Re: Slink to Potato
Hi Brad; unless Mutt is confused, you wrote:
> Hmmm... exactly 80-column lines, more or less. 72 or 76 is much better
> though, it leaves room for replies.
Ooops, sorry, I don't know how that happened; my vimrc files specs 76 columns,
maybe I need separate command in muttrc?
> I'm not sure what you mean here... There is a pretty good bit on the
> website, if you look in the right places. And if you're referring to the
> Perl changes, that was discussed and announced on -devel, which all
> developers are supposed to read. IIRC there's also some in the developer's
> section of the website.
Yes, I stand corrected. There is a bit of info if you look under release info.
My fault...
> I'd think they shouldn't put stuff on the webpage until they've made the
> decision, on debian-devel or debian-policy.
Considering the changes are so big (with egcs and libc2.1 and 2.2 kernel) that
it justifies the more flexible approach (not waiting until _all_ the details
are settled). And if upstream guys do their thing, we may be looking into 2.4
kernel pretty soon- does that mean another Debian stable release will be one
step back ( when potato become stable it'll rely on 2.2 kernel). Even though
the debian releases and kernel are not directly linked (I'm running 2.2.10 on
slink), it will give a wrong perception to the average user (like me).
> Still, there's the risk of major breakage. What do you count as a
> non-essential package? Gnome, which has 10,000 libraries and such that
> need to be properly managed?
> Remember that stable isn't just a collection of packages that work,
> everything works together as a unified system. If you start upgrading
> parts of that, you may end up breaking another part.
Maybe we should have another directory then for up-to-date-stable, which all
could download from at their own risk (which we do anyway, not like anyone is
guaranteeing anything in the first place). People who really want rock solid
system ( I like mine medium solid :-)) wouldn't upgrade anyways. I upgrade
something (say Enlightenment or wmaker), and if I don't like it, I pull out
old packages and reinstall old stuff. Thanks to Debian way of installing, I
have never ever damaged my system by doing this (until I did something
terribly stupid and deserved it). At worst, I have lost a bit of my time ( and
if I don't have the time in the first place, I don't play with upgrades).
And people can use real stable for fresh install or reinstall.
> Besides having to deal with possible breakage, what is it that makes
> stable better for you than unstable? Or is the possible breakage reason
> enough (it is a good enough reason)?
I am no computer wizard. And when I read about all the development currently
going on in Linux world, people are forgetting that semi-commercial
applications ( like StarOffice and netscape) still rely on libc6 and not new
libc. I happen to need those apps. But, spoiled and shallow as I am, I like my
wmaker to be at par with upstream, or my gimp or my tetex. None of those are
close to it in stable. Marcelo (wmaker maintaner) has been kind enough to post
his slink-based binaries for wmaker 0.60.0, but that is not the case with
0.61.0 any longer ( since he needs his time to be devoted to unstable branch,
along his regular life, I presume). My point is, I don't care if I d/l new,
say, gimp, and it craps out on me, and I have to go back to stable. I do care
if I attempt to upgrade to new libc6 and it fails, I get useless box that I
have to reinstall from scratch. Or if my system can't run netscape or
StarOffice.
> Absolutely! All apt does is download the packages and call dpkg to install
> them. For the next generation of Debian package managment, dpkg will be
> just another front-end to the underlying library, but rest assured it will
> still exist. It's way too useful to lose!
glad to hear that. Thank you.
> Personally, i've never used the --compile flag, since whenever i download
> the source i have need to modify something ;)
That is what I meant. Glad I'm not the only one.
> No problem.
Again, thanks for a good discussion points.
damir
Reply to: