Sorry for not responding to the thread sooner, but I've been busy the past few days. The underlying issue causing this regression has been identified and worked on a bit, I've posted a failing testcase to #catalyst-dev, etc. Its just a question of exactly what or how to do about it that's a bit complicated AFAIK - I haven't followed up on this in the past day or two to see what the final fallout was.
The true nature of the issue is that in previous Catalyst releases, if your Model directory contained Foo.pm and Foo/Bar.pm, Catalyst itself would always load both files, without question. In the release's way of handling the component loading, Foo.pm is always loaded first, and if it defines a "Foo::Bar" namespace, then Foo/Bar.pm won't get loaded by Catalyst (it assumes Foo.pm must have loaded it, basically). Some usages of some Models (in theory even DBIC::Schema, but the pattern is probably rare to nonexistant) rely on this older behavior to load layered "additional" definitions on top of Loader-generated in-memory definitions. -- Brandon _______________________________________________ List: [email protected] Listinfo: http://lists.rawmode.org/mailman/listinfo/catalyst Searchable archive: http://www.mail-archive.com/[email protected]/ Dev site: http://dev.catalyst.perl.org/
