The driver only checked if the corresponding value was in the config, so
no_dns: False
had the same effect as
no_dns: True
Change the check to take the value into consideration.
This makes false-y values like False (from YAML) and empty string
(from environment) work as if the value was not specified.
Clean up the .gitignore file:
- Remove no longer used .gitignore entries, like .bzr files
- Do not repeat autotools generated files over and over again
- Whitelist existent Makefiles in the repository
- Better separate the .gitignore entries
The ipa-client-install script and ipa-join use different methods
of resolving the hostname, the former uses gethostbyaddr() call,
while the latter reads the "uinfo.nodename".
This can result ipa-client-install failures in case of broken PTR
records.
https://fedorahosted.org/freeipa/ticket/4027
The framework had a concept of external hostnames,
which the controller uses to contact the test machines,
but they were not loaded from configuration.
Load external names from configuration.
This makes tests pass in setups where internal and external
hostnames are different, and the internal hostnames are not
initially resolvable from the controller.
Apparently when we relicensed to GPLv3 we missed these two spots.
The actual boilerplate was changed in these files but not the
license tag passed to python setup.
When AD administrator credentials passed, they stored in realm_passwd,
not realm_password in the options.
When passing credentials to ipaserver.dcerpc.fetch_domains(), make sure
to normalize them.
Additionally, force Samba auth module to use NTLMSSP in case we have
credentials because at the point when trust is established, KDC is not
yet ready to issue tickets to a service in the other realm due to
MS-PAC information caching effects. The logic is a bit fuzzy because
credentials code makes decisions on what to use based on the smb.conf
parameters and Python bindings to set parameters to smb.conf make it so
that auth module believes these parameters were overidden by the user
through the command line and ignore some of options. We have to do calls
in the right order to force NTLMSSP use instead of Kerberos.
Fixes https://fedorahosted.org/freeipa/ticket/4046
Add the server class name, such as [xmlserver] or [jsonserver_kerb] to
the server logs. This will allow easier debugging of problems specific
to a protocol or server class.
Modify ipalib.rpc to support JSON-RPC in addition to XML-RPC.
This is done by subclassing and extending xmlrpclib, because
our existing code relies on xmlrpclib internals.
The URI to use is given in the new jsonrpc_uri env variable. When
it is not given, it is generated from xmlrpc_uri by replacing
/xml with /json.
The rpc_json_uri env variable existed before, but was unused,
undocumented and not set the install scripts.
This patch removes it in favor of jsonrpc_uri (for consistency
with xmlrpc_uri).
Add the rpc_protocol env variable to control the protocol
IPA uses. rpc_protocol defaults to 'jsonrpc', but may be changed
to 'xmlrpc'.
Make backend.Executioner and tests use the backend specified by
rpc_protocol.
For compatibility with unwrap_xml, decoding JSON now gives tuples
instead of lists.
Design: http://freeipa.org/page/V3/JSON-RPC
Ticket: https://fedorahosted.org/freeipa/ticket/3299
The project's history is kept in Git. We used the spec changelog
for changes to the spec itself, which doesn't make much sense.
Downstreams like Fedora use their own changelog anyway.
A single entry is left for tools that expect a changelog.
For trusted domains base id is calculated using a murmur3 hash of the
domain Security Identifier (SID). During trust-add we create ranges for
forest root domain and other forest domains. Since --base-id explicitly
overrides generated base id for forest root domain, its value should not
be passed to other forest domains' ranges -- their base ids must be
calculated based on their SIDs.
In case base id change for non-root forest domains is required, it can
be done manually through idrange-mod command after the trust is
established.
https://fedorahosted.org/freeipa/ticket/4041
This object will allow splitting large translatable strings into more
pieces, so translators don't have to re-translate the entire text
when a small part changes.
https://fedorahosted.org/freeipa/ticket/3587
During the installation, copy the CA certificate to the systemwide
store (/etc/pki/ca-trust/source/anchors/ipa-ca.crt) and update the
systemwide CA database.
This allows browsers to access IPA WebUI without warning out of the
box.
https://fedorahosted.org/freeipa/ticket/3504
Due to a bug[0], python-ldap doesn't parse schema LDIF files correctly
if they use inconsistent capitalization.
This patch works around the bug in IPA schema files.
[0] https://bugzilla.redhat.com/show_bug.cgi?id=1007820
Note: git's --word-diff option is recommended for viewing these changes
The new schema updater only compares textual representations of schema
elements, as formatted by python-ldap.
This works well, but it is too strict for the current schema files in two ways:
- For attribute names in MAY and MUST, the correct letter case must be used
- AttributeTypes must specify explicit EQUALITY and SYNTAX fields even if
they are the same as its supertype's.
When these restrictions are not followed, the updater will always overwrite
the schema element. This is harmless but it fills up the log unnecessarily.
Modify the schema files to conform to these restrictions.
Part of the work for https://fedorahosted.org/freeipa/ticket/3454
Note: git's --word-diff option is recommended for viewing these changes
