On Fri, Nov 11, 2005 at 01:01:49AM +0200, Joona Kiiski wrote: > > Certainly not. If you want unstable packages, then use *unstable*. If > > you want to help test the next Debian release, then use *testing*. If > > you want something that will always work, then use *stable*. > > Yes, I've tried them all. > * Unstable was a bit too unstable for my taste. > * Stable is fine, but I don't really enjoy using only old software. > Often there comes new interesting software in testing, which really is > "stable enough" for me and installing it in stable is hard > (download+check dependencies+compile+install) and could easily lead to > bad problems (library incompatibilities etc.). > * So that's why my choice is and will be testing. 98% of the time it > fits my bill perfectly. And sometimes (I hope) I can file an useful > bug report which can help the development of debian. It's just sad > that rarely testing gets 'broken' as badly as it's now, but if it > can't be avoided then it can't be avoided and that's it. I can live > with it: just postpone 'dist-upgrade' long enough or change to > unstable for a while. > This is like the old saying: "Fast, Secure, Cheap. Pick two."
The reality is that "stable enough for you" may still not be stable enough to enter testing. I have written a howto [0] an customizing packages. You may want to take a look at it. Personally, I run stable since I don't have time to waste with my machines not working. However, like you, there are certain programs I would like to use. In those cases, I backport the packages back to Sarge and place them in my personal repository. You may want to give that a shot. > My purpose was not to complain. I have some coding experience (not > much but some) and I can only imagine what kind of mess so large > library transition could be with all those new bugs and problems > hiding around. I really appriciate that great job you volunteer for, > really. > > I just wanted to share my view to this subject (and I'm quite sure > there are hundreds of other testing users who think similar way as I > do), but as I wrote earlier in this message if this subject has > already been dicided by pros and 'breaking' testing once in a while > cannot be avoided, then let it be that way. > > Thank you for your answers, -Roberto [0] http://familiasanchez.net/~roberto/howtos/debcustomize -- Roberto C. Sanchez http://familiasanchez.net/~roberto
pgpjYDqm2BNt1.pgp
Description: PGP signature
