Just wanted to add that the goal of the process described below is to
automatically generate a pom suitable for deployment from the hadoop
source tree. I know that's not possible without at least a minimal
patch that fixes the jetty dependency where the groupId and artifactId
are transposed.

The alternative is to manually craft a pom from hadoop's provided
template, which is essentially what we would do if we deployed it
ourselves.

On Tue, Mar 9, 2010 at 7:56 AM, Drew Farris <drew.far...@gmail.com> wrote:
>
> Yet another alternative would be to figure out how, given a source
> tree of 0.20.2, to deploy a working artifact from it and post these
> instructions to the JIRA issue in the hopes that some soul from the
> hadoop project would follow them. I've messed around to this end
> briefly, but wasn't able to wrap it up between planes, trains,
> automobiles, children and thick stacks of forms that are begging to be
> completed.
>
> I have discovered that, 'ant -p' at the top level the hadoop sources
> reveals that there's a target that generates a maven pom file, running
> it once produced an error for me and running it a second time produced
> a pom that seemed inadequate and included the incorrect version
> number. Apparently this is some sort of ivy generated file instead of
> being a 'filled in' version of the pom template present in the source
> tree under the ivy directory.
>
> On Tue, Mar 9, 2010 at 7:35 AM, Sean Owen <sro...@gmail.com> wrote:
>> "What is, 'what do we do about Hadoop 0.20.2 not being in Maven yet'?
>> I'll take Potpourri for $400, Alex."
>>
>> I think Drew should just commit his copy of 0.20.2 under the Mahout
>> group ID and be done with it.
>>
>> 2010/3/9 Gérard Dupont <ger.dup...@gmail.com>:
>>> What was the question ?
>>>
>>> On Tue, Mar 9, 2010 at 13:13, Benson Margulies <bimargul...@gmail.com>wrote:
>>>
>>>> Options:
>>>>
>>>> 1. Email to their private@ list.
>>>> 2. Email to members@ seeking assistance.
>>>> 3. Sleuthing their archive to see who was the release manager for this and
>>>> try to persuade them.
>>>>
>>>> Opinions?
>>>>
>>>
>>>
>>>
>>> --
>>> Gérard Dupont
>>> Information Processing Control and Cognition (IPCC) - EADS DS
>>> http://forge.ow2.org/projects/weblab/
>>>
>>> Document & Learning team - LITIS Laboratory
>>>
>>
>

Reply via email to