Re: [PATCH v2 1/4] Feature test macro overhaul

[Corinna Vinschen <vinschen at redhat dot com>]

On Feb 23 20:33, Yaakov Selkowitz wrote:
> On 2016-02-23 12:27, Corinna Vinschen wrote:
> >>+#if !defined(__STRICT_ANSI__) && !defined(_ANSI_SOURCE) && \
> >>+  !defined(_ISOC99_SOURCE) && !defined(_POSIX_SOURCE) && \
> >>+  !defined(_POSIX_C_SOURCE) && !defined(_XOPEN_SOURCE) && \
> >>+  !defined(_BSD_SOURCE) && !defined(_SVID_SOURCE)
> >>+#define _BSD_SOURCE
> >>+#define _SVID_SOURCE
> >>+#endif
> >
> >Can you please explain this expression?  Why is _BSD_SOURCE + _SVID_SOURCE
> >the default?  Is that derived from some BSD default setting?
> 
> If neither __STRICT_ANSI__ nor any _*_SOURCE macros are defined, then the
> default is, as you noted, BSD, SVID, and POSIX.1-2008, the latter of which
> is turned on in the next chunk.
> 
> >The question is a bit dumb perhaps.  It's a result of me digging into
> >the Linux feature_test_macros man page which contains basically the
> >following information:
> >
> >- _BSD_SOURCE is deprecated and has the same effect as _DEFAULT_SOURCE
> >- _SVID_SOURCE is deprecated and has the same effect as _DEFAULT_SOURCE
> >- _DEFAULT_SOURCE is effectively
> >
> >     _BSD_SOURCE | _SVID_SOURCE | _POSIX_C_SOURCE=200809
> 
> I had thought _DEFAULT_SOURCE was really new, but time flies.  I'll look
> into incorporating it into the next draft.
> 
> >>+#if defined(_POSIX_C_SOURCE) && _POSIX_C_SOURCE >= 200809
> >>+#undef _ATFILE_SOURCE
> >>+#define _ATFILE_SOURCE
> >>+#endif
> >
> >Maybe better
> >
> >   #define _ATFILE_SOURCE 1
> 
> Fixed locally, and the same for the aforementioned BSD/SVID defines.

The other patches look good to me, but descriptive git log messages for
all patches are desperately missing, so please send a v3 with those
included.


Thanks,
Corinna

-- 
Corinna Vinschen
Cygwin Maintainer
Red Hat

Attachment: signature.asc
Description: PGP signature