ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ken Wood <>
Subject Re: onFailure
Date Mon, 29 Jan 2001 17:11:47 GMT

David Corbin wrote:
> What made me reconsider was message from Ken Wood that
> indicated ant should be focused on the what to build and how, and not
> all the periphery stuff around that.

Ah, but rarely are boundaries clearcut!

Strictly speaking, ant doesn't need to know
how to get stuff from a repository, any more
than it needs to know how to schedule itself to build...
Ant should just know how to do the build.

But, in fact, for many projects the first
step of a build is to get the latest
versions from a repository. Ant has
a variety of tasks for doing that via
different configuraiton management tools.
This happens frequently enough, and commonly
enough, that many people include accessing
a repository as the definition of what it
means to do a build. So, that boundary
of what defines a "build" is rather fuzzy.

Since I was dropping ant into an existing
infrastructure that already knew how to
get the latest files from the repository,
I've never used these features of ant. 
But I agree that they rightfuly have a place
within what we expect ant to be able to do.

Oh, and I like your idea. It's a little
odd that the listener can affect the behavior
of the build, but builds are often an odd problem
to deal with anyhow, so why not...

> Thoughts?
> --
> David Corbin
> Mach Turtle Technologies, Inc.
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

View raw message