You're mixing two issues here.
1. Maven *is* opinionated software. Just take a look at a pom and ask
yourself: how is it going to build that? Maven must come up with
something and thus it's opinionated. The problem here is that a lot of
the time I don't agree with what it comes up with.
2. Deterministic vs non-deterministic behavior. If Maven 2.0.8 really
uses a HashSet for its dependencies instead of a LinkedHashSet than it
shows non-deterministic behavior. That's just plain unacceptable and
supersedes regression.
Carlo
Andrew Lee Rubinger wrote:
This is not simply a problem for EJB3, it's proof that Maven
2.0.8 and
2.0.9 may lead to incompatible builds.
I've filed:
http://jira.codehaus.org/browse/MNG-3606
S,
ALR
Carlo de Wolf wrote:
> Yes, and no I don't want to talk about it.
>
> Carlo
>
> Andrew Lee Rubinger wrote:
>> EJB3 Build is currently hosed in Maven 2.0.8, but working in 2.0.9.
>> Running off the same machine, off a freshly-cleared local repository.
>>
>> The reason for this is dependency resolution differences between
>> these Maven releases, even though the Dependency Plugin is the same
>> version:
>>
>>
http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4154158
>>
>> Anyone seen issues like this?
>>
>> S,
>> ALR
>>
>> PS - Not an invitation for Maven bashing. I'm aware that this sucks.
>> _______________________________________________
>> jboss-development mailing list
>> jboss-development(a)lists.jboss.org
>>
https://lists.jboss.org/mailman/listinfo/jboss-development
>
> _______________________________________________
> jboss-development mailing list
> jboss-development(a)lists.jboss.org
>
https://lists.jboss.org/mailman/listinfo/jboss-development