On Fri, Feb 15, 2013 at 7:35 AM, Jesse Luehrs <d...@tozt.net> wrote: > > It could easily break things because first of all, it's a global change, > which could break other code that isn't expecting it, and second of all, > there's no way to guarantee that your code will continue to do the right > thing, since a new release could easily add additional functionality > into BUILD which your overridden version would no longer do (and this > sort of thing would be quite difficult to track down). Can you perhaps > override compose_namespace in your schema classes instead? Otherwise, > again, I highly recommend you ask the Catalyst or DBIC communities for > help in doing what you want, since they are much more likely to have > useful solutions. >
Ok, so it doesn't sound like it's a specific problem related to Moose, which was what most were curious about. But just a bad idea in general. I certainly agree with the risks you enumerate above. Yes, over the last few days I have asked both Catalyst and DBIC communities. #dbix-class response was essentially "No idea why the code is like that. Try the change it and report back." Seems ok so far. I couldn't get a clear understanding of compose_namespace so I was a little hesitant to override in the schema class -- not sure what else might call it. If the code was in an init() instead of BUILD then overriding init() would be the answer. Thanks very much for the feedback, -- Bill Moseley mose...@hank.org
