[conspire] wedged Debian Sid
rick at linuxmafia.com
Mon Oct 29 14:47:32 PDT 2007
Hello again, Bill. Just some general remarks.
Quoting Bill Moseley (moseley at hank.org):
> After a few years following Sid I've finally got a problem on one of
> my machines.
For background, and mostly for those accustomed to other Linux
distributions, Debian has, at any given time, a daily-progressively-changing
development branch called "unstable" (aka sid) and a closely related one
called "testing" (currently lenny). The unstable branch is defined by
the rolling contents of its directories on the Internet package
collections for unstable, and sprouts new packages immediately upon
their upload by package maintainers. The testing branch is those same
contents filtered through an automated quarantining filter, that does a
first-pass quality check.
Consequently, those two branches (unstable and testing) are very nearly
the exact same thing, and their packages are (at any given time) very
compatible with each other, such that the term "testing/unstable" is
Debian _also_ has a non-rolling branch that lags waaaaay behind testing
and unstable, called "stable" (currently "etch"). It is "non-rolling"
in the sense that it has deliberately unchanging contents over several
years, having only security and other critical fixes rolled out to it
from the package mirrors. At the end of its life, it gets retired,
the (then-) "testing" branch gets moved into place as the new "stable"
branch, and a new "testing" branch gets spun up.
Consequent upon the vast distance in package version numbers between
"stable" (at any given time) and the simultaneously extant
testing/unstable branches, it is problematic to fetch "stable"
packages into testing/unstable. It is likewise generally problematic
to fetch testing/unstable packages into stable.
(I expect I'll probably wait to see your comments, before saying more
along these lines.)
More information about the conspire