/srv/irclogs.ubuntu.com/2016/05/17/#ubuntu-motu.txt

persiaMaybe write a test case script?02:09
karstensragefor a PAM module?02:15
karstensragei mean what do backporters do for things like python?02:50
karstensrageor just anything ?02:50
persiaTypically more tighlyintegrated or deeper stack components are not backported: I do not mean this as discouragement, only apology foor the process.04:13
karstensrageits a really tough problem i totally acknowledge htat04:17
karstensrageif i could think of a test case i would04:17
karstensragebut pam modules are tricky business04:18
karstensragethey take over login type things, ssh, sudo. the source has a little app that you can use but you have to set it up04:18
Arshidhi05:29
=== tinoco is now known as tinocoff
karstensragefantastic14:36
karstensragethe admin for backporters no longer works with ubuntu14:37
karstensrageit strikes me here that this process is perhaps a bit broken and maybe it needs to be discussed at a higher level, the concern here is that no one is really doing it, there is no defined process for how someone trying to get a backport done is supposed to do it, there is no defined process for what a backporter is supposed to do and everyone is too busy anyways14:39
karstensrageis there a way to up-level this discussion, there is severe dissonance in xenial pulling from unstable and whatever backports is14:45
rbasakkarstensrage: there are four admins apparently: https://launchpad.net/~ubuntu-backporters/+members14:48
rbasakI believe at least one of them is still active.14:48
karstensragerbasak, although i will if i have to, throwing darts at that list has not really been fruitful14:49
rbasakIf there are delays, bring it up on the ubuntu-devel ML maybe? Perhaps more people on the backporters team are needed.14:49
karstensrageto be honest from the outside backporters seems dead14:50
rbasakSo...help reawaken it? :)14:50
karstensrageok seems like a very thin mailing list and im a bit frustrated at this point14:53
karstensragenot sure ranting will come off well14:53
karstensragei honestly dont know how to frame this correctly for the ubuntu process itself14:53
karstensrage"hi ubuntu-devel, backports sucks, please fix, k thanks bye"14:54
karstensrageso i guess there is a fundamental problem that i cant figure out how to fix, and thats that you cant expect a  backporter (or really anyone; same with debian maintainers) to switch in some random context of whatever is being backported and "evaluate" things15:00
karstensrageso either you come up with some "process" which is either severely anemic to handle every case or just arbitrary superficial stuff15:02
karstensrageor you "fiddle dee dee, ill think about that tomorrow" the problem and get what you have today15:03
karstensrageif the "process" for upcoming LTS is just "whatever unstable did" i see that as valid for backports as well, its justifiable15:05
karstensragethis is like the reverse bike shed problem15:06
karstensrageyoure trying to come up with a process to evaluate nuclear power plants15:06
rbasakThat's a reasonable point.15:06
karstensragethank you, rbasak the problem is if you dont address this, no one will volunteer to evaluate nuclear power plants... if you dont define some doable process, you cant expect backporters to do anything15:07
karstensragewhich is exactly what you have15:07
persiakarstensrage: So, the backports process has been created by the backporters: most things are self-organising, and waiting on some other thing tends not to work.15:11
persiareawakening sleeping teams is more likely to be successful than suggesting someone else do that.15:12
rbasakkarstensrage: the process is well documented, no? https://wiki.ubuntu.com/UbuntuBackports15:13
rbasakSpecifically https://wiki.ubuntu.com/UbuntuBackports#Approving_a_Backport15:13
karstensragerbasak, but those are all just "statements"?15:18
rbasakI don't follow.15:27
karstensragerbasak, just say "the package installs correctly."15:37
karstensrageno one is going to check that, or even knows what it means15:41
karstensrageright?15:41
=== tinocoff is now known as tinoco
karstensrageis that right rbasak ?17:53
rbasakkarstensrage: I would take "correctly" to mean "without errors". For each binary package, not just one.18:23
karstensrageyes of course18:25
karstensrageive already tested all that stuff including that it works but i know a lot about PAM18:25
karstensrageon all environments18:25
karstensragethe ppa's have been in launchpad forever... i just recently updated the ppa's with the latest that actually got into debian and xenial, and tested them AGAIN18:28
karstensragebut its all irrelevant to a backporter18:28
karstensrageif they existed18:28
karstensragepersia, suggested a test script.. im wracking my brain trying to figure out if thats even possible18:29
karstensragebut it would be a huge effort and with no commitment from someone that could actually backport stuff, its hard to justify that18:29
persiaMy idea was that it would be easier to get testers that way, but yes, it ultimately needs a backporter.18:30
karstensragepersia, the issue with a "test" if it could even be done (still thinking about it) is that who is it for. if its for me, ive already actually installed the ppa's and the xenial stuff18:32
karstensrageand configured it and it works18:32
karstensrageif its for some random backporter with no context of PAM or what it entails, the test just says "YES"18:32
karstensrageso its the same as me saying "YES"18:32
karstensrageof course the devil is in the details but the problem is working out those details is what the backporter actually wants to do, but it cant scale18:33
* karstensrage sighs18:34
persiaYep :/19:06
karstensrageso i unassigned micahg and ill try the mailing list to ask for someone to help21:16
karstensrageim willing to provide any technical assistance to a backporter to understand PAM, configure it21:16
karstensragebut given the ease with which xenial was just "done" im hesitant without some commitment on the other side21:17

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!