On Wed, May 14, 2008 at 2:57 AM, Brett Porter <[EMAIL PROTECTED]> wrote:
>
>  On 13/05/2008, at 8:21 PM, Trygve Laugstøl wrote:
>
>
> > What about compatability then?
> >
>
>  Can we not keep the old ones and deprecate them? That's what I did with the
> old goals.
>
>
>  On 14/05/2008, at 12:28 AM, Asgeir S. Nilsen wrote:
>
> > So when did you decide on changing the groupId to
> > org.codehaus.mojo.appassembler?
> >
>
>  Apologies, I thought I posted explicitly about that to this list, but I
> can't find it. I changed it in January and posted about the other changes.
>
>  Given it has multiple modules, I think it makes sense - any objections?

I've changed the webstart groupId 3 months ago.

As the old groupId (org.codehaus.mojo) was special, maven users lost
the ability to make simple invocations on the command line. They also
lost automatic upgrades as their POM is now not configured for the new
groupId.

As a result, I get
* issue reporteds by people still using the old version (issues
already fixed in latest release),
* people complaining for accidently running the old plugin on the
command line (it's still there in the repository)
* people having to use a too verbose (they don't know the settings.xml
groupId thingy trick, which I personally find unpractical - requires
local tuning, doesn't make it reproduceable accross developer's
environments).

A work-around would be for me to be able to publish some sort of
obsolete "transitional" package (in the Debian sense), either that
just fail to indicate the existence of the new groupId, either that
transparently execute the real plugins (not sure if that possible e.g.
using inheritance). I am not sure if that has been done, but I am
seriously thinking about implementing something like that.

As I see that several mojos have moved from one groupId to another, we
could maybe come up with a good strategy to avoid those pitfalls. Or
improve the maven support for this use case.

Cheers,

Jerome

---------------------------------------------------------------------
To unsubscribe from this list, please visit:

    http://xircles.codehaus.org/manage_email


Reply via email to