At 9:10 PM +0100 3/11/08, simon wrote:
>
>It's rather odd to want to inherit anything except (transient)
>dependencies from an "external" pom.
I agree - it is an odd case as well as an interesting one. :)
>Normally, you don't want an external pom to dictate how *your* project
>is built, or what reports *your* project contains, etc.
exactly.
>What is in pom X that you want T to have?
our test project is inheriting from a fairly complex external test framework,
which creates some test profiles that we are using, and declares
the dependencies for these profiles:
<artifactId>spring-osgi</artifactId>
<groupId>org.springframework.osgi</groupId>
<version>1.0</version>
/r
>On Tue, 2008-03-11 at 12:15 -0700, Russ Tremain wrote:
>> I have an interesting situation where we have one test project
>> that wants its parent to be an external project pom.
>>
>> I don't want that, because it means I can't have it inherit from a common
>> parent we have described locally.
>>
>> I.e. let X be the external pom, L be the local parent, and T be the
>> test project.
>>
>> What I want is:
>>
>> (X+L) -> T
>>
>> What I have is:
>>
>> X -> T
>> or
>> L -> T
>>
>> neither of which works for us.
>>
>> So, is there any way to inject the contents of X into T while
>> retaining L as T's parent?
>>
>> I have considered the following solutions:
>>
>> I. make X the parent of my entire project. This is undesirable
>> because I would inherit
>> X's dependencies in all projects, along with various other
>> properties unrelated to our
>> project (<developers/>, <scm/>, etc).
>>
>> II. duplicate L to L', and inherit X in L'. This is undesirable
>> because I would
>> be duplicating common definitions, which is error prone and
>> would require additional
>> training and maintenance.
>>
>> III. inherit L in L', duplicate selections from X in L', and inherit
>> L' in T. This is
>> probably the most workable solution, but would cause some pain
>> whenever we have to update
>> X.
>>
>> IV. an automated version of III, that would create a new project
>> installing L' before T
>> is executed.
>>
>> Has anyone else encountered this problem, or have other ideas?
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
