[Bf-committers] OSX Mavericks 10.9 (mostly) success

[Jake Kauth]

On 9/10/13, 2:11 PM, Jens Verwiebe wrote:
> For XCode5 dropped llvm-gcc complete, we are pushed to another solution, which could be
> Instead of using __attribute__((weak_import)) inside an extern "C" declaration, we
> should move these availablility checks to a discrete c file.
>
> I read clang trunk may already have a fix, whereas other say the issue still occures with -O2.
> My conclusion: lest check if a discrete C format file works here. Not too complcated.
>
> Can you or Mike try? I'am a bit short with time nowadays ...

Success!  I was able to factor out all of the "extern "C" calls so they 
are in plain C files and no longer ObjC++;  I was able to successfully
build a blender weak-linked against 3DConnectionClient framework.

Here is a "pasteall" of the diff: http://www.pasteall.org/45558/diff

I'm using cmake and I didnt' try to build via scons on this machine, so 
I don't know if changes to SConscript to build 
libbf_intern_ghostndof3dconnexion.a library file are correct.

jake
*****



On 9/10/13, 2:11 PM, Jens Verwiebe wrote:
> Hi Jake
>
> I investigated the ndof thingie and found following situation.
>
> It seems that with clang, setting attributes leads to producing mangled additional (c++-)symbols:
>
> $ clang++33 -c test.cc
> $ objdump -t test.o | fgrep foo
> 0000000000000010 l     F .text  0000000000000006 _Z3foov
> 0000000000000000         *UND*  0000000000000000 foo
>
> $ g++48 -c test.cc
> $ objdump -t test.o | fgrep foo
> 0000000000000000 l     F .text  0000000000000006 foo
> ..not exactly our problem, but i think the same outcome.
> I formerly simple circumvented this by just compiling ghost with llvm-gcc ( exception ).
> For XCode5 dropped llvm-gcc complete, we are pushed to another solution, which could be
> Instead of using __attribute__((weak_import)) inside an extern "C" declaration, we
> should move these availablility checks to a discrete c file.
>
> I read clang trunk may already have a fix, whereas other say the issue still occures with -O2.
>
> My conclusion: lest check if a discrete C format file works here. Not too complcated.
>
> Can you or Mike try? I'am a bit short with time nowadays ...
>
> Cheers ...Jens
>
>
>
>
> _____________________________________
>
> Jens Verwiebe
> Allerskehre 44  -  22309 Hamburg
>
> Tel.: +49 40 68 78 50
> mobil: +49 172 400 49 07
> mailto: info at jensverwiebe.de
> web:  http://www.jensverwiebe.de
> _____________________________________
>
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers
>