| Age | Commit message (Collapse) | Author |
|---|
|
If multiple ACL changes were made in the same second, the order of
results could be unstable. Sort by unique key after date to avoid
this. Based on work by macrotex.
|
|
We were sending stderr to stdout for remctl output, but sometimes
tests would fail because the remctl output was intermixed with the
test output and confused runtests. runtests only looks at stdout
for test output and sends stderr to /dev/null, so send remctl
output there.
This isn't a great solution since it interferes with debugging
tests. The correct solution is to implement the same logging and
polling that's used by C TAP Harness, although that's more difficult
to do in Perl. But it should make the immediate problem go away.
Continue to send the message about which remctl command line is in
use to stdout, but prefix it with # so that it is a TAP comment.
|
|
Update to rra-c-util 8.2:
* Implement explicit_bzero with memset if it is not available.
* Reformat all C source using clang-format 10.
* Work around Test::Strict not skipping .git directories.
* Fix warnings with perltidy 20190601 and Perl::Critic 1.134.
* Fix warnings with Clang 10, GCC 10, and the Clang static analyzer.
Update to C TAP Harness 4.7:
* Fix warnings with GCC 10.
* Reformat all C source using clang-format 10.
* Fixed malloc error checking in bstrndup.
|
|
If AUTHOR_TESTING wasn't set, the test would plan twice. Fix this
by restructuring it to delay printing a plan and convert the
Kerberos ticket test to a skip_all plan.
|
|
Add SPDX-License-Identifier headers to all substantial source files.
Collapse copyright years. Add some Emacs configuration for files
where the copyright notice is at the end. Add a test that every
file has SPDX-License-Identifier.
|
|
Mostly changing http eyrie.org URLs to https, but also remove my
old email address in one place and switch some tests away from my
old RRA_MAINTAINER_TESTS environment variable to use the Lancaster
Consensus variables properly. This uncovered a bug in skipping one
test unless Stanford Kerberos credentials existed.
|
|
Update to rra-c-util 7.2:
* Improve configure output for krb5-config testing.
* Define UINT32_MAX for systems that don't have it.
* Add SPDX-License-Identifier headers to all substantial source files.
* Fix new warnings from GCC 7 and Clang warnings.
* Require Test::Strict 0.25 or later to run those tests.
* Fix off-by-one error in return-value checks for snprintf.
* Use Autoconf to probe for supported warning flags.
* Fix running module-version-t -u with current versions of Perl.
* Use C_TAP_SOURCE and C_TAP_BUILD instead of SOURCE and BUILD.
Update to C TAP Harness 4.3:
* Add support for valgrind and libtool in test lists.
* Report test failures as left and right, not wanted and expected.
* Fix string comparisons with NULL pointers and the string "(null)".
* Add SPDX-License-Identifier headers to all substantial source files.
* Avoid zero-length realloc allocations in breallocarray.
* Fix new warnings from GCC 7 and Clang warnings.
* Use C_TAP_SOURCE and C_TAP_BUILD instead of SOURCE and BUILD.
|
|
|
|
|
|
|
|
Move options up with other keytab backend options except for the
bucket for keytabs, which does belong in the section on retrieving
existing keytabs.
|
|
|
|
|
|
It turns out that the length limitations apply to all keytabs, not
just service keytabs. This change creates unique ids for hostnames
that exceed the AD length limit.
|
|
The account name for a service keytab cannot exceed 20 characters.
The routine that was generating a unique id incorrectly attempted to
perform an LDAP query. This change fixes that problem.
|
|
The account name for a service keytab cannot exceed 20 characters.
The routine that was generating a unique id incorrectly attempted to
perform an LDAP query. This change fixes that problem.
|
|
* Make sure userPrincipalName is created for all keytabs and use it to
search for entries in AD.
* Allow the creation of any service principal. This requires making
sure that the cn used to create AD entries for service accounts not
be any longer than 20 characters.
|
|
Correct a variable reference that was causing AD keytab creation to
fail. Update the debugging for shell command execution that makes
debugging more rebust and highlights problems.
|
|
* This ad-keytab is useful in the initial setup of AD as a keytab
store for wallet.
* Change configuration variables to correctly reflect that some values
are relative distinguished names.
* Add a configuration variable for the base distinguished name for
ActiveDirectory.
|
|
with multiple enctypes specified, only the last one will actually take effect. If you wish to provide support for more then one, you need to add the values (0x04 + 0x08 + 0x10 = 0x1C).
replacing the 3 lines with one line to enable all three. Note that the keytabs generated will have 3 line for each principal (one for each enctypes).
See msktutil man page for further details on enctypes.
|
|
|
|
|
|
|
|
|
|
Flesh out recommends for more accurate dependencies for the Perl
modules. Pull the version from one of the Perl modules, now that
we have another test that ensures that those versions are all
consistent.
|
|
The versions of all of the wallet Perl modules now match the overall
package version except for Wallet::Schema, which is used to version
the database schema.
Import the test from rra-c-util 5.10 and exclude Wallet::Schema from
the tests.
Go through all Perl modules and standardize the syntax for setting the
version and indicating the required version of Perl. Fix a few other
syntax issues while I'm in there.
|
|
|
|
Also remove some configuration checks that aren't required, and
unify handling of some configuration options.
|
|
This requires changing the ACL verifier plumbing to pass object
type and name all the way through when verifying ACLs. Hopefully
I caught everything.
|
|
Conflicts:
NEWS
|
|
|
|
Failed kadmin commands were deleting the wallet database in the
test suite due to an END block in the test programs. Use _exit
to avoid this.
|
|
New versions of MIT now use the actual enctype in klist -ke output.
Also add 128-bit AES.
Also add some additional debugging that was useful when chasing
another problem.
|
|
Ubuntu precise and trusty don't have Net::Duo packages. Delay
loading to the constructor so that the modules will still pass
strictness tests. This also fixes Travis-CI testing.
Change-Id: I23f1fe6dbdddaac2040f459410a74be4a13b6755
|
|
Change-Id: I3a8b13a8b255522cff92910f8d99ec94dc020e6f
|
|
A new ACL type, external (Wallet::ACL::External), is now supported.
This ACL runs an external command to check if access is allowed, and
passes the principal and the ACL identifier to that command. To
enable this ACL type for an existing wallet database, use wallet-admin
to register the new verifier.
Change-Id: I21b72b4373eefc92985aca1505e2d1a1ec699602
|
|
|
|
The msktutil script does not always signal error conditions. This
change implements a check that examines the output from msktutil
and reports and error when the keytab creation fails to create
the keytab but does create a computer entry in the directory. If
an error is detected the directory entry is deleted leaving the
directory in a clean state.
Also, support has been added for output of debugging information
to syslog using the AD_DEBUG configuration variable.
Finally perltidy suggested changes were made to AD.pm.
|
|
Conflicts:
NEWS
|
|
|
|
When adding a new ACL, if creation of the verifier failed, we
reported a pretty minimal error message claiming that the
identifier was the problem. It can't possibly be the problem
when the constructor fails. Report the actual failure more
directly.
|
|
We need a fake NetDB server to test this stuff properly, but until
then, just avoid running the tests.
|
|
This version implements Active Directory as the store for keytabs.
The interface to Active Directory uses a combination of direct LDAP
queries and the msktutil utility. This version does not support the
wallet unchanging flag. Unchanging requires that a keytab be
retrieved without changing the password/kvno which is not supported by
msktutil.
|
|
Added a version of the LDAP attribute ACL. Like the root version for
NetDB, this requires that the principal end in /root, and then strips
off /root before doing matching against the given LDAP attribute.
Change-Id: I23119ef9c9ce3e0556f5d71a509815f2efc1bbe6
|
|
Change-Id: I842a7335a4b50c9c20b921ae2efc63aab571635e
|
|
Since we now check to see if something is a valid netdb node entry for
the ACL verifiers, we need to have a valid netdb setup to run.
Change-Id: Ic2651f8b8b306dfa1f426d91f329b5100a9a1d64
|
|
We needed a way to report on where all a specific ACL might be nested,
since we can't destroy an ACL until it's no longer being nested. For
the immediate this is part of wallet-report.
Change-Id: I41c11b73325d1eb3a28289eac3505bf965877be1
|
|
When destroying an ACL nested in other ACLs, we now fail with an
explanation rather than going through to remove all the places it's
nested. That's more in line with how we handle trying to destroy ACLs
that own things.
Change-Id: I8bc0530e37c54369ec52d9b369f8fabe98def77a
|
|
Removed some default text and explained why we grab the database handle
for future use.
Change-Id: I50b3ae06c1761453de3140d501830c245d550c04
|
|
There was an older mistake in sorting ACLs and entries, using && instead
of || when sorting.
Problem and fix pointed out to Chris Law.
Change-Id: Iab46b4bcbd842978f88a7d9f63958ebea4806413
|
