Excerpts from Eric Kow's message of Wed Nov 04 19:24:49 +0100 2009: > On Wed, Nov 04, 2009 at 19:06:01 +0100, Nicolas Pouillard wrote: > > Excerpts from Eric Kow's message of Wed Nov 04 18:38:33 +0100 2009: > > > > * Git ones are more UNIXy, first you have separated commands git-foo, > > > > they are > > > > maybe hidden in some libexec directory, they may reuse the core > > > > library (or > > > > not), and they are spawned by the main command. > > > > I think I was thinking more of the git-style plugins: > > > > Great, then. > > > > > http://bugs.darcs.net/issue1504 > > Actually, darcs plugins (in my original proposal) would be even less > ambitious than git style plugins in that Darcs will never assume that > there is a standard place to look into to grab a plugin. > > Instead, we want to use something I believe is called a 'forcing > function' (*) to make users consciously aware that they're installing a > third party plugin that isn't part of Darcs proper. The forcing > function in question would be to require plugins to be specified in the > user configuration, something like this: > > frobnicate /home/me/darcs-frobnicator.py > show index /usr/local/bin/hashed-storage-show-index > show authors /usr/local/bin/darcs-show-authors > delta-debug /usr/local/bin/even-better-than-bisect
I like this idea! > (*) If you're a real UI guy, please forgive me if I've misunderstood the > idea as explained by Donald Norman and abused the terminology > > > > I'll note also that I'm not terribly attached to the idea of darcs > > > plugins. I just thought it could relieve some pressure to implement new > > > functionality, and also in some sense extend the Darcs hacker community > > > beyond the Haskell-friendly world. But I imagine it also leads to some > > > interesting new problems that we may not be prepared to solve. > > > > What new problems do you have in mind? I don't see plugins causing > > troubles, > > providing an accessible core library on the other end will leads to > > interesting problems, but this is not really caused by plugins. > > I'm not very clear on this, but one thing I'm worried about is that in > the real world, nobody's going to pay attention to what is darcs and > what's a plugin. Because people are naturally inattentive, darcs > plugins will effectively *be* darcs, as far as users are concerned if > we're not careful about it. I worry that this would get us into trouble > somehow. I found Mercurial quite clear on that and seen the distinction. -- Nicolas Pouillard http://nicolaspouillard.fr _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
