On 25-May-08, at 3:21 PM, Brett Porter wrote:

Hi Jason,

I saw some commits over the weekend to the sandbox introducing mercury and merging your personal artifact branch in.

Is this just a prototype you're experimenting with that will be merged into maven-artifact piece by piece by a later proposal, or are you expecting to replace maven-artifact wholesale?


Eventually I would like to replace what we currently have with that, I need to write it up and provide visualization. That said wagon or commons-transport can plug into it. What I did with it is try to go simple, but robust and we'll see how far that gets us. Might not be far enough.

I went for simple HTTP first, and I enlisted the help of Greg/Jan and I doubt there is anyone who knows HTTP better then them. What is there is atomic, with configurable connection pooling and parallelization. I started with what I could absolutely guarantee could be atomic. So here for transport anything could be plugged in but I doubt it's going to be better then what Greg/Jan (and Jesse) whipped up. But wagon is still viable with an adapter for folks using FTP or SCP and they probably can be made as robust if they follow the same patterns Greg/ Jan created.

As far as the rest of it goes its layered as artifact resolver / metadata resolver / retrieve. Pure retrieval to a tree where the retrieval is dumb but safe, and then it gets interesting as version ranges are essentially an NP complete problem and so I think the best attempt we can make is borrow the SAT solver that is used in P2.

The main reason I ask is that as I've mentioned I'm working on finishing up a wagon 1.0 release. Though I'm mostly done with the HTTP parts there are patches for streaming, and there's the branch Don submitted with parallel download processing, which I wanted to get into trunk. I don't really want to spend a lot of time on that if it's going to end up being replaced.

It can stilled be used in 2.0.x and Wagon can even still be used in Mercury if it proves viable. The vast majority of folks only care about HTTP/S for retrieval and I believe that vast majority of people only care about DAV or deployment to a repository manager but Wagon can still be used for doing other protocols. I will also probably strip the Plexus requirement out of simply by providing constructors other IoC containers can get hold of. Basically the library for anyone wanting to talk to remote artifact repositories. That's the plan anyway.

I plan to have something working tomorrow, and hopefully this week have some visualization in Eclipse. Eugene made a little framework for visualization but I'd like to adjust it so that you can hit enter and watch another step in the resolution, enter enter enter enter and hopefully people will be able to see the potential problems. When that's done then it will be time to chat but I'm hopeful. Oleg and I will be working on it most of this week so we should have something to demo shortly.

So wagon will work nicely with it and for those using not-HTTP they will be fine. Though in the future I see everyone using HTTP/S for download, and DAV or just PUT to upload.



Thanks,
Brett

--
Brett Porter
[EMAIL PROTECTED]
http://blogs.exist.com/bporter/


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


Thanks,

Jason

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

Selfish deeds are the shortest path to self destruction.

-- The Seven Samuari, Akira Kirosawa



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

  • mercury? Brett Porter
    • Re: mercury? Jason van Zyl

Reply via email to