Mar 24, 2009

~arch will be testing again not stable

I think the idea of a meta distro on top of a stable base is the way to go, let people choose just what they want on bleeding edge. I like my kernel, gcc, and perl solid, sometimes I want firefox's beta's but that doesn't mean I want the rest of the system like that. At some point gentoo's ~arch became more stable that stable, this needs to end. Even if it means making stable a bit more unstable. I'd like to see more people running stable systems.

~arch is going to be aimed more at things that are stable enough, and pending stable. It's going to become more unstable than it is in a sense because I'm not afraid to hardmask stuff that was ~arch and is found to be broken. I'm also not afraid to release slightly broken stuff into ~arch, for example I'll release a new git into ~arch without an update noperl patch which means using -perl won't work. I won't release completely broken or overtly buggy things into ~arch though, I'm just not afraid to release with a missing patch (set) as long as it'll build and work in most cases and update later.

I've decided on a policy for stabilizing the kernel, first of regen2's kernel is vanilla-sources, if you're running another sources this doesn't apply. I will stabilize the previous kernel when the new major version is released so I will stabilize 2.6.28.x when 2.6.29 comes out, after that I will stabilize each new kernel a few days after it hits the tree (actually I may not wait), since the kernel's development model makes it highly unlikely that regressions will be introduced that late in the cycle. I may use a similar update scheme for other packages but it will be a package by package basis. But likely once a major version moves to stable it will be trivial to move new minor versions to stable.

I want to know what people use? what should be stable in regen2. 2 things I won't stabilize right now are qt-4.5 and kde-4.2. They work well enough but have too many bugs upstream, imho (should a kde developer look at this feel free to look at my buglist on your bug tracker, prtscrn must work before I'll mark kde stable, but hey at least dpms works with screensaver now). Anything else goes what do you use/need that's in ~arch and has been there for a long time. feel free to mention in comments, the dev list or the Bug tracker

Mar 23, 2009

Regen2 0.9.1.0 soon

Last night I decided I'd be rolling out the first release of Regen2 in the first week of April.

So as I'm sure everyone knows version are arbitrary they mean nothing... well ours are going to mean something. The first 2 numbers are the year, so 20(0.9).1.0. the Third number is the quarter, we are going to be mass stabilizing system once per quarter. I'm not sure that this will include new profiles. 4rth number is kinda arbitrary at this point, I'm not sure how/if I want to use it, opinions welcome. since releases will be 'stable' I could use 0 as kind of an alpha, first release. 1 as beta, 2 as rc, and 3 as final. I could also do a 0-alpha1, 0-beta1, etc. and when stage4's and other stuff are rolled out use the version number to denote updated non-system software. In other words I could do releases each month of say kde in stage4's but I wouldn't change system more than once a quarter.

So far my list of packages that'll be marked stable include.

gcc-4.3.3(maybe r1)
glibc-2.9_20081201-r2
openrc-0.4.3-r1
baselayout-2.0.0-r2
git-1.6.2.1
perl-5.10.0

I'll have to spend the next few days figuring out the rest.

I'm debating also on whether I'm locking down major or minor versions. meaning I might consider stabilizing gcc-4.3.4 in the middle of this quarter but not gcc-4.4.

I'm open to opinions on anything. This may be a dictatorship, but I still don't mind hearing from people not me, as with enough eyes all bugs are shallow.

Mar 18, 2009

postgresql initial setup (users and databases)

This guide assumes that you've managed to install and start the postgres server. I'm ignoring these because they are well documented in several places, probably including your distribution, however, what to do next isn't.

first you need to connect to the postgres database with psql as postgres, on gentoo* distributions you must be in the postgres group to do this. In the event this doesn't work you may have to su to the postgres unix user account somehow and run psql from there.

psql postgres postgres

you now have connected to what is essentially the 'root' user for postgres. as with all 'super user' accounts be careful and know what you are doing. also, as with all 'super user' accounts you don't normally want to run as them, so we need to create another user. I recommend creating the user with the same name as whatever shell account you may be connecting to it from (if applicable) because then you don't have to specify the username when typing psql.

CREATE USER xenoterracide WITH PASSWORD 'YeahLikeIdTellYou';

next you need to give the user a database. like the user account if you make it the same name as the shell account you won't have to pass it to psql.

CREATE DATABASE xenoterracide WITH OWNER xenoterracide;

now you're done with the setting up your users account so quit psql by tying \q at the prompt. now you can simply log in from your users shell by typing psql and SQL-ing away.

Mar 12, 2009

Tree hasn't been updated - why

So 2 days ago, when I went to sync to gentoo I didn't get any updates. I'm not really sure why as the process seemed to go alright. It could have something to do with the next bit.

My System appeared to have been compromised. Syslog-ng was disabled and all logs post dec 30 2008 were removed. So i've spent the better part of yesterday/today reinstalling my system. Tree should be synced and merged tonight.

Mar 7, 2009

SYNC_METHOD

I'm going to be pushing out an new feature to portage today, since testing has gone well.

SYNC_METHOD is a new portage variable, I've also gone to the trouble of defining it in make.globals. The default for Regen2. is git, however, rsync and cvs have also been defined although not tested.

the main benefit of SYNC_METHOD is to use one way of detecting the program you should use to detect the tree. previous variations were problematic, as rsync does not have a directory like .git and .git supports the rsync:// protocol.

I've added these defaults to make.globals

# Default sync method
SYNC_METHOD="git"

# Default sync mirror
SYNC="git://github.com/regen2/portage.git"

# default git method specific settings
PORTAGE_GIT_METHOD="checkout"
PORTAGE_GIT_REMOTE="origin/regen2.org"

you may of course override them with make.conf.

changes will officially be in portage-2.2_rc23-r3.

Mar 4, 2009

to Gentoo dev or not to Gentoo dev

So after a little post about Regen2 on the Gentoo-dev mailing list. I've been offered mentoring to become an official gentoo dev. I'm conflicted, my main reason for saying yes would be to learn stuff that I can't learn on my own. But I'm not sure I really want to be a gentoo-dev this means doing these other time consuming things, and I wouldn't be able to do nearly enough from the inside. I still have to proceed with regen2... as I don't believe gentoo (in spirit) can be saved from within. So should I work to become an offical gentoo dev?