Commits on Mar 12, 2003
  1. Added constraint on output channels shown in double list manager to o…

    garthwebb committed Mar 12, 2003
    …nly those
    in the same site as the destination being created.
  2. Putting the comment back.

    crucially committed Mar 12, 2003
  3. We only want to load the active element__site mappings, this was caus…

    crucially committed Mar 12, 2003
    …ing a bug
    when rereading the table and was not detected from the test script because I
    haven't found a way to flush the entire cache/state.
  4. Delay the saving of the Site collection until after the AssetType has…

    crucially committed Mar 12, 2003
    … been
    saved if the AssetType is new and doesn't have an id. This avoids the problem
    of creating the element__site map without a proper element__id.
  5. Change to test_sql as adviced

    crucially committed Mar 12, 2003
  6. Fixed exception handling to use throw_dp

    crucially committed Mar 12, 2003
    Documented new functions
    Added tests for Site collections
Commits on Mar 11, 2003
  1. Updated handling/use of ROOT_CATEGORY_ID

    garthwebb committed Mar 11, 2003
    This constant no longer exists and has been replaced by several methods:
    See Bric::Biz::Category for details on these methods
  2. Adding a version requirement for Exception::Class since versions prio…

    garthwebb committed Mar 11, 2003
    …r to 1.10 don't seem to support the 'alias' key.
  3. First step in multisite support for element/AssetTypes

    crucially committed Mar 11, 2003
    - New Site collection used by
    - Sites can be loaded based on element_id which is joined in from element__site
    - element__site is created and filled with default values / upgraded
    - AssetType can add and remove Sites from itself
  4. Fix testcase to match reality/API.

    crucially committed Mar 11, 2003
    We are testing href not list, and href is documented to always return a
    hash reference, even if it can be empty. This is also what
    Bric::Biz::Site returns.
  5. Added href() method. This is to that sites can be managed via a subcl…

    theory committed Mar 11, 2003
    …ass of
  6. Various site-related changes. Make sure to display the Site in the ca…

    theory committed Mar 11, 2003
    output channel, and destination managers. Added site_id accessors to
    Bric::Dist::ServerType synopsis. Added tests for site_id to server type test
Commits on Mar 10, 2003
  1. Added site_id to Bric::Dist::ServerType

    garthwebb committed Mar 10, 2003
    * Added site_id to package
    * Updated POD in package
    * Updated *.sql, *.con and *.txt files
    * Added upgrade script
    * Updated test script
    * Updated code that creates new ServerTypes
    * Updated UI to collect a site for all new ServerTypes
  2. Incorrectly checking site_id rather than just ID to determine if we h…

    garthwebb committed Mar 10, 2003
    …ave a new or existing category
Commits on Mar 9, 2003
  1. Changed default templates. The default "" template now uses the

    theory committed Mar 9, 2003
    key_name attribute of elements rather than the has_name() method. Although
    has_name() is still supported, it really should be deprecated. At least it now
    compares whatever is passed to the key_name, as well. The name is now merely
    used for display purposes.
  2. s/get_name/get_key_name/;. There were quite a few places in the conta…

    theory committed Mar 9, 2003
    widget where the old get_name() method was being called instead of the new
    get_key_name() method, and this naturally lead to things not working. There
    was also a spot in Bric::Biz::AssetType that was doing the same thing, and
    that's fixed, as well.
  3. s/depreciated/deprecated/g;

    theory committed Mar 9, 2003
  4. Reverted changes. Since the site_id parameter is already implicitly (…

    theory committed Mar 9, 2003
    …and soon
    to be explicitly) suppported by Bric::Biz::OutputChannel,
    Bric::Biz::OutputChannel::Element doesn't have to. All tests pass with this
  5. Changed "site__id" to "site_id". This is because the double underscor…

    theory committed Mar 9, 2003
    …e should
    only be used in the database. The API should always use a single
    underscore. Those exising places in the API that use a double underscore are
    wrong (and may be fixed one day).
    Also tweaked the docs here and there, and added some tests for "site_id" and
    "protocol" to Bric::Biz::OutputChannel::Test.
  6. Added missing permissions. I added tests for the permissions created …

    theory committed Mar 9, 2003
    …by the
    Bric::Biz::Site for its secret groups, and added tests for their presence in
    the default site, too. That caught me! :-)
    Also fixed old ToDo test failure in Bric::Util::Grp. It had a conflict between
    its get_secret() class method and its (missing) get_secret() instance
    method. This became an issue with the new secret sites in Bric::Biz::Site, so
    I finally fixed it. Now there is a new C<is_secret()> instance method that
    returns the secret status of a group object. C<get_secret()> still returns the
    default value of the secret attribute for new groups.
    Oh, and I also added support for the C<secret> and C<permanent> attributes
    to Bric::Util::Grp->new. These, too, are used by Bric::Biz::Site.
    All tests pass now, though not against an upgraded database. That should be
    fixed shortly.