[01:01] robert_ancell: hi, in case you didn't hear, I guess they want you to upload your ubuntu-welcome app soon :) [01:07] jbicha, thanks for the suspend-on-power fix BTW [01:08] duflu: maybe we'll bring it back next year :) [01:08] jbicha, if it's configurable, sure. :) [01:08] AFAIK it's not yet user-configurable for the login screen [01:09] right === ecloud is now known as ecloud_wfh [06:34] good morning desktoppers [06:37] Salut oSoMoN [06:38] salut jibel [07:11] good morning [07:13] bonjour didrocks [07:14] salut oSoMoN [07:15] salut didrocks oSoMoN [07:15] salut seb128 [07:15] Bonjour didrocks seb128 and all [07:15] lut jibel [07:15] good morning desktopers [07:16] salut seb128, jibel ! [07:37] andyrock, hey [07:37] andyrock, I'm backporting your udisks segfault fix to bionic, just as a fyi [07:38] hey hey [07:38] thx! [07:40] Morning oSoMoN, jibel, didrocks, seb128, andyrock [07:40] hey duflu [07:40] hey duflu! [08:00] morning [08:00] Morning willcooke [08:00] duflu, koza - are you doing the BT meeting today? [08:01] The BT meeting moved again. Depends if koza wants it [08:01] Or seb [08:01] Lovely timezones [08:01] Seb is out this morning for a little while [08:01] I have nothing [08:02] Alright. In that case I am putting the camera away. FYI willcooke, koza [08:02] & jibel ^ [08:02] duflu, ack. [08:02] * willcooke looks like crap anyway [08:02] moin [08:02] Morning Laney [08:03] The kids have given me some illness [08:03] hi Laney [08:03] Well, their immune systems get a good education [08:03] at least [08:05] hey willcooke, Laney [08:05] willcooke: I guess we can make a club of "be sick, thanks to the kids" this week :) [08:06] :DD [08:06] at least, you know how to do it, being sick during work days [08:06] for me, that was during extended week-end [08:06] less attractive :p [08:06] * didrocks needs to plan better next time [08:07] urg, that's the worst [08:07] actually, being sick on holiday is the worst [08:08] I guess I had my share of that since December. I should ask for reimbursement to "life"! :) [08:12] /o\ === niko is now known as upine === upine is now known as upmine [09:10] back [09:10] morning seb128 [09:10] hey willcooke, how are you today? [09:10] got a bit of a cold from the kids [09:11] One of thm was up a lot last night coughing [09:11] so tired as well [09:11] yay :) [09:11] wb seb128 [09:11] thx [09:12] willcooke, welcome to the club! [09:12] ha [09:27] i think there is a problem with the fontconfig cache on the live images [09:28] Hi ali1234, is that a new problem? In the last few days? [09:28] no [09:28] it's been around for a long time [09:28] it just took a long time to track it down [09:28] How is it manifesting itself? [09:29] basically, the files in /var/cache/fontconfig are not accepted, so the font cache gets rebuilt in ~/.cache/fontconfig for every installed font [09:29] ah. jibel - that could be a candidate for slow start up? [09:29] this takes like a minute or more and delays getting to the live desktop or installer [09:29] right, it's the slow startup [09:29] good finding if that's the issue! [09:30] looks like we owe you some beers ali1234 [09:30] if you update a font package it rebuilds the font cache and then the problem goes away [09:30] so it's only the cache included on the iso? [09:30] yes [09:30] do you know what it doesn't like about it? [09:30] a little bit [09:30] at xubuntu, we thought the problem was caused by bluez and we have a bug tracking that [09:31] that would be weird [09:31] last night i figured out bluez was a red herring: https://bugs.launchpad.net/ubuntu/+source/bluez/+bug/1754836 [09:31] Ubuntu bug 1754836 in bluez (Ubuntu) "Xubuntu: ERROR:dbus.proxies:Introspect error on org.bluez:/org/bluez: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.TimedOut: Failed to activate service 'org.bluez': timed out (service_start_timeout=25000ms)" [Critical,Confirmed] [09:31] see my comments at the bottom of that [09:32] i dont know whether to change that bug or make a new one, if a new one i don't know what to open it against :) [09:32] all the info is there though [09:33] one interesting thing is that ubuntu seems to rebuild the font cache during the installer so it is only affected in the live image. however, xubuntu does not so it happens again on the user's first login to the installed system [09:33] willcooke, could be but the bootchart should have shown IO generated by the creation of the cache. [09:34] jibel, sounds like it could well be a culprit, so interesting that the boot chart doesnt show it. [09:34] unless it stops too early [09:34] oh, yeah [09:34] it definitely gets rebuilt - just check ~/.cache/fontconfig. it should be empty except for maybe 1 file for the firefox emoji font [09:35] /var/lib/dpkg/info/fontconfig.postinst looks correct at first look [09:35] in the "trigger" case [09:35] well there are caches in the system location [09:35] but they are invalid [09:35] we should look for "fc-cache failed" in iso build logs [09:36] where is the iso build log? ;) [09:36] in case the invalid caches are from a fc-cache error [09:36] this is why I didn't put a link, I don't know :p [09:36] Laney, ^ help? :) [09:36] or maybe jibel knows [09:36] if anyone knows the structure of the cache files, i did a binary diff between working/non-working [09:36] details on the bug [09:37] the best case would be fc-cache generating that errors with the logs in /var/log/fontconfig.log [09:37] it's over my understand of fontconfig details [09:37] if it only generates invalid caches and don't error out… urgh, harder to understand [09:37] seb128, https://people.canonical.com/~ubuntu-archive/cd-build-logs/ubuntu/bionic [09:37] jibel, thx [09:37] then look at the build id in the first lines of the log [09:38] https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu/bionic/ubuntu/+build/128542 in this case [09:38] no fc-cache error line :/ [09:38] nothing useful in that log [09:38] Setting up fontconfig (2.12.6-0ubuntu1) ... [09:38] Regenerating fonts cache... done. [09:38] so the trigger happily passed [09:38] Processing triggers for fontconfig (2.12.6-0ubuntu1) ... [09:39] yeah [09:39] well not a surprise if a cache file is generated [09:39] and nothing below… [09:39] yeah [09:39] Setting up fontconfig (2.12.6-0ubuntu1) ... [09:39] Regenerating fonts cache... done. [09:39] yeah :) [09:39] /var/log/fontconfig.log ? [09:39] is that purged from the iso? (I don't have one handy [09:39] ) [09:39] in case that can help [09:40] as we run fc-cache in verbose mode and redirect to it… [09:40] didrocks, i'll have a look i'm testing beta2 [09:40] * didrocks crosses fingers log will help [09:42] Debian are fingering Noto as the problem, is that right? [09:42] no [09:42] Noto is really slow to generate because it is huge [09:42] but the issue is it shouldn't be getting generated because the cache is supposed to already be there [09:44] ahhhh [09:47] https://cgit.freedesktop.org/fontconfig/tree/src/fccache.c#n996 is the list of reasns why fontconfig will reject a cache file [09:48] the ifdef jumps out at me as something which could be represented by that mystery field [09:50] hm, latest iso doesn't boot if I go directly to the live session from syslinux [09:52] seb128: I got as far as noticing gsd-xsettings is slow when doing some fontconfig thing, this sounds further along than I got [09:52] or what did you want my help with? [09:52] Laney, getting the url of the iso build log, but jibel had it [09:53] ok [09:53] https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu/bionic/ubuntu/ this is a good index [09:53] thx [09:58] didrocks, http://paste.ubuntu.com/p/7YxNf7MsvY/ and it says existing cache is valid [10:00] it might be worth sending an email to the upstream fontconfig list with the cache bin diff [10:00] they might be helpful === pesari_ is now known as pesari [10:00] ali1234, would you like to do that? [10:01] jibel: :/ [10:01] ali1234, do you know which job trigger fc-cache on log in? [10:04] one day I'll dd an iso to my hard drive :/ [10:11] The GLib update doesn't look that interesting [10:11] interesting finding [10:11] gtk+3.0, probably is ok [10:11] seb128, not major update, minor debian revision [10:11] Laney, if you want, just ack and I'll upload in a ppa [10:12] right, the glib change seems a bit late in the cycle [10:12] I don't want to mess up with your iso generations or similar [10:12] seb128_: good catch! [10:13] seb128_, do you know what triggers fc-cache on log in? [10:13] no [10:13] I cannot find anything [10:13] likely fontconfig which when being used validates the cache [10:13] so apps themself [10:13] it breaks stuff too https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=894763 [10:13] Debian bug 894763 in libglib2.0-0 "libglib2.0-0: symbol lookup error: /usr/lib/arm-linux-gnueabihf/libgobject-2.0.so.0: undefined symbol: g_date_copy" [Grave,Open] [10:13] jibel: gsd-xsettings does something with fontconfig [10:14] and that is part of the startup [10:15] LocutusOfBorg: you can upload to the queue, it'll probably get accepted after the beta [10:15] bbiab, need to change location [10:15] and please make sure to do your changes in bzr too if you don't mind === muktupavels_ is now known as muktupavels [10:16] might want to check if jbicha_ hasn't started this work too, would be annoying to duplicate [10:16] ok nice === Beret- is now known as Beret [10:17] I think he is *happy* to stop merging it :) [10:18] ok, if you know that based on some evidence [10:18] last time he asked me to do the merge :) [10:19] I did a gtk2 and he asked me to also merge gtk3 [10:19] and after that also glib2 :) [10:19] he asked you to merge this glib upload? [10:20] nope [10:20] I'm talking about previous uploads [10:20] ok [10:20] done gtk3 in queue [10:30] * duflu closes eyes and logs off before finding any more mutter issues === oSoMoN_ is now known as oSoMoN [10:48] k, back at my desk/online [10:48] didrocks, ali1234, sorry, had to drop from IRC for a bit, was there anything else said on the topic meanwhile? [10:49] i read that bug report. seems to exactly describe what i am seeing [10:49] yeah, what I though as well from your binary diff/explanations [10:49] seb128: nothing more, no [10:50] ali1234, do you want to provide some details on that upstream report or should I? [10:50] jibel attached the log that was asked for [10:51] ah FC_DEBUG=16... i think i tried every other possible number :) [10:52] ah, nice [10:55] ah i think i understand this now from that log - "cache checksum 1522818676.601423551 dir checksum 1522818676.0" [10:56] so it is actually the font directory that has had the timestamp changed by going through squashfs [10:59] and that's why the mystery field changes to zero [10:59] right [11:00] didrocks, do you want that as a launchpad bug report or just IRC ping? [11:00] $ dpkg-deb -c libsysmetrics-dev_1.0.3_i386.deb | grep pc [11:00] -rw-r--r-- root/root 258 2018-04-03 09:51 ./usr/lib/i686-linux-gnu/pkgconfig/sysmetrics.pc [11:00] didrocks, the arch triplet is i386-linux-gnu no i686 [11:00] it makes ubuntu-welcome fail to build on i386 since it doesn't find the .pc [11:02] that looks like DEB_HOST_GNU_TYPE or something was used instead of DEB_HOST_MULTARCH [11:02] seb128: let me just fix it, no need for paperwork [11:02] yes [11:02] fixing right away [11:02] didrocks, thx [11:02] with right spelling of multiarch [11:02] :) [11:03] nw! thanks for pointing it! [11:03] Laney: :) === marathone is now known as marathon [11:10] seb128: fixed, built, tested and uploaded. You can take and build latest from github if you want to take that one instead of copying on your system :) [11:15] didrocks, thx, the build is from a ppa so I'm just going to wait for it to be in the archive and retry [11:15] my bionic system is amd64 :p [11:16] argh, I needed to keep one GNU style for build dir due to dh_golang… [11:16] oh, true, you reinstalled :) [11:18] seb's specialist dual boot system [11:19] :) [11:35] ok, the ppa ran worked this time (to avoid an useless upload), pushing [11:35] didrocks, :) [11:36] didrocks, thanks for fixing! [11:36] so yeah, I used the GNU type due to dh_golang and didn't think about build dir being different than install dir (don't understand exactly why) [11:36] yw! [11:36] the 2 tools, being debian specific, should use the same triplet format IMHO [11:37] yeah, I don't know either [11:37] Travis is having large roundtrip anyway to start a build, so I was able to slip that in between 2 builds :) [11:40] Is there a good LP bug for the fontconfig issue? I'd like to tag it rls-bb-incoming at least. I can open a new one based on the upstream one if needed [11:40] i don't know of any others [11:41] ali1234, the bluez one was about bluez or really about the slow boot/fontconfig? [11:41] in which case we just need to rename it [11:41] seems the timeout on bluez is due to slow boot [11:41] willcooke, we had one about slow boot on our side already which we can maybe use? [11:41] i opened it when we thought the slow boot was caused by bluez [11:41] there is https://bugs.launchpad.net/ubuntu/+source/casper/+bug/1749546 [11:41] Ubuntu bug 1749546 in casper (Ubuntu) "Booting a live session is slow " [Undecided,New] [11:42] reusing that seems fine to me [11:42] ah nice, let's reuse that one [11:42] ta [11:42] willcooke, thx [11:42] I'll update etc [11:42] it's great to have at least an idea where this may be coming from :) [11:42] willcooke, I'll reassign to fontconfig [11:42] jibel, I'm doing that now [11:43] I'm already assigned to it, it's fine [11:43] but if you want ... [11:43] :) [11:43] you can remove casper [11:44] oh you probably don't mean #1750846 [11:45] no it's different [11:46] is it though [11:46] i have a feeling that the fontconfig thing may cause a knock-on effect where other stuff times out [11:46] especially dbus - hence the bluez thing as well [11:46] +1 [11:46] i have no evidence for this though [11:47] sounds plausible [11:47] one easy way to figure out, fix the fontconfig issue and see if that resolves the other ones :) [11:47] :) [11:47] anyone want to own this one? [11:47] Would like to see if we can get it fixed for beta [11:47] do you know where is the team mapping list? [11:47] try at least [11:48] I wonder who owns fontconfig [11:48] be calm [11:48] I've been assigned to the bug I just referenced for some time, I'm working on it [11:48] wfm, thanks Laney! [11:49] I found out last night / this morning that gsd-xsettings was blocking the start up in fontconfig stuff [11:49] thanks Laney [11:49] so ali1234 coming in with more analysis is good timing [11:49] Laney, well it was just that if the issues turns out to be in fontconfig and that foundations own that then we might want to bounce to them at this point [11:50] but your call at the end [11:50] j_bicha was the last one to touch it ;)) [11:51] as for who owns the package, look at the subscribers [11:51] mapping is generated from that [11:51] desktop-packages and foundations-bugs are subscribed [11:51] so we both own it? :) [11:51] seems so [11:51] quick, press the unsubscribe button [11:52] I was going to say :p [11:52] ahah, first one unsubribing wins :) [11:55] * willcooke <- sandwich [11:56] willcooke, enjoy! [11:56] * seb128 just had some takeway pasta, that was good === jbicha_ is now known as jbicha === darkxst_ is now known as darkxst [12:54] can't think of a better way to fix that fontconfig thing than ignoring the nanoseconds if it's 0 [12:54] ali1234: got any ideas? [12:55] I tried a patch to do that, which works [12:55] boot speed still isn't exactly amazing though [12:56] ubuntu@ubuntu:~$ systemd-analyze [12:56] Startup finished in 24.412s (kernel) + 1min 23.971s (userspace) = 1min 48.384s [12:56] graphical.target reached after 28.680s in userspace [12:57] however it does fix the gsd problems [13:04] hmm that second try was way better, weird [13:05] ubuntu@ubuntu:~$ systemd-analyze [13:05] Startup finished in 12.572s (kernel) + 23.512s (userspace) = 36.084s [13:05] graphical.target reached after 23.473s in userspace [13:55] willcooke, seb128: libreoffice 6.0.3 will be out this week (currently in RC2, for which we have a snap in the candidate channel btw), and we have 6.0.2 in bionic today. How do you feel about pushing the update (supposedly bug fixes only, I will double-check that)? Or better to SRU after release? [13:55] hum [13:56] I dont suppose we've really given .2 a proper shakedown since not that many people are using it. [13:56] So shipping .3 isn't a massive problem [13:57] What do we do with, say, Firefox in this situation chrisccoulson? ^ [13:57] oSoMoN, I would push it to bionic, we can keep it in proposed for a while and decide to turn it into a SRU from there if we prefer [13:57] we're still ~3 weeks from release, we still do bug fix updates :) [13:59] willcooke, I normally just upload it and somebody approves it - it's usually already in the security pocket for other releases by the time I do that though [14:00] chrisccoulson, hey again, did you see my questions about those translations earlier? ;) [14:00] seb128, yeah. The answer is yes [14:00] jbicha, btw bug #1759538 has been ack-ed so you can make g-s recommends it with the next upload [14:00] bug 1759538 in bolt (Ubuntu) "[ffe] Install bolt by default for thunderbolt devices support" [Medium,Triaged] https://launchpad.net/bugs/1759538 [14:00] chrisccoulson, thx [14:01] good, so I'll double-check the upstream changelog and diff to make sure they didn't sneak new features in, and will coordinate with ricotz to prepare a 6.0.3 upload [14:02] thx, let me know if/when you need sponsoring [14:03] that'll most likely be early next week [14:06] k [14:07] looks like ipv6 came back, brb! [14:07] feels modern [14:08] is there a way to make softwares not removable in gnome-software? [14:08] bug #1756788 seems annoying [14:08] bug 1756788 in gnome-session (Ubuntu) "Removing "Startup Applications" in "Ubuntu Software" makes the system unable to launch GDM" [Undecided,New] https://launchpad.net/bugs/1756788 [14:08] I guess it removes gnome-session, DOH [14:08] yes, you can make it compulsory-for-desktop or whatever the thing is [14:09] k, thx [14:12] * ricotz wonders what openjdk-11 packages versioned as 10~46-4ubuntu1 are about [14:13] hey desktopers [14:17] hey ricotz [14:17] moin ricotz [14:18] don't think the desktop team particularly knows about java, better to ask the uploader? [14:18] +1 [14:18] seb128, Laney, hi [14:19] I guess the uploader already removed the source packages from the archive again [14:19] so hard to tell where it came from, I assume d_oko did [14:20] https://launchpad.net/ubuntu/+source/openjdk-lts/10~46-4ubuntu1 [14:20] jbicha, thx :) [14:21] Laney, oohh [14:21] Laney: i don't know, but can't you fix the squashfs, rather than patching fontconfig? [14:21] Laney, aptitude failed to show me the changelog [14:23] ali1234: how? [14:24] well the most obvious way would be to build it on a filesystem that doesn't have nanos, so that the cache remains valid after being copied through squashfs [14:25] manually stripping the nanos from /usr/share/fonts (recursive) before generating the cache would also work [14:26] chrisccoulson, hi, are you going to upload tb 52.7.3? [14:26] i suppose that would mess with the deb file timestamps [14:27] chrisccoulson, ah sorry, I thought they tagged it [14:28] not sure LP people would be happy with that, and I'd rather a more generic solution anyway [14:28] so either make fontconfig tolerate this situation or change squashfs [14:30] the way i see it, the timestamps are going to be truncated at some point. so you'd just be making it happen earlier [14:30] it wouldn't change the effective timestamp on any files managed by packages [14:32] but i have no idea how the image builders work [14:34] ricotz: did you want to try adding/moving the Thunderbird app icons LP: #1682455 ? [14:34] Launchpad bug 1682455 in thunderbird (Ubuntu) "thunderbird not available in GNOME Software" [Medium,Triaged] https://launchpad.net/bugs/1682455 [14:39] jbicha, tweaking the stable packages is more a thing for chris [14:39] jbicha, did you find a way to confirm that this works properly with the ff60 betas? [14:40] ask Lan_ey but I don't think appstream data from PPAs is usable in the GNOME Software app yet [14:41] so the only way I know to know for sure is to upload to bionic and see [14:50] jbicha, I could take a look if chris would push his packaging changes [14:50] should be similar to the firefox-esr changes I made [14:52] thanks [14:52] jbicha, e.g. https://paste.debian.net/plain/1018499 === ayan_ is now known as ayan === ayan is now known as Guest81868 === Guest81868 is now known as ayan [14:59] ricotz: maybe you could talk Mozilla into providing an ESR snap so you don't have to maintain that PPA :) [15:02] jbicha, I am not sure that people wanting firefox-esr are using snaps [15:04] it might not be a bad choice compared to a PPA [15:05] I'm just guessing, but I think we'll be hesitant to allow firefox-esr in to the Ubuntu archives [16:34] * Laney waves from the train [16:34] new 4g wifi thing works well! === simonquigley is now known as tsimonq2 [16:42] Laney, cool! =) [16:42] * Laney high fives xnox [16:43] I should probably turn some things off so I don't burn through all the data tho [16:44] * Laney ticks "restrict background data usage" [16:46] Laney, i wonder if unattended-upgrades respects the "restrict background data usage" is that an Gnome-shell thing, or a Network Manager thing? [16:52] xnox: It sets metered on the dbus interface I think, and GNetworkMonitor exposes that (g_network_monitor_get_network_metered()) [16:52] so it probably doesn't unless it uses one of those APIs [16:53] Laney, session or system bus? [16:53] NM is on the system bus [16:54] do something like busctl --system introspect org.freedesktop.NetworkManager /org/freedesktop/NetworkManager [16:56] although for me that is currently 4 which is apparently GUESS_NO [17:19] ok now it's 1 after reconneccting [17:32] xnox: https://paste.ubuntu.com/p/KBH6G3ZBGf/ [17:33] I think android and ios have a thing to set that automatically on tethering btw [17:42] goodnight from newark northgate [17:49] night all === sarnold_ is now known as sarnold === pstolowski is now known as pstolowski|afk === amano_ is now known as amano