pseudo-random HP-UX 11i compile things

[Mark_Andrews at isc.org]

	Known issue already fixed in CVS.

> 
> I've been messing about with the 9.3.0 snapshot (bind-9.3.0s20021115) as
> I was thinking of doing another round of perf work,and initially the
> 9.2.2rc went' belly-up on a compile - I figured that the 11.11 and later
> IFmumble ioctl source changes hadn't made it into that, hence the
> subsequent grabbing of a snapshot.
> 
> My initial ./configure went OK on both 11.11 (11i, ostensibly no IPv6)
> and 11.22 (11i v 1.6 again no IPv6) and the compiles (while generating
> _lots_ of storing ints into shorts and other warnings on 11.22) appeared
> OK, but the link of named went belly-up with an unsatisfied symbol on
> if_nametoindex().
> 
> I poked into the code and noticed that routine seemed to be associated
> with v6 things, so I redid the configure (after a make distclean),
> specifying --disable-ipv6.
> 
> Now the link of named fails - again with if_nametoindex, and with
> isc_net_probeipv6 as well. It was a bit of a surprise seeing a "v6"
> routine name after saying --disable-ipv6 :)
> 
> Probing that, I found that isc_net_probeipv6 is called in a number of
> places, regardless of the setting of "WANT_IPV6" I'm not sure if the
> "right" thing to do is start putting #ifdefs around the calls, or have a
> stub version created when WANT_IPV6 isn't set...
> 
> Deciding to route around that, I went to an 11i (11.11) system on which
> IPv6 _was_ installed. Again configure, with default settings, again an
> unsatisfied on if_nametoindex(). I went searching in /usr/lib and found
> a libipv6.sl with that symbol name in it.
> 
> If I set LIBS="-lipv6" before doing the configure, the compile and link
> for named appears to be happy and content.
> 
> Of course, if I set LIBS="-lipv6" and configure on a system where IPV6
> has not been installed/enabled, configure gets very cranky very early.
> 
> Does this mean that some enhancements to the configure script are
> indicated?
> 
> It also appears (?) that the if_nametoindex() stuff is called
> independent of WANT_IPV6. What I don't know is if "if_nametoindex" is
> supposed to be something associated only with IPv6, or if HP-UX 11i's
> pre-enablement of v6 without full v6 until a patch is applied is at
> fault. I'm sure though that I could find some opinions here :)
> 
> rick jones
> -- 
> Wisdom Teeth are impacted, people are affected by the effects of events.
> these opinions are mine, all mine; HP might not want them anyway... :)
> feel free to post, OR email to raj in cup.hp.com  but NOT BOTH...
> 
--
Mark Andrews, Internet Software Consortium
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: Mark.Andrews at isc.org