This reverts commit 8a5a8d492c.
Special-case /run/user/0, attempting to create it when resolving a
directory cache below it fails due to ENOENT and we find that it doesn't
already exist, either, before attempting to create the directory cache
(maybe helping, maybe just making things more confusing for #961235).
- update to 1.11.3
- drop patch for RT#7605, fixed in this release
- drop patch for CVE-2002-2443, fixed in this release
- drop patch for RT#7369, fixed in this release
- pull upstream fix for breaking t_skew.py by adding the patch for #961221
It's not a complete fix, and it may only muddy things further on systems
that are having the kind of trouble it's trying to avoid, so hold off.
For now, at least.
- pull in proposed fix for attempts to get initial creds, which end up
following referrals, incorrectly trying to always use master KDCs if
they talked to a master at any point (should fix RT#7650)
Add a patch to create /run/user/0 if we're trying to resolve a
DIR: ccache somewhere below it and neither the target location
nor /run/user/0 exist yet.
The better workaround is to set the location's owner to "linger"
via logind, since even after we do what we're doing here, if
the user logs in and logs back out, our location is still removed.
- pull in patches from master to not test GSSRPC-over-UDP and to not
depend on the portmapper, which are areas where our build systems
often give us trouble, too
In addition to basing the contents of an encrypted-timestamp preauth
data item on the server's idea of the current time, go ahead and do the
same for the times in the request.
- pull in upstream fix to start treating a KRB5CCNAME value that begins
with DIR:: the same as it would a DIR: value with just one ccache file
in it (RT#7172, #965574)
- pull the changing of the compiled-in default ccache location to
DIR:/run/user/%%{uid}/krb5cc back into F19, in line with SSSD and
the most recent pam_krb5 build
- when testing the RPC library, treat denials from the local portmapper the
same as a portmapper-not-running situation, to allow other library tests
to be run while building the package
- create and own /var/kerberos/krb5/user instead of /var/kerberos/kdc/user,
since that's what the libraries actually look for
- add buildrequires on nss-myhostname, in an attempt to get more of the tests
to run properly during builds
- go back to using reconf to run autoconf and autoheader (part of #925640)
- add temporary patch to use newer config.guess/config.sub (more of #925640)
drop the kerberos-iv portreserve file (long overdue), and drop the rest
on systemd systems, since we don't currently poke portreserve when we're
starting a service
- update to the 1.11 final release
- drop the rawbuild tag from a couple of patches which we don't actually
need to apply to get things to compile the way the package expects
- handle releases where texlive packaging wasn't yet as complicated as it
is in Fedora 18
- fix an uninitialized-variable error building one of the test programs
- move the rather large pile of html and pdf docs to -workstation, so
that just having something that links to the libraries won't drag
them onto a system
- actually create %%{_var}/kerberos/kdc/user, so that it can be packaged
- correct the list of packaged man pages
- drop backported patches to make keytab-based authentication attempts
work better when the client tells the KDC that it supports a particular
cipher, but doesn't have a key for it in the keytab
- drop backported fix for teaching PKINIT clients which trust the KDC's
certificate directly to verify signed-data messages that are signed with
the KDC's certificate, when the blobs don't include a copy of the KDC's
certificate
- add a backport of more patches to set the client's list of supported enctypes
when using a keytab to be the list of types of keys in the keytab, plus the
list of other types the client supports but for which it doesn't have keys,
in that order, so that KDCs have a better chance of being able to issue
tickets with session keys of types that the client can use (#837855)
- pull up patch for RT#7063, in which not noticing a prompt for a long
time throws the client library's idea of the time difference between it
and the KDC really far out of whack (#773496)
- undo rename from krb5-pkinit-openssl to krb5-pkinit on EL6
- version the Obsoletes: on the krb5-pkinit-openssl to krb5-pkinit rename
- reintroduce the init scripts for non-systemd releases
- forward-port %%{_?rawbuild} annotations from EL6 packaging
- selinux: hang on to the list of selinux contexts, freeing and reloading
it only when the file we read it from is modified, freeing it when the
shared library is being unloaded (#845125)
- go back to not messing with library file paths on Fedora 17: it breaks
file path dependencies in other packages, and since Fedora 17 is already
released, breaking that is our fault
- add upstream patch to fix freeing an uninitialized pointer and dereferencing
another uninitialized pointer in the KDC (MITKRB5-SA-2012-001, CVE-2012-1014
and CVE-2012-1015, #838012)
- fix a thinko in whether or not we mess around with devel .so symlinks on
systems without a separate /usr (sbose)
- backport a fix to allow a PKINIT client to handle SignedData from a KDC
that's signed with a certificate that isn't in the SignedData, but which
is available as an anchor or intermediate on the client (RT#7183)
- back out this labeling change (dwalsh):
- when building the new label for a file we're about to create, also mix
in the current range, in addition to the current user
- add explicit buildrequires: on 'hostname', for the tests, on systems where
it's in its own package, and require net-tools, which used to provide the
command, everywhere
- add a backport of Stef's patch to set the client's list of supported
enctypes to match the types of keys that we have when we are using a
keytab to try to get initial credentials, so that a KDC won't send us
an AS reply that we can't encrypt (RT#2131, #748528)
- when building the new label for a file we're about to create, also mix
in the current range, in addition to the current user
- also package the PDF format admin, user, and install guides
- drop some PDFs that no longer get built right
