testing vs unstable?

sweiss · 2005-07-05 07:49:19

I don't quite understand the difference between the testing repo and the unstable one, can anyone clarify?

CyberTron · 2005-07-05 08:03:45

if i recall correctly, testing is extremly testing of packages (highly unstable) while unstable are packages that are going into extra/current quite soon...

(I use unstable, but not testing)

dtw · 2005-07-05 08:25:35

nope, AFAIK it's the other way around:

[testing] is for new versions of pkgs already in [current], [extra] and new pkgs that are scheduled for those - latest kernels are in there now as is the latest xorg pkg, which I am using

[unstable] is for pkgs that are not in [extra] or [current] e.g fluxbox-devel

sometimes things get confused by, say, OOo2 for example. The reason it is in [unstable] and not [testing] is because in its current form it is a beta and will never go into [current] or [extra] in that state - when there is an official release it will probably it hit [testing] for a short time before going to [extra]

how's that?

tomk · 2005-07-05 10:02:41

So let's take as an example the recent problems with udev 060-1. If I had been using the testing repo, would I have seen these problems earlier, and been able to flag them to the devs before it went into current? If that's the case, it seems that the people who are using testing are not doing that.

Is there a set of guidelines somewhere on the proper use of testing?

dtw · 2005-07-05 11:53:12

not everything goes in testing (ahem)

tomk · 2005-07-05 12:05:47

Alright - that makes some recent events clearer. It also raises the question - why not?

LB06 · 2005-07-05 13:14:31

Because Arch wants to be bleeding edge. Testing is only done with packages that are more or less expected to cause some problems afaik. udev obviously wasn't one of those.

phrakture · 2005-07-05 14:35:25

tomk wrote:

Alright - that makes some recent events clearer. It also raises the question - why not?

It happens... hell, in my work experience I've logged on to production databases and added/removed columns before - no testing or even development changes...
Some things seem real minor and work 90% of the time when tested, but will hit corner cases every now and then.

Seeing as there are alot of posts about this sort of thing recently, maybe someone needs to propose a better way to test. As it is right now, not many people use the testing repo anyway.

Dusty · 2005-07-05 14:40:38

Somebody really needs to update the devland paper. I wonder if I will or will not have time for that. Hmmm...

Dusty

phrakture · 2005-07-05 14:42:17

Dusty wrote:

Somebody really needs to update the devland paper. I wonder if I will or will not have time for that. Hmmm...
Dusty

Dusty, Official Documentor !!

Dusty · 2005-07-05 15:14:50

phrakture wrote:

Dusty wrote:
Somebody really needs to update the devland paper. I wonder if I will or will not have time for that. Hmmm...
Dusty
Dusty, Official Documentor !!

Been there, done that. ;-)

Gullible Jones · 2005-07-06 00:15:52

Testing contains packages that are being tested and will be transferred to Extra or Current when proven not to trash you system. Using the testing repo on a system that must be stable is definitely not recommended.

Unstable contains packages of software that is still under development. Some of it (e.g. fluxbox-devel) is very stable; other stuff (e.g. waimea-cvs) is a bit buggy.

dtw · 2005-07-06 08:50:10

sorry gullible, was my explanation not good enough for you?

if only we could keep a "I obviously haven't read the thread before i posted score"

Gullible Jones · 2005-07-06 16:20:12

*bonk**bonk**bonk*

Sorry 'bout that.

Dusty · 2005-07-06 16:27:46

Dusty wrote:

Somebody really needs to update the devland paper. I wonder if I will or will not have time for that. Hmmm...

I'm working on it... any ideas are welcome... here's my first draft:

http://www.buchuki.com/misc/devland.html

Mostly I just added the section on AUR so far, and removed some of the sections no longer relevant. I've barely begun.

Dusty

Arch Linux

#1 2005-07-05 07:49:19

testing vs unstable?

#2 2005-07-05 08:03:45

Re: testing vs unstable?

#3 2005-07-05 08:25:35

Re: testing vs unstable?

#4 2005-07-05 10:02:41

Re: testing vs unstable?

#5 2005-07-05 11:53:12

Re: testing vs unstable?

#6 2005-07-05 12:05:47

Re: testing vs unstable?

#7 2005-07-05 13:14:31

Re: testing vs unstable?

#8 2005-07-05 14:35:25

Re: testing vs unstable?

#9 2005-07-05 14:40:38

Re: testing vs unstable?

#10 2005-07-05 14:42:17

Re: testing vs unstable?

#11 2005-07-05 15:14:50

Re: testing vs unstable?

#12 2005-07-06 00:15:52

Re: testing vs unstable?

#13 2005-07-06 08:50:10

Re: testing vs unstable?

#14 2005-07-06 16:20:12

Re: testing vs unstable?

#15 2005-07-06 16:27:46

Re: testing vs unstable?

Board footer