ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Peter Donald <>
Subject Re: TryCatchFinallyTask ?
Date Sun, 10 Feb 2002 20:41:49 GMT
On Mon, 11 Feb 2002 07:31, Erik Hatcher wrote:
> You don't have to write non-standard Java for your tasks.  The aspects
> themselves would contain the aspect-specific syntax.  And we could build
> with or without aspects, which has some interesting side effects of being
> able to turn on/off failonerror capability, logging, etc.  Why couldn't we
> rely on such a tool?  We rely on javac. Soon we'll rely on XDoclet.  Why
> not rely on something like AspectJ if it solves the issues we're
> encountering cleaner and more robustly than we could do without it?

It may solve it in a cleaner way but the cost of adopting it is prohitively 
high given the benefit we will receive from it. Much less complex to use a 
simple interface based on more familiar designs.

> > More importantly it can also handle special case needs like gump and user
> > preferences.
> I'm not sure what  Gump needs differently.  But user preferences shouldn't
> be considered a "special case" - it should be built in to the architecture,
> no?

Well special case or not - user preferences modify the task model prior to an 
object being instantiated. Gump needs to force tasks to use its ClassLoader 
etc. Anything that modifies the environment in which a task runs across a 
selection of tasks should be able to be modeled using an AspectHandler.



Be nice to your friends. If it weren't 
for them, you'd be a complete stranger.

To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message