On Aug 23, 2013 9:20 PM, "Igor Fedorenko" <[email protected]> wrote:
>
> Is scope=provided transitive? I believe it isn't [1], which means that
> all maven-core consumers will have to pick a version of guice to be able
> to use maven core and manually update guice dependency each time they
> update maven dependency version.

The idea would be that maven-core would then declare its own direct
dependency on whichever flavour of guice Maven wanted to use (which it
would need to do anyway if you wanted to exclude and swap in a different
guice edition), because the provided dependency via Sisu would not be
transitive. So this would only affect direct consumers of Sisu and not
their dependents.

That said in an ideal world there would be something like the "provides"
alias suggested by Stephen - but as this isn't available yet I'm looking to
see if there's a better approach than the current "pick a default and
exclude as necessary".

> At the same time, replacing guice with a different version does not
> require that many exclusions, at least this is the case for m2e embedded
> maven runtime [2].
>
> I believe current setup provides the best balance between easy of use
> and flexibility to customize and my preference is to keep it as is.
>
> [1]
http://maven.apache.org/guides/introduction/introduction-to-dependency-mechanism.html#Dependency_Scope
> [2]
http://git.eclipse.org/c/m2e/m2e-core.git/tree/m2e-maven-runtime/org.eclipse.m2e.maven.runtime/pom.xml
>
> --
> Regards,
> Igor
>
>
> On 2013-08-22 11:57 PM, Stuart McCulloch wrote:
>>
>> As one of the main downstream users of Sisu would you prefer it to
declare
>> a provided scope dependency to (sisu-)guice rather than the current
compile
>> scope dependency?
>>
>> Making it provided should make it easier to swap in alternative versions
>> while still documenting the dependency - and avoid lots of tedious
>> exclusions. The only downside I can see is that downstream users like the
>> Maven runtime would then need to explicitly remember to add the
>> (sisu-)guice dependency in their final application artifact/assembly (and
>> potentially in some tests) as it would no longer be transitively
included.
>> (though that might be a good thing documentation-wise)
>>
>> WDYT?
>>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>

Reply via email to