krb5-wallet.git/perl/Wallet/Object/Keytab.pm, branch debian/1.0-5

Fix several bugs in Wallet::Object::Keytab enctype handling

2013-11-05T20:34:28+00:00

There was a missing resultset() call in one place and the wrong resultset used in a different place, causing the enctype management code to not work. Change-Id: I796169c5968ec164f90f3cd75541dd346dd50fdf Reviewed-on: https://gerrit.stanford.edu/1070 Reviewed-by: Russ Allbery Tested-by: Russ Allbery

Use correct form of Stanford's copyright statement

2013-02-28T00:52:47+00:00

Change-Id: I06dd9ecca19315179bdd34d4b301548fe7604331 Reviewed-on: https://gerrit.stanford.edu/842 Reviewed-by: Russ Allbery Tested-by: Russ Allbery

Renamed dbh subroutines and variables for clarity

2013-02-01T00:35:01+00:00

In moving from DBI to DBIx::Class, we at first left the various variables the same. This goes through to update them for the proper names. * Wallet::Admin::schema was created to return the schema object (and similarly for Wallet::Server and Wallet::Report). * Wallet::Admin::dbh was modified to return the actual DBI handle again (and similarly for Wallet::Server and Wallet::Report). * Various places that used $admin->{dbh} were moved to $admin->{schema}. * Various places using $dbh for the schema object were changed to $schema. Change-Id: I00914866e9a8250855a7828474aa9ce0f37b914f Reviewed-on: https://gerrit.stanford.edu/733 Reviewed-by: Russ Allbery Tested-by: Russ Allbery

Fixed errors with Keytab object and its tests

2013-01-31T23:58:05+00:00

perl/Wallet/Object/Keytab.pm was using the wrong value for the database handle in some places (trying to load as a subroutine rather than part of the object). Also, the keytab.t tests were attempting to run against the DBIx::Class object rather than a direct dbh handle that they expected. Change-Id: Ifbb8b110d559f3ba867fc5b0dc3933fd2d4fd484 Reviewed-on: https://gerrit.stanford.edu/731 Reviewed-by: Russ Allbery Tested-by: Russ Allbery

Moved the Perl wallet modules and tests to DBIx::Class

2013-01-31T02:33:23+00:00

Moved all the Perl code to use DBIx::Class for the database interface. This includes updating all database calls, how the schema is generated and maintained, and the tests in places where some output has changed. We also remove the schema.t test, as the tests for it are more covered in the admin.t tests now. Change-Id: Ie5083432d09a0d9fe364a61c31378b77aa7b3cb7 Reviewed-on: https://gerrit.stanford.edu/598 Reviewed-by: Russ Allbery Tested-by: Russ Allbery

Fix POD stopwords and formatting

2012-08-30T21:57:10+00:00

Fix a formatting error in Wallet::ACL::LDAP::Attribute and add new stopwords required by the latest aspell.

Use L<> links instead of man page references for modules

2010-03-04T06:44:53+00:00

Do this only in the main text, not in the SEE ALSO section, since the latter is more for conventional man pages. This will produce better results for some POD to HTML converters (although not mine, yet).

Refactor Wallet::Kadmin keytab_rekey to return keytab

2010-02-19T06:06:17+00:00

Change the API for keytab_rekey to match keytab, returning the keytab as data instead of writing it to a file. This simplifies the wallet object implementation and moves the logic for reading the temporary file into Wallet::Kadmin and its child classes. (Eventually, there may be a kadmin backend that doesn't require using a temporary file.) Setting KEYTAB_TMP is now required to instantiate either the ::MIT or ::Heimdal Wallet::Kadmin classes.

Support unchanging keytabs with Heimdal without remctl

2010-02-19T05:31:10+00:00

Heimdal supports retrieving a keytab containing the existing keys over the kadmin protocol. Move the support for using remctl to retrieve an existing keytab into Wallet::Kadmin::MIT and provide two separate methods in the Wallet::Kadmin interface: one which rekeys and one which doesn't. Implement the non-rekeying interface for Heimdal. Expand the test suite for the unchanging keytabs to include tests for the Heimdal method.

Refactor sync handling in Wallet::Object::Keytab

2010-02-19T02:23:23+00:00

Pull the sync code out into separate methods to avoid a really long and awkward attr method. Document the limited object support for the sync attribute.