=== JanC is now known as Guest15498 === JanC_ is now known as JanC [08:05] hidehi [08:12] g'mornin' Laney [08:14] hi willcooke [08:14] what go? [08:14] last day of UOS [08:14] huzzah [08:14] think I'm going to get a hair cut this morning [08:14] changing the world one hangout at a time [08:14] you knowns it [08:14] knows [08:14] voted yet? [08:14] oh yes! that too [08:15] I shall do so on my hair to get a hair cut I think [08:15] i went to a hustings last night [08:15] I have seen none of the candidates canvasing at all [08:15] some protesters disrupted it at the end [08:15] yay! [08:15] other than that it was the worst [08:15] I'm going to spoil my ballot [08:15] * Laney looks at chrisccoulson [08:15] Awesome! [08:16] going for the classic "cock and balls"? [08:16] haha [08:16] I was thinking about appealing to the conscience of the counter [08:16] :D [08:16] "get up and leave the room now" [08:16] take the ballots with you and throw them in a pond [08:17] bloody PCCs [08:17] anything interesting is "operational" and not a matter for them [08:18] willcooke: he's a geek he can't throw, he'll wind up keeping hold of it and throw himself in the pond. Laney don't listen to willcooke he's just trying to get you wet [08:18] I meant the counter [08:19] I could put a steak on a stick and tempt them all into the pond [08:19] willcooke: see that's how a geeks brain works :D [08:20] or maybe my ballot on a rope dnangling over the pond would be more attractive to an election counter [08:21] * Laney votes for davmor2 instead [08:23] Laney: Europe is already broke they don't need my help ;) [08:30] going to get a haircut, bbl [08:30] hf [08:58] * Laney gets trello-y [09:03] bah, closed [09:04] I'll try again tomorrow === Drac0 is now known as Guest65030 [09:23] screw this, I'm buying an Intel wifi card [09:25] oh, 3 quid on ebay [09:25] should have done this a long time ago [09:35] morning all [09:38] xnox: NEO doesn't support 4096R? [09:38] hi andyrock! [09:39] hi andyrock [09:45] HI Laney [09:45] I'm not even bothering with that today - I think it's only PCC elections here. [09:46] I don't even know where my polling station is ;) [09:48] that's all we have too [09:48] nice walk in the sun if nothing else [09:57] man, what is up with the progressbars [09:58] the styling isn't being applied properly, and the background colour seems to be defaulting to transparent [09:58] I've checked the theme, and the CSS is correct and does apply in the inspector [09:58] so must be a specificity thing (again) [10:34] Laney: hey, so g-s is blocked from being sru'd because of this apparently: https://errors.ubuntu.com/problem/58f08ffc3281621f3df0ae8b636df69edadf0f02. it doesn't seem like it was introduced in the version just uploaded, but it seems like it might be occurring more often than before [10:37] attente: did you get one of those automated emails? [10:38] Laney: yeah [10:38] holy early morning [10:38] attente: doesn't look like a regressio nto me [10:38] I'd reply and say that [10:38] don't know what the problem is though, do you? [10:39] no TagF? [10:40] hard to tell... [10:40] might be sensible to check that [10:43] like maybe if the PkgF.Open fails [10:43] file:///usr/share/doc/libapt-pkg-doc/html/classFileFd.xhtml [10:43] it returns a bool [10:44] not sure how TagF can be invalid here [10:46] I think just check at 376 and fail with an error if the Open() doesn't work [10:46] can't think of another reason [10:46] wouldn't the backtrace fail in the Jump function? [10:46] other than trying to make a tagfile from a not opened FileFD [10:49] or a malloc failure or something === hikiko is now known as hikiko|off [11:54] Hi Laney, and chance you can sponsor bug #1575555? [11:54] bug 1575555 in fonts-noto-cjk (Ubuntu Xenial) "Chrome/Chromium use "Thin" as default font weight" [Medium,In progress] https://launchpad.net/bugs/1575555 [11:57] GunnarHj: I can look after lunch [11:57] is it just a case of backporting y to x? [11:57] Laney: Great. Yes, backporting or upload from PPA - your choise. [11:58] k [12:10] Good morning [12:17] morning pitti [12:20] hey willcooke [13:17] attente, around? [13:18] willcooke, attente: have you seen something like http://paste.ubuntu.com/16237898/ with gnome-software already? [13:18] dpm, ^ [13:19] it looks like I get no apt/dpkg results [13:19] willcooke: hey [13:19] heh, so basically, what dholbach said :) [13:19] I'm just running through some troubleshooting with dpm now [13:19] ah ok [13:20] dholbach, is it broken for both of you? cos wfm [13:20] can you try running it with a strace+grep to see what file it's trying to open? [13:20] attente, willcooke, is this the right version? http://pastebin.ubuntu.com/16237931/ [13:21] willcooke, yeah, we were on a call preparing the demo and it's broken for both of us [13:21] dpm, correct version [13:21] attente, http://pastebin.ubuntu.com/16237826/ [13:21] dpm, what does "snap find hello" return? [13:22] Name Version Summary [13:22] hello-world 6.0 Hello world example [13:22] dpm: does this still happen after killing it and re-starting? [13:23] attente, yes, tried to killall and restart multiple times [13:23] attente, output of strace -eopen,stat gnome-software (http://paste.ubuntu.com/16237969/) [13:23] dpm, what app are you searching for? [13:23] dpm: can you try doing an apt-cache update and restarting? [13:23] willcooke, I was just searching "calculator" [13:24] attente, restarting what? [13:24] dpm: restarting gnome-software [13:25] attente, I guess you mean 'apt update'? Tried that already [13:25] apt update and restarting gnome-software doesn't help in my case [13:25] dpm "snap find calculator" == "no snaps found for calculator" [13:25] yeah, sorry [13:25] apt update [13:25] I also tried to remove cache/config files of gnome-software, just to be sure it wasn't some old broken config [13:25] but that also didn't help [13:26] did we catch other bugs which said "failed to start fwupd" somewhere? [13:26] willcooke, I think that's a known bug with the 'find' CLI command, as the snap name is 'ubuntu-calculator-app', but up until today g-s found it by just searching for "calculator" [13:26] good afternoon desktopers! [13:26] salut seb128 [13:27] dpm, and so if you search for "ubuntu-cal" in g-s what happens? [13:27] hey dholbach [13:27] dpm, I joined in the middle, but substring matching for g-s/snap query never worked, if that's the topic [13:27] willcooke, that shows me 2 results now [13:27] seb128, tout est cassé :) [13:27] hey seb128, afternoon! [13:28] dholbach, oui, snapcrap fault! ;-) [13:28] haha [13:28] for me gnome-software shows no apt/dpkg apps at all [13:28] just the snaps I have installed [13:28] Laney, v4 yes, previous no. [13:28] seb128, it did work, I have proof! :) [13:28] Laney, hence it's my sub-key only at 2k [13:28] seb128, -> https://developer.ubuntu.com/en/desktop/get-started/ [13:28] xnox, so yeah. [13:29] xnox: no neo v4 though? [13:29] Laney, not yet [13:29] dpm, maybe before snappy->snapd [13:29] they switched the vendor and rumour has it, the nfc v4 (or i guess v5 will come) [13:29] dpm, in any case it's a store issue, not a g-s one [13:29] nice [13:29] dholbach, weird, do you have a debuglog to share? [13:30] the channel is more active that I though [13:30] today is not a bank holiday in the u.k I guess? [13:30] hi seb128 [13:30] we had that on monday [13:30] hey Laney [13:30] monday was for the 1st of may right? [13:30] dpm, can you install one of the two results you have now? [13:31] today is assumption day [13:31] ascension [13:31] doh [13:32] dholbach, isn't it a vac for you? [13:32] seb128, output of strace -eopen,stat gnome-software (http://paste.ubuntu.com/16237969/) [13:32] yes it's a replacement one [13:32] I see [13:32] dholbach, no normal log? [13:32] seb128, http://paste.ubuntu.com/16237898/ [13:33] willcooke, will try in a bit, need to jump on a call, will be filing some bugs :) [13:34] dholbach, did you try to kill the running instance? that log looks like it's not starting but exiting on favor or an existing one [13:34] seb128, yes, several times [13:34] dholbach, can you kill it and get a log with --verbose? [13:34] I also removed its cache and config [13:35] Laney, did you try helping dholbach already? (I just joined and start wondering how much debugging there was before that) [13:35] dholbach, is your issue the same as dpm - in that it used to find ubuntu-calculator if you just searched for "calculator" and now it doesnt? What happens if you search for "ubuntu-cal" - does it show up then [13:35] seb128: no, I don't want to join the fight here, I just came back from lunch [13:35] Laney, k [13:35] attente was involved [13:36] willcooke, 'ubuntu-cal' works for me, my problem is that no apt/dpkg packages are shown [13:37] dholbach, can you get the verbose log? ;-) [13:37] on it [13:37] but in a call [13:38] attente, seb128: http://paste.ubuntu.com/16238226/ [13:38] dholbach: anything in /var/cache/app-info? [13:39] http://paste.ubuntu.com/16238236/ === Drac0 is now known as Guest42382 [13:39] dholbach: /var/lib/app-info too? [13:39] http://paste.ubuntu.com/16238249/ [13:39] dholbach, do you have the U8 session installed as well? [13:39] yes [13:39] :) [13:40] I wonder [13:40] could be the first fall out from the PK thing seb128? When you pull in U8 [13:41] willcooke, I was starting to think the same [13:41] dholbach, dpkg -l | grep aptdaemon [13:42] http://paste.ubuntu.com/16238284/ [13:43] dholbach, your full log doesn't have the failed to open error and seems to list quite some content [13:43] they went to stderr [13:44] they are the same as what's in the log I pasted earlier [13:44] seb128, UOS is not a vac :) [13:44] dholbach, yeah, tell me :-/ [13:44] * dholbach hugs seb128 [13:44] * seb128 hugs dholbach back [13:46] attente, dholbach, everything in that log is kind "unknown" [13:46] dholbach, when did the issue start? [13:47] let me know if there's anything else I can try to do [13:47] I don't know - I haven't used it much [13:47] seb128: are you reproducing it too? [13:47] I just tried it earlier to prepare for a demo with dpm [13:47] attente, no [13:48] dholbach: and apt update, killing g-s, starting g-s didn't do anything? [13:48] no [13:48] I was up to date [13:48] and restarted my machine even [13:49] appstreamcli search . [13:50] http://paste.ubuntu.com/16238415/ [13:55] dholbach, what ubuntu serie do you use? [13:55] xenial [13:55] can you share syslog? [13:55] what's that u8 pk change that happened? [13:56] is it supposed to remove aptdaemon or something? [13:56] http://paste.ubuntu.com/16238523/ [13:56] attente, it removes the pkgkitcompat layer [14:00] so i'm fully up-to-date. what package should i install to try to reproduce? [14:01] attente, unsure if that's something where you install a package to reproduce, that was a random guess I think [14:01] those 'failed to call' messages are probably the thing to debug [14:02] right [14:04] dholbach, ls -ld /var/lib/app-info ? [14:05] drwxr-xr-x 4 root root 4096 Apr 19 15:05 /var/lib/app-info [14:06] attente, unity8-desktop-session-mir maybe(?), if it's about unity8 [14:06] dholbach, [14:06] appstreamcli dump vlc.desktop [14:07] http://paste.ubuntu.com/16238714/ [14:07] seb128, xnox: meh, got double-booked with a Mark session on our UOS session upstart [14:07] I asked to move it, but you might need to start without me [14:08] pitti, hum, k, if you can't be there we better reschedule I think? [14:08] Laney, attente, where is the gs "kind" info coming from? [14:10] plugins set it once they figure out what kind a component is [14:10] pitti, i noticed [14:13] dholbach, do you get the same issue in a guest session? [14:13] seb128, I won't be able to try today, to busy because of UOS [14:13] s/to/too [14:14] dholbach, sorry I've no idea about that issue [14:15] normally the appstream plugin is responsible for making application objects and setting them to be desktop applications (the ones we normally show) [14:15] yeah... I don't know either [14:15] gnome-software just shows snaps right now [14:15] so it all goes down to appstream failing to load on a no such file or directory [14:15] I think you want to attack this directly and find out why you're getting those messages [14:15] right, no idea about that though [14:15] the files are there on disk and permissions are ok [14:15] appstreamcli find them [14:15] and strace doesn't show anything weird [14:16] sure [14:16] that's why more debugging is required [14:16] failed to call gs_plugin_refine on appstream: Error opening file: No such file or directory [14:16] seb128, worth reinstalling the compat layer just to see what happens? [14:16] can't do that without being able to reproduce though [14:16] :( [14:16] do we know what file exactly it's trying to open? [14:16] that is the debugging [14:19] you need to trace gs_plugin_refine_app in src/plugins/gs-plugin-appstream.c and find out where the error comes from [14:20] dholbach, can you strace -f gnome-software and share the log? [14:22] the log is like 30M already and the UI didn't even come up yet [14:23] dholbach, ctrl-C, gzip? [14:23] or do it again but ctrl-C after 10s [14:23] Laney: A friendly reminder of that bug... [14:24] No need [14:25] seb128, http://daniel.holba.ch/temp/log.gz [14:26] dholbach, danke [14:26] de rien [14:31] dholbach: any chance you're on a metered wifi connection? [14:31] attente, metered? [14:31] I'm on wifi at home [14:31] dholbach: just speculating if the refresh is doing something differently in case you're on 3g [14:32] ah ok [14:32] but guess not [14:32] no, that shouldn't be the case [14:33] dholbach, is /var/lib/app-info/yaml/archive.ubuntu.com_ubuntu_dists_xenial_main_dep11_Components-amd64.yml.gz pointing to a valid file? [14:33] and how big is that file? [14:33] ah no [14:33] points to /var/lib/apt/lists/archive.ubuntu.com_ubuntu_dists_xenial_main_dep11_Components-amd64.yml.gz which does not exist [14:33] ah... I guess I moved from archive.u.c to de.archive.u.c at some stage [14:33] maybe that's why(?) [14:34] did you try to click on the "refresh" button in g-s? [14:34] on the top left [14:34] attente, Laney, when are those /var/lib/app-info/yaml links updated? [14:35] where is the refresh button? [14:35] if you go to the updates tab [14:35] should be during an apt update. refreshing with the button on the updates tab should trigger it though [14:35] in the top left [14:35] dholbach, can you copy the log of a sudo apt-get update as well? [14:35] see if you have warnings/errors there [14:36] I wonder if you have that bug where an invalid repo is breaking the update [14:36] the update is taking ages [14:36] in g-s [14:36] now it says "it's all up to date" [14:36] and the broken symlinks are still there [14:36] k [14:37] sudo apt-get update | pastebininit :p [14:37] http://paste.ubuntu.com/16239267/ [14:37] iz mvo's fault! [14:37] dholbach, does it fix it if you disable the spotify repo and do another update? [14:37] let me see [14:39] the symlinks are now up to date [14:39] and g-s works I guess? [14:39] yes, it took a while to load [14:39] but it looks good now [14:39] k [14:39] so we figured it out! [14:40] so is appstreamcli refresh not getting called when apt update fails? [14:40] dpm, issue found, tldr: apt repos with weaker security break appstreamcli's data update, g-s gets confused over missing data [14:40] attente, seb128: I'm sure that MANY users see this issue [14:41] Laney, ^ [14:41] yeah, ,likely [14:42] can't appstreamcli just take the apt lists which are good and up to date? [14:42] apt is not calling appstreamcli [14:43] because it thinks the update failed [14:43] mh [14:43] told yeah, iz mvo bog! :-) [14:43] lol [14:44] yeah->ya [14:44] g-s could also give some clue that the index refresh is failing [14:45] yeah [14:45] saying what? [14:47] "the package index update failed, check your sources?" [14:48] with maybe with a detail showing eventual E: [14:48] "E: The repository 'http://repository.spotify.com stable Release' does not have a Release file." [14:48] dunno [14:48] any error would be better than letting you hitting refresh again and again and not telling you it's failing [14:53] i don't understand why g-s wouldn't work when appstreamcli search . was clearly working at the time [14:53] I imagine there's something that could be done, like [14:54] - g-s knows there's no GsApps to display [14:54] - try to refresh [14:54] - if that fails, display some message [14:54] It sort of intersects with the first run work [14:54] Clearly needs design [14:54] and also clearly is not Ubuntu specific [14:54] right [14:55] In this case we should look for a solution that gets the index refreshed [14:55] attente: appstream-glib accesses the .yaml files itself [14:55] seems like dholbach had indexed in the apt dir [14:55] should the appstream plugin try to do a refresh itself? [14:55] .yml.gz, whatevs [14:55] why are we not using those directly? [14:55] it doesn't use that same database [14:56] and relying on symlinks that might be wrong/outdated [14:56] but the /var/lib/app-info/yaml ones are symlinks, not different db [14:56] they just pointed to the wrong files in his case [14:56] I didn't say that they weren't symlnks. [14:56] I said that appstreamcli search is looking at a different thing to appstream-glib. [14:57] ah ok [14:57] Usually they coinciide (so it is a useful test), but not always [14:57] And I would imagine it's because appstream-glib doesn't want to know about apt's private directories [14:57] Probably it's valid for other things to put appstream in that place [15:01] pitti, seb128: i'm double booked too.... shall we move/postpone the upstart user session ? [15:01] hm, meh [15:01] xnox, pitti, +1 [15:01] I guess we have to [15:02] Laney, attente, unsure if apt failing on invalid report is a bug or a feature? it could skip over those and get the update done anyway [15:03] need to check with mvo when he's around [15:05] maybe it's actually more appropriate for the appstream plugin in gnome-software to actually refresh the cache itself instead of depending on apt to do it [15:57] pitti: are you all set for your UOS session? [15:57] mhall119: sorry, no, double-booked with a Mark session on the sprint here [15:57] is there someone else who can run it? [15:59] I don't actually know much about the session upstart rationale/integration, so if seb128 and xnox want to go along, please do [15:59] otherwise we need to postpone [15:59] sorry, sprint craziness [15:59] pitti, mhall119, xnox is also double booked and I didn't prepare much, better to postpone [16:00] pitti: can you do it at 1800 UTC? [16:04] mhall119: WFM, if seb128 can do that too [16:05] seb128: ? [16:05] mhall119, pitti, no, sorry, I've dinner plans that conflict :-/ [16:05] I guess we'll talk about it next week then [16:05] but feel free to do it without me [16:05] pitti: can we schedule it for ubuntuonair.com next week? [16:05] this is all desktop-y work, not much point without desktop guys aroud [16:06] we probably should just do a three-day sprint (virtual or real) for this [16:06] Laney, seb128, xnox and me perhaps [16:07] that would work as well [16:07] let's talk next week [16:07] * Laney isn't sure of the scope yet [16:08] seb128, Laney: who would actually have some knowledge about how the current system works? [16:08] which part? [16:08] i. e. rationale for moving so much stuff to upstart instead of dbus activation, etc. [16:09] Laney: well, everything :) [16:09] i .e. how the session starts under upstart, and converting gnome-session to upstart jobs, etc. [16:09] stgraber drove most of it I think [16:09] tedg? [16:09] I know how but can't really speak to why [16:09] just "make things event driven" probably [16:10] Laney: ah, so mirroring the phases of g-sessoin? [16:10] Yes, event driven was important, but also having logging and control. [16:10] We had for instance every indicator implementing logic for managing restarts for debugging. [16:10] (previous Mark session is running over, but unsure when our's starts) [16:11] pitti, tedg yes [16:11] Where we were able to remove all that code and just use Upstart [16:12] DBus activation also has some interesting race cases that we never solved. Those are largely fixed by systemd, but I'd not recommend dbus activation without systemd managing the bus. [17:07] byeeEEee [17:07] see ya Laney [19:44] gnight all === hikiko is now known as hikiko|off