www-infrastructure-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Santiago Gala <santiago.g...@gmail.com>
Subject Re: [scm] Discussing vs. doing (Was: Best Practices so far?)
Date Sat, 03 May 2008 09:45:44 GMT
El vie, 02-05-2008 a las 16:05 +0300, Jukka Zitting escribió:
> Hi,
> On Fri, May 2, 2008 at 1:51 PM, Grzegorz Kossakowski
> <gkossakowski@apache.org> wrote:
> > As I previously said at @infra I'm not that much interested in discussions
> > with [dSCM] tag even if they are enjoyable to read from time to time. I
> > think that we should focus on _doing something_ and my choice is to let
> > existing Apache committers to try out dSCM tools in a full manner.
> The reason why at least I wanted to focus on use cases at first is to
> make it clear WHY we need better tools instead of focusing directly on
> WHAT those tools should be or HOW we should be using them. The latter
> questions are of course quite essential, but especially given the
> heated opinions about this topic I think it's very important that we
> make it clear that the need for better tools is real and that we're
> not just looking for shiny new toys.

+1 For me just the unconnected operation and speed improvements on
browsing and searching are worth it.

> I think now that we already have a good body of valid use cases (see
> the archives) we can start focusing more on the implementation
> details. Also, I think the discussions so far have quite clearly shown
> that we should be looking at incremental improvements on a number of
> fronts, including svn upgrades (1.5) and enhanced server infra (eu
> mirror, etc.), better interoperability with various scm clients
> (git-svn, IDEs), guidelines for patch and branch workflows, etc.

re: interoperability, some people is using mercurial too. I rejected it
for my personal use due to it requiring a subversion checkout, which is
typically bigger than a whole git repository + working copy, branch
management and the convoluted commit integration. I also prefer the git
gui. But I guess we should document about it too. I think someone at
shindig published a detailed guide to get a mercurial setup going.

re: IDEs, I have built egit a number of times, and I guess I could
provide an eclipse update site for it, in case people want it.

But I think the most important part of it is how we enhance the global
knowledge about the concepts and implications of SCM. Computer Science
curriculi have typically little to no exposure to the concepts (at least
here, just bits of it around life-cycle management), so this is
something that people is typically forced to learn "on the job". So
anything we can do to improve the quality of patches and branch
management is good for the ASF.

> So, I'm happy as we can move on to actually doing things and solving
> the detailed problems we come across, but I also hope people
> understand the value of having spent some time just discussing the
> issues in general. I believe that we now have a much better collective
> understanding of what the key issues are and how we should go about
> solving them.

+1 My understanding of all the issues is much better now.

> > (andd of course: infra is busy with higher-priority tasks so my wish is not
> > intended to be fulfilled right away)
> +1!
> BR,
> Jukka Zitting

Santiago Gala

View raw message