gump-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brett Porter <brett.por...@gmail.com>
Subject Re: [RT] module, project, target = repository, module, project...
Date Sun, 17 Apr 2005 11:37:19 GMT
> Hmm. One problem I see is that the rest of the world (ie make users, python
> developers) don't follow that model at all, and would have to make a
> significant adjustment to start thinking about "groups" and what are the
> groupings in their software.

group = product, project = thing that I build, artifact = thing that
comes out of my build

I actually thought that was a superset. Do you have an example from
those that doesn't fit? I thought also that python used a namespace -
equivalent to group, and packaging - equivalent to artifact, but I
don't know it well. Also bear in mind that the group doesn't need to
map directly to something, as long as whatever picked is consistent
with the spirit of the scheme and whatever gump is interacting with.

> Secondly, "everything has a unique id". It doesn't necessarily matter if its
> autogenerated, but its vital to do semantic-web-like stuff.

I'm not sure what you are quoting here, but if you want to have a
single unique id for each thing, I'd suggest it should be on the
artifact, not the project.

> Does that make sense?

I'm a little confused, but I think so.

Looking at this from the perspective of a Maven user (moreso than as a
developer), what I want is to not have to specify gump <-> Maven
mappings. I imagine this would be the same for any project that has
some sort of identification scheme of its own, rather than just
selecting them as they are added to gump.

What that means is that gump needs to internally take Maven IDs and
convert them to gump IDs (so a gump descriptor for Ant would need to
describe its Maven repository IDs, rather than the other way around as
it is now).

The other alternative is for the IDs to match - either by gump using
the Maven ones, or by Maven changing its repository.

I'm thinking the first option is the most realistic, and guards
against rogue naming by being able to add mappings.

Am I on the right page here for your goals?

Cheers,
Brett

---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@gump.apache.org
For additional commands, e-mail: general-help@gump.apache.org


Mime
View raw message