OT: Tape less Backup System

[Fairlight]

Is it just me, or did Brian K. White say:
> 
> http://chrootssh.sourceforge.net/

Thanks.

> Lot's of unofficial patches take a long time to get accepted but not because 
> there was anything wrong with them.

Yeah...they just take forever to get approved by the people rolling them
in.  -Adaptec's- story is that it took forever to get their i2o driver into
the linux kernel because the kernel maintainers wouldn't roll it in.  When
they told me that, I said, "Really?!  Well give me your name and number and
I'll give it to Alan and have him call you--we're on IRC together this very
moment."  (And we were.)  Alan is far more accessible than Linus, IMHO.  

> devfs is a pretty extreme example, since not only is it accepted, it's the 
> stock standard on at least linux and freebsd.

I don't like devfs for various reasons.  I really don't like devices that
pop in and out of existance, for one.

> ibcs was a patch forever and built-in the last couple years.

iBCS had stability issues with building for most of its
lifetime--especially after the a.out->ELF switchover and libc5's induction
into service.

If it's stability or security related, I don't like unofficial patches.  I
just plain don't like unofficial patches in the first place.  There are
development teams out there that know their software a hell of a lot better
than Red Hat does, for one--and yet RH insists on back-porting whatever
fixes went into later full releases into their own backward versions.  They
have a perl in RHEL3 that -to this day- (this makes it about what, seven
months now?) I can segfault at will.  Perl is not supposed to segv--ever.
And I have it narrowed down to one of three functions:  return(), int(), or
sort().  And I can tell you it's got to be sort() without even seeing their
source.  They just released a new version about 3 weeks ago.  I tested it.
It segv's yet, despite their being aware of the problem all this time (and
supposedly before, according to their own tech folks).  I have severe
issues with RHEL3, from what I've seen.  I'd never recommend it even for
non-production, much less real production.  I've seen too many problems
with it.  And they bring a lot of it on themselves with lousy design
decisions, and WAY too much in-house back-porting of unofficial patches
they dream up themselves based on what the problem is, rather than just
releasing a new version of the package as the original developer intended
it to be released.  As if the people that work with it day in, day out
don't know it better...  Right.  .

Hmm.  </rant>  Sorry.  Been ill since Friday night, lost most of my
Saturday, and I'm cranky.  Didn't mean to blow off so much steam.

But really, I do have many reasons for not liking unofficial patches.  No,
they're not always evil.  Sometimes they're the only way to get the job
done.  But if it comes down to security--well, I'd rather something like
this come from the openssh team themselves rather than some other team.
The openssh team is a known quantity.  Tacking into their work seems like
asking for trouble.

mark->
-- 
Bring the web-enabling power of OneGate to -your- filePro applications today!

Try the live filePro-based, OneGate-enabled demo at the following URL:
               http://www2.onnik.com/~fairlite/flfssindex.html