[News.eclipse.technology.packaging] Re: Is the Packaging tool Pde build

[Jochen Krause <jkrause@xxxxxxxxxxxxxx>]

Hi Pascal,

You have a very valid point. We should not reinvent the wheel and it
likely makes sense to collaborate on this. It seems like PDE build 
offers a lot of the functionality that we plan to provide.

I would propose to get the packaging project started and make it the 
first work items to research PDE build for packaging purposes, determine 
how we collaborate and where the different components live. Does that 
make sense to you?

Jochen

Pascal Rapicault wrote:
> Hi,
>
> Since the call we had yesterday I have been thinking about the packaging 
> tool. I deliver my thoughts here. Feel free to comment, ignore, etc...
>
> The more I think about the packaging tool to be built in the context of this 
> project and see what pde build packager offers, the less I see the need for 
> developing another technology.
> Of course I'm not saying that packager would fulfill all the needs of this 
> project (for example currently it does not fetch from update site), however 
> if we settle on building zips of various groups of plugins, then packager 
> provides all the core functionality required. More importantly, since no new 
> technology needs to be developed (to start with getting from an update site 
> is not mandatory) the project could start producing those zips in time for 
> Europa M5 which means that the community could be able to provide feedback 
> on the various drops at eclipsecon.
>
> For history, packager has been introduced in eclipse 3.0. The main two 
> clients I know are:
>     - SDK releng team which uses it to produce all the SDK downloads since 
> february 2006
>     - IBM to facilitate the consumption of various eclipse pieces by various 
> teams
>
> Note that I don't have any personal interest in promoting this technology 
> nor has IBM. In fact if packager is adopted it would just add work to my 
> plate :-). In the end I'm just looking for consistency, avoid reinventing 
> the wheel and minimizing the confusion for the end users.
>
> PaScaL