Re: Official sources vs. RCVS

[Karl Fogel]

I'd like to second Larry's response (though I'm less familiar with the
progress of the Renegade CVS project -- I haven't been following it
much).

As far as I can tell, the CVS development team generally responds
promptly and positively to useful patches.  Here, "useful" means
patches that follow the (imho quite reasonable) guidelines in HACKING,
including test cases, and which solve a problem that some significant
number of users care about.

(In fact, Larry himself seems to the one who most often picks up the
ball when such patches appear.)

But the majority of patches received don't fit these qualifications.
This isn't meant to be some sort of off-putting, elitist statement,
and I hope it doesn't sound that way -- the same thing can be said of
90% of the patches sent to 90% of the free software projects in the
world.

It would be nice if the CVS code base were a bit more modularized.
Then fitting in unusual functionality would be a matter of writing a
separable lump of code against an API (like an Apache module, for
example), and people wouldn't worry so much about whether or not their
patch made it into the central code base.  Unfortunately, the CVS we
have inherited is not like that, and keeping it maintainable and
comprehensible often must take priority over including every bell and
whistle.

... and if someone has a patch for *that* problem, please post! :-)

-Karl

larry.jones@sdrc.com (Larry Jones) writes:
> Paul Sander writes:
> > 
> > I seem to recall that the Renegade project was created specifically for the
> > reason that the official developers of CVS were unwilling to work with
> > Noel and others who needed various bug fixes and features.  At that time,
> > (a couple of years ago) the official developers were unresponsive to the
> > point of even failing to acknowledge receipt of bug reports initiated by
> > send-pr.
> 
> Not quite.  Renegade CVS was started by a couple of well-meaning but
> misguided people who were frustrated by the traditionally laze faire
> CVS development methodology compounded by the delays and uncertainties
> of the transition from Cyclic to Source Gear to Open Avenue.  (Anyone
> wanting the full story should consult the thread on Regegade CVS in
> the list archives at http://mail.gnu.org/pipermail/info-cvs/ from April
> 2000.)  In my opinion, it hasn't done much; it certainly hasn't solved
> any of the problems it was intended to.
> 
> > Should your statement be interpreted to mean that the official developers
> > are now more receptive to such "market-driven" forces?
> 
> The developers have always been receptive to market forces to some
> extent, but remember that most of us are volunteers with very limited
> time.  In general, bug fixes take priority over enhancements and things
> that are of personal interest to some developer take precedence over
> things that don't.  Things that are of interest to lots of people take
> precedence over things that are only of interest to a few, no matter how
> intense that interest may be.  Patches are always welcome, but please
> see the guidelines in HACKING for advice on strategy and tactics.  (Half
> thought out patches that don't quite work right all the time aren't
> particularly useful and that seems, in my admittedly biased opinion, to
> be what Renegade CVS has ended up specializing in.)  Likewise, new
> developers are always welcome (see DEVEL-CVS), but they're expected to
> show some long-term dedication to CVS and not just be dilettantes.
> 
> -Larry Jones
> 
> _______________________________________________
> Bug-cvs mailing list
> Bug-cvs@gnu.org
> http://mail.gnu.org/mailman/listinfo/bug-cvs