Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 11 Feb 1998 10:36:07 -0700
From:      Warner Losh <imp@village.org>
To:        Adrian Chadd <adrian@obiwan.creative.net.au>
Cc:        Jonathan Lemon <jlemon@americantv.com>, current@FreeBSD.ORG
Subject:   Re: Hollywood (Re: PATCH.M ) 
Message-ID:  <199802111736.KAA22807@harmony.village.org>
In-Reply-To: Your message of "Thu, 12 Feb 1998 00:55:15 %2B0800." <Pine.BSF.3.95q.980212005051.1744B-100000@obiwan.creative.net.au> 
References:  <Pine.BSF.3.95q.980212005051.1744B-100000@obiwan.creative.net.au>  

next in thread | previous in thread | raw e-mail | index | archive | help
In message <Pine.BSF.3.95q.980212005051.1744B-100000@obiwan.creative.net.au> Adrian Chadd writes:
: I assume Terry wouldn't be putting forward patches that didn't work. And
: hadn't been throughly tested.

For some parts of the kernel there are additional requirements to amke
sure that we're all haded in the right direction.  Terry's patches are
very interesting, but they are in an area that is hard to explain, as
well as hard to understand.  This presenets a barrier to having the
patches included.  I don't know the specifics of this, so I'll not
speculate further on the interactions at work here.

: If I mention VM code would someone kill me? :-) (Not taking a personal
: attack at anyone here, so please don't read it as such.. :)
: 
: Its not like stuff has been changed in -current before that hasn't
: horrendously broken things in a big way and take ages to fix.

My view of current is that things that should be in a future release
of FreeBSD go there.  If there are nit, bugs, crashses in the work
that is committed, so be it.  So long as those folks are comitted to
fixing it or backing it out in a timely basis, I don't care if things
are busted for a day or two here or there.

The VM code I think is a special case.  It is so complicated that it
has to have lots of testers as soon as possible to shake out the
problems.  Also, the person making the changes has a stellar
reputation for taking responsibility for making things work and has
done some very impressive things in the FreeBSD world for a long
time.

If I were to start making such extensive changes to the VM tromorrow,
I'd likely have multiple people ask me to prove myself (or at least my
code) before it went in (and stayed in).  However, no one seems to
blink much when I commit buffer overflow fixes (even in large numbers
of files) because I've done that many times.  Sometimes I have the
pointy hat thrust upon me for a commit that I've made, but in those
cases I've always backed out the changes until I could fix the issue
at hand.

-current is for experimental things, but not too radically
experimental.  -current is the place for new ideas and code to mature,
but isn't a dumping ground for all new code.  There is a balance that
needs to be maintained.  This balance isn't always 100% technical.
Sometimes there are polical reasons for doing things.  lpr/lpd and
sendmail come mind as programs that are known to be dangerous, but are
still the best solution to the problems available given all
considerations.

Anyway, I'm not a core team member.  These are just my views from
interacting with the CVS tree for a long time and some private release
engineerig philosophy that I've evolved in the jobs I've had over the
years.

Warner


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe current" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199802111736.KAA22807>