ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Erik Meade" <eme...@geekfarm.org>
Subject RE: Refactoring fork functionality into common (super) class
Date Sun, 22 Oct 2000 02:58:03 GMT
> -----Original Message-----
> From: Peter Donald [mailto:donaldp@locus.apache.org]
> Sent: Friday, October 20, 2000 7:57 PM
> To: ant-dev@jakarta.apache.org
> Subject: Re: Refactoring fork functionality into common (super) class
>
>
> At 06:54  20/10/00 -0700, you wrote:
> >I've been working on an ANTLR taskdef ( I should be in the position to
> >submit it soon ).  I had to implement forking in order to get some tree
> >stuff to work correctly, I took all of the code from the Java task.  So
> >I took a quick look around and I'm thinking it may be possible
> to refactor
> >the fork behavior into a superclass ( which would extend Task, I like the
> >name ForkingTask, but YMMV ) and have a bunch of the tasks which use
> >forking extend that.
> >
> >I may have a bit of time I could spend on this, but I would rather not
> >put in the time, before I heard what people thought of it.  So...
> >thoughts?
>
> I would be +1 on that ;) That is one of the things I want to address real
> soon now - and your method works for current ant so I like ;)
>
> Cheers,
>
> Pete

I took a longer look at how other tasks are handling forking.  JUnit
and ANTLR ( because I looked at JUnit to figure out how to do ANTLR )
handle the forking themselves, the EJB tasks handle it by using
the Java task ( they fork all the time ).

This makes me think that the superclass I proposed really isn't needed.
I can just change ANTLR to use a Java task, and setFork based on the
attribute in ANTLR.

Erik


Mime
View raw message