/srv/irclogs.ubuntu.com/2018/11/05/#snappy.txt

amurrayso am sure I am doing something daft but am trying to add autostart support to my snap and snapcraft is complaining that 'autostart' is not a valid property01:15
amurray(using snapcraft 2.43.1+18.04)01:15
amurrayIssues while validating snapcraft.yaml: The 'apps/indicator-sensors' property does not match the required schema: Additional properties are not allowed ('autostart' was unexpected)01:16
amurrayok so I guess this is some distinction between snapcraft.yaml and snap.yaml - how do I add autostart support to my snapcraft.yaml ?03:05
=== chihchun_afk is now known as chihchun
=== chihchun is now known as chihchun_afk
=== chihchun_afk is now known as chihchun
mborzeckimorning06:02
mvohey mborzecki - good morning06:02
mborzeckimvo: hey, aything interesting happened on thu/fri?06:03
mvomborzecki: I took those days off so I don't really know. a bit of debugging on vscode and why it was relatively slow to start, turned out to be a fontconfig issue06:04
mvoi.e. incompatible cache formats06:04
mvobetween library versions06:04
mborzeckimvo: mhm, saw that, i recall this being discussed before 18.04 as a potential issue, but it was never in the context of performance06:06
=== cpaelzer__ is now known as cpaelzer
amurrayfyi - I figured it out - need to use passthrough to be able to specify autostart in snapcraft.yaml06:24
zygagood morning06:56
mvozyga: hey, good morning06:57
mborzeckizyga: hey06:57
zygahey :)06:57
zygaone other interesting thing that happened yesterday06:58
zygasystemd maintainer produced base snaps for fedora29 :)06:58
zygausing parts of systemd06:58
zyga:-)06:58
mborzeckimkosi?06:58
zygayes06:59
zygait works, I will try to upload it to the store06:59
zygalast time I tried the store chocked on it06:59
zygawe also found three bugs:06:59
zygaI'll file them now06:59
mborzeckimvo: can you look at the latest changes in https://github.com/snapcore/snapd/pull/6039 ?07:02
mupPR #6039: snapstate: do not allow classic mode for strict snaps <Created by mvo5> <https://github.com/snapcore/snapd/pull/6039>07:02
zygahmm, no mup for launchpad bugs?07:04
zygaI filed: https://bugs.launchpad.net/snapd/+bug/1801663 https://bugs.launchpad.net/snapd/+bug/1801664 https://bugs.launchpad.net/snapd/+bug/180166507:05
mvomborzecki: sure, let me do this now07:05
mupBug #1801663: Cannot find snap-exec when fedora29 is the only installed base <snapd:Confirmed> <https://launchpad.net/bugs/1801663>07:05
mupBug #1801664: Snapd fails to work in pure v2 cgroup system <snapd:Confirmed> <https://launchpad.net/bugs/1801664>07:05
mupBug #1801665: Base snaps are not validated to be devoid of apps <snapd:Confirmed> <https://launchpad.net/bugs/1801665>07:05
mupPR snapd#6078 closed: ifacestate/ifacemgr: don't reload hotplug-gone connections on startup <Hotplug ๐Ÿ”Œ> <Simple ๐Ÿ˜ƒ> <Created by stolowski> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/6078>07:33
mupPR snapd#6081 closed: overlord/ifacestate: mark connections disconnected by hotplug with hotplug-gone <Hotplug ๐Ÿ”Œ> <Created by stolowski> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/6081>07:40
mupPR snapd#6092 opened:  tests,store,daemon: ensure proxy settings are honored in auth/userinfo too (2.36) <Created by mvo5> <https://github.com/snapcore/snapd/pull/6092>07:45
mborzeckilooks like #5897 is almost ripe for landing, i'll push some updates there and it should be good to go07:45
mupPR #5897: interfaces/builtin: add device-buttons interface for accessing events <Created by bergotorino> <https://github.com/snapcore/snapd/pull/5897>07:46
mwhudsonzyga: do you know if anyone still uses / cares for plainbox?07:50
zygano, it's been fused back into checkbox07:50
zyga(+1 to kill)07:50
mwhudsonzyga: would you be able to file a bug asking for removal?07:52
mwhudson(obviously i can too but it might look better coming from you!)07:52
zygaI can try, what is the three-letter-acronym for that?07:52
mwhudsonzyga: https://bugs.launchpad.net/ubuntu/+source/plainbox/+filebug07:52
mwhudsonzyga: there is no three letter acronym for ubuntu :)07:53
mwhudson(it would be RoM == request of maintainer in debian i think)07:53
zygaoh, just on launchpad?07:53
zygaI was thinking in debian07:53
mvo6070 needs a secnd review (should be simple)07:54
mwhudsonzyga: in that case https://wiki.debian.org/ftpmaster_Removals#How_to_request_removal07:54
mwhudsonzyga: but a launchpad bug would be great in addition07:55
mwhudsonzyga: (it fails tests with python 3.7)07:55
zygamborzecki: I'm double checking with CE but I suspect they don't need it anymore07:55
zygathey moved on to snaps now07:55
mwhudsonzyga: hmm there are some reverse-depends07:56
mwhudsonzyga: checkbox-ng, plainbox-provider-checkbox, plainbox-provider-resource-generic07:56
zygayeah07:56
zygabut all of those are equally dead as debian packages07:56
mvomborzecki: 6039 looks good, thanks for all the updates you did to this one!08:00
mborzeckimvo: np, heppy to move things forwards08:00
mupPR snapd#5996 closed: [RFC] daemon: support `snap logs snapd` to get snapd logs <โ›” Blocked> <Created by mvo5> <Closed by mvo5> <https://github.com/snapcore/snapd/pull/5996>08:01
=== pstolowski|afk is now known as pstolowski
pstolowskimorning08:09
mvohey pstolowski - good morning08:09
mvofwiw, I really love the new distro name08:09
pstolowskiindeed :)08:11
mborzeckidisco dingo?08:18
mvoyeah08:18
mborzeckizyga: i've updated #5897, we can either close it or land it, i still think the interface is useful, defining gpios as keys is super convenient and imo people use it (i used it for a couple of custom devices)08:31
mupPR #5897: interfaces/builtin: add device-buttons interface for accessing events <Created by bergotorino> <https://github.com/snapcore/snapd/pull/5897>08:31
zygamborzecki: yeah, I think we want jdstrand to review it and kola to ack that it still works on their hardware08:33
zygathank you for refreshing it08:33
kyrofazyga, I just a got a bug report from someone who's snap stopped working. Logs are filled with "cannot perform operation: mount --bind /snap/core/current//etc/alternatives /tmp/snap.rootfs_PA4OiD/etc/alternatives: No such file or directory"08:39
kyrofaAny ideas?08:40
zygakyrofa: yes, apparently there are no /etc/alternatives on the filesystem08:45
zygaunless we yanked them from core itself08:45
zygaand we indeed did just that08:46
zygahmmm08:46
zygakyrofa: what is the snap version?08:46
kyrofazyga, I'm checking. Stable hasn't updated though, right?08:46
kyrofaWonder if they're following a different channel08:46
zygastable is 2.35.5 which looks slightly updated08:47
zygabut not 2.3608:47
kyrofaWhen you say you did indeed just do that, what were you referring to?08:49
zygaI mean that on my system /etc/alternatives is no longer in core08:49
zygait looks like old snap-confine that is not matching the current core08:49
kyrofaDoesn't that get re-exec'd?08:50
zygain a way, we run the right snap-confine directly08:53
mupPR snapd#6067 closed: tests/main/snap-service-after-before-install: verify after/before in snap install <Created by bboozzoo> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/6067>09:03
mborzeckiwhat shall we do about #5946 ?09:04
mupPR #5946: cmd/snap: unhide --name parameter to snap install, tweak help message <Parallel installs โ›“> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/5946>09:04
mupPR snapd#6093 opened: daemon: spool sideloaded snap into blob dir <Created by chipaca> <https://github.com/snapcore/snapd/pull/6093>09:08
Chipacagood morning peeps09:08
pstolowskimornings Chipaca !09:08
Chipacai woke up early and got caught up in some snapd work instead of having breakfast09:09
Chipacaso now i'm going to go fix that, brb :-)09:09
mvohey Chipaca !09:10
Chipacamvo: o/ !09:19
Chipacamvo: how've you been?09:19
mvoChipaca: very well, thank you! nice and relaxing long weekend, just some debugging around fontconfig issues with snaps mixed in but beside this all great09:20
mvoChipaca: and you?09:20
Chipacamvo: no fontconfig debugging here09:21
Chipacamvo: but, wondering if we can't expose a "this snap uses fontconfig x.y.z" flag somewhere09:21
Chipacamvo: also wondering about per-user hooks09:22
mvoChipaca: yeah, I think there are some things we don't fully understand yet. for example on 18.04 I get a ~/.cache/fontconfig dir populated when I run with an older fontconfig. but that does not happen with the same snap on 18.1009:22
mvoChipaca: yeah, per-user cache with per-user hook would solve this09:23
kyrofazyga, look at that `snap version`: https://github.com/nextcloud/nextcloud-snap/issues/776#issuecomment-43580355909:24
kyrofaHow is that possible?09:24
mvoChipaca: but some work ahead, also versionizing the font-config cache files09:24
Chipacakyrofa: no re-exec09:25
mvoChipaca: if they are not already, the trouble is we don't really know (yet). but maybe james hensdrige knows more or someone from the desktop team09:25
Chipacahere's hoping the solution or workaround isn't  ubuntu-centric09:26
Chipacazyga: btw you semi-reviewed #6065, before I went and split it as you requested. Could use another semi review :-D09:27
mupPR #6065: cmd: make coreSupportsReExec faster <Created by chipaca> <https://github.com/snapcore/snapd/pull/6065>09:28
kyrofaChipaca, how would that get disabled in xenial?09:28
kyrofaDoesn't that require setting an environment variable in the systemd unit?09:28
Chipacakyrofa: pseudo-user on a pseudo-terminal09:29
kyrofaChipaca, that shouldn't effect which snapd is actually running then, right?09:32
zygakyrofa: looking09:33
Chipacakyrofa: that was a joke :-) AFAIK 2.20 already re-exec'ed, on ubuntu, so having it not do so requires user (sysadmin) action09:34
zygaChipaca: enqueued09:34
Chipaca2.10ish was when we flipped that particular switch iirc09:34
Chipacazyga: enthanked09:34
kyrofaChipaca, oh thank goodness. I respect you too much, I thought this was technology with which I was very much unfamiliar :P09:34
Chipacai should shave my beard and start a youtube channel, to get the respect dialed back a bit09:35
kyrofaAgreed09:35
mborzeckiChipaca: do you have any suggestions about https://github.com/snapcore/snapd/pull/5946 ?10:02
mupPR #5946: cmd/snap: unhide --name parameter to snap install, tweak help message <Parallel installs โ›“> <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/5946>10:02
Chipacamborzecki: what about?10:05
Chipacamborzecki: degville suggested things there already :-)10:05
Chipacamborzecki: and you have two +1's10:06
mborzeckiChipaca: thought you meant that we should add more changes10:07
mborzeckiChipaca: which in fact i'll do just now, the last paragraph is a bit off10:07
Chipacaheh10:07
mborzeckilast paragraph of the help message10:07
Chipacamborzecki: ok10:08
Chipacamborzecki: _the_ instance name?10:12
mborzeckiheh, let me force push that10:13
Chipacamborzecki: did you implement graham's suggestion? i can't see it but then i'm sleepy10:13
mborzeckiChipaca: yup, both of them10:14
Chipacak10:14
kyrofazyga, any thoughts on that?10:26
zygakyrofa: on the version?10:26
kyrofaYeah10:26
zygano, no idea, logs would confirm why snapd doesn't reexec10:26
zygaif you can collect logs from snapd.service it would tell us10:27
Chipacabrb, reboot10:27
kyrofazyga, alright, requested10:29
=== cpaelzer__ is now known as cpaelzer
* cachio afk10:32
ackkhi, I'm having an issue with bash completion in a (classic) snap. it doesn't seem to work, but if I snap run --shell and source the completion file, it does work there. is there perhaps something different in how snapd handles that?10:33
Chipacaackk: https://forum.snapcraft.io/t/debugging-tab-completion/419810:39
ackkChipaca, thanks10:39
sparkiegeekackk: can you share your snapcraft.yaml, and `snap version`?10:39
sparkiegeekwell, Chipaca's link is better :)10:39
Chipaca:-D10:39
ackksparkiegeek, https://git.launchpad.net/~sshoot/sshoot/+git/packaging/tree/snap/snapcraft.yaml10:40
=== chihchun is now known as chihchun_afk
Chipacaackk: please let me know how the debugging goes (specifically: at which step of the debugging things break down)10:41
sparkiegeekackk: sshoot-completion != shell-completion10:43
sparkiegeekackk: you have the former defined in the app, and the latter in your part10:44
ackksparkiegeek, shell-completion is a dir, in which sits sshoot which is copied as sshoot-completion10:44
ackkChipaca, should I run the "complete -p" inside the snap run or from my system?10:44
Chipacaackk: no10:45
Chipacaackk: in my defence it does tell you to exit the snap run --shell :-)10:46
ackkChipaca, sorry, missed it :)10:46
ackkChipaca, so, the -x shows something different: https://paste.ubuntu.com/p/h2TkVGxprm/10:52
Chipacaackk: that looks almost exactly the same10:53
ackkChipaca, yeah but see it fails?10:53
Chipacaah, yes10:54
ackkChipaca, there seems to be an error mesage like "cannot..." being parsed10:54
Chipacaackk: so go on to the next step :-)10:54
Chipacai'm starting to guess where the issue is10:54
Chipacabut go ahead10:54
ackkChipaca, so if I snap run and source the bash completion it doesn't work10:56
ackkChipaca, what do you think the issue is?10:56
Chipacaackk: looking at the completer, I guessed wrong10:56
Chipaca:-)10:56
ackkChipaca, fwiw the snap run --command=complete returns "default"10:57
Chipacaackk: when you say that sourcing the bash completion doesn't work, what do you mean?10:57
ackkChipaca, oh, wait10:58
ackkChipaca, n/m, so step 5 works10:58
ackkChipaca, so everything works, except it doesn't :)10:59
Chipacaackk: ok, let's rewind a bit10:59
Chipacaackk: what should it do, that it isn't doing?11:00
Chipacawhat should it complete with?11:00
ackkChipaca, the command options11:00
Chipacaackk: but it completes with filenames, here11:00
Chipacaalso11:01
Chipacaalso11:01
ackkright, because it's falling back to the default completer I guess11:01
Chipacabah, maybe i'm looking at an old version :-)11:01
Chipacayes probably11:01
Chipacaackk: what do you get when you do the 'snap run --command=complete yadda yadda' step?11:01
ackkChipaca, https://paste.ubuntu.com/p/NS3m5St8mz/11:01
Chipacaackk: that's the non-snapped one I presume11:02
ackkChipaca, no, it's the snap, from inside a snap run, as per step 511:02
Chipacaackk: what's the output on step 4?11:02
ackkChipaca, https://paste.ubuntu.com/p/7gbVwVM7Sv/11:03
ackk(I used the number from the bash -x from the step before11:03
ackkChipaca, I wonder if something in my bash setup is messing the compelter up11:03
Chipacahm, I don't get that11:03
ackkalthough that has never happened for any other completion11:03
Chipacabut I've installed the one in stable11:03
ackkoh?11:03
Chipacawhich might be different11:04
ackkChipaca, they're the same11:04
ackkChipaca, 72?11:04
Chipacaackk: https://pastebin.ubuntu.com/p/gFD7KsRVpR/11:04
Chipacaackk: and then all the files11:04
ackkChipaca, which version of the snap do you have?11:04
Chipacaackk: revision 7211:05
Chipacaackk: also, also:11:05
Chipaca~$ sshoot11:05
popeycjwatson: is there a timeline on when the "('Connection aborted.', gaierror(-3, 'Temporary failure in name resolution')) " store upload fails might be fixed?11:05
Chipaca/snap/sshoot/72/command-sshoot.wrapper: 2: exec: /snap/sshoot/72/usr/bin/python3: not found11:05
popeyhttps://launchpad.net/~build.snapcraft.io/+snap/b2218f767981315ff64384e4f403d972-xenial/+build/370290 for example11:05
ackkChipaca, https://paste.ubuntu.com/p/bVfBTFdVDW/11:05
Chipacaackk: I'm not sure what you're showing me :-)11:06
Chipacaackk: this is a classic snap, yes?11:06
ackkChipaca, yes11:06
ackkthis is really weird11:06
Chipacano not that weird11:06
Chipacajust classic weird11:06
mborzeckiopened a topic with cross-snap service ordering proposal https://forum.snapcraft.io/t/cross-snap-service-ordering/8319 feel free to comment11:06
Chipacaackk: do this11:06
ackkChipaca, does "sshoot list" work for you?11:06
Chipacaackk: get into 'snap run --shell'11:06
Chipacaackk: i mean 'snap run --shell sshoot'11:07
ackkChipaca, and then?11:08
Chipacaackk: /usr/bin/python311:08
Chipacaackk: and then11:08
Chipacaackk: import argcomplete11:08
Chipacaackk: and them11:08
ackknot found11:08
ackkbut why is it trying /usr/bin/python311:08
Chipacaackk: /usr/bin/env python311:09
Chipacaackk: unless your wrapper is rewriting PATH?11:09
ackkChipaca, so I need to put $SNAP/usr/bin in path first11:09
* Chipaca looks11:09
ackkno it doesn't11:10
Chipacaackk: also, the python3 you're shipping doesn't work here11:10
ackkwhat?11:10
Chipacaackk: bah, agian let me look at the wrapper11:10
Chipacaackk: so, the wrapper should be fine on the path front (it execs $SNAP/yadda), but the completion script needs the same treatment11:11
Chipacaackk: or tweaking the path11:11
Chipacaackk: but11:12
ackkChipaca, ok, I'll just include a copy of that script in the snap, it's one line anyway :)11:12
Chipaca/snap/sshoot/72$ ./usr/bin/python311:12
Chipacabash: ./usr/bin/python3: No such file or directory11:12
ackkChipaca, yeah it's just "python" I think11:12
Chipacaackk: nope, it's python3, a symlink to python3.611:12
Chipacaackk: but the libraries are messed up11:13
ackk$ ./usr/bin/python3 -c "print('hello')"11:13
ackkhello11:13
ackkChipaca, ^11:13
ackkChipaca, are you running it inside the snap run?11:14
Chipacaackk: you're on 18.04, running a core18 snap, i presume11:14
ackkChipaca, yeah11:14
Chipacaackk: yeah11:14
Chipacaackk: that doesn't exactly test the libraries are being picked up from where you expect them to be ;-)11:14
ackkI see11:14
ackksigh11:14
ackkok, I'll take a look, thanks for the help Chipaca11:14
Chipacaackk: fun11:15
mupPR snapd#6091 closed: packaging/fedora: Merge changes from Fedora Dist-Git <Created by Conan-Kudo> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/6091>11:15
ackkyeah...11:15
mborzeckimvo: can you cherry-pick that PR to 2.36? ^^11:15
Chipacaackk: classic is tricky this way (and others, but this way)11:16
Chipacaackk: good luck11:16
ackkChipaca, thanks :)11:16
popeyHm, is there some way to re-run console conf by removing some file on the sd card?11:18
mvomborzecki: sure, doing now11:19
mborzeckimvo: ta11:19
popey(I have moved my pi from home to office, and need to re-run console conf to connect to wifi)11:19
ograpopey, yes, there is a file ... one sec11:19
* Chipaca afk11:19
mvomborzecki: thank you! its in :)11:19
popeyogra: thanks11:20
ograpopey, /writable/system-data/var/lib/console-conf/complete11:25
popey$ sudo mount /dev/mmcblk0p2 /mnt/writable/11:26
popeymount: /mnt/writable: cannot mount /dev/mmcblk0p2 read-only.11:26
popey:(11:26
ograweird ... anything in dmesg ?11:27
popey[10917.675750] EXT4-fs (mmcblk0p2): write access unavailable, cannot proceed (try mounting with noload)11:27
popeyhuh, not sen that before11:27
popeyahh, [10917.675746] EXT4-fs (mmcblk0p2): INFO: recovery required on readonly filesystem11:27
ograwants an fsck it seems11:28
=== chihchun_afk is now known as chihchun
popeyDisk write-protected; use the -n option to do a read-only11:28
popeystupid sd card adapter :)11:28
ograyeah11:28
cjwatsonpopey: Not at this point11:29
cjwatsonIt's very very weird and has resisted understanding so far11:29
popeycjwatson: ok, so just keep re-triggering builds until it works?11:30
cjwatsonI don't have a better offer at the moment11:31
popeyogra: /writable/system-data/var/lib/console-conf/complete doesn't exist11:31
cjwatsonBut do tell us when it happens because it generally means we need to restart a thing11:31
popeyok11:31
cjwatsonOnce it's in that state there's a 1/3 chance of any given upload failing that way, until we restart it11:31
ograpopey, hmm, then console-conf should run ... unless something changed ... in which case mwhudson is your man i think11:32
popeycjwatson: i re-triggered the one above11:32
popeyogra: ok11:32
popeyogra: might start a thread on the forum, because this is something undocumented afaict11:32
cjwatsonI've requested that the cursed worker be killed11:33
popeycjwatson: thanks11:33
popeyfile:///home/alan/Pictures/Screenshot_20181105_113742.png https://usercontent.irccloud-cdn.com/file/1JYDSny9/aargh%20%3A)11:38
popeyogra: ^11:38
popeyusing an hdmi capture card as I don't have a monitor11:39
ograpopey, modern art ?11:39
popeythats - i assume - some message about a lack of ip address11:39
popeyI guess I need an ethernet cable11:39
cjwatsonpopey: Should be happier now, at least for the time being11:41
popeyUbuntu Core 16 on <no ip address> (ttyS0)11:44
popeyYou cannot log in until the system has an IP address.11:44
popey(Is there supposed to be a DHCP server running on your network?)11:44
popey(is what it says)11:45
ograpopey, well, console-conf should override this if the completion file is missing, not sure why you dont have it11:50
popeyI'll start a thread11:52
mupPR snapd#6092 closed:  tests,store,daemon: ensure proxy settings are honored in auth/userinfo too (2.36) <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/6092>11:52
zyga[   52s] snapd-2.33.1/gopath/src/github.com/snapcore/snapd/overlord/snapstate/autorefresh.go:182: Noticef format %s has arg int(maxPostponement.Hours() / 24) of wrong type int11:57
mborzeckizyga: wasn't that fixed in master already?12:00
zygathis is 2.33.1 :)12:00
mborzeckioh12:00
zygaone more failiure https://www.irccloud.com/pastebin/np5xC9Mc/12:01
zygaperhaps test keys?12:01
zygabut on the upsdie12:01
ograstop playing with that old cruft :P12:01
zygano more golang-packaging macros12:01
zygalogs are short12:01
zygacode is simple12:01
mborzeckizyga: opensuse?12:01
zygaerrors make sense12:01
zygayes, wrapping up that work12:01
mborzeckithat looks like a piece from rpmbuild log12:01
zygayes12:02
zygaI rewrote parts of the packaging to drop the golang-packaging macros12:02
zygato unbreak our builds12:02
=== cpaelzer__ is now known as cpaelzer
zygaon the upside master is green on suse now12:19
=== chihchun is now known as chihchun_afk
mborzeckihmm might have found a bug in generated service wrappers with Before=12:42
zygaoh?12:43
mborzeckisomething for .1 i belive12:43
mborzeckidamn, go templatles are both fun to use and really frustrating12:44
Chipacamborzecki: funstration is best stration12:44
ChipacaI mean, it beats defenestration at least12:44
mborzeckineed to look that up :)12:45
mborzecki'Defenestration is the act of throwing someone or something out of a window'12:45
mborzeckiwonder why 'someone' comes first12:45
zygamborzecki: wow really?12:45
zygaI mean, school has burned that word into my head12:45
Chipacamborzecki: because it's not usually used literally12:46
zygaI have 2.36 packaged for suse, if you want to review mborzecki12:46
zygaI'll figure out how to create a diff in osc12:46
mborzeckiisn't there osc diff? :)12:47
zygaI mean I want to push this12:49
zygabut I started in system:snappy12:49
zygaosc just sucks for workflow12:49
zygaand I already have "a branch" so I cannot branch again12:49
mborzeckijust imaging it's svn12:49
mborzeckiright, so when there's more than one serivce in before: ..., we generate Before=.. without whitespace separating service names12:50
zyga oops12:51
mborzeckimvo: ^^ i'll push a PR for this and we'll need to cherry pick it to 2.3612:53
zygaChipaca: hmm12:58
zygaaround?12:58
Chipacazyga: yes13:00
Chipacazyga: standup13:00
zygaoh13:00
zygasure13:00
Son_Gokumborzecki, unfortunately, obs vcs is derived from svn13:09
Son_Gokuzyga, you can branch again as much as you want, you just need a different branch project to use13:10
Son_Gokuzyga, you can also use copr to build for opensuse if you'd prefer git-ish workflows :)13:10
zyganah, that's fine - I'll just release it on OBS now13:10
Son_GokuJamieBennett, can we get RHEL 7 to wire up into spread?13:12
mborzeckiSon_Goku: centos is wip in my branch, but we hit some problems with MNT_DETACH and an old-ish kernel13:12
Son_Gokumborzecki, if you have a Fedora system, you can trivially boot RHEL 7 using RHEL Developer licenses on your computer in a VM13:13
Son_Gokuusing gnome boxes13:13
mborzeckiSon_Goku: do you know if there's any timeline for centos 7.6?13:14
Son_Gokucurrent expectation is sometime in the next 2-3 weeks13:14
Son_Gokuthey've already started chewing through a good chunk of it since the code drop last week: https://twitter.com/JohnnyCentOS/status/105835969162821632013:14
mborzeckiSon_Goku: thx13:16
amunizpHi all, any chance there are any IoT developers here?13:21
zygaamunizp: probably, what do you need?13:26
ograamunizp, how about you just ask a question ? :)13:26
mvomborzecki: uh, thanks, yeah, thats a show-stopper13:30
amunizpWell, the question might not be appropriate for the channel but here goes. I am looking for hands on mentors for a hackathon like event probably spring or summer 2019. The idea is to help develop IoT crowd system (size of about 4 million). The work will be paid for around 4 people managing a group of about 4 people. So my question is: what it the day rate for such a dev/hardware hacker? knowledge needed would be: is this a job for a beag13:31
amunizple bone or an FPGA device? what kind of encryption should be needed on the hardware? would it be best on the server? Can these devices consume low battery?13:31
amunizp*lower battery?13:33
amunizpDoes that answer your question zyga and ogra ?13:33
ograamunizp, sounds like the #beagle channel might be a good fit for you (at least for the beaglebone part of the question)13:35
zygaamunizp: I'll answer after a meeting I'm in13:37
mupPR snapd#6094 opened: wrappers: fix generating of service units with multiple `before` dependencies <Created by bboozzoo> <https://github.com/snapcore/snapd/pull/6094>13:37
mborzeckimvo: ^^13:37
amunizphi ogra, sorry my question is: Does the person I want to contract know about the different IoT hardware tech available? beagle bone vs fpga is an example.13:39
zygaamunizp: I think this is not the best question for things like that13:41
zygaamunizp: your question is very much dependent on the location and other parameters of the employer13:41
zygaer, I mean, not the best channel for questions like that13:42
ograamunizp, the moore knowledgeable people in #bagle are definitely capable of answering that ... though you need to likely be a bit more detailed about your use case "IoT" is just a marketing buzzword ans doesnt explain what your implementation really needs13:42
ogra*more ... #beagle13:43
ogra(geez my typing)13:43
amunizpThanks! Location is London. I'll pop by #beagle13:45
mvomborzecki: ta!13:51
popeythresh: https://github.com/snapcore/snap-store-badges :)13:58
kyrofazyga, these logs don't seem very helpful: https://github.com/nextcloud/nextcloud-snap/issues/776#issuecomment-43588417514:02
zygakyrofa: I'll check soon14:03
Chipacamborzecki: https://play.golang.org/p/Fv2i0k78LNK ?14:03
zygaI need to grab some food14:03
Chipacamborzecki: is the second one not what you want?14:03
mborzeckiChipaca: i want Before=one two\n14:04
Chipacamborzecki: y tho14:04
Chipacai mean i get it looks neater :-)14:04
mborzeckioff to pick up the kids14:06
* pstolowski lunches14:07
* cachio afk14:24
mupBug #1801735 opened: Typo in model assertions <Snappy:New> <https://launchpad.net/bugs/1801735>14:27
zygakyrofa: not much detali :/14:35
zygaDo you know more about that system? Is it a core device?14:35
Chipacazyga: kyrofa: added a comment there14:42
kyrofazyga, it's 16.04, but not much more detail than that14:45
kyrofaChipaca, thanks, although now I'm worried we'll never know how the machine got into that state14:45
kyrofaBut I suppose unblocking is really what they want14:45
Chipacakyrofa: I'm downloading 16.04.1 to see if i can reproduce it14:46
Chipacaok so 16.04.1 shipped 2.0.1014:48
* Chipaca tries .214:49
zygaThank you chipaca!14:49
Chipacawe've come so far since 2.0.1014:49
Chipacait had, like, 10 commands14:50
Chipaca:-)14:50
zyga10 commandments ;-)14:50
Chipaca$ snap14:51
Chipacaerror: Please specify one command of: abort, ack, change, changes, connect, create-user, disconnect, find, help, install, interfaces, known, list, login, logout, refresh, remove, run or try14:51
Chipacazyga: kyrofa: and 16.04.2 ships 2.2114:54
Chipacazyga: kyrofa: and I can confirm it re-execs into core just fine14:55
zygaSo manually disabled probably14:56
zygaI think there are two lessons14:56
zygaOne: core should โ€œassumes: snapd2.35โ€14:56
zygaAnd another, we need snap debug reexec14:57
kyrofaI wonder why it would be manually disabled14:57
Chipacakyrofa: they've tinkered with the environment, obviously14:58
zygaDoesnโ€™t have to be14:58
zygaBut probably is14:58
Chipacaum14:58
Chipacathose are DEBUG logs14:58
Chipacaso they _have_ tinkered with the environment14:59
Chipacawhat i was about to say was that it wasn't too much of a stretch to think that they'd done more than just set debug :-)14:59
Chipacawe can ask14:59
Chipacabut see rule #1 about users14:59
zygaback at the office15:05
ograseb128, yo ! ... there is a thing i noticed about the font handling in the dektop launchers ... wehn i build kiodk apps to run on ubuntu-core where there is no dekto i always have to patch the fonts.conf file to make fontconfig work at all ... (it has a hardcoded /usr/share a fonts search path), i noticed your fontconfig discussion and was wondering if that could be a hint ... afaik the desktop launcher codde ddoes not modify the in-snap fonts.conf at15:18
ograall (so shipped fonts are never in the search path)15:18
ograoh man ...15:18
ogramy broken kbd really showss15:18
seb128ogra, it's not the issue, but thx for pointing it out15:18
seb128the cache is by dir15:19
seb128if some dirs are missing it means some fonts would be missing15:19
seb128but not impact the start time15:19
ograseb128, well, if you set FONTCONFIG_FILE to $SNAP/etc/fonts/fonts.conf and that only has /usr/share/fonts, it will never consider the font files in $SNAP15:20
ograbut alway use the system dirs ... which go through an interface ...15:20
seb128most fonts are not bundled in the snap anyway15:21
ograk15:21
seb128ogra, also https://github.com/ubuntu/snapcraft-desktop-helpers/blob/master/common/desktop-exports#L24015:23
seb128 for ((i = 0; i < ${#data_dirs_array[@]}; i++)); do15:23
ograah, k on desktop you re-generate it anyway15:23
seb128right15:24
ograyeah, then you are likely fine ... (i cant do that on core ... well i could but for no benefit since there are no fonts ...) ... anyway, was just an observation15:24
jdstrandroadmr: hi! can you pull r1150?15:55
roadmrjdstrand: sure thing! coming up15:56
jdstrandroadmr: thanks!15:57
zygajdstrand: hello, welcome back15:57
jdstrandzyga: thanks! I was actually back since last Wed :)15:58
zygaI was off since Thursday15:58
jdstrandheh15:58
jdstrandhope you had nice time off :)15:58
mupPR snapd#6095 opened: packaging/opensuse: stop using golang-packaging <Created by zyga> <https://github.com/snapcore/snapd/pull/6095>16:15
zygajdstrand: thanks, it was nice to spend more time with the family16:15
roadmrjdstrand: hey a question: if I want to exempt 3 specific gadget snaps from being red-flagged/manually reviewed, they'd have to be added to redflagged_snap_types_overrides['gadget'], right? Is it OK to do so with gadget snaps that live in a brand store/16:17
roadmr?16:18
zygaroadmr: hey, do you remember that bug with fedora29?16:19
zygaroadmr: is there any update on that, I tried submitting it lately and it still fails on __all__16:19
* roadmr hides behind a bush16:20
jdstrandroadmr: yes and yes. I'm right there if you want to privmsg me16:20
roadmrzyga: the latest I heard, it was fixed :/ the store can remove the evil fedora 555 files, the review tools can as well, and those updates are in production16:20
zygaroadmr: let me try16:20
roadmrzyga: I assumed you were satisfied with it because I saw some published f29 snaps already :(16:21
zygaroadmr: those were done earlier with the hack16:21
zygahey, no worries though :-)16:21
roadmrzyga: oh I remember - I was supposed to push the f29 evil snap as root so snapcraft is also happy, and check what else the store says16:22
roadmrfwiw the hack should only be setting squashfs-root to 755, which doesn't sound horrible, does it? I mean, in fedora, is the mode for / 555? not writable by root? who are they fooling? :)16:23
zygaroadmr: yes16:24
zyga[zyga@faroe ~]$ ls -ld /16:24
zygadr-xr-xr-x. 18 root root 4096 10-19 11:58 /16:24
roadmrzyga: right... ok, that's probably why squashfs-root itself has that weird mode16:25
zygayeah16:25
roadmrzyga: I'll have another look in a bit16:28
zygaroadmr: thanks!16:28
zygathere's going to be a new build soon16:28
zyganow based on mkosi16:29
zygabut that should not change this16:29
zygaWimpress: FYI, I'm prepping for a 2.36 across opensuse16:29
zygait fixes many known issues16:29
zygaincluding unbreaking all of tumbleweed snap support16:29
Chipacapstolowski: do we have 'snapctl status'?16:48
zygacachio: hey16:48
zygaerror: cannot pack "/home/gopath/src/github.com/snapcore/snapd/tests/lib/snaps/basic-desktop": write /tmp/.snap-pack-exclude-116402540: no space left on device16:48
zygahave you seen any issues with insufficient space on amazon instances?16:48
pstolowskiChipaca: nope16:49
zygamvo: when do you anticipate 2.36.116:50
pstolowskiChipaca: interesting... did we have 'snap services' from the beginning, or was it added later? i remember snapctl was modelled after snap start/stop/restart16:51
Chipacapstolowski: services and logs were part of the same piece of work16:51
cachiozyga, hey16:52
Chipacapstolowski: any reason not to add the missing two?16:52
cachiozyga, yes, this is the one I mentioned today16:52
cachioin the standup16:52
zygaoh, sorry I missed that16:52
cachioI'll push a PR today16:52
zygathanks!16:52
Chipacaum16:53
Chipacaum16:53
Chipacahold on16:53
Chipacawhat's that doing?16:53
cachiozyga, I am still working on the spread issue but I'll back to that one later, thanks16:53
zygathank :)16:53
* Chipaca greps16:53
zygaChipaca: ?16:53
Chipacais it always that file?16:54
Chipacacachio: zyga:?16:54
zygaok, it's almost 18;00 and my head is dizzy from ABS fume16:54
Chipacazyga: dude16:54
zygaChipaca: first time I saw this16:54
zygaChipaca: not sure16:54
zygaI'll head upstairs until I can vent this room16:54
Chipacazyga: go smoke something healthy to give yourself a break16:54
Chipacalike, i dunno, bacon16:55
zygaabs shrinks and warps when cooled16:55
mvozyga: there is the open question about "snapd 2.36 API "system" vs "core" slot name" - I sent a mail about this last week, you are on the CC i have no reply yet16:55
pstolowskiChipaca: no, we should have status at least i think, not sure about logs. and they should be narrowed to services of current snap of course16:55
mvozyga: once we know what to do here we can continue16:55
zygamvo: the one for our customers?16:55
zygamvo: ok16:55
zygamvo: thanks!16:55
mvozyga: yeah, I ping htem now16:55
zygathanks!16:55
Chipacapstolowski: yeah not sure what purpose logs would serve16:55
zygaI'm asking because the base snap refresh bug fix is important for desktops16:55
zygaso perhaps I should cherry pick it16:55
Chipacacachio: dunno if you saw my question, is the out-of-space error always about .snap-pack-exclude?16:56
pstolowskiChipaca: i can only think of having 'logs' for consistency, but that's probably not enough16:56
mvozyga: cherry pick to 2.36? absolutely if its important16:56
cachioChipaca, no16:56
Chipacacachio: ah, phew16:56
zygayeah, I'll do that16:56
cachioI have seen it about other stuff too16:56
zygaChipaca: tomorrow: I'd like to shrink packaging and grow a makefile16:57
zygait's very annoying that we have so many things done in a build script that is distro specific16:57
zygaI'll bug you for reviews :)16:57
cachioChipaca, i.e. https://travis-ci.org/snapcore/spread-cron/builds/450694163#L250316:57
Chipacacachio: i was worried because .snap-pack-exclude could very easily be a static file somewhere (it was just a faff to get it right on darwin)16:57
roadmrzyga: hm, I updated https://bugs.launchpad.net/snapstore/+bug/1786071 with the traceback I got with the fedora snap - looks like the click reviewer tools are still unhappy, we fixed the case where files/dirs inside squashfs-root have funny perms but may have obviated the "squashfs-root itself has funny perms" case :(17:33
roadmrjdstrand: ^^ the above sounds similar to https://bugs.launchpad.net/click-reviewers-tools/+bug/1712476 , which was about files inside squashfs-root having funny perms but it may have obviated funny perms on squashfs-root itself :/17:42
zygaroadmr: that's great, thank you for checking17:44
roadmrnp, sorry we keep finding funny cleanup cases17:45
jdstrandroadmr: what revision of the snap? if seems the latest ones passed review17:53
zygajdstrand: when I pushed myself it was not even registered by the store17:54
zygaprior revisions were hacked to pass17:54
jdstrandzyga: is it the fedora29.2018.10.09.x86_64.snap you wormholed to me?17:56
zygaI honestly don't recall now, I can quickly build a daily to test17:56
zygabut it was simply chmod ugo-w / that broke things AFAIK17:57
jdstranddr-xr-xr-x root/root               269 2018-10-09 11:27 squashfs-root17:58
jdstrandthat passes here17:58
jdstrandroadmr: what are you looking at? ^17:58
jdstrandroadmr: it should be fixed in r113217:59
jdstrandroadmr: I commented in the bug18:28
cachioniemeyer, the reboot issue on spread I mentioned today is happening on some systems when the reboot command is executed18:31
cachioniemeyer, it is happening that the ssh hangs and then it finishes when the warn timeout hist18:32
cachiohits18:32
cachioI see this can be reproduced on systems which are rebooting so fast18:32
cachioi-e- debian 9 is rebotting in about 8 seconds18:33
cachioor less18:33
cachioI could fix it by running reboot on background18:33
cachioniemeyer, I'll create a PR with this18:33
cachioniemeyer, https://github.com/snapcore/spread/pull/7018:46
mupPR spread#70: Reboot on backgraound to avoid spread wait for long time <Created by sergiocazzolato> <https://github.com/snapcore/spread/pull/70>18:46
zygajdstrand: offtopic: https://github.com/snapcore/snapd/pull/5644#discussion_r22844437019:14
roadmrjdstrand: I'll check your comment now. Yes, I recall that was supposed to be fixed :/ and the store is on >1132 anyway... checking19:14
mupPR #5644: interfaces: add audio-playback/audio-record and make pulseaudio manually connect <Created by jdstrand> <https://github.com/snapcore/snapd/pull/5644>19:14
zygajdstrand: I wrote some POC code for that19:14
zygajdstrand: where snapd is a trust store itself that apps could query19:15
jdstrandzyga: I responded. I think that is interesting and possibly something for the future, but the patches already exist and are working their way in via SRU for pulseaudio (but see comment)19:25
jdstrandroadmr: if you have a specific snap where it is failing, I'm happy to look at it. I was going by a snap zyga wormholed me a while ago that had a 555 /, but maybe you have a different snap that causes trouble19:26
roadmrjdstrand: it's the same snap from zyga :/ I did have to unpack it and tweak it a bit so I could upload it to the store myself19:27
roadmrchanged the name and type mainly19:27
jdstrandroadmr: $ sha256sum ./fedora29.2018.10.09.x86_64.snap19:29
jdstrand15169bf33d889746a31828213ecf1adcd442ba82b13ad30d5b7c68eb164faf47  ./fedora29.2018.10.09.x86_64.snap19:29
roadmrjdstrand: yep, that's the one I based mine on19:30
jdstrandroadmr: it is working find here both as the review-tools snap and out of my bzr checkout of crt19:30
jdstrandfine*19:30
jdstrandroadmr: is it something in the store and not the review tools that is having trouble?19:31
jdstrandroadmr: can you give me your snap?19:32
roadmrjdstrand: sure! wormhole?19:32
jdstrandthat's fine19:32
roadmrjdstrand: also, what's barfing is the review-tools themselves, but I notice this is when doing unpack-package. Is that what you're testing?19:32
jdstrandroadmr: I am simply running click-review/snap-review19:33
jdstrandroadmr: let me try unpack-package19:33
jdstrandroadmr: worked fine19:34
roadmrinteresting!19:34
jdstrand$ PYTHONPATH=./ ./bin/unpack-package ~/Desktop/fedora29.2018.10.09.x86_64.snap /tmp/foo19:34
jdstrandSuccessfully unpacked to '/tmp/foo'19:34
jdstrandroadmr: is something wrong with your checkout?19:34
roadmrhm I don't think so :)19:35
jdstrandroadmr: $ PYTHONPATH=./ ./bin/unpack-package ~/Desktop/evil-roadmr-fedora-2018.11.05.01_amd64.snap /tmp/bar19:38
jdstrandSuccessfully unpacked to '/tmp/bar'19:38
roadmrjdstrand: ACTUALLY - I can try it on the very same store unit that had the error19:38
* roadmr does so19:38
jdstrandroadmr: $ PYTHONPATH=./ ./bin/click-review ~/Desktop/evil-roadmr-fedora-2018.11.05.01_amd64.snap19:39
jdstrandErrors19:39
jdstrand------19:39
jdstrand - lint-snap-v2:external_symlinks19:39
jdstrandpackage contains external symlinks: usr/lib64/libnssckbi.so, usr/lib64/p11-kit-trust.so19:39
jdstrand/home/jamie/Desktop/evil-roadmr-fedora-2018.11.05.01_amd64.snap: FAIL19:39
jdstrandroadmr: it not tracing back here19:39
jdstrandit's*19:39
roadmrinteresting... yes, if it were able to upload and then run the checks, I'd see something different19:40
jdstrandroadmr: maybe smething somewhere has an old crt?19:40
jdstrandroadmr: I need to step into a meeting but will watch backscroll19:41
roadmrjdstrand: sure! I'll triple-check everything. It's unlikely it's outdated but I'll leave no stone unturned19:41
roadmrBEHOOLD!19:58
roadmrjdstrand: instead of PYTHONPATH=./ ./bin/unpack-package ~/Desktop/evil-roadmr-fedora-2018.11.05.01_amd64.snap /tmp/bar try this19:59
roadmrjdstrand: 1- mkdir /tmp/baz19:59
roadmr2- PYTHONPATH=./ ./bin/unpack-package ~/Desktop/evil-roadmr-fedora-2018.11.05.01_amd64.snap /tmp/baz/quux19:59
roadmrthat will repro the crash. No rush, finish your meeting, I'll keep poking and maybe file a bug with a minimal reproducer snap19:59
robert_ancellcprov, did you expect /v2/find?section=gamesq=mmapper to return info on the mmapper snap (it doesn't)20:05
robert_ancellcprov, whoops, was missing the '&'. It does show.20:12
cprovrobert_ancell: it does, right ?20:14
robert_ancellcprov, yes, it was my mistake.20:14
cprovrobert_ancell: np20:14
jdstrandroadmr: ok, I can reproduce that20:17
roadmrjdstrand: the key is the extra directory level; this is what the store does, it creates a NamedTemporaryDir in /tmp to hold each snap and then e.g. says $THE_TEMP_DIR/unpacked for the unpack operation and so on. So that extra level is what's making things funky20:18
roadmrjdstrand: https://bugs.launchpad.net/click-reviewers-tools/+bug/1801788  with mini-snap exposing the problem20:26
mupBug #1801788: Unpacking a snap with mode-555 top-level squashfs-root in a subdirectory of a subdirectory of /tmp fails <Canonical Click Reviewers tools:New> <https://launchpad.net/bugs/1801788>20:26
zygaroadmr, jdstrand: woot, nice!20:36
jdstrandroadmr: can you pull r1152 of the review-tools? it has a fix for ^21:11
roadmrjdstrand: thanks for the quick fix! sure thing21:12
jdstrandroadmr: np. thanks for the reproducer21:14

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!