Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 17 Dec 1995 13:53:36 -0700 (MST)
From:      Terry Lambert <terry@lambert.org>
To:        hsu@clinet.fi (Heikki Suonsivu)
Cc:        bugs@freebsd.netcom.com, current@FreeBSD.ORG
Subject:   Re: FreeBSD-current-stable ??? (fwd)
Message-ID:  <199512172053.NAA09975@phaeton.artisoft.com>
In-Reply-To: <199512171817.UAA16727@plentium.clinet.fi> from "Heikki Suonsivu" at Dec 17, 95 08:17:13 pm

next in thread | previous in thread | raw e-mail | index | archive | help
>    Perhaps we need a 2.1-stable, 2.2-performance, and a 2.3-boom :-)  2.3-boom
>    could be the new wild west :-)  I know there is a disk space issue.
> 
> The problem is that people are committing broken code into source tree
> without testing it.  There should never be need for more than two threads,
> one which is being stabilized and one development.  Both of them should be
> workable when checked out.  As CVS makes doing separate testing trees easy,
> there should be no reason for untested code ever be committed?

Sorry, but this is inhernet in the CVS mechanism itself.

I have suggested before a reader/writer lock mechanism.  If the checkout
for general availability obeyed a reader/writer lock, it would be
impossible to get a checkout during an active checkin.

If committers built before rleasing the writer lock, the tree could
never be in a state where it could not be built.

There are situation where this isn't completely possible because of
lack of equipment (I thought this was what thud was for?), but in the
general case, it's really nothing more than a tree usage protocol
problem.


					Terry Lambert
					terry@lambert.org
---
Any opinions in this posting are my own and not those of my present
or previous employers.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199512172053.NAA09975>