT_OF_INDEXES for winrt

2016-04-20 Thread Hugo Beauzée-Luyssen
These macros are used to check for TlsAlloc & FlsAlloc success. Both these functions are allowed on winrt --- mingw-w64-headers/include/processthreadsapi.h | 9 + 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/mingw-w64-headers/include/processthreadsapi.h b/mingw-w64-header

Re: [Mingw-w64-public] __attribute__((__force_align_arg_pointer__)) and callbacks when SSE is enabled

2016-04-20 Thread Carl Kleffner
Hi, what about making "-mincoming-stack-boundary=2" as a default for i686? Carl 2016-04-20 14:28 GMT+02:00 Kai Tietz : > Hi, > > this issue has bitten me some times already, too. Nevertheless I > don't think that adding this attribute to WINAPI, APIENTRY, ... are > the right thing to do here.

Re: [Mingw-w64-public] __attribute__((__force_align_arg_pointer__)) and callbacks when SSE is enabled

2016-04-20 Thread Kai Tietz
Hi, this issue has bitten me some times already, too. Nevertheless I don't think that adding this attribute to WINAPI, APIENTRY, ... are the right thing to do here. This seems to me more like an assumption done by one compiler, which is obviously wrong in some conditions as it assumes that a spe

Re: [Mingw-w64-public] Adding a new thread model to GCC

2016-04-20 Thread lh_mouse
One more question about block-scoped static objects: >From compiled assembly I have learned that GCC uses a 64-bit integer guard to >ensure once-initialization of static objects with block scopes. Code in gcc/libstdc++-v3/libsupc++/guard.cc uses a global mutex to protect multiple threads from ra