| sean finney on 18 Aug 2004 20:41:03 -0000 |
|
On Wed, Aug 18, 2004 at 02:38:30PM -0400, Geoff Rivell wrote: > On Wednesday 18 August 2004 01:27 pm, Jeff Abrahamson wrote: > > > Testing is a nice compromise between stable and unstable. I've had a > > few problems, but they tend to get fixed. > > Testing is ok right now because its near stable release. But at one point you > couldn't get GNOME installed, and things like that. Also, no security > updates for testing (which they will start up soon now thats its near > stable). the quality is certainly much higher right now because of the RSN release of sarge, but i've found that typically if you already have a running system, you have to go out of your way to break it. i don't know what the story is with gnome, but istr a year or so back a situation with kde where for a couple months it was uninstallable. i wouldn't have noticed on my machines apart from the fact that i had a bunch of packages held back (since it couldn't meet the dependencies, it wouldn't upgrade them), but i had some friends who got burned pretty badly by dist-upgrading and losing kde. wrt security, testing is actually the worst place to be, because you don't have the excellent infrastructure that stable provides and at the same time have to wait for updates to trickle in from unstable (which can sometimes get the fixes before stable). typically, this means an extra 24 hours or so of waiting for an update, but if there are any problems with the package or its dependencies in sid, it can be much longer. this is why i usually recommend either a stable/testing mix (where a small number of packages are selectively pulled in from testing, and the remaining packages are still tracking security), or a testing/unstable mix (where packages are pulled in from unstable on an as-needed basis). sean Attachment:
signature.asc
|
|