[00:20] GunnarHj: interesting read: https://girinstud.io/news/2018/02/ibus-hangul-and-compose-key-the-incredible-journey-of-a-simple-patch/ [00:35] jbicha: Yeah.. Bug hunting may be truly frustrating sometimes. [00:55] jbicha: Btw, I can't reproduce that issue with compose:caps on artful. [00:57] are you using ibus? [01:55] jbicha: IBus is the selected IM framework, if that's what you mean. [05:54] Hey guys, I accidentally botched my radeon config file, and I'm stuck in Wayland and csan't get to the file due to its root restrictions [05:54] Any advice on how to restore the file to its defaults? [05:54] I know how to fix it, it's just a matter of getting to the root file [07:18] good morning desktoppers [07:18] Morning oSoMoN [07:24] Hmm Redback in the letterbox [07:24] Yep, it's summer [07:24] hey duflu [08:20] good morning [08:38] Morning didrocks [08:39] hey duflu, how are you? [08:39] didrocks, I am imperfect. But this made it a good day: https://www.youtube.com/watch?v=bCc16uozHVE [08:40] How are you didrocks? [08:40] Morning seb128 [08:40] good morning desktopers [08:40] hey duflu didrocks [08:40] salut seb128 [08:51] how to debug snaps that don't launch? [08:56] good morning didrocks, seb128 [08:56] "desktop-launch: not found" [08:58] Good morning all! [08:58] seb128: How about starting the day with a nice regex? :) [08:58] It's the patch at bug #1688994 [08:58] bug 1688994 in pkgbinarymangler (Ubuntu) "dh-translations: cannot build POT without intltool" [High,In progress] https://launchpad.net/bugs/1688994 [08:58] salut oSoMoN [08:59] hey GunnarHj, do you ever stop working/sleep? ;) [08:59] lut oSoMoN [08:59] tjaalton, that would be an issue :p [08:59] seb128: It happens. But don't tell anyone. [08:59] tjaalton, but usual way would be to try starting manually and see if it displays an error on stderr, then you can try to snap run --shell [08:59] that gives you a shell inside the snap [08:59] ok [09:00] GunnarHj, :) [09:00] GunnarHj, I'm still doing morning catching up but I look at that in 15 min or so [09:00] seb128: Great! [09:02] tjaalton, inside the snap when you use --shell the env is not fully set, e.G you can't start the command directly, usually you want to use /snap//current/command-.wrapper [09:04] seb128: that's what I tried and it said "desktop-launch: not found" [09:04] willcooke, seb128, hangout? [09:04] darktable [09:04] tjaalton: in your snapcraft.yaml, do you have the desktop part? [09:04] where is that? [09:05] this is your snap, correct? [09:05] no [09:05] you authored snapcraft.yaml [09:05] ah [09:05] installed from the store, darktable 2.4.1 [09:05] maybe the snapcraft author didn't pull the correct part [09:05] first, do as seb told, using snap run --shell [09:05] tried to launch from the store page [09:05] and try to find in $SNAP desktop-launch [09:06] sorry duflu, didnt open my calendar and got sucked in to email [09:06] looks like I can't access anything while on the shell [09:06] you should [09:06] or you are not in $SNAP [09:07] "can't open directory /share/locale" [09:07] ok, the snap is probably busted, where to file bugs? [09:08] flexiondotorg: is there a field people can use to contact snap's author? ^ [09:09] duflu, willcooke, sorry, I still didn't get used to that one being shifted to be one hour earlier ... do you need me to join? [09:09] nah [09:09] if so give me a few minutes [09:10] tjaalton, email to the snap author? [09:10] there is no unique location like launchpad for snap reports [09:11] bah [09:11] ok [09:17] heh, is there no limit to how many instances of a snap app can be launched? installed wallpaperdownloader, and have a few copies running : [09:17] didrocks Yes, the contact field which is exposed via `snap info` and the store page. [09:20] flexiondotorg, hey, do you know if that snap is known to be buggy? [09:25] seb128 I've not tested darktable myself. I'll take a look when I'm back at work tomorrow. [09:45] ok, people who like regexps and maybe perl, how does https://launchpadlibrarian.net/356033584/pkgbinarymangler-configure.ac.debdiff look to you? [09:45] it's to catch cases where " AC_SUBST([GETTEXT_PACKAGE], [name])" is used [09:46] instead of " GETTEXT_PACKAGE=" [09:46] was there really a "like" in your sentence? :p [09:46] the regexp looks fine to me but I wouldn't say no to have somebody else to comment [09:46] didrocks, oh, I see you raise your hand, thanks for stepping up :p [09:46] there was indeed a like… [09:46] so doesn't match ;) [09:47] haha [09:47] hum, I wonder if it shouldn't be matched in one expression instead of 2 [09:50] you mean? [09:51] .*GETTEXT_PACKAGEname [09:51] ? [09:51] where somethingcomplex is "=" or "],/s*"? [09:51] \s [09:52] yeah, I found an expression, but it's way less readable [09:52] so, let's use the or [09:53] my only remark is that I would use \S+ for both [09:53] (first and second expression) [09:53] to ensure $1 isn't empty [09:53] and expand [09:53] the rest looks good to me [09:53] GunnarHj: ^ [09:56] didrocks, thanks for having a look [09:57] didrocks: \S+ is in both. [09:58] Or did I misunderstand you? [10:00] GunnarHj: not in the debdiff I saw at least: + $domain = $1 if /^GETTEXT_PACKAGE\s*=\s*(\S*)/ or [10:00] \S*, not \S+ [10:02] didrocks: Ah, see now. Actually there are several other regexes in the script which use \S* as well. Is it motivated to make that change all over? [10:03] didrocks: (i.e. I didn't introduce \S*) [10:04] GunnarHj: just have some consistency, either use \S* for both regexp or \S+ [10:04] I wouldn't change the existing code [10:04] so yeah, maybe use \S* for the new line? [10:04] so that $1 is consistent between packages having GETTEXT_PACAKGE= or AC_SUBST() [10:05] wfm [10:05] seb128, didrocks: Ok, I'll upload a new patch with that change. [10:05] thx didrocks & GunnarHj [10:05] yw! [10:07] building webkit… [10:08] it's I cold day, I can understand :) [10:08] true [10:09] my laptopt isn't warm enough yet though :p [10:09] and I'm sure I'll have a gjs build after the webkit one… [10:11] :) [10:11] what bug are you chassing? [10:12] just building the whole stack for GNOME design to test the Shell with volume overdrive [10:13] on jhbuild [10:16] seb128: New patch uploaded. Can you sponsor please? [10:18] GunnarHj, yes, I guess you tried it on packages using both syntax to verify it works as it should? [10:18] didrocks, I see, fun [10:19] seb128: I tested the loop via a throw-away script and a fake configure.ac. Think that's sufficient. [10:20] right [10:20] thx for fixing that [10:20] Tested both syntaxes, of course. [10:20] I had it on my backlog for a long time [10:21] seb128: You provided the real solution. I just did last bit. :) [10:25] :) [10:26] GunnarHj, oh btw I don't remember now, you said that the remmina upload/tweak worked? [10:26] seb128: Yes, it did. [10:27] good :) [10:31] seb128: I'm also looking at bug #1730793. It's a really bad one - the .po files seems to not have been imported to LP in 7 years. [10:31] bug 1730793 in util-linux (Ubuntu) "Translations not imported from upstream" [Undecided,New] https://launchpad.net/bugs/1730793 [10:32] GunnarHj, is that only some locales? do we strip the translations/does it lead to have that software not translated at all for 7 years (we would have noticed no?)? [10:33] seb128: There are translations in LP and in the langpacks, but they are dated. [10:34] ah [10:34] let me know if you figure out anything or if you need help [10:35] Wonderful. I can't fix one part of mutter because another part is too bad at math to support the fix. Will have to do a separate fix... [10:35] Tomorrow. [10:35] Later... [10:35] night duflu [10:35] seb128: The latest build log claims that the tarball util-linux_2.30.2-0.1ubuntu1_amd64_translations.tar.gz was created. I'd like to see what it contains, but don't remember where to look. [10:43] GunnarHj, wait [10:46] GunnarHj, doing it with lp apis worked before [10:46] I had that in my notes [10:46] $ lp-shell [10:46] ubuntu = lp.distributions['ubuntu'] [10:46] uploads = ubuntu.getSeries(name_or_version='bionic').getPackageUploads(name='util-linux', version='2.30.2-0.1ubuntu1 ', custom_type='raw-translations') [10:47] for upload in uploads: [10:47] print(upload) [10:47] but that doesn't work on this one [10:49] seb128: There were actually a couple of SRU uploads after the latest bionic upload. Can it have been replaced with any of those? [10:49] I don't think so [10:49] maybe try asking cjwatson on #ubuntu-devel [10:50] seb128: Ok. [11:05] seb128: Did you see Colin's link at #ubuntu-devel? Anyway, I checked [11:05] https://launchpad.net/ubuntu/bionic/+upload/16957223/+files/util-linux_2.30.2-0.1ubuntu1_amd64_translations.tar.gz [11:06] GunnarHj, I did [11:06] and found that it contains the updated .po files. [11:09] GunnarHj, my lp-shell line works as well, I just had an extra space in the version I didn't notice :/ [11:10] seb128: Maybe there is a template confusion. There is a util-linux-locales in the tarball (the name of the package which contains the translations at Debian), but at https://translations.launchpad.net/ubuntu/bionic/+source/util-linux/+pots/util-linux/+admin the template name is util-linux. [11:11] seb128: s/util-linux-locales/util-linux-locales folder/ [11:12] GunnarHj, how is named the .pot? [11:12] seb128: util-linux.pot [11:12] so that's the correct domain [11:14] seb128: So why isn't it imported? [11:15] seb128: It seems that the package builds fine. [11:15] GunnarHj, https://translations.launchpad.net/ubuntu/bionic/+source/util-linux states " This source package is sharing translations with Util-Linux-ng head series. " [11:15] that might be the issue [11:16] it's importing translations from https://code.launchpad.net/~vcs-imports/util-linux-ng/trunk [11:16] which fails to import since 2015 [11:17] GunnarHj, one way would be to fix that import, another would be to disable the sharing and redo a source upload so files from the package get imported [11:18] seb128: The latter seems to be easier. [11:18] let's do that then [11:18] GunnarHj, I'm giving it a try [11:19] seb128: I just dropped the share link. [11:19] I did a minute ago [11:19] if we both did it it should be in place for sure :) [11:19] seb128: Then it ought to be safely gone! [11:19] indeed [11:19] I'm doing an upload [11:19] let's see [11:20] Great. [11:26] didrocks: I just use 'sudo apt build-dep' then 'jhbuild buildone' on the components I actually need instead of trying to build everything [11:26] GunnarHj, hum, the pkgbinarymangler upload failed, I did a retry but there might be issues to resolve (it didn't look related to the change though so maybe just other parts changed in bionic that create issues) [11:26] good morning [11:26] hey jbicha [11:26] jbicha: I have a bunch of components to build with particular patches [11:26] jbicha: so, that's why I have my jhbuildrc [11:26] seb128: :( [11:26] which only builds what is needed [11:26] didrocks: but webkit?? [11:27] sounds like it's on gnome-shell path… [11:28] seb128: But it's currently building. [11:34] didrocks: Ubuntu doesn't needs GNOME Tweaks' Sound panel, right? [11:34] GunnarHj, I retried the build in case it's a transient issue [11:35] seb128: I see. [11:35] GunnarHj, it failed again, https://launchpadlibrarian.net/356106245/buildlog_ubuntu-bionic-amd64.pkgbinarymangler_132_BUILDING.txt.gz [11:35] jbicha: no, we have the g-c-c patch [11:36] jbicha: I'm still hoping to get aday reviewing the volume thing for 3.28 and get upstream approving, but we'll see [11:36] didrocks: GNOME is applying your volume patches everywhere for 3.28 except for g-c-c, right? [11:37] jbicha: that's the goal, indeed [11:37] jbicha: we may have also the dialogs trigger that was made by GNOME design [11:38] (but won't be applied in GNOME, unfortunately) [11:38] seb128: Saw it. Looking at debian/rules. This is over my head. [11:39] GunnarHj, I'm going to have a look don't worry [11:39] didrocks: ha, it confuses me when GNOME comes up with patches like GNOME bug 688320 but refuses to actually apply them! :) [11:39] Gnome bug 688320 in general "dconf config option for disabling the activities hot corner" [Normal,New] http://bugzilla.gnome.org/show_bug.cgi?id=688320 [11:39] seb128: Ok, great. [11:39] jbicha: yes ;) [11:53] Hey all [11:54] morn…afternoon andyrock :) [11:56] seb128: https://translations.launchpad.net/ubuntu/bionic/+source/util-linux/+imports :) [11:57] GunnarHj, good, my theory was right :) [11:57] GunnarHj, I approved the template, it should import in a bit/while [12:03] seb128: if you didn't upload the udisks debdiffs yet maybe backport is better then upstream in the Origin dep-3 field [12:03] hey andyrock [12:04] andyrock, so can you refresh me what you were saying yesterday ... you prefer to go for the hack in gdu? no need of the udisk upload then? [12:04] there is no need but we can keep it [12:05] in theory you can still use this to hide devices from gdu [12:06] and maybe they solve the problem of existing mounted snaps directly in snapd [12:06] and we already have the code there [12:06] we'll just drop the second patch [12:11] k, so I can upload udisks2 then? [12:13] andyrock, ^? [12:13] yep [12:13] and please check the update patch for gdu [12:16] andyrock, looks good, I'm going to add the bug reference to the changelog and upload [12:16] kk [12:16] thanks seb128 [12:16] andyrock, thanks for all the work on that! [12:16] andyrock, one down, back to livepatch? ;) [12:16] yeah [12:17] i'm trying to understand if it's possible to run snapd inside the chroot [12:17] during the installer [12:18] because if we login during the live that macaroons are not going to be valid once the system is installed [12:18] because they have infos on the system used to log in which are going to be different? [12:18] that might be something worth asking mvo about [12:20] jbicha, do you know if there is any reason we didn't follow the 3.27 versions of glib/libsoup/gvfs and from leaf apps during the cycle so far? [12:22] sooo for what I understand snapd has an intern state (that is saved in /var/lib/snapd/state.json) [12:23] a "field" of its state is the authenticated users + macaroons etc [12:25] every time you try to authenticate a request, snapd will validate it also checking if the userr is "already in .../stage.json") [12:25] andyrock, please try asking on the #snappy channel if they can help you to understand/recommend what to do, easier than spending days to trying to figure out the pieces yourself [12:34] seb128: glib tends to be more Laney's responsibility [12:34] right, I just wondered if you had a clue [12:34] he uploaded to Debian experimental mid-january [12:35] he's off this week, I ask him next week when he's back [12:35] do you want 3.27 libsoup and leaf apps uploaded to bionic now? [12:35] not especially "now" [12:35] I was just curious if we were holding on those for a reason [12:35] usually you do update "non risk" apps [12:36] like gedit, I don't expect surprises to come from it since [12:36] the last few cycles I mostly waited until the .90 milestone to upload those apps, and the .90 milestone is only this week [12:36] since there is no active maintainer/planned features|new design [12:36] k [12:36] that makes sense [12:36] thx [12:37] I'd upload webkit now but it's blocked on security MIR review of brotli/woff2 [12:38] those have been bounced to the security team only recently [12:39] it's a bit earlier to nag them yet [13:59] Trevinho, hey, could you respond to Omer's question re: fractional scaling here: https://plus.google.com/u/0/+WillCooke/posts/dmxF68Pha2a?cfem=1 [14:00] willcooke: ok [14:00] thanks Trevinho [14:33] Trevinho: are huge emojis in Thunderbird a known issue in bionic? https://imgur.com/a/SVpgj (thanks popey!) [14:34] oh wow [14:34] Saviq: oh... Nope... :| [14:34] Saviq, is that an hidpi issue? [14:34] Saviq: pretty funny [14:34] Once again, me throwing random Emjoi in places uncovers bugs in surprising places! [14:34] jbicha: you know anything? ^ [14:34] *emoji [14:34] * seb128 wonders why Saviq as to Trevinho who hasn't been working on emojis nor tb [14:34] I guess it's just thunderbird not supporting properly the font [14:35] seb128: he was the last to write here, and I knew he'd be amused ;) [14:35] (and know who to throw this at) [14:35] :) [14:35] and no, no hidpi [14:35] k [14:35] 2 screens, 1080p and 1920x1200 [14:37] Saviq: give a try removing the noto font... [14:38] oh, well it will fix it for sure :-D [14:38] no need to try [14:38] but, maybe trying other emoji fonts, but to mee it looks like thunderbird isn't uspporting such fonts properly... Sooo, possibilities are various [14:41] GunnarHj, that util-linux upload worked well, the french translations have been updated, going from 3k untranslated to 200 [14:41] I'll file a bug with Thunderbird, then [14:42] seb128: Yep, same for Danish (the OP's language) and Swedish. [14:42] GunnarHj, good job! [14:43] Saviq, upstream best if you can, I don't think we have anyone reading lp bugs/having time to work on it [14:43] seb128: I mostly made some noise; you spotted the problem. [14:43] GunnarHj, result is there :) [14:43] Right. :) [14:44] jbicha, is shotwell something you usually look at/plan to update? [14:49] seb128: feel free to work on shotwell if you like :) [14:49] jbicha, that's a "no"? :) [14:49] the Debian Shotwell maintainer is frustrating to work with [14:50] FWIW https://bugzilla.mozilla.org/show_bug.cgi?id=1360862 was filed already, as it happens [14:50] Mozilla bug 1360862 in Graphics: Text "Huge Google Noto emojis since recent Firefox update" [Normal,Resolved: fixed] [14:50] I confirmed with Shotwell upstream that 0.28 is targetted to land at the same time as GNOME 3.28 [14:50] Saviq, good to know, thx [14:50] but that's Firefox, and fixed [14:50] jbicha, ok, I think I'm going to start by updating to the current stable and maybe try to have a look to 0.27 later [14:51] seb128: 👍 [14:51] jbicha, thx :) [14:53] oh shotwell isn't too bad now, I guess it's simple-scan that has the ignored patches right now [14:53] ignored by the Debian maintainer [14:53] wonder if it's fixed in more recent thunderbird [14:55] hmm there isn't a more recent thunerdbird :/ [14:56] Thunderbird builds its official releases from the Mozilla ESR branches [14:57] If you just want an easy workaround, you can try Beta: https://launchpad.net/~mozillateam/+archive/ubuntu/thunderbird-next/ [14:58] That bug suggested that it wouldnt be in the ESR :( [15:13] andyrock, I'm going to upload the xenial diff to 17.10 rather than the bionic one if that's ok for you ... what do you think? [15:13] ok for me [15:13] safer and easier I guess [15:13] andyrock, I don't think we want to SRU the udisks change/new API [15:13] right [15:13] thx [15:14] ricotz, FYI, I just pushed a patch to the ubuntu-bionic-6.0 branch for bug #1696250 [15:14] bug 1696250 in libreoffice (Debian) "Please hide Start Center and Math" [Unknown,New] https://launchpad.net/bugs/1696250 [15:15] mvo: I'm trying to allow store login in the installer. One possible solution would be to: [15:15] 1) stop/disable snapd.service/snapd.socket (the one running in the live session) [15:15] 2) manually start snapd inside the /target chroot [15:15] 3) a client can now talk with the chrooted snapd through /run/snapd.socket [15:15] Do you think that (1) and/or (2) could create any side effect? [15:16] ops wrong channel [15:16] mvo let's move to #snappy [15:19] oSoMoN, ack === mmstick_ is now known as mmstick [15:53] seb128, do you know if we're likely to get GStreamer 1.14 in for Bionic? [15:54] Bryan pointed to a slide deck from FOSDEM which said MP3 codecs are moving to good [15:54] So if we are likely to get it, I'll need to speak to legal again [15:55] bah, bionic has older firefox than xenial :P [15:56] willcooke, funny than you ask, I was poking gstreamer people about that earlier and yes I think it's likely [15:56] willcooke, they should have their pre-version for the new stable out next week [15:56] tjaalton, that's usual, dev series always have toolchain etc issue and usually updates are security ones which focus on stable series [15:58] 58 fixed slow horizontal tab scroll [15:58] kinda miss that already :) [15:58] you can probably install the 17.10 deb [15:59] it's in -proposed, ppc64el ftbfs [15:59] thanks seb128 [16:14] kenvandine, hi, could you take a look at https://launchpad.net/~gnome3-team/+snap/gedit or disable it if it won't work [16:16] ricotz, i plan to fix that [16:17] ricotz, sorry about that, it's the relative git submodule for libgd [16:17] kenvandine, alright, I see [16:18] haven't figured out how to solve that :) [16:18] but i want those daily builds [16:18] ok, maybe disable it until then [16:18] ok [16:18] i'll do that [16:18] at least the automated builds [16:19] thanks, less launchpad spam ;) [16:19] i'm really good at ignoring LP spam :) [16:19] sorry about that [16:19] ricotz, disabled [16:20] thanks [16:49] Hey guys, I accidentally botched my radeon config file, and I'm stuck in Wayland and csan't get to the file due to its root restrictions [16:49] Any advice on how to restore the file to its defaults? [16:49] I know how to fix it, it's just a matter of getting to the root file [16:50] ulysses: maybe I'm missing something... but use a VT? [16:52] I have but one machine and I need to fix thing so I can get back on Xorg or Mate [16:52] Wayland is great and all, but it's not exactly up to par yet [16:52] A vt? [16:53] ulysses: what's the file you are trying to edit and how are you currently trying? [16:53] I've been trying to get into it through root nautilus [16:53] but Wayland doesn't support that [16:55] ulysses: what is the filename and path? [16:55] you could just open a terminal and do sudo nano path-to-filename [16:56] I'm trying to remember the directory [16:56] IIRC you can open admin:/// in nautilus (via ctrl-l), this lets you edit files as root [16:56] How does one do that? [16:57] Pardon my ignorance, I've been out of the game for a bit [16:57] launch nautilus, press ctrl-l, type admin:///, press Enter [16:57] you get, effectively, a root nautilus that works on wayland [16:58] that's neat [16:58] Dude that's really awesome [16:58] Thank you so much [17:24] kenvandine, seb128 do you know if we plan to enable macsec support for wpa_supplicant for 18.04? [17:25] i have no idea [17:25] kenvandine, who controls that package? [17:26] for classic [17:26] i think cyphermox [17:27] cyphermox, can you confirm? [17:34] not really [17:35] whomever can upload it, I think Julian was doing a merge though [17:36] yeah, merge is done [18:26] cyphermox, any idea if we'll support macsec with wpa_supplicant for 18.04? [18:28] jhodapp: there are no specific plans [18:29] cyphermox, any reason why or just from omission? [18:29] no reason, and not omission, it's just not come up [18:30] cyphermox, ok [18:30] I do not routinely make wpa changes [18:31] cyphermox, yeah, besides taking in the latest package version [18:52] night all [18:52] can someone please tell me whether/how snaps are integrating with the running desktop session? I have installed the gnome-characters snap, but I can't seem to find it in the applications list, same for Telegram [18:53] installed snaps don't seem to be executable from the gnome search thing? [18:53]