maven-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Paul Benedict <pbened...@apache.org>
Subject Re: [DISCUSS] Version Comparison Edge Case
Date Wed, 23 Sep 2009 02:34:40 GMT
I vote for option #2. Maven 3 should fix all things Maven; and if we need to
introduce backwards compatibility flags to help mitigate those problems, I
would do that too.

Paul

On Tue, Sep 22, 2009 at 9:26 PM, Brett Porter <brett@apache.org> wrote:

>
> On 16/09/2009, at 8:18 PM, Benjamin Bentmann wrote:
>
>  Ignoring for the moment that neither "3.0-!" nor "3.0.0.!" are
>> OSGi-conformant versions, the OSGi version ordering also implies
>>
>>  3.0 == 3.0.0 < 3.0.0.20090915-152839-2 < 3.0.0.GA < 3.0.0.SNAPSHOT
>>
>> So I'm not sure how much we can borrow from OSGi for Maven.
>>
>
> IIUC, The code that Hervé pointed to now has a number of workarounds so
> that OSGi works in most of its current uses, and Maven versions work in most
> of its current uses - but it doesn't conform to the spec of either. ie,
> alpha, beta, rc, cr, milestone and snapshot are all before the release, the
> rest after. I think this would have been a good idea if we'd done it to
> start with, but I'm sure there are weird versions out there that might cause
> confusion (only mediated by the fact that people don't use ranges all that
> often).
>
> I think there are 3 options here:
> - keep adding more special cases in the version comparer for things like !
> - declare a break in compat (this may result in having to make repository
> changes to preserve functionality though in a mixed environment)
> - restore the Maven 2 behaviour and make each place a range is used
> configurable so that OSGi can be used as an alternatives - for example -
> osgi:[x.y.z.q,) vs mvn:[x.y.z,)
>
> Cheers,
> Brett
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@maven.apache.org
> For additional commands, e-mail: dev-help@maven.apache.org
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message