www-repository mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Brian E. Fox" <bri...@reply.infinity.nu>
Subject RE: releases using incubator artifacts
Date Mon, 12 Nov 2007 13:14:08 GMT
I don't think in this case it would normally be transparent that
active-mq is using an incubator. If we didn't have a firewall this would
not be apparent. 

I guess this is one reason it's not a best practice to define other
repos in your poms.

Anyway, saying that it's good this breaks so people add it to their
repoMan doesn't fly because not everyone uses a repoMan and how would
they know? Not everyone is going to completely inspect every dependency.

It simply boils down to do we trust the incubators or not? And if not,
then we shouldn't allow releases that use them. If we trust them, then
put them in central. The assumption here is that if active-mq tested and
trusts this component, than I as a user would trust their judgement, but
I shouldn't have to jump through arbitrary hoops to use it.

-----Original Message-----
From: Brett Porter [mailto:brett@apache.org] 
Sent: Monday, November 12, 2007 1:37 PM
To: repository@apache.org
Subject: Re: releases using incubator artifacts

On 12/11/2007, at 7:15 AM, Brian E. Fox wrote:

> I agree that releases shouldn't use incubator artifacts, but I can  
> also
> see that if this is the case, the incubator projects will never find
> users and thus would never get out of the incubator. The only thing  
> that
> makes sense then is to publish them to central. If they are official
> releases signed off by the incubator PMC, then what's the harm? The
> current method of not publishing them AND allowing them to be used is
> troublesome.

Robert Burrell Donkin said...
"AIUI the issue wasn't the officialness or not of the releases but the
inability for users to make an active and informed choice to use
incubating software (or not)"

Which kind of makes sense. Though by the same token, you would hope  
that their active and informed choice to use activemq (no pun  
intended) was transitive to activemq's active and informed choice to  
use activeio. (hmm, I think I may be channelling James Strachan).

Personally, I'm happy to press the 4 or 5 buttons it took to proxy  
the incubator repo to Archiva whitelisting the projects I know about  
as they come in :)

Perhaps we just need to explain this on the maven site and let  
projects that use incbuating stuff link there? I would hope it's not  
a very frequent (or at least permanent) occurrence.

- Brett

Brett Porter - brett@apache.org
Blog: http://www.devzuz.org/blogs/bporter/

View raw message