On Mon, 22 Oct 2001, Conor MacNeill <[EMAIL PROTECTED]>
wrote:
> Another fascinating Ant2 thread.
I can't pretend that I didn't know what I'd start 8-)
> There are some constructs in the buildfile which should be
> recognized and handled by the parser and not delegated.
Point regarding classloaders taken, but that doesn't necessarily mean
that the parser has to handle that stuff itself, "privileged" tasks
could be loaded from the same classloader as the core for example.
> We need something similar to the servlet engines which carefully
> structure their classloader hierarchies and, require clear
> separation of classes (responsibilities).
Servlet containers typically don't provide as much control over
themselves as we've been discussing it lately.
Declaring data types and tasks and context and whatever would simply
require a registry service of some sort, but I'm still not sure
whether I see all consequences of our blurry "aspect" concept.
> I'm not sure why Peter believes we won't need <taskdef> at top level
> anymore.
<taskdef> at project scope is used to avoid proxies for the tasks in
Ant 1.x - as soon as we switch to an all proxy system, there is no
difference between <taskdef> in or outside of targets.
> With regard to undefined properties causing a failure, rather than
> becoming the value ${ljljlkj}, I agree with this.
That makes Pete, you and me, which means it is accepted for Ant2
unless anybody else tries to veto it.
> Not sure if my answer to 2 is a different option.
If it does, it gets a +0 from me (like the old option (2) did), maybe
even a +0.1.
That makes two -1s and two +0s for you - any preference?
Stefan
