Stevan Little <stevan.lit...@iinteractive.com> writes:
> On Jul 18, 2009, at 8:18 AM, Hans Dieter Pearcey wrote:
>> On Sat, Jul 18, 2009 at 08:48:06PM +1000, Daniel Pittman wrote:
>>> Which code is in the "wrong" here — MooseX::Daemonize, or MooseX::Declare?
>>>
>>> By "wrong" I mean, which of the two is making an assumption that isn't
>>> supported by Moose, or are they both doing fine, and just not compatible?
>>
>> They're both doing fine. Use 'with' inside your class definition, as usual
>> for Moose, so that it's composed immediately, rather than in the class
>> declaration, which delays it until the end of the class block.
>>
>> There's no single right thing to do here right now; some roles break if you
>> compose them too early, and some if you do it too late.
>
> This is a MooseX::Declare problem.
>
> MX::Declare is still relatively new and there are still bugs, issues and
> various annoyances that need to be shaken out. Pure Moose will just do the
> right thing almost all of the time (the few cases where it doesn't tend to
> be very pathological).
Thank you for the clarification. I am a little curious, Hans, about which
roles you are thinking of that fail if composed too early?
I suspected that the composition model of MX::Declare was at fault; the main
reason I was playing with it was the automatic variable binding from
signatures is very nice, and because it is always interesting to see the
future.
Regards,
Daniel
--
✣ Daniel Pittman ✉ dan...@rimspace.net ☎ +61 401 155 707
♽ made with 100 percent post-consumer electrons
