Moving to Samba. I rebuilt it in Rawhide.

Build is running: https://koji.fedoraproject.org/koji/taskinfo?taskID=59630156

I was expecting a samba rebuild, but was also thinking there should have been a conflict when I tried to do the original update. Shouldn't dropping a symbol trigger a soname version change so that people would know rebuilds are necessary?

The commit that removed these says:

commit a3a5c9b31cd9b421941eff9cd11dbc3876f4ff90
Author: Robbie Harwood <rharwood>
Date:   Thu Jan 23 13:09:00 2020 -0500

    Remove private mutators for context enctypes
    
    krb5_set_default_in_tkt_ktypes() and krb5_set_default_tgs_ktypes() are
    not part of the API and are only used in test programs, so remove
    them.  Also remove the now-unused in_tkt_etypes field from
    krb5_context.  Update test suite consumers.
    
    Fix a minor bug wherein the etinfo executable would not correctly
    print its usage text.
    
    [ghudson: adapted some tests rather than remove them]


In Samba we have this:

/* With MIT kerberos, we should use krb5_set_default_tgs_enctypes in preference
 * to krb5_set_default_tgs_ktypes. See
 *         http://lists.samba.org/archive/samba-technical/2006-July/048271.html
 *
 * If the MIT libraries are not exporting internal symbols, we will end up in
 * this branch, which is correct. Otherwise we will continue to use the
 * internal symbol
 */
 krb5_error_code krb5_set_default_tgs_ktypes(krb5_context ctx, const krb5_enctype *enc)
{
    return krb5_set_default_tgs_enctypes(ctx, enc);
}

So a rebuild is required to get rid of a reference to an internal symbol that is removed.

Thanks for the explanation. I'll let you know if the samba rebuild works for me. If it finishes in the next hour or so, I should be able to try it out today. Otherwise I'll try it tomorrow.

Thanks. We are waiting for both ARM targets to complete, all others are done.
aarch64 is almost done with compilation and linking, armv7hl is roughly half way through.

Thanks for catching that Alexander!

To expand on the commit message:

These functions were present in libkrb5.so, but we didn't ship them in any header.  krb5 considers such functions internal, and we don't support their use.  The reason we have such functions is that libkrb5 gets reused in other parts of krb5 (like GSSAPI).

Because this practice is unsupported, it means that any program that uses them does so at its own peril.  We provide no stability guarantees for them.

FEDORA-2021-974a7310df has been pushed to the Fedora 34 stable repository.
If problem still persists, please make note of it in this bug report.

I tested the rebuilt samba package and it fixed the problem for me. Thanks.

FEDORA-2021-247e62f0df has been pushed to the Fedora ELN stable repository.
If problem still persists, please make note of it in this bug report.

*** Bug 1916609 has been marked as a duplicate of this bug. ***