=== chihchun_afk is now known as chihchun | ||
mup | PR snapcraft#1312 opened: state: fix the name of the source details <Created by elopio> <https://github.com/snapcore/snapcraft/pull/1312> | 04:58 |
---|---|---|
mup | PR snapd#3290 closed: add support for `snap install foo --channel=3.4` <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/3290> | 06:24 |
mup | PR snapd#3307 opened: tests: abstract common dirs which differ on distributions <Created by morphis> <https://github.com/snapcore/snapd/pull/3307> | 06:51 |
Chipaca | *yawn* | 06:56 |
=== chihchun is now known as chihchun_afk | ||
mup | PR core#39 opened: add version-script <Created by mvo5> <https://github.com/snapcore/core/pull/39> | 07:04 |
mup | PR snapd#3308 opened: tests/lib: introduce pkgdb helper library <Created by morphis> <https://github.com/snapcore/snapd/pull/3308> | 07:04 |
Chipaca | ohhhhh | 07:13 |
* Chipaca figured out why so many tests failed | 07:13 | |
Chipaca | MATCH *must* take the thing it's matching from via stdin | 07:13 |
Chipaca | ah well | 07:13 |
Chipaca | school run now, fix later | 07:14 |
=== chihchun_afk is now known as chihchun | ||
zyga | re | 07:16 |
zyga | good morning | 07:16 |
=== JohnAgosta is now known as JohnAgosta-afk | ||
morphis | zyga: morning! | 07:21 |
morphis | zyga: pushed two PRs this morning which start introducing things necessary for fedora/suse spread testing | 07:21 |
zyga | morphis: I saw, I'm reading one now! :) | 07:27 |
morphis | zyga: :-) | 07:29 |
zyga | morphis: commented on 3308 | 07:32 |
pstolowski | morning | 07:36 |
zyga | morphis: commented on 3307 now | 07:37 |
zyga | o/ | 07:38 |
zyga | pstolowski: if you remove this part of https://github.com/snapcore/snapd/pull/3282/files#r115504625 we can land that branch right away | 07:39 |
mup | PR snapd#3282: hooks: default timeout <Created by stolowski> <https://github.com/snapcore/snapd/pull/3282> | 07:39 |
zyga | pstolowski: then the extra change can be made in a separate PR | 07:39 |
pstolowski | zyga, the change from 5 to 10? | 07:40 |
zyga | pstolowski: yes | 07:40 |
pstolowski | zyga, ok, doing | 07:40 |
zyga | thank you! | 07:40 |
zyga | morphis: can you do a quick review of (tiny) https://github.com/snapcore/snapd/pull/3262 | 07:40 |
mup | PR snapd#3262: cmd/snap-confine: aggregate operations holding global lock <Created by zyga> <https://github.com/snapcore/snapd/pull/3262> | 07:40 |
morphis | zyga: sure | 07:41 |
zyga | mvo: does https://github.com/snapcore/snapd/pull/3111 need more work or is that a definite +1? | 07:41 |
mup | PR snapd#3111: snapd: initial implementation for systemd software watchdog for snapd <Created by mvo5> <https://github.com/snapcore/snapd/pull/3111> | 07:41 |
morphis | zyga: replied on both PRs | 07:41 |
zyga | thanks | 07:41 |
mvo | zyga: it fails in tests currently | 07:41 |
mvo | zyga: in myserious ways | 07:41 |
mvo | zyga: let me have a look at it now | 07:42 |
morphis | zyga: done | 07:42 |
mvo | zyga: oh, looks like test-failure are gone | 07:43 |
mvo | zyga: strange | 07:43 |
zyga | morphis: replied | 07:45 |
zyga | mvo: I spent some time last night on gardening PRs, maybe that fixed it | 07:46 |
morphis | zyga: my point for the package installation barely is, when I add a new distribution I just have to edit a single file and not have to touch hundreds of test cases. If a test case needs a specific package there it should express that in a common way "netcat" and the layers below will map that correctly. The really becomes a maintenance burden. Also respect variants of distribution with different package names etc. | 07:48 |
zyga | morphis: that assumes that it is sufficient. My point is that it is a magic assumption that may not hold. What if you need two packages? | 07:49 |
zyga | morphis: I bet you will need to edit all files anyway | 07:49 |
morphis | why that? | 07:49 |
zyga | morphis: and this introduces extra layer that feels wrong | 07:49 |
zyga | morphis: because something may be installed by default on one distro and not in another | 07:49 |
morphis | zyga: I take this really like an "interface" which has different implementations | 07:50 |
zyga | morphis: I think there is no such thing because those packages can have different content | 07:50 |
zyga | morphis: (maybe not snapd but certainly true for others) | 07:50 |
zyga | morphis: anyway, I think another review is needed | 07:50 |
morphis | zyga: so if you express "netcat" and that is shipped by default the apt-get call will either do nothing or we tweak the install function to detect that an do nothing | 07:51 |
morphis | right now it falls back to the supplied name but we can also ignore any errors | 07:51 |
morphis | zyga: actually it feels wrong to me to express all these distribution specifics in the test cases we have | 07:51 |
morphis | zyga: but yeah, lets have another review :-) | 07:52 |
mvo | zyga: re the next cloud update - this only happens inside lxd apparently | 07:53 |
mvo | zyga: just fyi | 07:53 |
zyga | mvo: thanks for the hint! | 07:53 |
mvo | zyga: I could not reproduce on a regulra system and kyrofa also could not reproduce the nextcloud failure on a regular system, but it failed for him inside lxd. I will update the title of the post | 07:54 |
zyga | mvo: I honestly don't have a good feeling for lxd; I don't use it daily and we don't test it at all so there may be dragons of any sizes lurking inside | 07:55 |
Chipaca | TFW you're looking at a test and wondering how it ever passed before | 08:01 |
zyga | Chipaca: https://github.com/snapcore/snapd/pull/3289 needs a 2nd review if you are in review mood | 08:04 |
mup | PR snapd#3289: daemon: do not allow to install ubuntu-core anymore <Created by mvo5> <https://github.com/snapcore/snapd/pull/3289> | 08:04 |
Chipaca | zyga: got a lot of MATCH-related state right now | 08:05 |
zyga | Chipaca: stay there then :) that branch would be great to land too :) | 08:05 |
* zyga -> breakfast | 08:19 | |
mup | PR snapd#3262 closed: cmd/snap-confine: aggregate operations holding global lock <Created by zyga> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/3262> | 08:19 |
morphis | zyga: do you have time today to check https://github.com/snapcore/snapd/pull/3222 again? | 08:22 |
mup | PR snapd#3222: many: fix test cases to work with different DistroLibExecDir <Created by morphis> <https://github.com/snapcore/snapd/pull/3222> | 08:22 |
zyga | morphis: yes, I certainly will; is the discussion there resolved? I recall gustavo had some questions/comments | 08:27 |
morphis | zyga: he didn't responded yet, will ping him as well when he comes online | 08:27 |
morphis | zyga: when you have a few min: https://build.opensuse.org/request/show/494796 | 08:34 |
zyga | morphis: done, you may want to disable leap 42.1 on your branch | 08:37 |
morphis | zyga: yeah, this branch will be removed anyway once you merge :-) | 08:38 |
morphis | zyga: you had any time for the Yocto review? | 08:39 |
zyga | morphis: not quite, it's still open it my tab; I'll do my best to review it before the end of week today | 08:42 |
morphis | zyga: we're not in a hurry | 08:42 |
morphis | zyga: https://github.com/morphis/meta-snappy/pull/10 will only take a minute :-) | 08:43 |
mup | PR morphis/meta-snappy#10: snapd: update to 2.25 release <Created by morphis> <https://github.com/morphis/meta-snappy/pull/10> | 08:43 |
morphis | zyga: thanks! | 08:52 |
* zyga switches to coding mode | 08:56 | |
Chipaca | zyga: ping | 09:40 |
zyga | Chipaca: yes? | 09:46 |
Chipaca | zyga: were you aware that in one of the tests, what used to be “! grep expr /sys/some/file” worked because /sys/some/file did not exist? | 09:47 |
zyga | nope | 09:47 |
zyga | Chipaca: are you uncovering a can of worms just now? | 09:47 |
Chipaca | zyga: /sys/fs/cgroup/devices/snap.test-snapd-tools.env/devices.list in particular | 09:47 |
Chipaca | zyga: should I change the test to check for nonexistance directly, or is it conceivable that the file could sometimes exist, depending on test ordering? | 09:48 |
zyga | Chipaca: I don't know, let me look at the test quickly | 09:48 |
Chipaca | zyga: main/security-device-cgroups/ | 09:49 |
Chipaca | I think the file is removed by the cleanup, but I don't know if that is mandatory or incidental | 09:49 |
Chipaca | and “if [ -e /sys/yadda ]; then MATCH -v expr /sys/yadda; fi” DTRT AFAIK | 09:50 |
zyga | Chipaca: that's cgroup, not a real file | 09:50 |
zyga | Chipaca: and I don't know how it behaves very well honestly | 09:50 |
* zyga looks | 09:50 | |
Chipaca | sure looks like a file to sh :-) | 09:50 |
Chipaca | zyga: meh, then i'll leave the if [ -e | 09:50 |
zyga | sure, my point is that cleanup that removes it is whereR? | 09:50 |
zyga | I actually didn't read the cgroup code much | 09:51 |
* zyga looks | 09:51 | |
Chipaca | zyga: in the same yaml | 09:51 |
Chipaca | udevadm --potato | 09:51 |
zyga | udev-support nees some refactoring love | 09:53 |
zyga | we spawn a shell for each device /o\ | 09:53 |
zyga | Chipaca: so this looks like a layered can of worms, can you please add a FIXME there and I will get back to it | 09:54 |
zyga | this code has not changed at all since ubuntu-app-launcher | 09:54 |
zyga | and that particular test was written by jdstrand | 09:54 |
Chipaca | zyga: # FIXME: this is, apparently, a layered can of worms. Zyga says he needs to fix it. | 09:54 |
Chipaca | zyga: (or tell me what to put so you remember) | 09:55 |
zyga | so I need to look at it for some time before I can make heads or tails | 09:55 |
zyga | Chipaca: yes, that's fine :) | 09:55 |
morphis | ogra_: what is the state on https://bugs.launchpad.net/snappy/+bug/1650688 ? | 10:08 |
mup | Bug #1650688: timedatectl set-timezone fails on UC16 <hwe> <Snappy:Confirmed for ogra> <https://launchpad.net/bugs/1650688> | 10:08 |
ogra_ | morphis, well, i'm not sure mvo's suggestion will work (symlinking a bind-mount) ... | 10:10 |
zyga | ogra_: symlinking a bind mount? | 10:11 |
morphis | ogra_: just wondering as it was set to be fixed for 2.25 and we now released 2.26 and it is still not fixed and two customers are already asking for it | 10:11 |
ogra_ | morphis, i guess what could work is to patch systemd to not handle it as link at all and write the data into it ... | 10:11 |
ogra_ | zyga, https://bugs.launchpad.net/snappy/+bug/1650688/comments/25 | 10:12 |
mup | Bug #1650688: timedatectl set-timezone fails on UC16 <hwe> <Snappy:Confirmed for ogra> <https://launchpad.net/bugs/1650688> | 10:12 |
morphis | ogra_: so what prevents us from patching systemd? | 10:12 |
ogra_ | having to carry that patch ... | 10:13 |
ogra_ | we're just at the point where we can drop systemd from the PPA | 10:13 |
morphis | something we could sent upstream? | 10:13 |
ogra_ | (having it there caused quite some probs regarding libusb (there is a fixed version dependency) | 10:13 |
morphis | ogra_: yeah, ran into that problem multiple times | 10:14 |
ogra_ | morphis, right, it must either be upstreamable or at least suitable for a distro patch we can hand to foundations | 10:14 |
morphis | ogra_: argument for distro patch would be already that Ubuntu Core is a first class citizen in our ecosystem, isn't it? | 10:15 |
ogra_ | yes, but it must still be suitable :) | 10:15 |
morphis | sure, so what is the plan, get it fixed for 2.27? | 10:15 |
ogra_ | we can work towards that, yeah | 10:15 |
morphis | just saying, we should fix it by 2.27, this is now delayed for two releases .. | 10:16 |
ogra_ | right | 10:16 |
Chipaca | so, there's an important difference between "grep foo | wc -l" and "grep -c foo" | 10:26 |
ogra_ | morphis, mvo .... oooh ... http://paste.ubuntu.com/24559979/ we already carry a patch, i guess we just need to adjust it | 10:28 |
mup | PR snapd#3309 opened: interfaces/mount: keep track of kept mount entries <Created by zyga> <https://github.com/snapcore/snapd/pull/3309> | 10:34 |
mup | PR snapd#3310 opened: interfaces/mount: spell unmount correctly <Created by zyga> <https://github.com/snapcore/snapd/pull/3310> | 10:34 |
Chipaca | ogra_: “Forwarded: OMGno, this is a rather nasty hack until we fix system-image to get a writable /etc” | 10:39 |
* Chipaca giggles | 10:39 | |
ogra_ | Chipaca, well :) | 10:39 |
ogra_ | the point is that we have a patch and it doesnt seem to work .... | 10:39 |
mup | PR core#39 closed: add version-script <Created by mvo5> <Merged by zyga> <https://github.com/snapcore/core/pull/39> | 10:39 |
zyga | mvo: I just merged the version script, cannot wait to see this in snap list! | 10:40 |
ogra_ | and i think the reason is: | 10:40 |
ogra_ | 96 if (r >= 0 && startswith(realfile, "/etc/writable")) { | 10:40 |
ogra_ | because ... | 10:40 |
ogra_ | ogra@pi3:~$ ls -l /etc/localtime | 10:41 |
ogra_ | lrwxrwxrwx 1 root root 18 May 12 04:20 /etc/localtime -> writable/localtime | 10:41 |
ogra_ | it looks for an absolute name | 10:41 |
Chipaca | ogra_: what package is that patch for btw? | 10:44 |
ogra_ | systemd | 10:44 |
Chipaca | ah | 10:45 |
Chipaca | ogra_: note that the filename is passed through readlink_and_make_absolute first | 10:48 |
* Chipaca looks at how that one works | 10:48 | |
ogra_ | yes, i''m just digging into that ... it calls readlinkat() in the end | 10:48 |
Chipaca | ogra_: maybe it should be changed to readlink_and_canonicalize? | 10:50 |
ogra_ | because that calls chase_symlinks() ? | 10:51 |
ogra_ | i have the feeling just making the link itself absolute would be sufficient | 10:52 |
ogra_ | funnily the subsequent link *is* absolute | 10:53 |
ogra_ | ogra@pi3:~$ ls -l /etc/writable/localtime | 10:53 |
ogra_ | lrwxrwxrwx 1 root root 27 May 11 06:35 /etc/writable/localtime -> /usr/share/zoneinfo/Etc/UTC | 10:53 |
Chipaca | ah, fair enough | 10:54 |
ogra_ | https://github.com/snapcore/core/blob/master/live-build/hooks/08-etc-writable.chroot#L14 | 10:55 |
mup | PR core#40 opened: make links to /etc/writable absolute <Created by ogra1> <https://github.com/snapcore/core/pull/40> | 11:00 |
ogra_ | :) | 11:00 |
* zyga requests a review for one-liner https://github.com/snapcore/snapd/pull/3310/files | 11:01 | |
mup | PR snapd#3310: interfaces/mount: spell unmount correctly <Created by zyga> <https://github.com/snapcore/snapd/pull/3310> | 11:01 |
zyga | pstolowski: I think you could review this quickly https://github.com/snapcore/snapd/pull/3309/files | 11:02 |
mup | PR snapd#3309: interfaces/mount: keep track of kept mount entries <Created by zyga> <https://github.com/snapcore/snapd/pull/3309> | 11:02 |
mup | PR snapd#3311 opened: difs,interfaces/mount: add support for locking namespaces <Created by zyga> <https://github.com/snapcore/snapd/pull/3311> | 11:02 |
pstolowski | k | 11:02 |
zyga | mvo: and this is for you (no C :) https://github.com/snapcore/snapd/pull/3311/files | 11:02 |
mup | PR snapd#3311: difs,interfaces/mount: add support for locking namespaces <Created by zyga> <https://github.com/snapcore/snapd/pull/3311> | 11:02 |
pstolowski | i'm not sure what's going on with travis test on my #3282.. seems to be infra issue: | 11:06 |
pstolowski | Error preparing linode:ubuntu-core-16-64:tests/regression/ : kill-timeout reached, cannot reconnect to linode:ubuntu-core-16-64 (Spread-15) after reboot: dial tcp 96.126.110.77:22: i/o timeout | 11:06 |
mup | PR snapd#3209 closed: interfaces/mount: add partial implementation of Change.Needed <Created by zyga> <Closed by zyga> <https://github.com/snapcore/snapd/pull/3209> | 11:06 |
zyga | pstolowski: I:m looking too | 11:06 |
zyga | just restarted | 11:06 |
zyga | ogra_: question about https://github.com/snapcore/core/pull/40 -- why were those relative in the first place? | 11:06 |
mup | PR core#40: make links to /etc/writable absolute <Created by ogra1> <https://github.com/snapcore/core/pull/40> | 11:06 |
pstolowski | zyga, I restarted it an hour ago.. with smiliar failure again | 11:07 |
ogra_ | zyga, i have not idea :) | 11:07 |
ogra_ | zyga, that patch came originally from pitti irc | 11:07 |
zyga | ogra_: maybe he remembers, could you ask him? | 11:08 |
zyga | ogra_: just in case we're missing something magic | 11:08 |
ogra_ | zyga, but i assume because he expectes readlink_and_make_absolute() to read a link and make it absolute :P | 11:09 |
* zyga hates expect tests :( | 11:09 | |
ogra_ | *expected | 11:09 |
* zyga goes for lunch and small break | 11:09 | |
ogra_ | zyga, i'm just ruling out the potential that it doesnt with this change ... | 11:10 |
ogra_ | there is no guarantee it fixes anything | 11:10 |
ogra_ | and it wont do any harm either (nothing else cares if the link is relative or absolute) | 11:10 |
zyga | fgimenez: - linode:ubuntu-16.04-64:tests/main/interfaces-openvswitch failed here: https://travis-ci.org/snapcore/snapd/builds/231504228?utm_source=github_status&utm_medium=notification | 11:30 |
zyga | fgimenez: there's no good information there | 11:30 |
zyga | fgimenez: do you know if that is a massive package that has lots of IO during installation? | 11:30 |
mup | PR core#40 closed: make links to /etc/writable absolute <Created by ogra1> <Merged by ogra1> <https://github.com/snapcore/core/pull/40> | 11:31 |
fgimenez | zyga: i don't think so according to https://travis-ci.org/snapcore/snapd/builds/231504228#L3487 it is just 1820kb | 11:32 |
fgimenez | zyga: thanks a lot, i didn't have luck trying to reproduce, will try again | 11:33 |
zyga | fgimenez: it's just random, I don't know if the cause is the same | 11:34 |
fgimenez | zyga: ok, now with the seed it may be easier to confirm/discard the ordering cause, let's see | 11:35 |
zyga | note that it died on timeout so we don't really know what happened :/ | 11:35 |
zyga | Chipaca: do you know MATCH sometimes works with an argument? | 11:50 |
zyga | Chipaca: could it be something as silly as missing -- ? | 11:50 |
Chipaca | zyga: stdin handling in that shell might be wonky | 11:51 |
Chipaca | zyga: like, if you do a match, the second match finds stdin closed (and so reads the args) | 11:51 |
Chipaca | that's an unproven hypothesis right now | 11:51 |
Chipaca | but you asked :-) | 11:51 |
zyga | thanks | 11:53 |
Chipaca | zyga: also, note grep always reads the args | 11:53 |
zyga | I think we should grok this, we rely on match too much :) | 11:53 |
Chipaca | the problem is that MATCH uses 'cat' to read stdin unconditionally | 11:54 |
Chipaca | so if cat is not connected, it hangs | 11:54 |
Chipaca | if cat is closed, it doesn't hang | 11:54 |
Chipaca | i mean if stdin is closed | 11:54 |
Chipaca | because cat returns immediately | 11:54 |
Chipaca | zyga: makes sense? | 11:55 |
Chipaca | needs testing but makes sense to me :-) | 11:55 |
zyga | mhm | 11:55 |
* zyga would like to see MATCH in C | 11:55 | |
Chipaca | wat | 11:55 |
Chipaca | no dude | 11:55 |
Chipaca | i mean, it can be improved, sure | 11:55 |
pstolowski | #3282 failed again, this time on create-key test | 11:55 |
Chipaca | and probably should be, but "write it in c" is a whole new class of pain | 11:56 |
Chipaca | zyga: my pet peeve with it right now is that sometimes we really need the expressiveness of pcre, but we're stuck with -E | 11:59 |
=== alan_g is now known as alan_g|lunch | ||
Chipaca | this stdin is just a quirk | 12:00 |
Chipaca | (compared to that) | 12:00 |
zyga | me nods | 12:01 |
Chipaca | heh, somebody kicked off the test suite again | 12:01 |
Chipaca | but i'm working through the bugs :-) | 12:01 |
Chipaca | (just took me a bit because i didn't have a 32-bit image for local qemu) | 12:02 |
zyga | Chipaca: spread.zygoon.pl | 12:02 |
zyga | :D | 12:02 |
Chipaca | zyga: 1205993472 bytes transferred in 451 seconds (2.55 MiB/s) | 12:03 |
Chipaca | my network is lazy today | 12:04 |
zyga | lazy as in slow? | 12:04 |
zyga | my network is never this fast | 12:04 |
Chipaca | yeah, it should be about 4 times that | 12:05 |
Chipaca | (on paper i mean; in practice it's usually 2x) | 12:06 |
zyga | heh | 12:16 |
zyga | so my one character change PR keeps failing | 12:16 |
zyga | snap-sign this time | 12:16 |
zyga | because expect sucks :/ | 12:16 |
zyga | what I find surprising is that travis fails way way more often than autopkgtests | 12:18 |
zyga | I wonder why that may be | 12:18 |
zyga | since the infra is actually dedicated, not shared | 12:18 |
zyga | I wonder if I should merge the private-interfaces branch now | 12:19 |
zyga | (now == just after release) :) | 12:19 |
Chipaca | zyga: which is your one character change PR? | 12:21 |
zyga | https://github.com/snapcore/snapd/pull/3310 | 12:23 |
mup | PR snapd#3310: interfaces/mount: spell unmount correctly <Created by zyga> <https://github.com/snapcore/snapd/pull/3310> | 12:23 |
Chipaca | aw, i wanted to see the expect failure | 12:24 |
zyga | Chipaca: just killed by timeout | 12:25 |
Chipaca | zyga: expect timeout is how expect fails, always | 12:25 |
Chipaca | it times out waiting for a match | 12:26 |
Chipaca | (well, it can also fail with a syntax error :-) ) | 12:26 |
Facu | sergiusens_, elopio, reading this https://forum.snapcraft.io/t/in-progress-snapcraft-2-30/347 and wondering if it should mention something about https://bugs.launchpad.net/snapcraft/+bug/1670471 or https://bugs.launchpad.net/snapcraft/+bug/1686162 | 12:27 |
mup | Bug #1670471: Bad message after failed release <Snapcraft:Fix Committed by facundo> <https://launchpad.net/bugs/1670471> | 12:27 |
mup | Bug #1686162: Support "branch"es in Store responses <Snapcraft:Fix Committed by facundo> <https://launchpad.net/bugs/1686162> | 12:27 |
jdstrand | zyga, Chipaca: iirc, once a cgroup is setup the dirs and files stay until after a reboot. I recall trying to cleanup but being unable to. aiui, all you can do is reset the cgroup but the files are still there. it's possible there is a way that I didn't find | 12:27 |
zyga | jdstrand: interesting, I will look into this after working on namespaces | 12:29 |
pedronis | mvo: hi, I updated the forum entry about the repair assertion to match the state of your branch | 12:31 |
mvo | pedronis: thank you very much | 12:31 |
Chipaca | man, tests/main/interfaces-openvswitch needs to be hacked into shape | 12:33 |
Chipaca | it takes _ages_ | 12:33 |
Chipaca | (to the point that it often times out entirely) | 12:33 |
zyga | Chipaca: yep | 12:34 |
Chipaca | in _prepare_ | 12:34 |
* Chipaca looks | 12:34 | |
mup | PR core#41 opened: limit the version string to 32chars (this is what the store allows) <Created by mvo5> <https://github.com/snapcore/core/pull/41> | 12:36 |
zyga | Chipaca: apt-get install | 12:38 |
Chipaca | zyga: maybe it's just that the base prepare takes ages | 12:39 |
Chipaca | and this will push it over the edge | 12:39 |
Chipaca | zyga: because that apt-get got the stuff in 0s | 12:39 |
Chipaca | then, sure, it needs to dance the install dance | 12:39 |
Chipaca | but, we probably should bump the prepare timeout overall if it's this close | 12:39 |
zyga | can we do that for a particular test case? | 12:40 |
mvo | pedronis: also thanks a bunch for your review, I will add the missing test now | 12:40 |
Chipaca | zyga: yes | 12:41 |
Chipaca | niemeyer: is kill-timeout for the whole test, or separately for prepare and execute? | 12:41 |
niemeyer | Yo | 12:44 |
niemeyer | Chipaca: for each script individually | 12:44 |
Chipaca | niemeyer: and prepare and execute are scripts? | 12:45 |
Chipaca | niemeyer: hello! good morning :-) | 12:45 |
ogra_ | mvo, so that rsyslog removal ... | 12:47 |
ogra_ | mvo, any opinion yet ? | 12:47 |
Chipaca | huh, another test that was psasing who-knows-how | 12:53 |
zyga | ogra_, mvo: tests broken because of the version script on core snap | 12:53 |
zyga | mvo: specifically tests/main/listing | 12:53 |
kyleN | hi ara | 12:55 |
Chipaca | mvo: fgimenez: question about tests/main/ubuntu-core-create-user | 12:56 |
zyga | hmm | 12:56 |
zyga | or maybe not? | 12:56 |
Chipaca | mvo: fgimenez: the test currently purports to check that the output of create-user on a managed device (without the --force-managed flag) is “error: while creating user: cannot create user "nosuchuser@example.com"” | 12:57 |
Chipaca | mvo: fgimenez: but that's not the output from that command (and it hasn't been that for a while) | 12:58 |
Chipaca | I don't pretend to understand how this test passed in the past | 12:58 |
Chipaca | but i want to know if the current message is what it should be, in which case i will make the test check for that | 12:58 |
Chipaca | and if it ins't, i'll fix the message | 12:58 |
Chipaca | mvo: fgimenez for the record the message currently is “error: while creating user: cannot create user: device already managed” | 12:58 |
MrGeneral | howdy | 12:59 |
MrGeneral | Is it possible to run snap @ debian jessie? | 13:00 |
zyga | MrGeneral: hello | 13:00 |
zyga | ah standup time | 13:00 |
* zyga never remembers debian codenames | 13:00 | |
zyga | testing is OK | 13:00 |
MrGeneral | All good. | 13:00 |
zyga | before is too old | 13:00 |
MrGeneral | hmhm I see | 13:00 |
MrGeneral | so I need to enable testing I guess. | 13:00 |
MrGeneral | too old? | 13:00 |
MrGeneral | hmhm | 13:00 |
pedronis | Chipaca: the test is wrong it seems, it should try the wrong user before it sets up a real one | 13:00 |
pedronis | Chipaca: I mean it seems to be mixing up failure modes | 13:01 |
zyga | MrGeneral: yes, kernel and systemd and perhaps something else | 13:01 |
fgimenez | Chipaca: something is indeed wrong there, is a "$" missing https://github.com/snapcore/snapd/blob/master/tests/main/ubuntu-core-create-user/task.yaml#L39 ? | 13:02 |
Chipaca | fgimenez: hah! | 13:02 |
Chipaca | omg | 13:02 |
Chipaca | fgimenez: yes, that is missing a $, but, even that does not fail | 13:03 |
=== alan_g|lunch is now known as alan_g | ||
Chipaca | fgimenez: it never reaches there | 13:03 |
Chipaca | fgimenez: (or if it reaches, it isn't enough to time out) | 13:03 |
Chipaca | to error out I mean | 13:03 |
mup | PR snapd#3312 opened: DO NOT MERGE YET: add coveralls.io integration <Created by mvo5> <https://github.com/snapcore/snapd/pull/3312> | 13:05 |
MrGeneral | Got it, thank you zyga :) | 13:05 |
mup | PR snapd#3312 closed: DO NOT MERGE YET: add coveralls.io integration <Created by mvo5> <Closed by mvo5> <https://github.com/snapcore/snapd/pull/3312> | 13:08 |
mup | PR snapcraft#1305 closed: Use architectures field which existed in older LXD <Created by kalikiana> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/1305> | 13:13 |
mup | PR snapd#3313 opened: send things to codecov.io <Created by mvo5> <https://github.com/snapcore/snapd/pull/3313> | 13:14 |
mup | PR snapcraft#1249 closed: Add Linux Mint support <Created by nefelim> <Closed by sergiusens> <https://github.com/snapcore/snapcraft/pull/1249> | 13:16 |
ogra_ | morphis, do we have any snap that uses the timedatectl dbus call that i could use to test https://bugs.launchpad.net/snappy/+bug/1650688 ? | 13:28 |
mup | Bug #1650688: timedatectl set-timezone fails on UC16 <hwe> <Snappy:Confirmed for ogra> <https://launchpad.net/bugs/1650688> | 13:28 |
ogra_ | calling it locally will not use dbus but directly call the tool whihc works fine | 13:29 |
morphis | ogra_: not really | 13:29 |
ogra_ | i'm pretty sure the missing line for /etc/localtime in the interface will also have some effect here | 13:30 |
mup | PR snapcraft#1306 closed: recording: record global build-packages installed on the host <Created by elopio> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/1306> | 13:31 |
mup | PR snapd#3314 opened: tests: allow 16-X.Y.Z version of core snap <Created by zyga> <https://github.com/snapcore/snapd/pull/3314> | 13:32 |
mup | PR snapd#3315 opened: rename host's http proxy env vars <Created by fgimenez> <https://github.com/snapcore/snapd/pull/3315> | 13:33 |
mup | PR snapd#3316 opened: make /etc/localtime writable in timezone_control <Created by ogra1> <https://github.com/snapcore/snapd/pull/3316> | 13:36 |
morphis | ogra_: I got a basic qemu snap going which gives me latest git with raspi2 machine and got the kernel from the pi2-kernel snap booting | 13:53 |
morphis | ogra_: however didn't got uboot started yet and because of that the kernel fails to find its rootfs | 13:53 |
ogra_ | push it to edge .... happy to poke at it | 13:54 |
diddledan | what is that error supposed to mean?: [Error 21] Is a directory: '/build_corebird/prime/usr/lib/locale' <-- I told it I wanted 'usr/lib/locale' in the prime section of my yaml | 13:55 |
ogra_ | jdstrand, i'd appreciate a review of https://github.com/snapcore/snapd/pull/3316 | 14:02 |
mup | PR snapd#3316: make /etc/localtime writable in timezone_control <Created by ogra1> <https://github.com/snapcore/snapd/pull/3316> | 14:02 |
mup | PR core#41 closed: limit the version string to 32chars (this is what the store allows) <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/core/pull/41> | 14:03 |
zyga | mvo: can you review https://github.com/snapcore/snapd/pull/3314 to unbreak master | 14:09 |
mup | PR snapd#3314: tests: allow 16-X.Y.Z version of core snap <Created by zyga> <https://github.com/snapcore/snapd/pull/3314> | 14:09 |
pedronis | niemeyer: this is the first issue(s) about auto-connect to address right? https://forum.snapcraft.io/t/what-should-the-auto-connect-logic-be-like/312 | 14:24 |
zyga | pedronis: one thing that is perhaps related is that I want to add a small tweak to interfaces to give them a way to share more meta-data | 14:26 |
zyga | pedronis: we could then share more useful things, like auto-connect 1-no-N flags or anything else we need internally | 14:27 |
zyga | pedronis: (and useful things like human readable descriptions we just keep as code comments) | 14:27 |
ogra_ | niemeyer, in case you feel like reading it https://lists.ubuntu.com/archives/ubuntu-devel/2017-January/039634.html ... (for a boring afternoon or so) | 14:27 |
ogra_ | niemeyer, thats the discussion about dropping syslog from the distro | 14:28 |
niemeyer | ogra_: Can you please mention that in the forum thread? It's good background. | 14:28 |
ogra_ | yep | 14:28 |
niemeyer | pedronis: Yeah | 14:28 |
Chipaca | looks like we ran out of travis coin | 14:30 |
pedronis | zyga: that's interesting but as niemeyer says on the forum, I think we should start fixing the initial bug before adding stuff | 14:30 |
niemeyer | Chipaca? | 14:30 |
zyga | pedronis: yes, as I said it's not directly related (my motivation is to get the user visible interface descriptions available in the API) but once we have a hold of the problem and know what to do it may be a handy mechanism | 14:31 |
pedronis | at the moment all auto-connecting interfaces are 1-to-N afaik | 14:31 |
pedronis | I mean from slot (1) to plugs (N) | 14:32 |
mup | PR snapd#3317 opened: many: start implenting "base" snap type on the snapd side <Created by mvo5> <https://github.com/snapcore/snapd/pull/3317> | 14:33 |
zyga | pedronis: perhaps | 14:34 |
zyga | pedronis: things like gpio and other hardware might not be | 14:35 |
zyga | (it may be but might not be) | 14:35 |
zyga | we need to land https://github.com/snapcore/snapd/pull/3314 to fix master | 14:38 |
mup | PR snapd#3314: tests: allow 16-X.Y.Z version of core snap <Critical> <Created by zyga> <https://github.com/snapcore/snapd/pull/3314> | 14:38 |
zyga | please review | 14:38 |
* zyga just lands it | 14:42 | |
mup | PR snapd#3314 closed: tests: allow 16-X.Y.Z version of core snap <Critical> <Created by zyga> <Merged by zyga> <https://github.com/snapcore/snapd/pull/3314> | 14:43 |
ogra_ | niemeyer, hmm, is there a way that a non-creator of a topic could set a category in discourse (perhaps via a config option) ... | 14:48 |
ogra_ | niemeyer, i.e. https://forum.snapcraft.io/t/3g-usb-modems-compatible-with-modem-manager-snap-on-ubuntu-core/582 would fit into "device" but the original creator didnt set it ... would be a cool feature if others could add a category in that case | 14:49 |
pedronis | ogra_: let me see | 14:49 |
mvo | zyga: nice one, thank you | 14:49 |
pedronis | ogra_: done | 14:49 |
pedronis | ogra_: I can change categories, it's one of those things discourse grants after some use afaict | 14:50 |
ogra_ | pedronis, oh | 14:50 |
pedronis | part of one of the badges | 14:50 |
* ogra_ thought he did earn enough badges yet ... seems not :P | 14:51 | |
pedronis | ogra_: I got it with "Regular" | 14:51 |
pedronis | you should have it too | 14:51 |
pedronis | afaict | 14:51 |
pedronis | https://forum.snapcraft.io/badges/3/regular | 14:51 |
zyga | mvo: OK to land https://github.com/snapcore/snapd/pull/3295 ? | 14:52 |
mup | PR snapd#3295: interfaces/builtin: make all interfaces private <Created by zyga> <https://github.com/snapcore/snapd/pull/3295> | 14:52 |
pedronis | ogra_: "You can now recategorize and rename topics..." | 14:52 |
niemeyer | ogra_: Yeah, moderators can do it | 14:52 |
pedronis | niemeyer: also Regulars | 14:53 |
niemeyer | pedronis: Ah, indeed! | 14:53 |
niemeyer | discourse++ | 14:53 |
pedronis | ogra_: click the pencil on the right of topic titles | 14:53 |
ogra_ | oh man | 14:53 |
ogra_ | yeah, i can | 14:53 |
diddledan | how do I get snapcraft to prime a directory? it's complaining that: [Error 21] Is a directory: 'path I want included' | 14:53 |
ogra_ | it didnt strike me that i need to edit the headline for it | 14:53 |
pedronis | you should click all that is shiny :) | 14:54 |
morphis | ogra_: can we change https://bugs.launchpad.net/snappy/+bug/1650688 to be in-progress? | 14:54 |
mup | Bug #1650688: timedatectl set-timezone fails on UC16 <hwe> <Snappy:Confirmed for ogra> <https://launchpad.net/bugs/1650688> | 14:54 |
ogra_ | diddledan, catching the snapcraft guys might be easier on rocket.ubuntu.com | 14:54 |
ogra_ | morphis, done | 14:54 |
morphis | ogra_: thanks | 14:55 |
morphis | niemeyer: you have time to have another look on https://github.com/snapcore/snapd/pull/3222 today? | 14:55 |
mup | PR snapd#3222: many: fix test cases to work with different DistroLibExecDir <Created by morphis> <https://github.com/snapcore/snapd/pull/3222> | 14:55 |
diddledan | really? why have they moved over therE?! | 14:55 |
niemeyer | morphis: Yeah | 14:55 |
diddledan | I hate this replacing with shiny just because | 14:56 |
pedronis | morphis: the icon here seems broken (or a cache issue?): https://forum.snapcraft.io/badges/101/fedora | 14:56 |
morphis | niemeyer: thanks! | 14:56 |
diddledan | there wasn't anything wrong with IRC! </rant> | 14:56 |
morphis | pedronis: hm, didn't created that badge so not sure where it takes the icon from | 14:56 |
morphis | niemeyer: ^^ | 14:56 |
morphis | pedronis: https://snapforum.s3.amazonaws.com/original/1X/bfb1e013ef4b1f1276a1ca3d38a40a789a056bb8.png -> doesn't exist .. | 14:56 |
niemeyer | Wow.. strange | 14:57 |
niemeyer | Let me check that | 14:57 |
ogra_ | diddledan, nobody says there is anything wrong with it (and they are still idling here but probably dont check it as often as rocket) | 14:57 |
niemeyer | Apparently I screwed up by removing a reference to the image, and Discourse is smarter than I expected and removed the image that had no references | 15:02 |
niemeyer | morphis: Fixed, thanks for the note | 15:07 |
niemeyer | pedronis: ^ | 15:08 |
pedronis | thx | 15:08 |
pedronis | I just noticed because I went to the badges page | 15:08 |
* niemeyer => lunch | 15:18 | |
pedronis | niemeyer: we should remember to remove upcoming from stuff that is done (I think you listed that in the process post) | 15:24 |
pedronis | mvo: Chipaca: is this resolved https://forum.snapcraft.io/t/hashsum-failures-during-tests/198 ? | 15:31 |
Chipaca | I don't think so | 15:32 |
pedronis | niemeyer: mvo: I removed the "upcoming" tag from some stuff that was done | 15:37 |
morphis | pedronis: are you planing to solve and implement https://forum.snapcraft.io/t/gadget-snap-config-defaults-dont-work/409/4 for 2.27? | 15:47 |
pedronis | morphis: yes (https://forum.snapcraft.io/t/next-snapd-2-27/572 ) | 15:48 |
pedronis | morphis: have a plan, should have PR(s) on Mon or Tue | 15:48 |
morphis | pedronis: awesome! | 15:48 |
=== chihchun is now known as chihchun_afk | ||
pstolowski | pedronis, do you know from top of your head where do we create 'update-aliases' task? for some reason I can only find it in tests... but not where it's created | 16:19 |
niemeyer | pedronis: +1 | 16:20 |
niemeyer | pedronis: Thanks for cleaning it up | 16:20 |
pstolowski | 2017-05-12 15:40:57 Failed tasks: 1 | 16:26 |
pstolowski | - linode:ubuntu-16.04-32:tests/main/create-key | 16:26 |
pstolowski | 2017-05-12 15:40:57 Failed task prepare: 1 | 16:26 |
pstolowski | - linode:ubuntu-16.04-32:tests/main/completion | 16:26 |
pstolowski | #3282 failed again ^ ... | 16:26 |
pedronis | pstolowski: there is no such task | 16:38 |
pedronis | pstolowski: aliasesv2.go has a comment listing all the aliases tasks | 16:38 |
pedronis | pstolowski: why the question? | 16:39 |
pedronis | pstolowski: ah, update-aliases, it's a backend operation, not a task | 16:40 |
pstolowski | pedronis, cause I unexpectedly get this in of the my failing tests; and grepping master gives a bunch of those.. | 16:40 |
pedronis | lots of aliases tasks end up creating those | 16:40 |
pedronis | pstolowski: I fear I need to see the test to help | 16:41 |
pedronis | pstolowski: anyway under normal circucmstances (refresh/install etc) it comes from setup-aliases | 16:42 |
pstolowski | pedronis, i see. thanks | 16:42 |
pedronis | also I misremembered the comment listing tasks is in handlers.go | 16:43 |
mup | PR snapcraft#1313 opened: Meta: Version from deb <Created by piso77> <https://github.com/snapcore/snapcraft/pull/1313> | 16:46 |
zyga | 2017-05-12 16:53:57 Cannot allocate linode:ubuntu-16.04-64: cannot create Linode disk with ubuntu-16.04-64: you do not have enough unallocated storage to create this Disk (608 requested, but only 0 available) | 16:59 |
=== alan_g is now known as alan_g|EOW | ||
niemeyer | Fixed.. that was Spread-06 FWIW | 18:28 |
niemeyer | All others are clean | 18:28 |
mup | PR snapd#3295 closed: interfaces/builtin: make all interfaces private <Created by zyga> <Merged by zyga> <https://github.com/snapcore/snapd/pull/3295> | 18:47 |
mup | PR snapd#3310 closed: interfaces/mount: spell unmount correctly <Created by zyga> <Merged by zyga> <https://github.com/snapcore/snapd/pull/3310> | 18:47 |
=== JanC_ is now known as JanC | ||
mup | PR snapcraft#1314 opened: catkin plugin: add support for rosinstall files <Created by kyrofa> <https://github.com/snapcore/snapcraft/pull/1314> | 21:43 |
nacc | sigh, i think i've found a relatively serious issue with the store, on a friday :) | 22:44 |
nacc | oh nm, it's a bug in my snapcraft.yaml in the process of renaming a snap | 22:44 |
nacc | it does feel like the store should not try to cross-build/upload snaps: https://launchpad.net/~nacc/+snap/git-ubuntu/+build/37953 | 22:44 |
nacc | snap is git-ubuntu, but the generated snap is usd-nacc, and yet it tried to upload it :) | 22:45 |
nacc | hrm, but it does seem like lp is a bit confused: https://launchpad.net/~nacc/+snap/git-ubuntu/+build/37953 | 22:53 |
nacc | the 'manage this package in the store' url ends up pointing to the wrong snap (the one built, admittedly) | 22:54 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!