jonathan at jonmasters.org
Wed Mar 3 08:03:20 UTC 2010
On Tue, 2010-03-02 at 21:07 -0500, Seth Vidal wrote:
> On Tue, 2 Mar 2010, Jesse Keating wrote:
> > Ok... removing deprecated uses is a questionable at best update, but
> > here is the kicker. The perl in F11 is perl-5.10.0-82.fc11. So these
> > functions aren't actually deprecated in F11. So... why is this update
> > going out? What possible benefit does the user get from this? Does
> > anybody see this as a reasonable update to publish on F11?
My own personal opinion is that stable updates should only fix serious
issues, or security problems. Fedora has such a short lifetime as it is,
I really can't see the value in pushing features to F11 when it will die
soon. I think it's far better to leave the churn in rawhide.
> the suggestion I had made at fudcon went something like this:
> 1. all packages being put in as updates would need to be marked as per
> the type of update. the default is 'trivial'. Options might include: new
> pkg, trivial, feature, bugfix, security
> 2. We would issue security updates whenever they happened. Issue bugfix
> updates once every 2 weeks. Everything else once a month.
Far better and more predictable. Even better would be to explicitly call
out the security updates in to a separate repo feed like $other distros.
The packages are the same (not a separate buildroot - I realize there
are non-trivial dependency issues) in my utopia, but they're easily
distinguished from non-security related features.
As it is, I agree with various blog postings by people here over the
last few days. I very rarely update my (non world facing) Fedora systems
these days unless I know I can reboot and have time to fix things. I
have rawhide systems for rawhide but I know if they break I can just fix
them later because they're not needed to get other stuff done and I can
always use another VM, or whatever. The point is, one expects rawhide to
"break", but one does not expect stable to break.
This isn't $Enterprise_Linux, it doesn't come with a guarantee and does
expect to be a moving target, but that doesn't mean there can't be a
predictable update cycle and a reasonable expectation that updates are
necessary and won't break systems.
More information about the devel