On Mon, 8 Nov 2010, Simon Wilkinson wrote:
On 8 Nov 2010, at 12:00, Niklas Edmundsson wrote:
That build was against the IBM krb5 (--with-krb5=/usr/krb5), so it shouldn't
have touched the heimdal stuff at all. As I have understood it the IBM krb is
MIT-based.
If the system has Heimdal installed in a standard search path, then
we'll try and use the libroken from that installation. You can work
around this by passing --with-roken=internal which forces the use of
an internal libroken.
Ah, that might be it.
Perhaps it would be best if I removed the installed heimdal altogether
on the buildbot LPAR? Even though it provides for interesting tests of
the build systems ability to disqualify old prereqs it has found I
suspect that it's more interesting with an environment similar to the
one that you might find on a machine with the bare minimum prereqs
installed for building an openafs release tarball...
/Nikke
--
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Niklas Edmundsson, Admin @ {acc,hpc2n}.umu.se | [email protected]
---------------------------------------------------------------------------
"...to boldly go where no mallard has gone before!" - Darkwing Duck
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
_______________________________________________
OpenAFS-devel mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-devel
