On 15 May 07, at 9:57 AM 15 May 07, Shane Isbell wrote:

Just to clarify, the implementation is now as follows: NMaven uses the
default repo format remotely and then transforms locally; this is the most pragmatic approach, and I don't have any immediate concerns. The problem, however, is that we are exposing the internal schema to the client; this creates a fair amount of confusion as people look for a general schema that satisfies the various languages, as opposed to a general API, say through
REST or SOAP. Although HTTP GET with a URL may qualify as an API,

I don't think it does. If you're referring to how an actual artifact is retrieve then it should only be via a set of parameters like artifactId, groupId, version, type (optional classifier) and the provider should deal with fetching the said artifact and giving it back to the client. Using an URL as the only way to retrieve an artifact is problematic.

So specify the set of parameters you can get an artifact from a Gem repository, CPAN, a Maven repository, or anything else provided you delegated to appropriate provider to deal with repository specifics.

The call for a artifact for assemblies or JARs would look the same once you have a reference to the provider. So you would have something like Maven Artifact eventually I would imagine.

under its
current form its really implementation (file-system) specific. I would be surprised if this issue doesn't keep coming up, as people become interested
in using Maven for other languages.

Shane

On 5/15/07, Jason van Zyl <[EMAIL PROTECTED]> wrote:


On 6 May 07, at 9:13 PM 6 May 07, Carlos Sanchez wrote:

> I didn't have a chance to talk about this with Shane but the idea in
> the end is to make the repository agnostic on how things are stored
> and how the client uses them.
> Right now is a simple directory, but could be a database with a web
> front end or anything like that.
> It shouldn't matter how NMaven artifacts are stored, as long as the
> client handles them correctly. A solution we talked about time ago was > to put them as any other artifacts and either developers could choose
> what format their local repo is in or the pom could say how they
> should be stored
>

It all boils down to packaging that's important. It doesn't matter
how they are stored. What matters is how they are transformed and I'm
sure someone can find a work around for having the name in the
assembly manifest being burned in and breaking the linker when the
file name and manifest entry doesn't match.

The repository can theoretically be stored in anything Wagon supports
but it's unlikely we'll stray very far from file-system based
mechanisms.

> But this is a total different discussion
>
> On 5/6/07, Daniel Kulp <[EMAIL PROTECTED]> wrote:
>>
>> Shane,
>>
>> Honestly, it sounds like the NMaven stuff will need a complete new
>> set of
>> repositories for NMaven artifacts.   There isn't any way, IMO,
>> that the
>> repo layout can change for the normal maven 1 and maven 2
>> repositories.
>>
>> Incubator or repo1.maven.org is relatively irrelevant in that
>> regards.
>> The layout is pretty much set in stone. There are too many plugins
>> (deploy, etc...) that rely on it, there are too many other apps
>> (several
>> different proxy applications, etc...) that rely on it, etc...
>>
>> If the current layout is inadequate for NMaven, the NMaven team
>> should
>> figure out an appropriate place for a new repository. My personal
>> suggestion is to work with the Maven team and create a new area at
>> repo1.maven.org/nmaven or similar.   But that's me.  In either
>> case, I
>> think that discussion is separate from where the m2 artifacts go. It >> make make sense to put the nmaven stuff in dist/incubator for a while >> until the layout is finalized, then move to central. However, the
>> layouts for m1/m2 are finalized.  Thus, they can/should go to
>> central.
>> (IMO)
>>
>> That said, I don't know the NMaven details.     But my #1 concern
>> is your
>> line:
>> > I
>> > would expect that an incubator release repo would be more
>> amendable to
>> > such changes.
>>
>> No chance, IMO.   Once an artifact is released, it's SET IN
>> STONE.   That
>> includes the layout of the repository it's sitting in.  Once
>> theres the
>> possibility that another project is relying on a particular
>> artifact to
>> be living at a particular location, it needs to stay there.   The
>> incubator m2 release repository is no different from central in that
>> regard.
>>
>>
>> Dan
>>
>>
>>
>> On Sunday 06 May 2007 14:11, Shane Isbell wrote:
>> > [ ] use standard repositories
>> > [ x ] relocate repositories under /www.apache.org/dist/incubator
>> >
>> > My reasons are as follows: First, NMaven does not follow the
>> standard
>> > repo layout; second, the repository layout structure is still in a
>> > state of flux, meaning that there is a need for potentially
>> changing
>> > the layout for .NET artifacts, while still doing releases.
>> >
>> > Getting more into some more specifics, with NMaven, there is no
>> > version information contained within the artifact file name and the
>> > version must follow a standard 0.0.0.0 format. This precludes
>> the use
>> > of "incubator" within the version itself. As mentioned above, at
>> this
>> > early stage, it's also not 100% clear on exactly how NMaven .NET
>> > artifacts will reside within the repo. For instance, there is an
>> open
>> > question as to where pom files will reside when we add the
>> concept of
>> > classifiers to the repo. Also, given the repository layout
>> structure
>> > for NET artifacts may change over time, as the incubator project
>> > evolves, I have concerns whether any of the standard maven repos
>> would
>> > accept - and with good reason - an NMaven incubator release at
>> all. I
>> > would expect that an incubator release repo would be more
>> amendable to
>> > such changes.
>> >
>> > Shane
>> >
>> > On 5/6/07, Eelco Hillenius <[EMAIL PROTECTED]> wrote:
>> > > [ x ] use standard repositories
>> > > [ ] relocate repositories under /www.apache.org/dist/incubator
>> > >
>> > > Eelco
>> > >
>> > >
>> -------------------------------------------------------------------- >> > >- To unsubscribe, e-mail: general- [EMAIL PROTECTED]
>> > > For additional commands, e-mail: general-
>> [EMAIL PROTECTED]
>>
>> --
>> J. Daniel Kulp
>> Principal Engineer
>> IONA
>> P: 781-902-8727    C: 508-380-7194
>> [EMAIL PROTECTED]
>> http://www.dankulp.com/blog
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: [EMAIL PROTECTED]
>> For additional commands, e-mail: [EMAIL PROTECTED]
>>
>>
>
>
> --
> I could give you my word as a Spaniard.
> No good. I've known too many Spaniards.
>                             -- The Princess Bride
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
>
>

Thanks,

Jason

----------------------------------------------------------
Jason van Zyl
Founder and PMC Chair, Apache Maven
jason at sonatype dot com
----------------------------------------------------------




---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



Thanks,

Jason

----------------------------------------------------------
Jason van Zyl
Founder and PMC Chair, Apache Maven
jason at sonatype dot com
----------------------------------------------------------




---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to