svn 14838

[Rob Landley]

On Friday 14 April 2006 4:07 am, Bernhard Fischer wrote:
> That #define _GNU_SOURCE suggests that it isn't really too portable.
> We do have a drop-in (libbb/vdprintf.c) for (__GLIBC__ < 2), but
> since {v,}dprintf isn't used too often we should drop it for the
> sake of portability, i think.

We don't support libc5.  We support uClibc and glibc 2, and we're leaving the 
door open for newlib and whatever macos X uses.

Both uClibc and glibc 2 support dprintf just fine, by the way, and have for a 
good ten years now.

> If this would happen, we can also remove
> static void fdprintf(int fd, const char *ctl, ...) from crond.c
>
> Current users of {v,}dprintf are:
> coreutils/tail.c:1
> coreutils/watch.c:1
> miscutils/crond.c:4
> shell/ash.c:2
> sysklogd/syslogd.c:2
>
>
> Sounds ok?

What's the objection again?  This has been there for how many years now?  
Could we have a specific objection, like "MacOS X has a name that conflicts 
with this feature of glibc?"

"Being portable" is not a specific objection.  "This breaks on MacOS X" is a 
specific objection...

Rob
-- 
Never bet against the cheap plastic solution.