[05:18] hmm are debian installer images at the first bootslpash supposed to revert to normal mode if you change the mode and then move on to the menu and select check disk for defects then go back over to install on server or alternate installers? === chihchun is now known as chihchun_afk [09:48] fgimenez, hello, good morning. [09:48] hi rhuddie good morning [09:49] fgimenez, I noticed that you have removed the MultipleSignalReactor from your branch [09:50] rhuddie, yes, sorry are you using it? i can put it back [09:50] rhuddie, I used that in my wait-for-download branch, so wondered if I should just move it onto that branch? [09:50] either way is fine. [09:51] If you weren't using it on your branch, maybe it makes more sense to include it on my one? [09:51] rhuddie, in the apply update branch it doesn't make much sense, there's only one signal expected [09:51] fgimenez, ok. I'll add it to mine, using the base class that you have defined. [09:52] rhuddie, ok, thanks, i've also introduced some changes suggested by veebers in the test_services file, have you seen them? [09:53] fgimenez, no, not yet. I'll look through the changes [10:13] fgimenez, it seems to me that there is not much difference between the single and multi signal reactor, just how the signals get returned by the run function [10:14] fgimenez, so, how about just returning self.signals? then let the calling code check how many records there are etc? [10:14] that way you would only need 1 class for single and multi signals [10:15] rhuddie, that's right, barry pointed out that when more than one signal may arise you cannot only return the first one [10:15] fgimenez, for the download case, there are multiple signals, and in the reactor I just return the full list [10:15] then the calling code can check through them, or just look at the last one (which is what I actually do) [10:17] rhuddie, if it works for you go ahead with it, i tried to dig in si code and was not sure of how the signals array was actually filled with more than one signal [10:18] fgimenez, it's just an array which gets a new record added for every callback it gets. you can see what I have here: https://code.launchpad.net/~canonical-platform-qa/ubuntu-ota-tests/wait-for-download/+merge/252759 [10:19] fgimenez, so I think we could probably do away with the single and multi array classes, just make BaseReactor run() return self.signals [10:20] fgimenez, let me try that out [10:20] rhuddie, of course :) [10:23] fgimenez, yes that works fine for my multisignal case [10:23] rhuddie, ok perfect then :) [10:24] the only thing, imo given the asynchronous nature of the process we shouldn't assume any order when multiple signals may come [10:28] fgimenez, great, yes. Well, in that case the reactor is just passing the list of received signals back. the caller could then check the order if necessary? Or you could add a derived run() method to check the order? [10:29] rhuddie, yes the caller should check, maybe the happy-path signal is not even in the array [10:31] fgimenez, I'll add my comments then to your mp about returning self.signals from base class run () and removing SingleSignalReactor [10:33] rhuddie, ok thanks, probably barry can tell us if there may be any corner case not covered by this approach === chihchun_afk is now known as chihchun [11:24] rhuddie, i'm not sure if calling self.quit in the _do_ methods of the reactor is right in the multiple signal case [11:24] rhuddie, see http://bazaar.launchpad.net/~ubuntu-system-image/ubuntu-system-image/client/view/head:/systemimage/reactor.py#L98 [11:26] rhuddie, ok, i've seen that you removed it from _do_UpdateProgress === chihchun is now known as chihchun_afk [11:36] fgimenez, yes, I just left it in the case where the operation had either failed or succeeded === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [14:45] ubuntu-qa, autopilot3 appears to be broke on vivid atm. I'm assuming an update to vivid broke it, and autopilot2 remains unaffected. [14:45] Hmm [14:45] balloons, i'm vanguard today i believe [14:45] balloons, do elaborate [14:46] brendand_, I'll get you a log and I'm digging in myself atm [14:46] brendand_, http://paste.ubuntu.com/10610073/ [14:46] I've tried calendar and clock and both fail this way.. [14:47] I'm surprised the backend code is really that different, but we'll see [14:47] balloons, on device or desktop? [14:47] brendand_, desktop [14:47] balloons, okay can you put your exact repro steps as well [14:51] brendand_, sure, since this isn't a known thing. But more or less try and run any autopilot3 test on the desktop. They all seem to fail trying to instantiate a backend [14:52] balloons, did you install an -autopilot package or run from source? [14:53] brendand_, here's a full reproduction: http://paste.ubuntu.com/10610117/ [14:53] balloons, dpkg -l | grep autopilot would be good too [14:53] brendand_, let me add that [14:54] balloons, although it does list the version in your log, but anyway [14:54] right [14:54] it's sad the vivid testrunner hasn't been running recently, could have pinpointed when this broke [14:57] http://paste.ubuntu.com/10610145/ [14:59] balloons, dunno about you but trying to run that just crashed my session [14:59] brendand, heh, fun times [14:59] I'd like to try on device now [15:00] balloons, afaik everything is fine on the device, folks have been running sanity tests today [15:00] right, I would assume so, and it's backend stuff anyway [15:03] brendand, ohh the python-autopilot package wasn't updated, so that's why it still works. Makes sense now as the backend code is the same [15:03] balloons, is that from looking at dpkg -l | grep autopilot? [15:05] brendand, yes and looking at the source it's different. A new version was not pushed [15:05] balloons, can you paste it? [15:06] good mmorning. [15:07] morning elopio ! elopio since I have you, I'm interested in talking QA for scopes. Who's the best person to talk to about that? I want to get some docs of the best practices on developer.u.c and I need an example scope and tests to do it [15:08] balloons, ahem, vanguard :P [15:09] brendand, ohh I get to bug you for that too? brillant [15:09] brendand, what did you want me to paste anyway? [15:09] balloons, yeah - we just want to get out of the habit of elopio being the bottleneck for all qa requests [15:09] balloons, dpkg -l | grep autopilot [15:10] brendand, right it's in the paste above [15:10] http://paste.ubuntu.com/10610145/ [15:11] balloons, ah i missed that one - session crashy remember :) [15:11] ahh right, sorry mate [15:12] * balloons suspects https://bugs.launchpad.net/autopilot/+bug/1297595 [15:12] Ubuntu bug 1297595 in Autopilot "Multi-touch gestures fail if the application is launched before the input devices are created (Qt)" [Critical,Invalid] [15:14] balloons, open a bug and veebers can have a look later today [15:15] balloons, as for the scopes thing, elopio probably does know best there (cop out i know) [15:15] brendand, I will just gathering some more data. [15:15] thanks for the scopes answer :-) I'll await Leo [15:22] balloons: hello [15:22] balloons: we need to get together with pete-woods to define those things. [15:23] balloons: do you want to set up a meeting? somebody from qa, pete, and yourself [15:23] elopio, sure can. Who might you recommend from qa? ask jfunk ? [15:24] balloons: yes, invite jfunk and he'll take somebody else if he thinks it would be good. [15:24] thanks. [15:24] elopio, ty yt [15:24] dkessel: https://github.com/magne4000/quassel-webserver [15:24] this looks promising for using quassel on the phone. [15:25] brendand: balloons: I can't reproduce the autopilot issue on my updated vivid. [15:25] I see a comment on my branch from veebers who says he was going to ask about it to CI. [15:26] elopio, oh really? [15:28] filed https://bugs.launchpad.net/autopilot/+bug/1432700 [15:28] Ubuntu bug 1432700 in Autopilot "Autopilot3 on vivid desktop fails: "RuntimeError: Unable to instantiate any backends"" [Undecided,New] [15:41] hey balloons - quick one ... question of course :p [15:41] http://cdimage.ubuntu.com/daily-live/current/ [15:41] does that look right to you ... am I missing something [15:41] pretty slick there eh? [15:42] it's correct insomuch as the new builds all fail automated testing [15:42] aaah [15:42] slick... as in slickymaster? [15:43] I don't think slickymaster had anything to do with it.. [15:43] now knome on the other hand :-) [15:43] right - so one's you could grab from tracker as daily are ok - but as it's failing auto testing - the daily there isn't updating - is that right? [15:44] would i ever break anything? [15:45] balloons: so - that is expected from that link? just need to answer some people [15:48] elfy, -pending images don't get linked to the tracker as the are completely broken and useless for testing. However, 10 days is quite a long time to not have a new image [15:48] normally you might see a day or 2 gap in the image if something like that happens [15:48] sometimes it's as simple as making a .1, and you never get the .0 [15:49] elfy, and yes you are correct. the tracker lists the last good image to test and won't update until a new good image is produced [15:54] balloons: so thew tracker is listing today - that direct current shows the 6th [15:57] yep [15:57] ohh [15:58] ahh I see.. well that does make more sense for it not to be 10 days old, but . . . [16:11] balloons: yea - bit broken there [16:11] balloons: also - is -release the right place to raise - umm chaps there's something wrong with the dailies? [16:14] ping project-team vanguard [16:15] elfy, yes they are a good place to start [16:15] thanks balloons :) [16:16] ty! [16:16] so - back to that link - not roight? should be same as tracker? [16:17] elfy, it's a bit weird.. what I explained to you is correct, but indeed it doesn't match what we see [16:18] ok - I'll post in the +1 thread for them then :) [16:27] fgimenez: it's your turn on the vanguard, right? [16:28] elopio, i think that brendand took over, the rotation's been messed up after the friday [16:29] brendand: jfunk pinged ^ [16:29] jfunk, hey [16:30] seems like vanguard notification needs a little work [16:30] brendand: there's a bug https://launchpad.net/bugs/1427439 [16:30] Ubuntu bug 1427439 in urfkill (Ubuntu RTM) "Urfkill saved wrong WWAN state after enabling/disabling flight mode" [Critical,Confirmed] [16:30] which I think we need to add a test for [16:30] it should be simple enough [16:31] jfunk, well - to be fair it's in the channel topic - ping ubuntu-qa. if you don't think that's right suggestions are welcome, we can change it [16:32] jfunk, i suppose the flight mode sanity test ought to catch that [16:33] brendand, no, because we test only once [16:33] jibel, we enable, then disable. do we need to do that again in order to trigger it? [16:34] brendand, yeah, we'd want to do it like 100 times [16:34] jibel, oh it's a race condition (seems to be from reading the description though not 100% clear) [16:35] what we need is a regression test for the issue, at the low level. First they need to find the cause, and make a scenario to reproduce it. [16:36] elopio, well that, yes [16:36] elopio, agreed. But we can also test the other direction for issues that are not known [16:38] jibel: we can, but we need to think about the suites we will automate, make sure they are covering holes left by the low level suites. [16:38] if they fix this issue, and make a regression test for it, we will waste 100 cycles because all the runs will be green. [16:38] a long running test with common tasks that a user would do sounds nice. [16:39] but for that to work, first we need to make the random long running test to work. [16:39] elopio: I guess I am worried that this problem could get out to users and want to get something in place to make sure we catch it [16:40] jfunk: they must make a test that catches it. [16:40] the issue has nothing to do with the indicator, so that test must not be through the UI. [16:41] elopio: they can make a test that catches what they *think* causes it (as they already have) but I ran into it on my own, which proved they hadn't found all the causes of it [16:41] jfunk, automated testing is the perfect place for that, yes. whether it should be a sanity test... that depends [16:42] elopio: right, I am not suggesting the test must be AP on the UI, only that the test must run regularly as part of regular testing [16:42] jfunk: right. But in order to write a good test, we first need to have a good scenario to reproduce the issue. That's where they must spend their time first, before writing the test. [16:42] jfunk, we'd have to get the balance between frequency/number of times run and severity [16:43] jfunk, if we ran it 100 times every image that would add a lot of time to the sanity run [16:43] if we know with 90% certainty that the issue will occur if the action is repeated 100 times, that's not good enough to get good automation out of it. [16:43] we caan use that knowledge to collect more knowledge. [16:45] add logs, debug, until the problem is understood. [16:46] then it will be easy to write a single test that when run only once will catch any regression of the same problem. [16:47] jfunk, by all means add a card to the backlog so we don't forget about it [16:54] brendand, should I add a task for looking at the sanity failures? [17:23] brendand or rhuddie: this needs one more review: https://code.launchpad.net/~canonical-platform-qa/autopilot/fix1266601-Pointer-pressed-move-2/+merge/252932 [17:37] elopio, I can take a look, but it will have to wait until later as it seems quite large === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === om26er_ is now known as om26er|food === om26er|food is now known as om26er [18:32] * knome stares at om26er who changed their nick for 15 minutes to tell they were away [18:32] (O.O) [18:32] knome, heh, it was a really quick dinner [18:33] well i'm really surprised that there really isn't any other way to tell you are away than changing your nick [18:33] oh but there is, /away ! :P === roadmr is now known as roadmr_afk [18:35] roadmr_afk, et tu, brute? [18:39] /away is far less visible, and depends on the client to notice and ... grey out the nick or something... a nick change is immediate and hard not to notice (especially if tab-completing a name) [18:39] ToyKeeper, but why would you want to know if everybody is away or not? [18:40] for example, why would you care if i'm away or not if you don't have anything to discuss with me? [18:40] if you have something to discuss with me, why couldn't you see my away message? [18:41] In general, /away status tends not to mean much (especially auto-away), and /away messages tend not to get seen by anyone. [18:41] fwiw, it's even on the ubuntu IRC channel guidelines: [18:41] !away [18:41] Please do not use noisy away messages and nicks in Ubuntu channels. It is annoying and unnecessary. Use the command "/away " to set your client away silently. See also «/msg ubot5 Guidelines» [18:41] ToyKeeper, i couldn't disagree more... [18:42] But it depends on the person. I don't generally set away messages, others do. [18:42] The nick change tends to bug me because it can make logs weird or break privmsg sessions. [18:42] i don't mind if you do or not (i can probably see if you are away or not judging by the idle time) [18:43] but i do care about awaynicks because i most often do not need to have the immediate information that you are eating lunch... [18:44] I also think one should never ping without data (i.e., just ask the question, don't ping first), but I have trouble convincing others to do the same. [18:45] What I'm getting at is ... I can't change human nature, so I choose not to get upset about it. [18:45] well, what i'm getting at is that setting an awaynick is actually just as much work than an awaymessage... [18:49] Yes, but one can't fight every battle. :) [18:50] but i can win this battle ;) [18:50] knome: i'm going to go drink a gallon of water and /nick every 5 minutes to let you know i had to go use the toilet :P [18:53] knome: what you should do, is every time someone uses a nick change to show their status, pipe an !away to them [18:54] you could even write a bot to do it :) [18:54] dobey, nah, it's better to leave a disturbing, shameful, public personal message ;) [18:54] eh, shame is for the weak :P [18:55] :P [19:21] la_juyis, have you taken a look at what the help app looks like so far? would you like to? [19:22] hullo balloons, what's up? [19:23] hey knome.. Trying to get work for the week all sorted, yourself? monday is in the books for you, I'm jealous [19:23] kicking up the action again after a long weekend too [19:24] and continuing with personal projects on the evening :P [19:45] balloons: ping hey I saw your bug go past in my email. I'll be looking at that today, once I get a VM up. [19:45] I suspect a new vivid something is happening [19:48] veebers, aye me too. Sadly the jenkins testing bot hasn't had a run this month (seems most things are running on utopic again for some reason) [19:48] that would have made it easy to see when it stopped working :-( [19:51] balloons: yeah, a little annoying. Oh well. [20:11] balloons, hey :) [20:11] Letozaf_, buonasera [20:12] balloons, I am having problems with autopilot3, I get errors running tests on filemanager app and calendar app, are there any issues or is it just me ? [20:15] balloons, http://paste.ubuntu.com/10611753/ [20:16] balloons, http://paste.ubuntu.com/10611758/ [20:17] Letozaf_, you are hitting the lovely bug: https://bugs.launchpad.net/autopilot/+bug/1432700 [20:17] Ubuntu bug 1432700 in Autopilot "Autopilot3 on vivid desktop fails: "RuntimeError: Unable to instantiate any backends"" [Undecided,New] [20:17] just found it this morning [20:18] balloons, ah! ok thanks, so I am stuck :P [20:18] balloons, but can I use autopilot2 ? [20:19] Letozaf_, yes you can, if the testcase supports it [20:19] balloons, how do I find out ? === roadmr_afk is now known as roadmr [20:26] Letozaf_, by trying it :-) no harm in running it with autopilot [20:26] the only one I know of that won't work is reminders and calendar [20:26] balloons, :) yes just found out that filemanager app does not run also with autopilot2 [20:31] Letozaf_, ahh.. plugin issues? [20:32] Letozaf_, you can also install the old version [20:32] knome: I've disabled my nick changer, but not because of your attempt at shaming me (can hardly be ashamed of something so widely done) but because you're right that it's in the channel guidelines. Plus I'm not important enough that I need to broadcast the fact that I'm away :) [20:32] ubuntu-qa: I'm having some trouble with adt-run failing to find fakeroot: http://pastebin.ubuntu.com/10611845/ The package builds on my computer [20:33] * knome offers roadmr cookies and milk [20:33] knome: thanks :) [20:34] balloons, the old version of autopilot ? or what ? [20:35] Letozaf_, yes. http://mirrors.kernel.org/ubuntu/pool/universe/a/autopilot/python3-autopilot_1.5.0+14.10.20140812-0ubuntu1_all.deb [20:36] balloons, ok thanks I will now [20:36] * balloons hasn't tried that just yet, so do give it a whirl :-) [20:36] * Letozaf_ is giving it a whirl :) [20:36] * balloons feels dizzy [20:36] round and round . . .. [20:36] :D [20:41] balloons, unfortunatly filemanager fails also with the older version: http://paste.ubuntu.com/10611904/ [20:42] balloons, also calendar app: http://paste.ubuntu.com/10611911/ [20:43] balloons, I have a pomodoro in 10' if you wish to talk about that :) [20:43] la_juyis, mmm.. [20:43] balloons, ? [20:44] ohhh sorry la_juyis is a nik [20:44] la_juyis, sorry :P [20:44] la_juyis, vi piace cibo di casa? [20:44] Letozaf_, :p [20:44] 10 minute tomatoes? [20:44] la_juyis, balloons ohhh così la_juyis è Italiano [20:45] balloons, si mi piace il cibo fatto in casa :D [20:45] Letozaf_, no, actually she isn't heh.. We have an Italian restaurant locally here called pomodoro, and that's there motto, so I was responding to her with it :-) [20:46] well.. cibo di casa is the motto anyway, heh [20:46] balloons, oooh cool :D [20:46] balloons, Italian food is the best :P [20:46] otherwise I don't know what pomodoro is :-) [20:46] balloons, pomodoro is tomato [20:46] balloons: it's an app [20:46] balloons: http://pomodorotechnique.com/timer/ [20:47] dobey, mmm real tomatoes are better :P [20:47] oh, well, a "time management technique" [20:47] wild.. [20:47] Letozaf_: only in marinara [20:47] or ketchup [20:47] dobey, I also like tomato salad [20:48] beyond that, meh [20:48] dobey, tomato sauce, we have tomato everywhere here in Italy [20:48] yea.. probably quite good.. not factory farmed under lights :-) [20:48] yeah [20:48] well, depends [20:49] and even the ones outdoors are farmed under lights :P [20:49] Letozaf_, :) [20:49] :D [20:49] balloons, mi piace moltissimo! [20:49] it's just a very, very big light bulb that runs on fusion [20:50] :P [20:50] la_juyis, ti piace moltissimo cosa ? il pomodoro ? :D [20:50] balloons, sorry, i though you'd knew the pomodoro tecnique :D [20:50] Letozaf_, anche! [20:50] la_juyis: it's more fun when they don't though :P [20:50] Letozaf_, ma il cibo fatto in casa. il cibo che faccio io :P [20:50] la_juyis, no, but all this food discussion is more fun eh? [20:51] la_juyis, ma che cibo fai in casa? pasta al pomodoro ? [20:51] balloons, certainly :P [20:51] Letozaf_, tantissime cose! ieri ho fatto biscotti :D [20:52] la_juyis, biscotti al pomodoro :P :D [20:54] balloons, shorts app tests work with the older autopilot version [20:55] balloons, sorry I mean with autopilot2 not autopilot3 older vers. [20:55] Letozaf_, hahaha ! per carita, no! biscotti di burro, mele e avena [20:55] la_juyis, lol was just teasing :P [20:55] la_juyis, mmm sound delicious [20:56] * balloons was worried Letozaf_ ate tomatoes with sugar [20:56] la_juyis, anyways; people.canonical.com/~nskaggs/help/index.html [20:56] balloons, lol [20:58] balloons, by the way there is a receipt for tomatoes with sugar : http://ricette.giallozafferano.it/Pomodori-confit.html [20:58] my italian is not as good as my spanish [20:59] recipe not receipt :P [20:59] hmm, or maybe it is, if only talking about food [20:59] dobey, unfortunatly I do not know Spanish :D [21:00] Letozaf_, i had a salad with those the other day. delicious! [21:00] balloons, how are you dealing with translations? :) [21:01] * Letozaf_ is starting to feel hungry [21:01] balloons, maybe it's worth noting that the FAQ and the Community pages are .en only? [21:04] la_juyis, launchpad for translations. We'll have a big push to get them done soon. [21:05] and la_juyis yes, only German has much in the way of translation at the moment. [21:05] But I thought I would share what the web pages look like atm.. The offline click app looks similar on the phone [21:06] Letozaf_, glad at least some of it can be worked around [21:06] balloons, sure, just asking :) [21:07] balloons, yes I have a bug for shorts to fix so... [21:07] balloons, lmk if you need help with the spanish side of things [21:18] la_juyis, thank you. Do you have additional content (or vice versa, would you like some of this content?)? [21:19] balloons, did you add those to AU? [21:21] I'm not sure if you saw the embedded video; http://people.canonical.com/~nskaggs/help/settings.en-us.html [21:21] la_juyis, I was wondering if you had more things like that. And no, I didn't add anything to AU === chihchun is now known as chihchun_afk [21:23] chihchun_afk, didn't we just talk about awaynicks today on the channel :P [21:23] * balloons notes knome is on the crusade! [21:24] think as i do or get decapitated! [21:25] balloons, oh, I thought you'd be adding stuff there, and consuming the contents from there later [21:27] la_juyis, the idea was to get the basics into the app and link out to AU for harder questions or complex questions that might change rapidly [21:28] la_juyis, the idea to scrape from AU didn't work out.. partially because of translations (we can translate this app) === chihchunl is now known as chihchun [21:48] nuclearbob: you are probably missing -B [21:48] or --unbuilt-tree [21:48] or both. [21:50] nuclearbob, elopio: ugh sorry I missed the ping to u-qa :-\ [21:51] nw === chihchun is now known as chihchun_afk [23:26] balloons: are you still around perchance? [23:50] barry: are you around? :-)