On Mon, Jul 04, 2016 at 06:48:32PM +0200, Kamil Cholewiński wrote: > > I think we're confusing two different concepts here... Latest stable > release (with most recent security patches) vs following bleeding edge. > > Former is almost always what you want in production. There might be only > one exception: known new bug in the patch. If such a thing is common, I > guess that's really a problem with the process, not with the packaging > infrastructure...
There are always problems with every process. That's why there are patches/bugs/missing features. That's why there are programmers, mops, brooms, etc. Don't want to rebuild your production web/business servers? Look up Mtier in the mailing list. Nice helpful infrastructure. > > Bleeding edge has its place too. Even reasonable people do it in > non-critical environments, to check out changes before they're surprised > in the next release, or give feedback to the OS developers / packaging > team. > > > That's the large problem with the current devops/container culture. > > If you're unhappy with this culture, do something to affect it in a > positive way. Devs at my $WORK know not to curl | sudo bash, because > we've had several chit-chats about what's OK and what's not, and why. > Not to sound like I'm kissing ass, but you really should look at the incredible work Marc Espie has done for OpenBSD. His work and dedication deserves respect. You should look at how things have improved tremendously since OpenBSD forked from NetBSD. The mailing lists go way back. I've read some of the older stuff. Impressive history. OpenBSD has it's own, distinct culture. I like it. Chris Bennett
