Fallout from the CVS discussion

[Adrian Chadd]

My personal take on all of this is that it's all very badly "directioned."

As in - personally, I think CVS as a package is fine. However:

* I'd like to first see a roadmap for doing this - eg, "we're adding a
NO_CVS option; CVS will become a port, you can migrate to the CVS port
with your next build/installworld";
* if you're that way inclined, backport the NO_CVS option (if it
doesn't exist) to -9;
* Ensure all of the stuff that uses CVS is migrated beforehand, and
publish all of that effort somewhere;
* Make sure you're doing it for reasons that aren't coming across as
"GPL free! at all costs!"

The last is the most important to me. I am beginning to feel that the
push for clang, no CVS/RCS, migration to BSD licenced tools, etc is
for reasons other than _a sound set of technical and long-term goal
reasons_. You run the risk of falling afoul for the same kinds of
stupid crap that GPL zealots fall afoul of - you first pick a
political/philosophical stance, then you base all your technical
decisions on that. Then stir in a bit of cognitive dissonance and
suddenly you're coming up with justifications to remove things - when
the underlying honest reason is "because it's not BSD licenced."

Now, to stir up trouble, I hereby suggest that if you're going to
remove CVS because it's no longer used for FreeBSD's project stuff, we
should obviously import subversion into the base because _it_ is being
used for the FreeBSD project stuff. Think of why you're not doing that
(likely because it's already a port/package and there's just as much
inertia to introduce something to the base system as there is removing
it and making it a port) and see if that helps refocus your reasons
for and against doing things.

2c,



Adrian