| slangasek | RAOF: because it's a binary sync, there's no room for SRU bugs to be added; so it's out-of-process | 00:32 |
|---|---|---|
| RAOF | slangasek: Fair enough. I guess you shepherd it through? | 00:32 |
| slangasek | I can, yes | 00:33 |
| RAOF | I was more thinking the post-acceptance shepherding, but that's good too :) | 00:37 |
| slangasek | :) | 00:45 |
| RAOF | slangasek: Oh, thinking of binary syncs - trusty-unapproved has a unity-control-center sync from a landing PPA. I went to review it, but as far as I can tell the source is inaccessible to me (the landing PPA has had all its contents wiped many times over since that sync was proposed). | 01:03 |
| RAOF | I presume that launchpad has the relevant bits refcounted somewhere; do you know how to access them to do a review? | 01:04 |
| slangasek | RAOF: if I'm not mistaken, you can do a 'queue fetch' to grab the source? | 01:04 |
| cjwatson | I'm not even sure it's processable any more in that case | 01:04 |
| slangasek | you just can't get a diff | 01:04 |
| cjwatson | Copies aren't properly refcounted, which is part of the slightly-amorphous "Soyuz queue redesign" item on the LP stakeholders agenda | 01:04 |
| cjwatson | But you can try "queue fetch"; if that doesn't work the sync is probably dead | 01:05 |
| cjwatson | And somebody should poke the CI Train folks and/or the silo owner for allowing that silo to be cleaned in advance of the sync being processed | 01:05 |
| slangasek | robru: ^^ the silo process should already provide guidance regarding this case, I think? | 01:08 |
| infinity | Yeah, they're aware of this. They often prod directly for reviews when silo space is getting low. | 01:09 |
| infinity | So someone cleared that one by accident, I'd guess. | 01:09 |
| * RAOF learns about the queue tool | 01:10 | |
| robru | slangasek: what are we talking about? An sru? If the silo isn't present it was likely freed or abandoned... | 01:10 |
| cjwatson | I can't check the citrain logs for who did it, since I think it was before the last redeployment | 01:10 |
| RAOF | So, queue only seems to be able to grab the _source.changes; is it expected to grab the full source package? | 01:10 |
| slangasek | robru: yes, an SRU; silos aren't meant to be freed before the SRU is accepted, I believe? | 01:10 |
| cjwatson | RAOF: Yes, but it's not going to be able to here. The only thing to do is to reject it. | 01:10 |
| infinity | https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-001/+packages?field.name_filter=unity-control-center&field.status_filter=&field.series_filter= | 01:10 |
| infinity | It might still be there, however. | 01:10 |
| cjwatson | Hm, really? | 01:11 |
| * cjwatson looks | 01:11 | |
| cjwatson | That's just a historical record, no still-present files attached. | 01:11 |
| infinity | Nope, it's dead. | 01:11 |
| infinity | Yeah. | 01:11 |
| robru | slangasek: RAOF: yeah the silo should not have been freed. If the silo is gone if say just skip it, is the landers responsibility to make a new silo | 01:11 |
| slangasek | robru: I mean: it's important that the silo not be freed for an in-process SRU until the SRU is accepted; and I believe this is the understood process already | 01:11 |
| * slangasek nods | 01:11 | |
| infinity | RAOF: Reject it, it's all you can do. | 01:11 |
| RAOF | Yeah, I could find the build logs, but they had no active links. | 01:11 |
| RAOF | infinity: Already done. | 01:11 |
| RAOF | Also, yay for learning about the queue command. | 01:11 |
| RAOF | That's going to make processing SRU syncs much nicer! | 01:12 |
| cjwatson | As far as Soyuz internals go, the unapproved copy is basically not a lot more than an archive, series, package name, and version | 01:12 |
| robru | RAOF: what silo was this? It's been over a month since i freed an SRU due to being low on silos | 01:12 |
| cjwatson | robru: ubuntu/landing-001 | 01:12 |
| RAOF | robru: silo 001 | 01:12 |
| slangasek | having queue diffs for syncs would be nicer yet, but yes, 'queue' does in a pinch :) | 01:12 |
| cjwatson | And this is basically the same reason that proper queue diffs for copies are currently a hard problem | 01:13 |
| robru | Hmmmmmmm yeah i think it was me that freed that one, strange you guys are just noticing now. | 01:13 |
| RAOF | slangasek: Well, my previous method was to go to the relevant PPA and fiddle around to wget the links. :) | 01:13 |
| infinity | queue diffs will come with soyuz redesign. Feel free to put more pressure on people outside Foundations to understand the importance of that. :P | 01:13 |
| cjwatson | LP would have to basically process the first half of the copy to track down the publications involved | 01:13 |
| infinity | RAOF: wget? Not even dget? Heathen. | 01:13 |
| RAOF | infinity: Last time I tried it, dget fails from PPAs because of annoying url reasons? | 01:14 |
| infinity | RAOF: Also, depending on your bandwidth, you still might want to navigate-and-dget/wget, since "queue fetch" will grab the *entire* copy, including binaries. | 01:14 |
| cjwatson | queue fetch --source | 01:14 |
| cjwatson | HTH | 01:14 |
| RAOF | Learning all round! | 01:14 |
| infinity | Oh, yeah, there are some LP pages that fail miserably at dget. | 01:14 |
| robru | slangasek: i freed that one back when we only had 30 silos and there was a huge crunch. It'll need to be redone unfortunately. | 01:14 |
| robru | Whoever was on charge of that did a poor job of following through with it, i freed it because it sat on a silo for 2 months | 01:15 |
| cjwatson | infinity: The thing is that it's almost impossible to communicate to people without doing the "write up requirements properly" part of the redesign. I hear that you and William were making some progress on that ... | 01:15 |
| infinity | cjwatson: We were making progress on scheduling a time to make progress on it... | 01:15 |
| infinity | cjwatson: We might need to form a committee to help find a time. | 01:16 |
| cjwatson | Ha. Pls to be doing it soon so that I stop having to indirectly fend off annoyed stakeholders who think it's too high up our agenda :P | 01:16 |
| infinity | cjwatson: I was kinda hoping you and little Willy would flesh it out without much need for input from me, since you've been on both sides of the fence and know what sucks, and he knows how it needs to be unsucked. | 01:17 |
| cjwatson | I'm happy to try to help, but every time I ask I hear you have the baton :) | 01:18 |
| cjwatson | So if it's OK to have a go at it on the principle that the best way to get answers on the internet is to produce something incorrect, then sure | 01:19 |
| infinity | cjwatson: Hah. I trust you not to be Lennart. | 01:24 |
| slangasek | robru: right; so not that it's a common occurence, but when those silos get freed it's probably a good idea to follow up with the sru team and have us remove it from the queue at the same time, or else nag us into processing it from the queue so it /doesn't/ have to get freed prematurely | 01:27 |
| robru | slangasek: yeah fair point. I try to never do that but we were really crunched at the time and there was a big push to get some landings in for whatever milestone was happening at the time | 01:36 |
| RAOF | Hey super release friends! Does anyone have the policies for what we can upload to landing PPAs destined for vivid-overlay? | 05:49 |
| RAOF | Specifically - the Mir landing PPA needs a new glmark2 with some changes required to build against the new Mir. Can we be lazy and just upload the wily glmark2 (which has those changes, plus some others), or do we need to do the sensible thing and add a minimal patch to update it? | 05:51 |
| RAOF | For bonus points - where is this documented? ☺ | 05:51 |
| infinity | RAOF: The Ubuntu Release Team has zero influence over the phone overlay PPA. | 07:01 |
| RAOF | Grr | 07:02 |
| infinity | RAOF: Arguably, the whole point of it is to circumvent things that ubuntu-release and ubuntu-sru policy would forbid. :P | 07:02 |
| * RAOF thinks you can drop the “Arguably” :P | 07:02 | |
| infinity | RAOF: Anyhow, looking at how people have been doing vivid-overlay/wily dual landings, I'd assume that "just use the wily version with a backport version number and a rebuild" would probably be fine, if the changes aren't drastic. | 07:04 |
| RAOF | Eh, sure. Why not. Substantially less than half of it changed :) | 07:09 |
| rbasak | How do I go about getting an answer from the release team about nginx? And is this something that is within the mandate of the relese team or does it need to go to the TB? | 09:07 |
| === pete-woods1 is now known as pete-woods | ||
| === alex-abreu|off is now known as alex-abreu | ||
| === kickinz1|afk is now known as kickinz1 | ||
| mdeslaur | could python-django be migrated please, the two test failures were failing before | 14:59 |
| kickinz1 | infinity: any eta for docker.io sru? | 14:59 |
| infinity | kickinz1: I'm out until Monday. If I get bored, I might look, but otherwise, you'll want another SRU person, or to bug me next week. | 15:00 |
| kickinz1 | infinity, OK, sorry. | 15:01 |
| === pgraner is now known as pgraner-lunch | ||
| balloons | infinity, do we have an official EOL date for utopic? I haven't seen it, and it should be soon yes? | 16:08 |
| Odd_Bloke | balloons: FYI, "16:00:35 < infinity> kickinz1: I'm out until Monday." | 16:23 |
| balloons | Odd_Bloke, ty | 16:34 |
| Riddell | anyone know what's stopping ktp-* from transitioning? it all installs fine for me | 16:39 |
| === pgraner-lunch is now known as pgraner | ||
| Laney | Riddell: At least ktp-call-ui krfb (http://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt search for "Trying easy from autohinter: telepathy-logger) | 17:09 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!