Josh Metzler wrote:
On Tuesday 09 August 2005 06:56 pm, Erik Steffl wrote:
well, the fixes take forever to get to testing
That is because they need to go through testing and bug fixes in unstable.
well, what does it matter? The bugs take forevr to fix so testing is
not really usable...
so while testing seems like a good idea in general it doesn't seem to
be very appealing in its current incarnation... (I started to use
testing but gave up)
What you are asking for is to have experimental be what unstable is now, and
unstable be what testing is now. If the gcc transition were going on in
no I don't.
experimental right now, rather than unstable, unstable would go without
updates to gcc, X, etc. etc. for months, just as testing is right now.
which is not what I was pointining out in my mini-rant and I already
explicitly stated that I don't see how to avoid problems during big
upgrades like that. As long as it's announced and visible it's OK (after
all that IS what unstable is for).
I think if you got what you are asking for, you would switch to using
experimental.
Like you, I enjoy using and testing the latest versions of everything, so I
also run unstable on my desktop. But to do that, you need to be willing to
suffer the occasional breakage and pay close attention to everything that
is going on with Debian.
repeat: note that I am not complaining about c++ abi changes etc.,
that can't be avoided, I think. I was specifically talking about
_unneccessary_ problems that get weeks to fix for unknown reasons like
jackd bug #318098 (perhaps there is a good reason for it, don't know).
what I was suggesting is doing experiments in experimental, put
release candidates into unstable, which is what seems to be current
policy anyway.
I also wish the "it's unstable, deal with it" was banned because it's
used as a blank excuse for number of otherwise inexcusable problems that
can be easily avoided. That's why I'm trying to persuade people that
unstable is the only viable alternative for number of desktop users
(well,so far I have only proven that that number is at least one:-)
erik
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]