Ah, we have a few AIX systems, but only one is at 5.3 level and has PAM, it also had a broken native compiler, hence me using gcc. Our bigger system, and the one we're actually building imapd for will be getting upgraded, I think over this weekend, so I won't be able to try a native compile until next week, so please don't trouble yourself about it quite yet, I'll post back if I fail to get it running using a41 and the native compiler.
___ ___ /___\ herbie <at> uic <dot> edu /___\ (o\|/o) There is no such thing as illegal immigrants, (o\|/o) U^^^U only illegal Governments. - (Ambalavaner Sivanandan, 1997) U^^^U On Thu, 27 Jul 2006, Mark Crispin wrote: > We are getting rid of our AIX systems (hooray!) and none of them had PAM. > So I don't know much about how IBM implemented PAM; there are subtle > differences (e.g., the need to use use PASSWDTYPE=pmb on Solaris) in > different ports. > > The other issue is that you're using gcc; we always use IBM's compiler. > The ga4 build was a contributed build and I can't vouch for its > correctness. > > You're using gcc4 which is quite a bit different from earlier versions of > gcc. I doubt very much that ga4 was for gcc4. gcc4 explains the > > warning: pointer targets in assignment differ in signedness > > warning: pointer targets in passing argument differ > messages. This behavior is new to gcc4; it is less permissive with > implicit casting between char* and unsigned char* that other compilers > (including earlier versions of gcc). > > On top of that, I'm really suspicious of your report that you can't get > 2004g to compile without SSL. This suggests a systemic problem, such as > improper compile flags; because people regularly compile without SSL (in > spite of recommendations not to do so!). I would not recommend running > 2002e in production in any case; it is old and has known problems. > > This, however, is quite serious: > > osdep.c: In function 'checkpw': > > osdep.c:148: warning: assignment from incompatible pointer type > osdep.c is an assembled file; very likely this is in ckp_pam.c. This > points to some incompatibility in AIX's implementation of PAM. > > So, the first thing that I would try is using IBM's compiler with 2004e or > 2006. If all is well, then call it good. If not, then send me a complete > transcript of the build (from scratch - do a "make clean" first) and I'll > see what I can determine. I may not be able to get to it until Monday. > > One last thing: do you actually intend to run in production without SSL? > Or are you building without SSL just to simplify testing? > > -- Mark -- > > http://staff.washington.edu/mrc > Science does not emerge from voting, party politics, or public debate. > Si vis pacem, para bellum. > _______________________________________________ Imap-uw mailing list [email protected] https://mailman1.u.washington.edu/mailman/listinfo/imap-uw
