Re: RC severity for Python 2.6 related bugs

[Yaroslav Halchenko <debian@onerussian.com>]

On Sun, 28 Feb 2010, Sandro Tosi wrote:
> I absolutely agree with this (even though, for those packages that
> byte-compile the files they install, it's a smaller problem) and I
> fear there are several situations where there are hidden bugs only
> discovered with (long) *usage* of a system with 2.6 as default:
> waiting to do the switch, doesn't help to release a better squeeze,
> only a worst and buggier one.
Well, to catch those we have an extended army of beta-testers:
users of recent Ubuntu releases with 2.6 as default -- for me they
managed to hit 2.6 issues with fail2ban so upstream could resolve them
promptly year(s) before 2.6 to become the default in Debian ;)

So, it might be worth for python package maintainers to have a peek at
their packages in Ubuntu to see if they got something fixed without
reporting upstream/

> Additionally, as a side note, unstable is "unstable" by definition:
> its users knows it, and if something breaks in it, it will either be
> fixed or not in stable, so "break users apps" problem is less
> appealing (even though it exists).
although I totally agree with description of 'unstable', in reality, the
path from unstable into stable, closer to freeze it is, is shorter --
just 10 days to migrate from unstable into testing.  For the packages
with low users count it becomes more important since it becomes unlikely
that an important bug would get detected that rapidly.... So....
lets switch to 2.6 now so we have more time!

-- 
                                  .-.
=------------------------------   /v\  ----------------------------=
Keep in touch                    // \\     (yoh@|www.)onerussian.com
Yaroslav Halchenko              /(   )\               ICQ#: 60653192
                   Linux User    ^^-^^    [175555]