On Tue, 2014-01-28 at 19:04 +0100, drago01 wrote:
> Second, give people what they *do* care about: choices of
> above the base level, and a layer of separation so that there isn't a big
> impact when the base layer changes. To quote someone I talked to: No
> distribution does that well, so if you can, you'd really have something
> valuable to me.
This is again "hand wavy"(sorry for overusing this term). I can
already have multiple language stacks
for instance python, java, ruby and php on fedora (or pretty much any
other distribution) just fine today.
And I don't expect it to break when the "base layer" (whatever that
means ... kernel? glibc? systemd?) changes.
So in that case I didn't even get the problem itself so I cannot
comment much on the solution.
From my perspective it isn't about the language stacks, its about
stacks within the languages. Fedora has a 18 month window. My servers
run for years. I don't have time to update the entire server every 18
months. It just isn't feasible. (I'm also not fully virtualized and
automated with ansible and friends yet). So given that I use CentOS
since I know it will remain stable. The downside is that php is never
updated on it. It has a really really old php 5.3, I want 5.4 but so
many things are compiled against the 5.3 so its a huge amount of work.
I'm SUPER intrigued by Fedora.next if it is trying to solve this issue.
I can have a base OS, but pick the PHP 5.4 stack or 5.5 stack and/or
upgrade *just* that part on a live server. I'd be super happy. I could
upgrade my php stack but leave the os running happily. Less work.
Eventually I'd expect bringing up a cloud instance with ansible
provisioning and all that would also allow that.