On Sun, 26 Sep 2010, Brad Jorsch wrote:
> On Sun, Sep 26, 2010 at 05:40:42PM +0200, Tamas TEVESZ wrote:
> >
> > + if (uname(&un) != -1) {
> > + wwarning(_("%s is not implemented on this platform; "
> > + "tell [email protected] you are
> > running "
> > + "%s release %s version %s"), __FUNCTION__,
> > + un.sysname, un.release, un.version);
> > + notified = 1;
> > + }
>
> No fallback if uname() fails? Not that it should ever fail, but...
my line of thinking, if it fails the first time (which, as you noted,
shouldn't, really), it will succeed the second time.
if i provide fallback, that will either "flood" (haha), or (by virtue
of falling back to something less precise) lose information.
i don't think anything more fool-proof is needed in this particular
case. just think of the next run as the fallback :)
--
[-]
mkdir /nonexistent
--
To unsubscribe, send mail to [email protected].
