[00:00] Blah, smartmeter installation [00:00] No power for an hour [00:01] you have solar? [00:01] No [00:01] i only got my smart meter because i got solar [00:01] They're replacing them nation-wide eventually. And state-wide now [00:29] wgrant is going to feel like he's had an arm cut off with no interwebs [00:36] * StevenK peers at Jenkins [00:43] wgrant: StevenK: +7/-4 https://code.launchpad.net/~wallyworld/launchpad/edit-sharing-policies-1040336/+merge/120894 [00:43] You can't keep but add lines :-P [00:44] i know :-( [00:44] little choice here though [00:47] wallyworld: r=me, with one thing to keep in mind. [00:47] * wallyworld looks [00:47] thanks [00:48] StevenK: i looked at the other choice source popups which have descriptions (there are only a few on bugtask), and they are all ok [01:17] wgrant: BAH, tharwted by your factory.makeBug(target=) rather than product= [01:31] StevenK: Oh? [01:34] wgrant: Muscle prompted makeBug(product=product) [01:34] Muscle memory, rather [01:53] * wallyworld needs to go buy coffee. !caffeine === bad [01:54] see you in 5 :) [02:46] Hmmm, wasn't test_getAllPermissions_constant_query_count failing about a week ago? [03:24] StevenK: Yes [03:42] wgrant: It failed in my destruction of security_contact with 19 != 20 [03:44] Hmm [03:47] Hmm, it may not be failing locally. [03:48] wgrant: I can paste the query log from the ec2 failure if you wish. [03:56] Probably spurious [03:56] Land :) [03:59] wgrant: Hmmm, I may have introduced an unintended change. [03:59] I think that the bug supervisor will be subscribed to private bugs. [03:59] Er, PRIVATESECURITY bugs [03:59] That would be bad [04:02] Right, the bug supervisor is subscribed to PRIVATESECURITY bugs if the pillar has private_bugs. [04:02] Still bad? [04:04] If it's a regression, yes. [04:04] But it might not be [04:06] wgrant: In createBug it used to be if information_type in SECURITY_RELATED, grab security_contact, else if pillar.private_bugs, grab the bug_supervisor [04:07] Hum hum [04:08] I'd restore the old behaviour [04:08] Subscribe the supervisor if pillar.private_bugs and it's not in SECURITY_RELATED [04:08] RIght [04:18] wgrant: xx-upstream-bug-privacy.txt is also dumb. I had to change it from name12 to name16 because name12 was no longer subscribed === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away [05:03] wallyworld, wgrant: https://code.launchpad.net/~stevenk/launchpad/sanity-for-rafa/+merge/120912 [05:06] StevenK: i think it looks ok. i think we want a business exception rather than a ValueError though [05:07] actually, the logic may be a bit out [05:08] ah, no, ok [05:11] wallyworld: Happy to do so, I just couldn't think of a name, so went for ValueError [05:11] yeah, i'm not sure if we have one already [05:12] raise SharingViolation("wgrant is very displeased with you.") [05:12] it's more a SharingIntegrityError perhaps? [05:14] wgrant: ^ [05:16] I think an assertionerror is probably correct [05:16] It's not a valueerror or typeerror [05:16] something is just broken [05:16] It certainly doesn't deserve its own Exception subclass [05:17] It'll hopefully never happen [05:17] Oh, then I'll just assert [05:17] No [05:17] Do not just assert [05:17] Raise AssertionError [05:17] asserts are skipped sometimes, eg if running optimised [05:17] Which we don't but still [05:17] We prefer to raise AssertionError explicitly for important sanity checks, I believe [05:26] StevenK: i should have asked this up front sorry - could we include the info type in the message? not sure if we want to list the pillar names also? [05:27] Hmmm [05:28] Just trying to figure out how to phrase it [05:28] raise AssertionError("oh shit") :) [05:28] raise AssertionError("All pillars require access policies.") [05:28] It's more for us so we can look into the problem easier [05:28] Which is what I have now. [05:29] Right [05:29] So a list of pillars and APs might be useful [05:29] These pillars (ccc, xxx, ccc) required access policies for information type xxxx but do not have one [05:29] s/required/require [05:29] or something [05:29] Well, I don't know which one is missing, I just know one is. [05:30] you can figure out which one(s) [05:30] the returned polices have pillars [05:30] and compare that to the pillars parameter [05:45] wallyworld: http://pastebin.ubuntu.com/1162099/ === almaisan-away is now known as al-maisan [05:47] StevenK: look good. perhaps ", ".join(... [05:47] wallyworld: http://pastebin.ubuntu.com/1162104/ [05:47] StevenK: great, thanks for the changes. +1 from me [05:56] wallyworld: Pushing it up now, since I got distracted. [05:56] that's ok, i already +1 [05:56] Yeah, I just saw, thanks. === al-maisan is now known as almaisan-away [07:44] good morning === almaisan-away is now known as al-maisan [08:29] StevenK: I thought I'd testfixed test_getAllPermissions_constant_query_count. [08:29] StevenK: Try sticking an extra list(self.main_archive.getAllPermissions()) before the record_two_runs call and see if that tickles it into appearing locally. [08:31] cjwatson: So far, I'm ignoring it hoping that buildbot doesn't get infected. [08:33] The previous failure depended on test ordering. [08:33] I'll be away for the next week, so your problem to fix it if it breaks. :-P [08:34] heh [08:37] cjwatson: random non LP question. when I plug my laptop into charge, it does start to charge, however the screen brightness doesnt change, if I reboot machine it picks this up, any ideas ? [08:38] No idea. Not a kernel guy [08:38] cjwatson: you're also not a LP guy either but know a fair deal, so just thought I'd chance my arm :) [08:38] The kernel is a total blind spot for me [08:52] czajkowski: check to see if the acpi events are picked up in scripts in /etc/acpi/ [08:52] night all [08:52] bigjools: nn [08:52] will do === al-maisan is now known as almaisan-away === almaisan-away is now known as al-maisan === Guest81255 is now known as jpds === jcsackett changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: jcsackett | Firefighting: - | Critical bugs: 4.0*10^2 === al-maisan is now known as almaisan-away === almaisan-away is now known as al-maisan [13:58] sorry wallyworld. I see you have discovered that anything that adds a reference to the Person table require code to manage merges [14:00] deryck: abentley adeuring I forgot to ask int he stand up, any JS issues? Thinking of flipping 3.5.1 for ~launchpad? [14:00] rick_h_: I haven't noticed any issues. [14:00] rick_h_: i haven't noticed any [14:00] rick_h_none that I've found [14:01] ok thanks [14:01] deryck: any reason to hold off on increasing the range on the FF then? [14:01] I think we've been on it 1.5wk [14:02] rick_h_no, I think it's safe. flacoste, any concerns about turning lp on to yuk 3.5 for everyone on launchpad? [14:02] yui rather [14:03] deryck: JFDI. We can always change the flag if we are proven wrong [14:04] sinzui, ah, good advice. Thanks :) [14:04] rick_h_^^ [14:04] * sinzui wishes how could say the same about sharing flags [14:04] ok, will do [14:14] jcsackett: once again fate pulls the chair out from under sharing [14:15] sinzui: oh no. [14:15] jcsackett: Lp still does not permit structural subscriptions on ubuntu. It is not possible for the security team to subscribe to private security in ubuntu [14:17] sinzui: can we rectify that? what is/was the reason for no structsubs on ubuntu? [14:17] I have hacked a url: https://bugs.launchpad.net/ubuntu/+subscriptions [14:20] sinzui: so it's possible, we just don't have the UI in place? [14:21] jcsackett: I do not have permission to compete the change [14:21] sinzui: ah, yes. i just did that same experiment. [14:21] So the question, is, If a webops did that, can he select the ubuntu-security, and can he save [14:23] Well, I will try this in a dev instance to verify if the security contact role role can create its own subscription [14:23] then I will try admin [14:23] then I will ponder the value of going on immediate vacation [14:27] * jcsackett laughs [14:27] bugger. Dev permits anyone to subscribe to ubuntu [14:27] Ubuntu is configured different;y in production [14:27] oh goody. [14:35] deryck: +1 to what sinzui said [14:38] gary_poster: benji: gmb: bac: can any of you remember what controls/blocks structural subscriptions on Ubuntu? Alas our own stories in dev show it can be done. I think a config option is preventing me (well I want ubuntu-security team) from creating a subscription to all private security bugs in Ubuntu === al-maisan is now known as almaisan-away [14:40] sinzui, I don't but maybe if I tried to duplicate what you want to do it would ring a bell. [14:42] gary_poster: there is not link to subscribe to bug on ubuntu. not on the front page or on the bugs page. I can hack the url https://bugs.launchpad.net/ubuntu/+subscriptions and I can say what I want, but I get permission denied on save [14:42] http://people.canonical.com/~curtis/configure-ubuntu-security-notifications.png [14:43] * benji hears bells far, far in the distance. [14:45] sinzui, I duped. ISTR something hard coded about subscribing to all ubuntu bugs. I don't think we touched it, and I could be making it up. IOW, I'm not much help. [14:45] gary_poster: Indeed that is what I thought...but our dev instance lets me do it. I think I will try API or SQL to prevent ubuntu-security from getting a regression [14:46] sinzui: Looks like you need to be in the bug supervisor [14:46] if IDistribution.providedBy(self) and self.bug_supervisor is not None: [14:46] if subscriber is None or subscribed_by is None: [14:46] return False [14:46] elif (subscriber != self.bug_supervisor [14:46] and not subscriber.inTeam(self.bug_supervisor) [14:46] and not subscribed_by.inTeam(admins)): [14:46] ah, thank you [14:46] return False [14:46] Oh, not just you, but the subscriber [14:46] I suspect an admin will have to do it through the API [14:47] Except ubuntu-security is in ~ubuntu-bugcontrol [14:47] fab [14:47] So an ~ubuntu-security admin can probably do it [14:47] They can do it. [14:47] I will show them my picture of what I recommend [14:49] sinzui: so we can pick the chair pack up for sharing? [15:15] jcsackett: anyone with a clear head actually. wallyworld set a pre-emptive testfix to rollback a bad revision in db-devel. But since buildbot was running, it was ignored. Do I want to force a build to land a trivial testfix change to make the buildbot restart [15:17] sinzui: i think so, yes. i mean, right now db-devel buildbot isn't running, right? [15:17] not running [15:29] ^ jcsackett does forcing a build pul the latest revision? if not I want to land a trivial testfix change to poke buildbot [15:29] sinzui: i think you're right. forcing doesn't pull, it just restarts the known build. [15:29] okay. I will find a trivial text change to act as the test fix. [15:43] flacoste: Do you approve enabling writable sharing for everyone in production? https://pastebin.canonical.com/72918/ [15:43] sinzui: of course! [15:43] This is just a formality [15:58] flacoste: sharing is one for everyone. I am making change to /launchpad now to repeat my qastaging activities to very all is well [15:59] \o/ [16:03] ha, yes, I can finally get rid of all the deactivated user and structural subscription spies on ~launchpad's projects [16:04] sinzui: congrats [16:05] wtf? lifeless baby alarm go off? [16:06] I think I could celebrate by writing a api script that will share with a team, then unshare with the team members to cleanup any project's shares [16:09] sinzui: lifeless' alarm baby went off [16:09] Its like a baby alarm but organic. [16:13] :0 [16:13] :) [19:44] What is the 'findReferencedOOPS' API call on projects about? === almaisan-away is now known as al-maisan === al-maisan is now known as almaisan-away [21:09] cody-somerville: it finds reference OOPS. [21:09] cody-somerville: OOPS that are not referenced are gcable by oops-tools. [21:10] if we had a generic fast RE engine across all project content, that method could go. [21:13] lifeless, What does it mean for an OOPS to be 'referenced' in this context though? I didn't think Launchpad itself stored information about OOPs? [21:14] If the text OOPS-123456 etc is in a text field in an asset relevant to the project. [21:15] ah, ok. So if someone mentions an OOPS in, for example, a bug report? [21:16] yes [21:17] oops.c.c sees I dunno, 40/50K per day [21:17] we toss most [21:17] ones folk put in bugs we keep [21:17] we keep them long enough for short term datamining [21:17] just long term we toss [21:17] after (IIRC) 5 days [21:18] Ok, Cool. [21:42] does anyone know if there is a team representing former canonical staff? [21:45] I have written a script that shares a project with a team, the revokes the direct access for the team members so that it is easy convert from subscriptions to shares [22:07] sinzui: https://code.launchpad.net/~jcsackett/launchpad/kill-lazr-1/+merge/121079 === jcsackett changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: - | Firefighting: - | Critical bugs: 4.0*10^2 [22:12] sinzui: there is [22:14] http://lmgtfy.com/?q=launchpad+not+canonical :P === Ursinha` is now known as Ursinha [22:21] thank you lifeless [22:21] * mwhudson learns from that about a couple of people who've left... [22:36] not-canonical includes people who never were