=== jzheng is now known as Guest91633 === strycore_ is now known as strycore === CrazyMelon is now known as CrazyLemon === ksamak_ is now known as ksamak [07:55] morning all [08:00] Morning [08:01] hi!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! [08:01] morning flexiondotorg Laney [08:01] Welcome flexiondotorg! [08:01] Laney, o/ [08:01] willcooke, Thanks :-) [08:02] flexiondotorg, poke me once you have access to the inner sanctum [08:02] OK [08:02] hey flexiondotorg [08:02] welcome! [08:02] hey Laney, hey willcooke [08:02] Thanks Laney [08:02] hello flexiondotorg, welcome to the desktop team [08:02] morning pitti [08:02] pitti, Thanks [08:02] And morning [08:02] you will have your mandatory reeducation shortly [08:03] hi pitti [08:03] how's berlin? [08:03] I have purged childhood memories in readyness [08:03] flexiondotorg: la plus important chose : on parle français ici ☺ [08:03] Non [08:03] I saw the weekend kernel fun :/ [08:03] Laney: kernel> ugh, yes [08:03] * willcooke prepares the poetry appreciation chairs [08:04] Laney: awesome! we went to the Tempelhof airfield yesterday, and walked around a lot [08:04] hablo un poco de español [08:05] pitti: ah nice, we went there too [08:05] me too! [08:05] ah larsu [08:05] ♥ [08:05] hey Laney ♥ [08:05] hey willcooke, flexiondotorg, larsu, larsu! [08:05] when's the conference? === davmor2_ is now known as davmor2 [08:06] Morning didrocks larsu [08:06] morning didrocks, didrocks! [08:06] * Laney pouts at didrocks didrocks [08:06] welcome flexiondotorg :) [08:07] Thanks didrocks :-) [08:07] flexiondotorg: so, you now have to learn French [08:07] hi flexiondotorg! I guess I can't say welcome anymore... but you know ;) [08:07] flexiondotorg: whatever you do, do. not. learn. french. [08:07] you are wearing larsu's shoes [08:07] now about that geonames project... [08:07] * larsu will not give up the nick [08:07] the previous maintainer was a JERK! [08:08] actually, I might. The 'u' is starting to bother me [08:08] * flexiondotorg looks around. Can't see his GCSE French books [08:08] make it mean something else [08:08] I think you can expense some online lessons [08:08] Laney: I heard he's quite nice actually [08:08] über alles === popey_ is now known as popey [08:45] flexiondotorg: yes, Canonical pays the full price of the French duolingo course :) === infinity1 is now known as infinity [09:01] happyaro1: what's the status of bug 1575614? It currently marks the 1.2.2 SRU as failed, and blocks further SRUs of NM (I need to work on one) [09:01] bug 1575614 in network-manager (Ubuntu Xenial) "[SRU]Can't select secret key for TLS auth for 802.1X authentication" [Low,Fix committed] https://launchpad.net/bugs/1575614 [09:04] hei flexiondotorg, welcome!! :-) [09:04] Trevinho, o/ [09:04] Thanks :-) [09:05] huh ? [09:05] flexiondotorg, duuude ! [09:05] pitti, happyaro1 is in SFO this week, so expect a bit of a delay. Maybe drop him an email [09:05] willcooke: ah, thanks [09:05] * flexiondotorg hugs ogra_ [09:05] * ogra_ hugs flexiondotorg and dances [09:11] Laney, would you be able to take a look at this when you get a mo: https://bugs.launchpad.net/bugs/1627190 [09:11] Ubuntu bug 1627190 in livecd-rootfs (Ubuntu) "livecd-rootfs: .pyc files stripped from desktop ISO livefs, makes python startup slower" [Undecided,New] [09:11] I think for "desktop" it doesn't really matter. We have all the horsepowers. Maybe flexiondotorg has an opinion here too, wrt RasPi etc. ^ [09:13] Shipping .pyc will improve start up time. [09:13] That said, I've had mixed feedback from guys in Debian about shipping .pyc in packages. [09:14] Some reviewers have suggested removing other prefer to keep them. [09:14] willcooke, well, effectively it only matters on readonly media ... and only for the first start of apps ... on rw media the files are created on the fly at the first app start [09:14] openSUSE do pre-compile all python and ship .pyc. [09:14] i think we do so too [09:14] It's done at install-time [09:14] by default [09:14] I think [09:14] is it? [09:14] * Laney spins [09:15] If it is a proper python module, and dh-python is used you'll get .pyc [09:15] well, i definitely have the files after adding ubuntu-core to the exception in livecd-rootfs [09:15] they must come from somewhere ... :) either a postinst snippet or build time [09:17] IMO, it makes sense for desktop, server, cloud and core to all do the same thing. So I'd be +1 for leaving them in on that basis alone [09:17] Anyway. [09:17] I would guess it was to save space [09:18] it was [09:18] it surely was [09:18] You would want data about how much space it saves [09:18] well, do a build with ubuntu-desktop in the exception :) [09:19] I can imagine how to arrange to get the data, but thanks :P [09:43] vigo, blank app scope on yakkety bug 1627668 [09:43] bug 1627668 in unity-scope-click (Ubuntu) "/usr/lib/x86_64-linux-gnu/unity-scopes/scoperunner:11:QObject::disconnect:QObject::disconnect:QDBusConnectionPrivate::closeConnection:QDBusConnectionManager::run:QThreadPrivate::start" [Undecided,New] https://launchpad.net/bugs/1627668 [09:59] willcooke: flexiondotorg: ogra_: Laney: it's a dpkg trigger creating them at install time. We don't ship it as part of the package by default because we don't want to rebuild the world when a new python version (with a different path) is shipped. So, the pyc are only compiled for the python version the users has installed on his machine [10:01] didrocks, ah, interesting. Would you mind commenting on the bug? [10:02] willcooke: will do it this afternoon (but that's the basic python mecanism) [10:02] thx didrocks [10:03] ubuntu core case is different (as we don't install packages, but ship an installed rootfs, with some filtering). But ogra now ships the .pyc due to the long Python Speed Discussion thread(Tm) [10:05] Yeah, that's how it works, thanks [10:05] didrocks: Doesn't matter for that bug though [10:05] It would be the same whatever way they were shipped [10:06] except that if it were in the binaries then the re-byte-compiling would need to be reproducible [10:06] not sure if it is nor not [10:06] or [10:06] and doesn't matter, because it's not done that way :) [10:31] does anyone know a way to run evolution without clutter (i.e. plain gtk3) without recompiling ? [10:31] i'd love to use it under Mir but it errors out with a clutter error currently [10:32] ogra_: just run it in libertine and be done ;) [10:33] eeek, no... that even kills my XPS13 ... already tried ... i want to run the native deb on the actual host ... [10:33] i dont like my cpu fans to spin all the time ... and my RAM is precious :) [10:34] ogra_: then use claws :P [10:34] :P [10:34] hey pitti [10:34] i'll just wait for the mutt gtk3 port :P [10:34] hey Trevinho [10:34] pitti: can you comment on https://code.launchpad.net/~seb128/hud/systemd-for-session/+merge/300430? [10:34] Trevinho: just stepping out of the door, will look later [10:35] ogra_: doubt it, gtk-clutter uses a different gtk_init [10:35] ah, sad [10:35] clutter-gtk* [10:35] i know it used to be possible to build it without clutter though ... perhaps that still works [10:35] it is [10:35] * ogra_ will have to check next weekend [10:36] also https://git.gnome.org/browse/clutter-gtk/log/?h=wip/mir dunno what that is [10:36] ;-) [10:36] pitti: ok, no worries... there's no rush [10:37] oh [10:43] Trevinho: seb's away, so if you want that in it might be a good idea to fix it :) [10:43] Trevinho: also, hi! [10:44] hey Laney [10:44] how's it going? [10:44] Laney: sure, I know that... that's why I'm asking... [10:44] Laney: good, good... Doing some qml work these days [10:45] O_O [10:45] And staying some days in Turin, now I'm in a coworking space near to the Politecnico... Quite nice. [10:45] Laney: you? [10:46] not too bad thanks [10:46] reviewing FFes [10:46] but at least I have a coffee [11:25] morning all [11:35] hey andyrock [11:48] omgz the world is down [11:53] * Laney stockpiles essentials [11:55] Laney: I asked some GC-related questions at the D forums: https://forum.dlang.org/thread/tgypnluirhwmeqwuzcgb@forum.dlang.org [11:56] so far, no reply, but they are rather special, so let's see [11:56] (for another question I asked I got a faster reply which was basically "your workaround is good" and "don't use the GC and use refcounting instead" :P [11:56] heh [11:57] * RAOF resists the urge to be a pedant. [11:57] thanks for filing [11:57] I was just trying to recompile with LDC and then everything on earth went down [11:58] ximion_: you could run with the ubuntu backend ... [11:58] or if it's in the downloading stuff, use a remote mirror [11:59] good job I decided to set this up in staging isn't it :-) [11:59] Laney: jup, but who knew that this would be an issue? [12:00] I don't see any off the bugs you describe, so it's either something in the Ubuntu backend causing them, or a GDC weirdness [12:00] unknown unknowns, that's the point of testing [12:04] Laney: I fixed something super-weird at the weekend: https://github.com/ximion/appstream-generator/commit/6ab455ed9c33c2e014a52d8c9539ba5169a6131d [12:04] Lanes: I don't know why storing a pointer in the class and using it later results in wrong behavior there - according to the fontconfig docs, that shouldn't have been the case [12:05] Laney: so either some bad optimization was happening there, or Fontconfig simply sucks. Both is equally probable, IMHO [12:05] all these bugs are weird [12:06] like the LMDB one we can't explain [12:08] countdown to the C rewrite... :) [12:09] happy monday [12:09] resistance won [12:10] Laney: uh, I hope I can get around doing that ^^ - multithreaded C is even less fun [12:10] GThread! [12:10] hi desrt [12:10] well done [12:10] hey laney :) [12:10] Laney: for the fontconfig bit, *everyone* uses the library that way, so I kind of wonder why - probably because fontconfig is just buggy there [12:12] Laney: I am pondering over writing a MongoDB backend for asgen... Shouldn't be too hard and would allow us to get rid of the mmap and also store massive amounts of data, which is interesting for the "I have thousands of PPAs" case === cpaelzer_ is now known as cpaelzer [12:17] ximion_: Ya we should think about the PPA thing [12:18] Might want to make the asgen more cloudy to work with that [12:18] so you can have 10 asgens that take requests from a pool, or something [12:18] Laney: Mongo is the hipster thing to go with - can't be more cloudy ^^ [12:18] Trevinho: replied to the hud MP [12:18] you'd need to do a work queue thing too [12:19] and maybe other stuff [12:19] LMDB will run out there, as it becomes pretty inefficient if you have too much stuff in there [12:19] like make it a daemon that is polling the queue [12:19] yes, will be quite an effort to plan and implement properly [12:19] Trevinho: ... or was trying to, until LP timed out :) [12:22] Laney: especially plan it, I haven't thought enough about it yet, so if you have ideas - since Ubuntu/LP will - at time - be pretty much the only user of this, any feedback on how this should look like to work for you would be useful [12:22] * ximion_ knows nothing about LP [12:23] not sure right now [12:24] basically LP submitting to a queue (amqp or something), then fetching from some well known place (maybe just the existing rsync) [12:25] cloudy is going to be no good if each instance needs 32GB though ._. [12:29] Laney: yeah, we need to get to the bottom of this [12:30] although [12:30] if it's split by arch too then the leak won't be so bad [12:30] Laney: when I asked https://forum.dlang.org/thread/uihemwyqzycxgvilhjbv@forum.dlang.org, the first reply was use refcounting, so in case the GC really is the culprit, we could certainly do that (and then need to check for reference loops...) [12:30] pitti: thanks, i'll fix that next days then... I guess it's better not to remove all the upstart stuff too [12:31] pitti: also, about that landing silo with indicators, shouldn't that only land in yakkety (while it was created with a triple landing) [12:31] causing troubles, I guess [12:31] but tbh, I don't think it's the GC - I kind of suspect LMDB or zarchive to be the problem here, nothing else could fill up 32G quickly enough [12:31] Trevinho: if you can get away with landing it in y, sure; but I thought some of these packages should be kept in sync between v/x/y [12:32] ximion_: we should make a small archive and valgrind it [12:32] it grows with each arch [12:32] pitti: mh, i see.. I just would like to avoid the landing to delay more because of SRU bits [12:34] Trevinho: yeah, me too [12:34] Laney: I am using almost exactly the same code (TagFile and archive stuff) in another project, and memory usage is fine there [12:34] Laney: I also did Valgrind it at some point when the project was still younger, w/o finding memory leaks [12:34] would be worth trying again [12:34] Trevinho: I don't know what the status of that is or what blocks it; but from my POV, anything that unblocks it is good obviously :) [12:35] and with the ubuntu backend / with remote repo [12:35] Valgrind got a bit confused by the GCs memory pools though, and GLib also sometimes has extra memory allocated (but there is some env var to work around that) [12:36] Laney: you can alter the LMDB mapsize to something much smaller: https://github.com/ximion/appstream-generator/blob/master/src/asgen/datastore.d#L131 [12:36] then you can valigrind the stuff, but also only can process small archives [12:36] yeah saw that [12:36] i'll try it some point in the future [12:37] * Laney is trying ldc build now [12:37] Laney: you will run into https://github.com/ldc-developers/ldc/issues/1774 [12:38] that's what you get from having a Git snapshot compiler in Debian :-/ [12:38] (markos, what did you do!!) [12:38] Laney: if they manage to fix that quickly, then we are in luck and I can upload a fixed version asap [12:39] I got a build ... [12:39] the last beta version also works [12:39] of the snapshot version, backported to xenial [12:39] and that works to build asgen [12:39] Laney: it just can't build itself :P [12:40] whatever was in xenial can build it :) [12:40] if it builds, it's fine ^^ [12:40] are you using Meson? [12:40] not atm [12:40] too many backports [12:41] ^^ [12:42] * ximion_ goes through the D world writing Meson build files at time === mhall119_ is now known as mhall119 [13:13] willcooke: Hi Will, the guest session feature seems to be broken in Ubuntu (with Unity); please see bug #1627304. It's not clear yet which package(s) are the culprit. Guest session works fine in Ubuntu MATE, so it's probably not lightdm. Anyway, just wanted to call your attention to the issue, since I suppose it's something which should be fixed before the release. [13:13] bug 1627304 in xubuntu-default-settings (Ubuntu) "User locking problems - guest login crashing" [Undecided,Confirmed] https://launchpad.net/bugs/1627304 [13:46] * Laney late lunch [14:07] erk, LP down? [14:07] me 2 [14:07] ta [14:07] * willcooke makes tea [14:10] erk, it's not just LP -- http://people.canonical.com/~ubuntu-archive/ [14:13] luckily I work on spreadsheets! [14:13] ruh roh [14:13] :) [14:14] didrocks, could you make a presentation about it? [14:14] I just thought "let's do some archive admin while I wait for this bisect"... [14:14] ha [14:14] willcooke: indeed, with some Action items associated :) [14:14] didrocks, \o/ === charles_ is now known as charles === jzheng is now known as Guest62467 [14:48] oh noes! [14:48] nothing to do now, might as well go to the beach [14:50] :) [14:50] Laney, lol =) === JanC is now known as Guest94585 === JanC_ is now known as JanC [14:54] blehhhhhhhhhhhhhhhhhh [14:54] you did realize that it's raining outside? :) [14:54] (says the weather website :p) [14:54] as if that ever stopped a brit from going to the beach [14:54] i'll even have a bbq [14:54] ahah, different standards! [14:55] that should remind you of those glorious days of launchpad update :) [14:55] just wedge the umbrella into the handle [14:55] no matter if you get a little bit of plastic in your food [14:55] adds to the flavour [14:56] oh man, the pre FDT days [14:56] FDT? [14:57] fastdowntime [14:58] Trevinho: can we make unity have Unity;Unity7 in XDG_CURRENT_DESKTOP plz? [14:58] s/;/:/ # argh [15:02] yeah, pre FDT, with an hour and half downtime ;) [15:17] I just installed 16.10 from the daily ISO. works like a charm so far [15:23] :) [15:23] shut it down now so nothing goes wrong [15:23] lulz [15:25] eyelash: Did you try to enter a guest session? [15:25] GunnarHj: no [15:26] omg there's a bug!!!!! [15:26] eyelash: Please do (bug 1627304). [15:27] Error: Could not gather data from Launchpad for bug #1627304 (https://launchpad.net/bugs/1627304). The error has been logged [15:27] launchpad is temorarily unavailable :( [15:28] I wouldn't bother trying [15:28] Not sure why he's trying to get you to experience a bug [15:29] Laney: ?? [15:29] Why are you telling someone who had a good experience to go have a bad one? [15:30] Laney: Mostly because I wanted you to have something useful to do. ;) [15:31] Laney: yeah, not sure is defining it though [15:31] Ok, thanks for your concern - be assured that the bug is known :) [15:31] Trevinho: It's in the session file - DesktopNames [15:32] is there anything else I should test? I have a Radeon X1300 which might be quite an unusual card. will test unity8 in the coming days [15:32] Laney: Great, and to be serious, that's what I wanted to make sure. [15:37] I assume the theme bugs will be fixed before the release!? [15:37] depends if launchpad comes back before release :) === daniel3 is now known as Odd_Bloke [15:50] Laney: any idea why compiz would go awol if you select the video scope? [15:50] 0 [15:50] did it crash? [15:51] Laney: nope just used 98.2% cpu until the spinner stopped then dropped [15:52] Laney: this is a fresh install so no actual videos on the system [15:52] don't know, you would be better off asking a compiz developer [15:52] Laney: hmmm okay [15:52] soz [15:52] Laney: no worries it was a longshot anyway :) [16:07] jdstrand: https://bugs.launchpad.net/ubuntu/+source/lightdm/+bug/1627304 <- looks to do with the apparmor profile (setting it to complain makes login from greeter work); could you put it on your (or someone's) stack please? [16:07] Ubuntu bug 1627304 in xubuntu-default-settings (Ubuntu) "User locking problems - guest login crashing" [Undecided,Confirmed] [16:08] * Laney got called by pagerduty /o\ [16:10] Laney: someone from the security team will look at that this week (it may not be jdstrand) [16:10] tyhicks: righto, thanks [16:10] tyhicks: I made it affect apparmor now [16:11] Laney: thanks! [16:11] I mean, if the fix is in the profile then it's lightdm, but this presumably gets the right people to look [16:20] tyhicks: re that ^^ bug - you asked for apparmor=Denied - yup that is there - will comment on bug [16:20] flocculant: thanks, that'll help a lot [16:22] tyhicks: welcome [16:23] been a few mid air collisions on there :P [17:03] hmmm [17:03] crashy [17:10] willcooke: welcome to my world [17:41] * flexiondotorg is still subscribing to mailing lists... === alan_g is now known as alan_g|EOD [17:57] Right, dinner time. [17:57] night all