| === Ursinha is now known as Ursinha-afk | ||
| pitti | Good morning | 03:45 |
|---|---|---|
| === chihchun_afk is now known as chihchun | ||
| === jpds_ is now known as jpds | ||
| === jpds is now known as Guest68298 | ||
| === Guest68298 is now known as jpds_ | ||
| pitti | jibel: I'm looking again into setting $http_proxy and friends in the autopkgtest VMs | 06:56 |
| pitti | jibel: this feels like a déjà vu, I think we tried this like three times; remind me, why did we switch that off again? or did it never work? | 06:57 |
| jibel | pitti, IIRC it works almost fine but makes some tests behave weirdly. | 06:59 |
| pitti | jibel: didrocks wanted it for ubuntu-developer-tools-center, and some others like gem2deb need it; but I can't remember any more which other tests fail with a proxy | 07:01 |
| pitti | jibel: perhaps I'll enable it again, send a message to devel-announce@, and we'll watch out for bugs? | 07:01 |
| pitti | jibel: just trying to remember whether it broke something generic, like apt-get | 07:02 |
| pitti | but I'll set no_proxy=localhost,ubuntu.com,launchpad.net | 07:02 |
| pitti | (which we didn't last time) | 07:02 |
| jibel | pitti, aptdaemon failed for sure, there was also a problem with gnupg | 07:02 |
| pitti | jibel: ah thanks, I'll try those then | 07:02 |
| jibel | pitti, well, not autopkgtest of gnupg but the call to gnupg | 07:03 |
| pitti | in adt-run? | 07:03 |
| pitti | we don't need that in the DC | 07:04 |
| pitti | (it now uses a pre-generated key; we didn't in the past) | 07:04 |
| pitti | jibel: but anyway, it would now only use the proxy in the VM, not outside where adt-run is | 07:04 |
| pitti | that was more complicated in the old world where everything ran in the VM | 07:05 |
| pitti | I think I'll just give it another shot | 07:05 |
| jibel | pitti, agreed, it's better to re-enable it and investigate tests that are failing with a proxy | 07:06 |
| pitti | jibel: is http://d-jenkins.ubuntu-ci:8080/view/Utopic/view/AutoPkgTest/job/utopic-adt-setup-testbed/configure generated from some branch/template file, or should I just edit it? (need to add proxy vars) | 07:25 |
| jibel | pitti, it's updated manually | 07:26 |
| pitti | jibel: ack, thanks; I do a local test on alderamin with my patch to autopkgtest, check gem2deb and aptdaemon, then deploy this | 07:26 |
| jibel | pitti, we should probably save it in bzr just in case jenkins explodes | 07:27 |
| pitti | jibel: jenkins errors? unheard of! | 07:27 |
| * pitti crawls into a corner to weep | 07:27 | |
| pitti | jibel: ah, while watching the utopic-adt-setup-testbed testbed and seeing that wazn takes three times as long as the other three: what again does it take to wean us off the old prepare-testbed stuff? | 07:57 |
| pitti | jibel: i. e. what do the old VMs have that the new VMs are missing? | 08:10 |
| pitti | jibel: lxc, sbuild, and aptdaemon succeed after proxy changes | 08:33 |
| jibel | pitti, I don't think new VMs are missing anything. I didn't test DKMS tests with newer VMs that's all we miss to remove prepare-testbed | 08:52 |
| pitti | jibel: ah, splendid; please let me know if there's anything missing; I'd rather reduce the number of ways we are doing things | 08:53 |
| pitti | jibel: (not urgent, of course; RTM first) | 08:53 |
| jibel | pitti, ideally I'll port DKMS tests to autopkgtest, I might be able to dynamically generate a debian/test/ and run the tests against that. | 08:54 |
| === jpds_ is now known as jpds | ||
| === _salem is now known as salem_ | ||
| === Ursinha-afk is now known as Ursinha | ||
| pitti | jibel: so update-manager was fallout from the proxy change; I fixed the test and uploaded a new version | 12:50 |
| jibel | pitti, thanks! | 12:51 |
| === chihchun is now known as chihchun_afk | ||
| === Ursinha is now known as Ursinha-afk | ||
| === Ursinha-afk is now known as Ursinha | ||
| === roadmr is now known as roadmr_afk | ||
| === roadmr_afk is now known as roadmr | ||
| === salem_ is now known as _salem | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!