2011/5/18 sebb <seb...@gmail.com>

> On 18 May 2011 08:31, Antonio Petrelli <antonio.petre...@gmail.com> wrote:
> > 2011/5/17 Henri Yandell <flame...@gmail.com>
> >
> >> Looks useful to me.
> >>
> >> Download the ant task. Use this standard file. Execute.
> >>
> >> Save us maintaining a pointless build.xml.
> >>
> >
> > I don't think we need it at all. Don't you think that developers are able
> to
> > download Maven like they did for Ant?
>
> Maven requires a lot more downloading than just the product.
> The first time you use Maven for a build, it will download a lot of
> other stuff, much of which is not (yet) needed.
>

I don't want to start a Maven vs Ant war, but we need to choose only one of
those.
Providing two ways of building a project is not choice, it increases
confusion.
What is important to OGNL users is the possibility to use it in a Maven,
Ant, Whatever build tool.
Since Apache is committed to Maven we *should* use Maven. But Ant users can
use Ivy. Other build tools are able to use Maven repository.

> I think that presuming that developers want always to use Ant for their
> > builds is an insult to their intelligence :-D
>
> What has intelligence to do with it? This is about providing choice.
>

What I meant is that developers are able to download and install Maven like
they did it for Ant. After some minutes of ranting, they will be able to
build the project from source, even if we don't provide an Ant build file.


>
> > Don't misunderstand me, Ant was a great tool, but now we have to move on,
> > like when most of us moved away from Java 1.4.
>
> Ant still is a great tool. But it has different goals from Maven.
>

But surely we won't use it as a primary way of building OGNL. What I am
trying to say is that it is superfluous to use it even as a secondary way.
Adding a plugn for Maven that generates a build file, adds a (minimal)
amount of maintenance effort that, IMHO, can be avoided because it does not
give a real benefit to final users.

Antonio

Reply via email to