Re: [9fans] broken updates

Re: [9fans] broken updates

[Russ Cox]

> 	I updated from the bell replica today and,
> as has happened several times before, the successive
> build fails with errors. Most of the time its a
> 'name not declared', etc. It would be nice to have
> replica branches (e.g. CVS -current -stable) so I
> can still track the "bleeding edge", yet, maintain
> a stable environment. Any ideas on this?

some day i will set things up so that the
whole world gets built every night, just
to catch these things.

as for the difference between current
and stable, i'd say sources is somewhere
in between.  it's not quite like current
because we don't put unfinished or
terribly buggy programs out there -- we
work on them internally and then push them.
it's not quite like stable because the code
that's out there has only been tested by
us internally, not necessarily by the world
at large.  i'd call it "mostly stable".

having two trees is just more work for us.
we have too much as it is.

Re: [9fans] broken updates

[Fco.J.Ballesteros]

The stable releases are the ones in the CD images from the updates
page, IIRC.

If you pull often from replica (as I do), you get what you
call `the current brach'.

[9fans] broken updates

[north_]

Heyas,
	I updated from the bell replica today and,
as has happened several times before, the successive
build fails with errors. Most of the time its a
'name not declared', etc. It would be nice to have
replica branches (e.g. CVS -current -stable) so I
can still track the "bleeding edge", yet, maintain
a stable environment. Any ideas on this?
Don
p.s. I went to a lecture, recently, on the analysis
of TCP based on fractal patterns. The speaker was
from AT&T Labs. I am wondering if any of that
research was initiated before the split with Lucent
(I havent had time to read the .ps...), and, if so,
is there any theoretical application for generating
new calculations and improvements to streamline
cross-network communications?