Re: [imp-dev] Proposed reorg of IMP SVN repository structure

[Jurgen Vinju <jurgen@xxxxxxxxx>]

Hi Ed,

You can get a project set from a wiki page for example, and then all you need to do
is "import project set" from Eclipse.

Discussion: 

I agree with you that the proposed setup would simplify things from a certain perspective,
but I know from experience that it will complicate things too. At least where IMP's architecture is supposed to 
support variability, like using the run-time without having to use the meta-tools, is where its repository should 
support independent branching and tagging. To alleviate some of the overhead of modularity, we might go for 
some middle-ground where each "feature" (collection of "projects") is put into one main repository folder with 
its trunk,branches and tags. I wouldn't do that either, but nevertheless its an idea. Its a matter of guestimating which
parts of IMP are expected to evolve (semi-)independently and which are going to be in complete lock-step.

One other thing;  my developers only use and depend on two or three of IMP's projects. If we reorganize the
code as proposed, they will be presented (unnecessarily) with all of IMP if they check out the trunk.

Cheers,

Jurgen

On Fri, Apr 10, 2009 at 10:18 AM, Ed Willink <ed@xxxxxxxxxxxxx> wrote:

Hi

Seems like a really good idea to me.

I found the existing structure a major pain.

[What use is a project set; I've never learnt to use them, and
surely you can't use the project set till you've checked it out?]

       Regards

               Ed Willink

> -----Original Message-----
> From: imp-dev-bounces@xxxxxxxxxxx
> [mailto:imp-dev-bounces@xxxxxxxxxxx] On Behalf Of Robert M. Fuhrer
> Sent: 09 April 2009 22:15
> To: IMP Developers List
> Cc: Emmanuel Geay; Beth R. Tibbitts; Cal Swart
> Subject: [imp-dev] Proposed reorg of IMP SVN repository structure
>
>

> Hi Folks,
>
> I'd like to reorganize the SVN repository structure. Right now, it
> looks like this:
>
>    org.eclipse.imp
>        proj1
>            branches
>            tags
>            trunk
>        proj2
>            branches
>            tags
>            trunk
>        ...
>        projN
>            branches
>            tags
>            trunk
>
> It seems more conventional (and easier to deal with) for it to look
> instead like this:
>
>    org.eclipse.imp
>        branches
>            proj1
>            proj2
>            ...
>            projN
>        tags
>            proj1
>            proj2
>            ...
>            projN
>        trunk
>            proj1
>            proj2
>            ...
>            projN
>
> This way, for the most common case (checking out trunk or a set of
> projects with a given tag/branch), one has a better chance of being
> able to select a contiguous set of projects, rather than having to
> expand each one in order to find the appropriate trunk/tag/branch.
>
> Obviously this will cause short-term pain as we all rebuild our
> workspaces, but I think in the long run it will cause less confusion
> to newcomers and less manual work for veterans.
>
> Comments?
>
> --
> Cheers,
>    - Bob
> -------------------------------------------------
> Robert M. Fuhrer
> Research Staff Member
> Programming Technologies Dept.
> IBM T.J. Watson Research Center
>
> IMP Project Lead (http://www.eclipse.org/imp)
> X10: Productivity for High-Performance Parallel Programming
> (http://x10.sf.net
> )
>
> _______________________________________________
> imp-dev mailing list
> imp-dev@xxxxxxxxxxx
> https://dev.eclipse.org/mailman/listinfo/imp-dev
>


_______________________________________________
imp-dev mailing list
imp-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/imp-dev