Title: #958: ldap_child: do not try PKINIT
Fine, as I wrote, this was really insignificant comment thus patch is accepted.
I do not think this is matter of coding *style*.
Configure check is an interesting idea but I do not think it justifies an effort.
Having explicitly mention this in the documentation is preferred, of
course, but even then it is only documentation and the implementation might have issues
that it can be safely assumed that "krb5 functions that free
data all handle NULL" and that it is a bug in libkrb5 if it is not the case?
From my point of view, only having this info stated in krb5 docs explicitly would justify
such a statement (kind of "liability shift").
But then again, I agree this is not something that have to be seriously addressed.
See the full comment at https://github.com/SSSD/sssd/pull/958#issuecomment-565392480