[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [p2-dev] Why is the director needed during install?

There is two approaches:
- When you are building your repo, the aggregator tool you use could detect old style update site (maybe plug-in a new repo factory?) and download every jar from every feature so it can perform a complete generation for the IU.
- Iterate over the repo looking for the partial IUs, download artifact into a temp place, generate the IU and replace it in the repo.


Inactive hide details for Thomas Hallgren ---09/24/2009 11:13:14 AM---Hi Pascal,Thomas Hallgren ---09/24/2009 11:13:14 AM---Hi Pascal,


Thomas Hallgren <thomas@xxxxxxx>


P2 developer discussions <p2-dev@xxxxxxxxxxx>


09/24/2009 11:13 AM


Re: [p2-dev] Why is the director needed during install?

Hi Pascal,
That was interesting news.

My use-case is that I first aggregate a repository. The result of the
aggregation is always a P2 repository. Seems to me, it would be
beneficial if the aggregator could resolve this issue so that the
resulting P2 repository is complete.

How would I do that? I guess I would start by looking for the partial
IU's. But then what?

Thomas Hallgren

On 09/24/2009 04:45 PM, Pascal Rapicault wrote:
> This dependency is from the eclipse touchpoint to the publisher. It is
> necessary in order to generate the complete IUs for the bundles being
> installed from the legacy update sites (becasue when we are generating
> metadata for UM site we don't pull down all the bundles, but just the
> features).
> The code is written this way so we can honor the fact that the dependency
> on the publisher is optional. For example if you know you only contact p2
> repos.

p2-dev mailing list

GIF image

GIF image