On Thu, Oct 16, 2008 at 5:00 PM, Trent W. Buck <[EMAIL PROTECTED]> wrote:
> "Jason Dagit" <[EMAIL PROTECTED]> writes: > > > If we're going to use franchise for darcs, why do we need to exclude > > the use of autotools and cabal? > > Because maintaining build infrastructures is a pain, and thus > maintaining more than one is both undesirable and (at least > theoretically) unnecessary. I have heard the above argument before and I don't find it compelling in this context. Let me explain why. This would be a good answer to the question, "Why should we minimize the number of build systems?" But I feel like we're in a different context (hence my big list of points in the previous email) and so I'm asking a different question above. Briefly, we're in a situation where autotools currently meets our needs on all but one platform and Cabal compliance is a common request and fills the need on the platform that autotools misses. So it seems that regardless of which build system we use, cabal support is going to continue to be requested and the pain of redundant maintenance on the build system(s) will still be taken on by someone. It seems as though we would be smart to accept that work into the official repository so we can ensure people are not duplicating the effort of said potential redundancy. Even if it's only maintained by volunteers. As an example this would parallel some of our build targets like api-docs. David doesn't build the api-docs so he doesn't worry if that support breaks. But, those of us who do use the haddocks submit patches when/if it breaks. I haven't heard any complaints about this way of working. I suspect someone will respond with "If you don't like it, please maintain a branch with that feature." Which is actually the case now. People who want cabal support are sharing patches in unofficial darcs repositories. I think the community would be better off with a unified resolution. Thanks, Jason
_______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
