[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]
Re: fixing devel/gmp for MacOS X/clang
Am 20.09.2013 um 14:09 schrieb Joerg Sonnenberger
> On Fri, Sep 20, 2013 at 01:58:03PM +0200, Jens Rehsack wrote:
>> Am 20.09.2013 um 13:52 schrieb Adam Ciarciński <adam%NetBSD.org@localhost>:
>>>> Well - no :)
>>>> $ ls -l
>>>> total 168
>>>> -rw-r--r-- 1 root wheel 2214 19 Sep 13:58 string
>>>> drwxr-xr-x 57 root wheel 1938 20 Jun 2012 tr1
>>>> -rw-r--r-- 1 root wheel 5929 19 Sep 13:58 typeinfo
>>>> -rw-r--r-- 1 root wheel 2789 19 Sep 13:58 utility
>>>> $ $ find
>>>> -name tuple
>>> Well, yes. These includes are for libc++, not libstdc++. Provide
>>> -stdlib=libc++ and it should compile just fine.
>> Okay - the symbolic link isn't needed, just the -stdlib=libc++ flag
>> Does that patch looks better for you?
> Please hold off on this. I strongly believe that it should become the
> default after the branch and be handled in a unified fashion.
Yes - this got mixed between devel/googletest and devel/gmp.
The devel/gmp thing doesn't use c++ and failes because of obstack.h is missing
(which happens fairly often on MacOS X that some libc exports are available
without prototypes to "support" old ABI compatibility).
The longer discussion with adam@ was about devel/googletest and I think
doing a right fix and merge it later is fair enough. OTOH we can do the
bulk build using Xcode 4 and gcc.
Main Index |
Thread Index |