No problem, I can add a new option something like --useFallbackRepositories.
Just to make things clear what I suggested is to change the default value of
maven repositories from a list with 4 repositories to just one. No code
changes.

On Sat, Jul 18, 2009 at 6:52 AM, Niclas Hedhman <[email protected]> wrote:

> Alin,
>
> You know my general position of depending on something that might disappear
> overnight.
>
> Your suggestions make sense, with the above risk hanging over it.
> I will +1 it IFF you maintain the current mechanism as a fallback, possibly
> in separate bundle and activated by command-line option. Having a single
> point of failure in the hands of VCs isn't acceptable.
>
> Guys, please realize that even though we like what Sonatype and
> SpringSource are doing, strong financial interests are in control and can
> pull the rug from under them and us if we depend on them. Always assume it
> will happen, and come up with the contingency plan for "then what" question.
>
> Cheers
> Niclas
>
> On Jul 18, 2009 12:00 AM, "Alin Dreghiciu" <[email protected]> wrote:
>
> Hi guys,
> back on the performance issue with pax exam/pax runner we discussed a while
> back. One of the reasons finding something as a profile goes slow is because
> of the resolution process, To figure out where an artifact is located mvn: 
> handler has to go to all the repos and look for metadata determine highest 
> version (in some cases) and then get the artifact. In case that there are 
> more repositories out there , as in current case where we have 4 (central, 
> ops4j, spring external, spring releases) this
> process could be very slow. An simple solution to speed this up is to have
> just one repo to look into it.
>
> This is the reason why I asked for this:
> https://issues.sonatype.org/browse/OSSRH-50 . And this is already done and
> can be accessed at: http://osgi.sonatype.org/content/groups/pax-runner/
> the advantage in this case is that Nexus will pre merge the metadata from
> all of this repositories and give it back in one shot so the resolution
> process should be faster. I did not made actual measurement but is look also
> faster doing a manual test.
> As this is an indirection, another advantage of the approach is that in the
> future we can easily add/remove repos on the fly without having users
> download a new version of pax runner.
> Nice is that if we do that I can remove the special handling of profile
> repository that I added to the list just in case that --profiles was used,
> fro the same reasons. To not search yet another repository if there are not
> profiles involved. But now, the profiles repository is already in the
> composite repository.
>
> So, what I'm doing here is to ask for a vote if I should change the default
> list of repositories to this one. Please vote with a +/-. The vote should be
> open till our next pax runner release at least.
> For now you can try this out by using the --repositories=
> http://osgi.sonatype.org/content/groups/pax-runner/
>
> --
> Alin Dreghiciu
> Software Developer - Looking for new projects!
> My profile: http://www.linkedin.com/in/alindreghiciu
> My blog: http://adreghiciu.blogspot.com
> http://www.ops4j.org - New Energy for OSS Communities - Open Participation
> Software.
> http://www.qi4j.org - New Energy for Java - Domain Driven Development.
>
> _______________________________________________
> general mailing list
> [email protected]
> http://lists.ops4j.org/mailman/listinfo/general
>
>
> _______________________________________________
> general mailing list
> [email protected]
> http://lists.ops4j.org/mailman/listinfo/general
>
>


-- 
Alin Dreghiciu
Software Developer - Looking for new projects!
My profile: http://www.linkedin.com/in/alindreghiciu
My blog: http://adreghiciu.blogspot.com
http://www.ops4j.org - New Energy for OSS Communities - Open Participation
Software.
http://www.qi4j.org - New Energy for Java - Domain Driven Development.
_______________________________________________
general mailing list
[email protected]
http://lists.ops4j.org/mailman/listinfo/general

Reply via email to