Jesse Keating wrote:
Which vision is that? The one where we should produce a generally
usable stable operating system every 6 months, one that users can
confidently use throughout the life of that release? Making sure our
updates given to users are better tested seems to be working toward that
vision, not away from it.
I was debating on whether I needed to clarify my last sentence and it
looks like I need to. I really hoped I didn't have to quote you, of all
people, Fedora's Foundation[1]. Friends, Features, and First being the
most important ones here.
Fedora Rawhide/Fedora N+1
Yes, these need rigorous testing and QA. Policies and tests are being
set in place that will make a better and brighter Fedora future.
However, it seems that the same release-level criteria are eroding down
to stable updates.
Fedora N
Testing is nice, but mandatory on all updates? Before you point at me
claiming I wish to destroy Fedora with bad updates, lets re-hash
everything that has been hashed out the past few months. If someone puts
out a "bad" update, they should be allowed to correct their mistake as
soon as possible - karma or not. Time, nor testing, can catch
everything. This must be accepted.
Fedora N-1
It seems there is an unwritten rule that this release level only
receives bug-fixes only, especially right before EOL. Testing of these
packages is extremely minimal because most QA folk have moved on to N or
even N+1. Are bug fixes going to be missing after EOL due to inadequate
testing to meet the stringent requirements you wish to set?
Fedora's contributors, myself included, are not at a level where we can
provide simultaneous testing of all Fedora releases (including rawhide)
for all packages. Requiring all Fedora releases to sit under one QA
doctrine is asinine. You want suggestions? There need to be multi-level
policies. No two Fedora versions are the same, so why should updates all
be subject to the same policies?
[1]
http://fedoraproject.org/wiki/Foundations