On Fri, 2007-01-05 at 14:24 -0500, Ken Hornstein wrote: 
> Okay, this is a bit late of an answer, but I've been busy.
> 
> >I was wondering, are there any patches available for getting qpopper 4.0.8 or
> >4.0.9 to compile with Kerberos 1.4.3 (or any Kerberos >= 1.3)?  Has anyone
> >submitted these patches to qualcomm?  I saw one posting of a possible patch
> >here:
> 
> Are you sure you _need_ those patches?
> 
> That is to support the older "KPOP/sendauth" style authentication.
> Newer versions of qpopper support GSSAPI via SASL (using the Cyrus-SASL
> library), which really means Kerberos when you get down to it ... but
> the protocol is a lot more reasonable, and more mail readers support
> SASL/GSSAPI.  I think we retired our KPOP support over a year ago as
> everyone had a mail client that could do GSSAPI.  The only big user of
> KPOP was Eudora, but that can now do GSSAPI as well.

Well, I was able to create my own patches that seem to work, but I
haven't been able to fully test them yet.  I was just wondering if
anyone had patches before I went and re-invented the wheel.  (I am
attaching them with this email.)  Of course, what you're saying brings
up some *really* interesting questions.  Are you saying that qpopper
does *not* need the --with-kerberos flag when running configure in order
to use GSSAPI protocol?  If not, I take it that the purpose of the
--with-kerberos flag is to support the KPOP type authentication.
(Please pardon me if that is a re-statement of what you said above.)  If
one does not need to use the --with-kerberos flag when running the
configure script, where can I find documentation on how to use
Kerberos/GSSAPI authentication?

The attached patch fixes the configure script to only add in the
krb5util library if it exists, and can find the function
krb5_compat_recvauth() in it.  Otherwise, it leaves it out.  Also, if it
can't find the function, then the patch has qpopper using only krb5
functions.  This will allow machines using Kerberos >= 1.2.8 to compile
when the --with-kerberos flag is given.  It should not have any impact
on machines with older version of Kerberos.  (I still have not been able
to fully test all of these statements.  So more testing would definitely
be needed.)  Note, I have provided patches for configure and
configure.in in case we would want to regenerate the configure script.

Also, is this list the best place to post patches?

Thanks.

-- 
John Guthrie
[EMAIL PROTECTED]


Reply via email to