On 04/16/2015 at 05:34 PM, Patrick Wiseman wrote: > On Thu, Apr 16, 2015 at 5:25 PM, David Wright > <deb...@lionunicorn.co.uk> wrote: > >> Quoting Joe (j...@jretrading.com):
>>> However long the wait, the result will be the same. In fact, the >>> longer the wait, the more upgrades there will be in one go. >> >> This may be true, but there's a difference. If you wait a few >> months in jessie before moving to stretch, a lot more people will >> have tried the latter and discussed, maybe fixed, the bugs that >> crop up. >> >>> Once the floodgates open into the new Testing, there will be a >>> similar upheaval here in Sid, as software which has been kept >>> back because it won't be compatible with the initial new Testing >>> (which has to be smoothly upgraded from the present Testing) will >>> then be dumped into Sid and pushed into Testing as soon as the >>> magic two weeks have passed without complete disaster. >>> Interesting times everywhere, except hopefully in the new >>> Stable. >> >> Lisi and I both suggested to stick with codenames, and jessie. This >> means that the OP can choose exactly when to move distribution >> instead of being at the mercy of the Debian release schedule. >> >>> This upgrade of old Testing to new Testing is the first >>> real-world preview of the *next* Stable upgrade, or at least some >>> of it. I'm sure a lot of thought has gone into preparing for it. >> >> Sure, but we're not all gagging for it. > > I've had these lines in my sources.list for years: > > deb http://ftp.us.debian.org/debian/ testing main contrib non-free > deb-src http://ftp.us.debian.org/debian/ testing main contrib non-free > > deb http://security.debian.org/ testing/updates main contrib non-free > deb-src http://security.debian.org/ testing/updates main contrib non-free > > Yes, things get interesting after the current testing stabilizes, > with sometimes hundreds of updates at a time, but things don't get > into the testing distribution unless they mostly work, and I've very > rarely had to wait long for any wrinkles to work out. (I do NOT, of > course, do this on a production server, but only on my personal > boxes.) Same here - with minor syntax tweaks, but the point is that I have sources.list pointing to 'testing' by name. (And of course, yes, a production server stays with stable or even oldstable until manual testing has confirmed that something newer is ready to use for production.) Prior to that, I actually did the same thing with sid, again for years straight. _That_ was a mistake; it ended up with my primary system in a state which, while it did mostly work, I can only describe as "broken" from a design perspective. With testing, however, I haven't had major issues - and the ones I have had have mostly been of the sort where the fix is to install a slightly newer package version, either from sid or from experimental, which is needed for my config but hasn't made it into testing yet. I do generally wait a week or three after a new stable release before dist-upgrading, and I _always_ review the apt-listchanges report for any dist-upgrade before giving the go-ahead to proceed with it, but that's about the limit of how far I go to avoid new-testing chaos. -- The Wanderer The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man. -- George Bernard Shaw
signature.asc
Description: OpenPGP digital signature
