> On Wed, 2003-12-10 at 11:33, Jason van Zyl wrote:
> > Howdy,
> >
> > I'm moving along with the maven-project component and I
> have arbitrary
> > levels of inheritance working for properties and the model.
> >
> > Now I'm starting to look at the property interpolation
> mechanism and I
> > would like to propose for 1.1 that values in properties files be
> > interpolated in the model values but ${pom.foo} values not be
> > interpolated in properties values.
I think I'm confused. Do you mean not permitting property file values
to exist in the POM xml? That makes total sense to me because the POM is
Supposed to know nothing about plugins or other processes.
> > The first reason is that it makes the process a lot more clear and
> > second I'm not sure if having model values interpolated in
> properties
> > values is really of any use. I think when we have cases like:
> >
> > maven.xdoc.version = ${pom.currentVersion}
> >
> > We could really just use the ${pom.currentVersion} directly in the
> > relevant plugin instead of putting it in a property and
> then sucking
> > it back in.
> >
> > At any rate, to make the project processing in 1.1 highly
> > deterministic I would like to remove ${pom.foo} value
> interpolation in
> > properties files.
>
> Another point that I wanted to raise was that properties may
> ultimately come from other sources such as database, or ldap
> and how much sense does it make for those sources to have
> access to the POM. I honestly don't think a whole lot. I
> think properties should just be flat name/value pairs with no
> ${pom} interpolation.
In my opinion, I agree with you in that the statement above
maven.xdoc.version = ${pom.currentVersion}
doesn't offer any more clarity than just using ${pom.currentVersion}.
However I think there might be times where you might want to declare
a local variable in a plugin that contains several references to the POM,
e.g.
maven.someVar = //steven-${pom.foo}/${pom.bar}/garcia/${pom.ruu}
So it's easier to use ${maven.someVar} than the ugly thing it equates to.
I would also think that normal, Ant style substitution within properties
should remain (and I don't think you've said anything to suggest the
alternative.)
> --
> jvz.
>
> Jason van Zyl
> [EMAIL PROTECTED]
> http://tambora.zenplex.org
>
> In short, man creates for himself a new religion of a
> rational and technical order to justify his work and to be
> justified in it.
>
> -- Jacques Ellul, The Technological Society
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
