Re: I resign as Guile maintainer

["Greg J. Badros" <gjb at cs dot washington dot edu>]

Nicolas Neuss <Nicolas.Neuss@IWR.Uni-Heidelberg.De> writes:

> > From: Michael Livshin <mlivshin@bigfoot.com>
> > 
> > best.  and I also agree with Dirk's point about the perfectionism wrt
> > feature additions -- there is a stable version out, after all, so let
> > the CVS version be as shaky as it needs to be.
> > 
> > --mike
> > 
> 
> My experiences with shaky CVS versions is very bad.  Then, many people
> prefer the stable versions, and you end up in having less testers for
> new features.  Probably, it's the main duty of a maintainer to ensure
> a good compromise.  Here a proposal (which I made some years ago in my
> group, and which was not accepted:-(), if one does not have a central
> maintainer: if several people have writing access to the repository,
> one could set the rule that at least two persons of that group take a
> look at code before it is checked in.

Having two people look at code before checking in is not a good plan,
IMO.  Too many headaches:  CVS commits are a great way to share
code... emailing around patches stinks.

The repository need not be perfect at all times (though ideally it
should always build and nightly regression tests would be beneficial).
It'd be a great goal to have someone look over commits of others, but
demanding that others do so before the commit just makes things
unnecessarily tedious.

Greg