[Nut-upsuser] make errors on solaris express dev 02/07

Charles Lepple clepple at gmail.com
Wed Mar 14 19:24:21 CET 2007


On 3/14/07, Arjen de Korte <nut+users at de-korte.org> wrote:
> Charles Lepple wrote:
>
> > Arjen, could you please take a look at this?
> >
> > from 'svn blame' on server/access.c:73:
> >
> >   737 adkorte-guest            return (IN6_IS_ADDR_V4MAPPED(ip6) &&
> >   737 adkorte-guest                    ((((const u_int32_t *)ip6)[3]
> > & prefix) == net->s_addr));
> >
> > Is there a more portable way of matching the embedded IPv4 address?
>
> Ironically, this is a portability macro defined in RFC 2553. But with a
> little Google'ing around, I found out that the implementation in glibc
> 2.1 is broken. We'll follow the same path as the KDE, Apache and Bind
> developers followed and (conditionally) override this.

I thought the error that Zoltan saw might have been due to the next
line (the "const u_int32_t" bit), but I don't have a Solaris box to
test.

-- 
- Charles Lepple



More information about the Nut-upsuser mailing list