On 9/17/07, Derrick Brashear <[EMAIL PROTECTED]> wrote: > On 9/4/07, Suzanna <[EMAIL PROTECTED]> wrote: > > > Why the 1.5 branch? It doesn't look like amd64 support was merged. > > > > Trying to be up to date. I wasn't aware of the missing amd64 code. > > 1.5 isn't the recommended unix release, anyway, so "up to date" if you want > bleeding is the CVS head and not 1.5 anyway.
My solution is --disable-pam. That's what the new version in pkgsrc has set for the OpenAFS package, that should now support the 3.x releases and the 4.x pre-release versions of NetBSD on amd64 & i386. (OpenAFS 1.4.5, patches in RT, #75602, I suppose I should test CVS head too.) > > > Are you using process.amd64.s? What happens when you add -fPIC to > > > the compile options? > > > > I am. Adding -fPIC doesn't change anything as it is not passed to > > as(1). Apparently, as(1) has no options to force the creation of > > PIC compatible code. Adding it to cc -E before doesn't change > > anything, either. > > > > The error occurs when pam_afs.so.1 is linked... I was wondering about > > linking a lib.a into a lib.so, anyway. > > It's doable if your dynamic linker supports it, but most amd64 linkers > don't. I'm working a bit, when I'm up in the middle of the night, with someone who understands this better than I do. And I have some hope of help finishing the client support on NetBSD soon, too. -Tracy _______________________________________________ OpenAFS-devel mailing list [email protected] https://lists.openafs.org/mailman/listinfo/openafs-devel
