On Mon, 22 Oct 2001 17:42, Stefan Bodewig wrote:
> On Sun, 21 Oct 2001, Peter Donald <[EMAIL PROTECTED]> wrote:
> > The simple fact of the matter is that you have yet to show a
> > reasonable use case for having a Declarable interface.
>
> Before we start getting unreasonable in tone again ...

who said it hadn't already ? ;)

> Declaring TaskContexts (whatever that will be), declaring and
> attaching build listeners from within a build file, configuring
> aspects (do we fail on error and all that, this probably is the
> taskcontext stuff).

those examples work. For these things to be implemented as tasks we are 
definetly going to need the "privlidged" tasks I have talked about in past 
where the container provides specialized services that allows them to do all 
manipulation stuff.

I suspect we are going to have to define a set of runtime tasks that are much 
more coupled to container that access all the runtime data. We should somehow 
mark these aas special and any that are not developed by ant-dev group should 
be considered unsupported and possibly break in future. Not sure really. I 
started to do this in myrmidon but there was a number of issues that arose 
aout of how granular the services are. Fine grain services allow tasks to do 
more interesting stuff but also could be a maintanence nightmare and overly 
constraining as we evolve. Large interfaces can be inflexible.

-- 
Cheers,

Pete

-----------------------------------------------------
First, we shape our tools, thereafter, they shape us.
-----------------------------------------------------

Reply via email to