| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2007-12-05 | Various coding style fixes and cleanup based on a much-appreciated | Russ Allbery | |
| code audit by Simon Cozens. I didn't take all of his advise, and he shouldn't be blamed for any remaining issues. | |||
| 2007-12-05 | Coding style improvements. | Russ Allbery | |
| 2007-12-05 | Bump version number for previous changes. | Russ Allbery | |
| 2007-12-05 | Maintain a global cache of ACL verifiers in Wallet::ACL and reuse them | Russ Allbery | |
| over the life of the process if we see another ACL line from the same scheme, rather than only reusing ACL verifiers within a single ACL. | |||
| 2007-12-05 | Determine the class for object and ACL schema implementations from the | Russ Allbery | |
| database rather than a hard-coded list and provide Wallet::Schema methods for adding new class mappings. Add a missing class mapping for the netdb ACL schema verifier. | |||
| 2007-12-04 | Add a subclass of the NetDB ACL verifier that requires the principal | Russ Allbery | |
| have an instance of "root" and strips that instance before checking NetDB roles. | |||
| 2007-12-01 | Fix the test number in the SKIP block for enctype restrictions. | Russ Allbery | |
| 2007-12-01 | Add a MySQL method to check the tables remaining after a drop. | Russ Allbery | |
| 2007-12-01 | Document that all tests are now run against the configured database if | Russ Allbery | |
| configuration is present. | |||
| 2007-12-01 | Run the full schema test with the configured database. It should work | Russ Allbery | |
| fine with MySQL. | |||
| 2007-12-01 | Following the DBI documentation, turn on AutoCommit after RaiseError so | Russ Allbery | |
| that we have some hope of getting error messages if it fails. | |||
| 2007-12-01 | If we're using MySQL, force all the tables to InnoDB so that we get | Russ Allbery | |
| transactions. | |||
| 2007-12-01 | Clean up the database after running this test. | Russ Allbery | |
| 2007-12-01 | Don't disconnect on destruction of the server object if the handle is | Russ Allbery | |
| marked with InactiveDestroy. | |||
| 2007-12-01 | Set InactiveDestroy to the right thing. Doh. | Russ Allbery | |
| 2007-12-01 | Make sure we always have a valid object and access to our database | Russ Allbery | |
| handle when calling kadmin so that we can do our database handle mangling. | |||
| 2007-12-01 | Don't use die when forking kasetkey fails and similarly don't close the | Russ Allbery | |
| database connection on a failed fork. | |||
| 2007-12-01 | Don't close our open database handles after a failed fork. | Russ Allbery | |
| 2007-12-01 | Use reinitialize instead of initialize. | Russ Allbery | |
| 2007-12-01 | Expiration dates are now expressed in YYYY-MM-DD HH:MM:SS instead of | Russ Allbery | |
| seconds since epoch and returned the same way. Timestamps are now stored in the database as correct date and time types rather than seconds since epoch to work properly with MySQL. | |||
| 2007-12-01 | Don't do an in-place edit of the SQL inside the object, since we then | Russ Allbery | |
| can't reuse the object with another database handle. | |||
| 2007-12-01 | Actually use the username and password from the database configuration. | Russ Allbery | |
| 2007-12-01 | Hide kinit errors. No one cares. | Russ Allbery | |
| 2007-12-01 | Move wallet.conf to /etc/wallet since we're going to want to put other | Russ Allbery | |
| things in that directory as well. | |||
| 2007-12-01 | The wallet backend test suite now supports using a database other than | Russ Allbery | |
| SQLite for testing. Also start a new Util.pm module for the test suite and move the contents sub into that module. More to follow. | |||
| 2007-11-30 | Initial work on supporting testing with MySQL. | Russ Allbery | |
| Add a drop() method to Wallet::Schema to destroy the wallet database. Add a test suite for it. Add a reinitialize() method to Wallet;:Server that drops the database before creating it. Modify the wallet object test cases to call reinitialize() to create the initial database and drop() to clean up the database after the test is complete. Fix a bug preventing Wallet::Schema from being initialized multiple times. We now stash the schema in a class static variable and reuse it for subsequent initializations, since re-reading DATA doesn't work. | |||
| 2007-11-20 | Attempt to create the object with a default owner on get and store | Russ Allbery | |
| when the object doesn't exist. | |||
| 2007-11-20 | Add support for running a user-defined function whenever an object is | Russ Allbery | |
| created by a non-ADMIN user and using the default owner ACL returned by that function provided that the calling user is authorized by that ACL. This permits dynamic creation of new objects based on a default owner ACL programmatically determined from the name of the object. | |||
| 2007-11-15 | Fix the configuration files for remctld, which had gotten inverted. | Russ Allbery | |
| 2007-11-15 | Use the right remctld configuration for the ACL verifier. Don't produce | Russ Allbery | |
| Perl warnings if the NetDB roles are empty. | |||
| 2007-11-15 | Add an ACL verifier that checks access against NetDB roles using the | Russ Allbery | |
| NetDB remctl interface. | |||
| 2007-10-11 | Add an acl_history method to the server layer and test it. | Russ Allbery | |
| 2007-10-10 | Add a history method to the Wallet::ACL API. | Russ Allbery | |
| 2007-10-10 | Use variables for the user and host in the history traces. | Russ Allbery | |
| 2007-10-10 | Adjust for the new behavior of showing the former value of an attribute | Russ Allbery | |
| when it is set. Use variables for the user and host. | |||
| 2007-10-10 | Add tests for the history method and its ACLs. | Russ Allbery | |
| 2007-10-10 | Add the history method (for objects). | Russ Allbery | |
| 2007-10-10 | Add tests for the history trace, which tests history records for | Russ Allbery | |
| attributes. | |||
| 2007-10-10 | Make enctype changes in sorted order so that the history trace is reliable | Russ Allbery | |
| and can be tested. | |||
| 2007-10-10 | Redo history tracking for attributes. Currently, all of our attributes | Russ Allbery | |
| take multiple values, so log them like flags. Hopefully this will continue to work; if not, we'll have to revisit it later. Improve logging for unsetting of attributes. | |||
| 2007-10-10 | Test rollback when attempting to set an invalid enctype. | Russ Allbery | |
| 2007-10-10 | Start checking the history. | Russ Allbery | |
| 2007-10-10 | Pass a timestamp in for trace information as well and check the timestamp | Russ Allbery | |
| on show rather than zeroing it. | |||
| 2007-10-10 | Add an object history method and tests for it. | Russ Allbery | |
| 2007-10-09 | Fix foreign key reference in the history table for the object name. | Russ Allbery | |
| 2007-10-09 | Bump version number for the keytab changes. | Russ Allbery | |
| 2007-10-09 | Add NEWS entry for enctypes. Clarify in the documentation the behavior | Russ Allbery | |
| of the enctypes attribute when the unchanging flag is set. | |||
| 2007-10-09 | Add user documentation for the enctypes attribute on keytabs. Mention | Russ Allbery | |
| that the enctype must be in the database in the keytab object API documentation. The README and NEWS files are now done. | |||
| 2007-10-09 | Flesh out the rest of the tests for enctype handling. | Russ Allbery | |
| 2007-10-09 | We have to check the validity of enctypes ourselves since SQLite doesn't | Russ Allbery | |
| enforce referential integrity. | |||
 |
index : krb5-wallet.git | |
| Unnamed repository; edit this file 'description' to name the repository. | www-data |
| aboutsummaryrefslogtreecommitdiff |