[OpenAFS] AKLOG build error...

Derek Atkins warlord@MIT.EDU
Fri, 28 Apr 2006 16:28:35 -0400


Derrick J Brashear <shadow@dementia.org> writes:

> On Fri, 28 Apr 2006, Halter, Jonathan wrote:
>> I've done several tests here, and it seems that we may be getting bit by
>> some hardcoding of CC somewhere in OPenAFS's configure. No matter what I
>> specify for CC, by the end of configure, CC always ends up being set to
>> /opt/SUNWspro/bin/cc. This is a problem on our systems, since our
>> compilers are not located there.
> Yes, that's not surprising. There are autoconf macros 
> (src/cf/solaris-cc.m4, i think) which could be hooked in, or better yet a 
> test which tries to determine if a gcc which supports compiling Solaaris 
> kernel modules (and uses the requisite flags instead of cc flags in 
> libafs/MakefileProto.SOLARIS.in) would be nice, if someone wants to do the 
> work on that and contribute it I will certainly find some time to work 
> with them on it, but at the moment I don't have time to just do the work.
> There's no point in testing. We know it really wants /opt/SUNWspro/bin/cc. 
> If you want it to not want that, it's going to need some help.

I think the problem is that the configure script swaps over to
/opt/SUNWspro/bin/cc too early, before it runs the krb5 checks,
and therefore causes the tests to fail.

So..  Perhaps we need to move all the kernel tests later in the
configure script and have all the user-space tests earlier, so we
don't wind up with this problem?

       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord@MIT.EDU                        PGP key available