[06:06] hello [07:01] Bonjour tout le monde ! === ljp is now known as lpotter [08:11] morning [08:52] morning all [08:54] morning willcooke and all! [08:54] willcooke: did you spend a good week-end ? [08:54] (I don't know if it's the right answser) [08:55] :) yes thanks alexarnaud [08:55] did you? [08:56] willcooke: the right question is : did you have a good weekend? [08:58] My week-end was studious. I've working for a study project : beeslide.fr [09:01] Neat! [09:01] hey hey [09:02] what up Laney [09:02] hey Laney, how are you? good weekend? [09:05] hey willcooke and pitti! [09:05] not bad thanks, went out for some birthday drinks and ended up staying out late [09:05] so did a lot of sleeping yesterday, followed by quiz & tapas at a local vegetarian restaurant [09:06] what about you? [09:09] Laney: we visited a colleague of Annett the whole Saturday, and yesterday loots of household stuff: clearing drains, cleaning the appartment, cutting trees in the garden, etc. [09:09] Laney: uh, I have your bday for this Wednesday, is my calendar wrong? [09:10] pitti: wanted to have some drinks on a weekend, and the next one I'll be in .be :) [09:10] ah, good point :) [09:10] * pitti is now faced with an impossible task -- pick an UTF-8 char for excuses.html for "retry this test" [09:12] pitti: still a couple more days of matching /2\d/ [09:12] :) [09:12] umm [09:12] is there a unicode recycling symbol? [09:13] 🔃🗘🛠 [09:13] moin! [09:13] hey Sweet5hark1 ! [09:14] ♲ [09:14] although 🛐 kind of fits too, "I hope that it works this time!" :-) [09:14] willcooke: finished building a snap: 353MB on amd64 ... [09:14] * Laney doesn't have any of those ones in this font/charset/whatever [09:14] * Laney stabs screen a bit again [09:14] Sweet5hark1, woot! Thanks! [09:15] Laney: have a look for this one U+267C ♼ seem the most ubuntu-esque [09:15] Laney: ⚒ is nice too [09:16] pitti: heya, testing people irc clients to support wild utf-8 codepoints this morning? [09:16] Sweet5hark1: no, finding a nice symbol for "retry this test" for excuses.html :) [09:17] the hardest part of implementing it! [09:17] * pitti likes ⚒ and will go with that for now [09:17] hrhr [09:23] ➥ pitti [09:24] ↪ pitti or this more elegant version [09:24] hm, that doesn't connote "retry" for me, it's more like "forward"? [09:25] ♻ might be best there indeed (also a bit stronger than the above ♲ ) [09:30] good morning! [09:30] * Laney loves that there is ♻ and ♲ and ♼ [09:30] hi larsu! [09:30] hey larsu! [09:30] hi Laney and pitti! How's life? [09:31] schönes wochenende? [09:31] larsu: it's nice, thanks! had a full but nice weekend [09:31] good! [09:31] went to the cat cafe (H) [09:31] larsu: how about you? [09:33] oh nice :) [09:33] mine was nice as well, but very quiet :) [09:33] watched "the danish girl". good movie! [09:34] good morning everyone. reading backlog, it's all about that utf-8, 8, 8, 8 no ascii. [09:34] haha hi xnox [09:34] what up xnox [09:35] * Laney flashes some gang signs [09:35] mostly been reading my bug mail... and i feel like starbucks and not fixing any of my assigned bugs [09:36] my condolences [09:36] larsu: ah, that's on our list still, too; last time we picked "The Revenant", that was also great [09:38] i liked danish girl, and cringed and looked away during The Revenant. [09:38] hey xnox, how are you? [09:38] i guess there are people who did the opposite too. [09:38] "The Big Short" is also on my list still; not enough weekends to go to the movies! :-) [09:38] also the horse scene was a lot like Alien vs Predator to me.... [09:39] pitti, yeah. I think i read the book about it too. No idea if movie is at all based on said book or not. [09:39] but Big Short is a funny movie, or so i take from the trailer. [09:40] pitti, well, i need to run 4km to get a free cinema ticket voucher.... [09:40] xnox: wow, you have a "free cinema tickets" booth 4 km from you? [09:41] dumpster diving @ cineworld [09:41] haha [09:41] Laney, you do too, no? from vitality? [09:41] nah I don't have that [09:41] * Laney NHS [09:42] Laney, they still redirect to nhs. I'm not sure how they make money, cause i think i'm getting more out of them than the premium. [09:43] pitti, if i log/sync enough steps or 30min of active time through one of the apps, i get one free cinema ticket a week & a starbucks coffee - http://www.pruhealth.co.uk/vitality/partners/cinema/ [09:43] you counting what you pay or the full price? [09:43] xnox: oh, wow [09:44] Laney, i'm counting the full price. [09:44] Laney, but obviously i beat the bit i pay. [09:45] k [09:45] then part of the answer is probably that vitality negotiated a discount [09:45] and some economies of scale [09:52] Laney, yeah, i suspect these partners pay vitality to make these offers. To generate additional in-house revenue, once the person is through the door. === vrruiz_ is now known as rvr [10:03] Laney: http://people.canonical.com/~pitti/tmp/excuses.html [10:03] Laney: the package name/releases are bogus of course, but you should be able to click and SSO-auth [10:03] (it will then say "no such release" or so) [10:04] it's a page generated by britney's tests [10:04] pitti: is it expected that the team membership thing is unchecked by default? [10:04] I bet if I forget to click that it doesn't work :) [10:05] Laney: expected yes, desired no, this needs an RT or something to "bless" autopkgtest.u.c. for getting this checked by default [10:05] there's no API parameter or so to have this on by default? [10:05] huh [10:05] Laney: no, this needs to be done on the SSO side [10:05] I'm sure there is a good reason for that... [10:05] robru: ^ who did you ask for getting the team checked by default? [10:05] Laney: also, the link isn't https:// yet, as that's firewalled (RT pending) [10:06] so you'll get some "unsafe blabla" warning [10:06] I won't anncounce this until ssl is on [10:06] where's the code for this? [10:06] just in case :) [10:08] Laney: https://git.launchpad.net/~canonical-foundations/+git/autopkgtest-retrier [10:08] ah [10:08] pitti: thanks! [10:09] pitti: I suppose if you move to the uploader check then you don't need the team thing [10:09] or, since this is a public team, the app itself could check membership too [10:10] if this takes long and people are forgetting to check it all the time [10:22] willcooke: do you know when seb128 come back? [10:23] probably not until tomorrow afternoon [10:23] actually, later than that [10:23] Wednesday morning [10:23] Laney: yeah, indeed; but hopefully whitelisting a.u.c. will be quick [10:29] pitti: sometimes when "hope" meets RT the good team loses :P [10:29] lol [10:31] * Laney likes it when you file a bug and find it's fixed before submitting it [10:41] willcooke: OK, thanks === hikiko is now known as hikiko|ln === hikiko|ln is now known as hikiko [12:13] * Laney slowly turns glib's tests green [13:09] xenail daily build iso fails to start here [13:14] should nautilus in xenial support google drive OOTB, as omgubuntu claimed it (3.18) would? [13:15] with goa I think, not uoa [13:15] try with gnome-control-center.real [13:15] no guarantee [13:21] ah [13:21] bummer then [13:24] you could try to hit mardy up to see what work would be required [13:40] ok [13:47] andyrock: in what way? [13:48] http://people.canonical.com/~laney/weird-things/hi.png [13:48] :) [13:49] not that "works for me" is very helpful [13:49] !wfm [13:49] Common Sense: Just because you can, does not mean you should (and especially recommend to others). Think before you do. "Works for me" does not mean it is ok. The latest version of everything is not always useful if you aim for stability. [13:49] always thought that was a weird factoid [13:51] Laney, How goes the GNOME Software transition? [13:52] ask robert_ancell [13:52] he posted a ppa on the list [13:52] try that [13:52] Thanks [13:59] "Hello, and good morning." [13:59] hi desrt [14:00] hello will cooke. [14:05] desrt, do you know where gdkscreen gets its information about screen rotation from? [14:05] xrandr [14:06] thx [14:06] * willcooke goes further down the rabbit hole [14:08] https://git.gnome.org/browse/gtk+/tree/gdk/x11/gdkscreen-x11.c#n616 might be a good place to start digging [14:08] XRRGetMonitors basically gets you everything you need [14:09] cool, thanks === maclin1 is now known as maclin [15:34] Laney: it's a kernel failures [15:34] *failure [15:35] looks like something related to nouveau drivers [15:35] but not sure [15:46] Are resize-handles handled differently now in gnome-shell, versus a few months ago? I use x2x to merge two desktops, and the screen edge that transitions to the other display isn't magic any more if there is any window close to it. [15:46] I didn't use Shell for a few months. I don't know when more precisely. [16:09] pitti: have you looked into udisks2/ppc64el? [16:10] (did I ask about this before?) [16:35] Laney: I didn't look at it yet, no [17:17] https://plus.google.com/+AmartyaDattaGupta/posts/EouWaWmibwS [17:30] DAMN YOU Gtk CSS! [17:30] Why wont my bottom borders show up on notebook tabs?!!?!?!? [17:31] I think it's there but transparent. In that the area of transparency increases when I change the size of the border, but I just can't get it to be visible === alan_g is now known as alan_g|EOD [18:15] willcooke, cyphermox: I have a crash trying to install oem 64bit against secure booted kvm instance in the installer about to file the bug if it lets me now [18:17] oh on 14.04.4 btw [18:22] willcooke, cyphermox: bug 1537858 [18:22] bug 1537858 in ubiquity (Ubuntu) "Crash happened during the install of oem mode" [Undecided,New] https://launchpad.net/bugs/1537858 [18:23] says there's a package conflict [18:24] I'll try an install here in a bit [18:25] cyphermox: just trying a normal install rather than oem see if it is only triggered in that mode [18:25] ack [18:25] I'm zsyncing the image [18:33] I am very new to this so excuse my ignorance. I am looking for general help with a new install of Ubuntu 14.04 it is running Super Slow is this the right place for this convo [18:39] BruBeer: probably better to ask in #ubuntu [19:01] cyphermox: looks like it is specific to the OEM version the normal version has just installed [19:02] cyphermox: so possibly one of the oem specific files maybe? [19:05] cyphermox: fwiw I checked the 3rd party and download updates boxes during the install I'll add that to the steps === DanChapman is now known as DanChapman_ === DanChapman_ is now known as DanChapman [19:17] davmor2: oh, indeed I suppose it could be why [19:22] When you're trying to work out a Gtk theming issues and you're reading the source SCSS for Adwaita... [19:22] """ [19:22] notebook { [19:22] // Through me you go to the grief wracked city; [19:22] // Through me you go to everlasting pain; [19:22] // Through me you go a pass among lost souls. [19:22] // ... [19:22] // Abandon all hope — Ye Who Enter Here [19:22] padding: 0; [19:22] """ [19:23] sounds like time to call it a day :) [19:23] :D [19:24] willcooke: That sounds suspiciously like the place you need to be to fix stuff, you know it's like the Warning No Entry sign ignore and press on what could possibly go wrong...will.......willlllll...........wiiiiiiiiiiiiiiiiiiiiillllllllllllllll [19:25] I prefer sarnold's idea :) [19:25] and once robert_ancell gets up that's exactly what I'm gonna do [19:26] willcooke: wrap it up nicely with a little bow and label He'll never see the trap till it's too late :) [19:27] :D [19:28] willcooke: oh and remember the gag the admiral Ackbar he has on his shoulder that will forewarn him [19:37] morning robert_ancell [19:38] willcooke, hi [19:38] lookout! it's a trap! [19:39] willcooke: did I not warn you to gag him ;) [19:39] * willcooke remembers to speak to Boba Fett about sarnold # [19:39] robert_ancell: https://www.youtube.com/watch?v=-bzWSJG93P8 if you hear something like this it's just the party next door ignore it :) [19:40] uhhh, /me has to do something else suddenly... [20:03] robert_ancell: i think i got the reviews to work now [20:07] attente, nice! otp - be back soon [20:08] robert_ancell: no worries, tyt :) [20:08] robru, Laney: got the "can requestor upload trigger or package" check implemented now, and completely dropped the Teams bits [20:09] that's indeed easier, and circumvents the "team not checked by default" issue [20:09] pitti: well "team not checked by default" issue is resolved by talking to sso people to get on the whitelist, but yeah, checking upload rights is better anyway [20:10] hey pitti, another night shift? [20:10] robru: right, it's more like a nice side effect than an insurmountable obstacle of course, but as we planned to do that anyway, we can do it right away :) [20:10] sarnold: I don't plan to, I just wanted to figure this out [20:10] aha :) [20:11] pitti: where's the code? can I see it? [20:11] robru: I stumbled over the fact that nobody was allowed to upload to e. g. trusty Release, so it needs pocket=Proposed [20:11] robru: I guess you found that too [20:11] robru: just committing, hang on [20:12] pitti: http://bazaar.launchpad.net/+branch/cupstream2distro/view/head:/citrain/recipes/base.py#L352 did you see the train implementation? [20:13] robru: yes, that's what I modelled it after [20:13] pitti: ok cool ;-) yeah we do pocket=Proposed as well [20:13] robru: except for the New bit, as that isn't necessary for retries [20:14] pitti: are you sure though? does autopkgtest not run for packages that are not in ubuntu archive yet? [20:14] pitti: like you could have a package that is in the silo but not in ubuntu archive, and the test could fail, and somebody could want to retry that [20:14] robru: it can run for PPAs [20:14] robru: but people who do that can already request runs via the "normal" AMQP route [20:15] robru: so, I do want to implement this eventually [20:15] robru: but this is even less dependant on team membership, but we need to ask for upload privs to that PPA [20:15] robru: but then again, the package must already exist in the PPA [20:15] pitti: what do you mean "it can run for PPAs"? if you check upload rights to a train ppa you'll just get trainguards (eg, me) [20:15] robru: this is *re*trying a failed test, therefore it must have run at least once for an existnig package [20:16] pitti: yes, but it could have run on a package that's in a ppa and not in the ubuntu archive. [20:16] robru: I can upload to silos too :) [20:16] pitti: if your code is checking against the upload rights to the main archive, the package might not be there [20:16] robru: that's correct; but so far that's all that it did anyway [20:17] robru: you couldn't retry PPA runs with the previous version either, the validation would have errored out way before [20:17] that feature is just completely missing still [20:17] ok [20:18] attente, did you talk to dobey? [20:18] robru: https://git.launchpad.net/~canonical-foundations/+git/autopkgtest-retrier/commit/?id=4fe5b6e9 FYI [20:18] coool [20:18] robert_ancell: no, but i saw some of the backlog from last night [20:19] attente, so it sounds like ubuntu-sso-client is ultimately the wrong long term solution, but I guess there's no replacement yet? [20:19] yeah, i'm not sure how to proceed. if a rest api exists, we can try using it directly, but it sounded like it doesn't? [20:19] Depending on ubuntu-sso-client will mess up the dropping Python 2 goal. [20:19] robru: how would we verify that a person can retry a silo test? [20:20] robru: if relevant devs can't actually upload to the silo? (that's quite surprising to me) [20:20] attente, you can use the REST API, but that means that g-s needs the UI and data storage which is suboptimal (but was my previous plan_ [20:20] pitti: well, the same way we handle NEW packages in train publications: if that person has permission to upload flashplugin-nonfree into universe. [20:21] pitti: right, only a small group can upload to PPAs because the train does PPA uploads for them. [20:21] robru: ah, so you still want to limit it to ~ubuntu-dev (makes sense) [20:21] pitti: I don't think there's a meaningful way to identify a silo owner programmatically [20:21] so i guess we either put that ui directly in g-s or write a new service [20:22] attente, does ubuntu-sso-client integrate with u-c-c? [20:22] robru: ok, so we'd need to extend the "is this package/version valid" check for PPAs, and then this binNEW trick with a hardcoded universe package name [20:22] pitti: yeah [20:22] robert_ancell: not sure. grep says "no" [20:23] pitti: damn son, I *almost* like your raw REST calls more than lplib... [20:23] attente, that was the bit I was confused about - what the relationship between libaccounts and ubuntu-sso-client [20:23] oh. but there is a plugin that integrates with it, right? [20:24] robru: well, lplib is just syntactic sugar around those :) I just want to stay out of the business of all those backports and potential bugs [20:24] pitti: yeah exactly, but I mean the way you make the calls manually is almost nicer than the 'sugar' [20:24] robru: while looking into bug 1153671 I found at least eight bugs in py3-lplib :) [20:24] bug 1153671 in apport (Ubuntu) "Port to python3-launchpadlib" [Low,Triaged] https://launchpad.net/bugs/1153671 [20:26] robru: yeah, "def lp_request()" is a bit low-level, I split that out now; using it looks reasonably non-ugly [20:26] pitti: despite those bugs we are using it with great success in the train for many months now [20:26] robru: yeah, apport is using quite a lot of API :/ it auto-creates bugs, attachments, needs to deal with arbitrarily weird data, etc. [20:27] ah [20:27] robru: I really like the "100% test coverage" enforcement! (at least for a tiny project like this) [20:27] pitti: yeah, it's really great. I consider 100% coverage to be a minimum for python [20:27] now I just need SSL un-firewalled, then this is ready to go [20:28] attente, so, what works now? [20:28] pitti: one time I wrote a null testsuite for a project and coverage reported it was 70% covered just by importing the code, so python coverage numbers are a bit skewed. [20:28] robru: yes, it doesn't tell you anything about the quality of your assertions of course [20:29] robru: like, by my standards this is actually relatively bad as it has almost no black-box integration tests, just unit tests [20:29] but oh well, cf. "simple enough" :) [20:29] robert_ancell: should be able to post a review with the g-s ui you added [20:29] pitti: yeah, exactly. that's why 100% is a minimum, because once you get 100% you also need to write integration tests ;-) [20:29] attente, nice! [20:30] if the user isn't logged into u1, then ubuntu-sso-client will prompt, but you're right, we have to replace it [20:30] attente, so I guess the next goal is to solve the ubuntu-sso-client issue - fix, replace or put that UI into g-s (for now) [20:30] attente, you've seen the REST API docs for that right? [20:30] are there other more urgent things we should look at? [20:31] haven't, and not sure where the docs are for it [20:31] attente, the other things are software-properties-gtk (I started hacking on that) and Unity integration [20:31] And anything else we might have forgotten [20:31] dobey: hi, do you know where the docs are for the rest api of ubuntu sso? [20:31] attente, I've got them, hang on [20:32] attente, http://canonical-identity-provider.readthedocs.org/en/latest/ [20:32] * willcooke has just emailed andyrock about the Unity integration (if that's the same thing we just talked about robert_ancell?) [20:32] willcooke, yep [20:32] attente, so it seems pretty straight forward to log in. [20:33] FWIW - if U-O-A isn't going to be fixed by those guys, then I'm all for adding that functionality in to G-S until such time as they do fix it (which might be never) [20:33] attente, and my blogpost http://bobthegnome.blogspot.co.nz/2015/12/accessing-webservice-using-libsoup-and_10.html [20:33] That said, mardy did think it would be low effort to fix it properly, so perhaps that's the better option [20:33] willcooke, it's the pragmatic solution but probably wont be able to go upstream like that [20:33] robert_ancell: ah, nice [20:34] robert_ancell, ahh. [20:34] willcooke, it also breaks the internal model of g-s where the plugins don't have any UI access [20:34] might be worth poking mardy again tomorrow and asking him for a list of steps to fix. I can do that in the morning once he's back online if that will help [20:34] so it sounds like we should basically write a better u-sso-client [20:34] attente, I had some old code for plugins to report what authentication they wanted (see git history). That's what I was thinking to push the request to the UI. [20:35] lemme see if I can find mardy... [20:35] attente, I think that is by far the best solution if we can get it done in time [20:35] robert_ancell: i saw, you might be right [20:36] attente, so I would say if you think it's do-able then that's the top priority to work on. It will give a better UX and it will be upstreamable like that. [20:37] ok, sure [20:37] We should also keep in mind how to support login with GOA as that's what the Ubuntu GNOME people will want to use. Probably wont have the time for us to do it but if we know what needs to be done someone else might step up and do it. [20:37] attente: i'm not really here today/tomrorow (swap days), but if python is ok, then you can probably work from the code that xnox wrote a few years ago to add u1 login to the installer. it uses the newer REST API version. [20:38] i guess python isn't good for UOA/GOA though [20:38] dobey: ok, thanks [20:39] * pitti waves good night [20:39] robert_ancell: are we going to file an FFe for this? [20:39] pitti, bye! [20:39] night pitti [20:39] attente, are we past that point already? [20:40] attente: too bad i didn't realize this was the case last week. we could have talked in austin :) [20:41] dobey: ah! i didn't know you were there! [20:42] robert_ancell: i guess we have time, i'm just worried since the deadline is coming up [20:43] attente, yeah, if the click stuff / appstream support is too delayed we'll have to do a FFe [21:05] meh, gtk wins this round. [21:05] But I'll be back tomorrow [21:05] g'night all [22:09] dobey, omg