[Dbix-class] Status update, plans

[Matt S Trout]

Dave has committed a change that uses IxHash for the _primaries hashref
so ->retrieve($pri1, $pri2) works reliably, and some tests. More tests for
basic functions using composite keys would be most appreciated given it's
something Class::DBI often falls down on.

AndyG has added some tests for commit/rollback

Draven (Macrus Ramberg) has added extra tests for the live object index.

Thanks very much to committers 2, 3 and 4, and welcome aboard.

Next time I get a chance to hack on it I'm going to build a cursor
(iterator) class and move all search/retrieve stuff over to using it
(although in array context all the objects will still be returned, of course).

Updatable cursors we can think about later; the aim here is to get something
that provides the same style of interface as Class::DBI::Iterator but only
does the $sth->fetchrow when necessary, and doesn't force you to keep the
inflated objects for the lifetime of the iterator. I'm thinking probably
a cache_objects constructor parameter or similar, and at some stage something
that causes ->count to do a COUNT(*) on the db rather than fetch all records.

This will only be a first cut (and I'll probably just test it against the
cdbi iterator tests) so we can refactor (or rewrite from scratch if necessary)
later, but comments and suggestions are still welcome.

-- 
     Matt S Trout           Website: http://www.shadowcatsystems.co.uk
  Technical Director        E-mail:  mst (at) shadowcatsystems.co.uk
Shadowcat Systems Ltd.