[02:35] jdstrand: I think he won it when we had the term husks in the snapd code base ;-) [02:42] PR snapcraft#1272 closed: tests: initial setup for the snapcraft snap tests with spread [02:45] PR snapcraft#1279 closed: misc: rename the snap dir variable to prime dir === chihchun_afk is now known as chihchun [04:19] jdstrand: you can now go back to use normal snapcraft-preload as fix for you issue has been merged upstream [05:30] PR snapcraft#1238 closed: beta [07:03] mvo: hi, I got a +1 from Gustavo for snapd#3220, maybe you can finish the review you were doing? some points I addressed in snapd#3245 [07:03] PR snapd#3220: many: implement snap prefer (aliases v2) [07:03] PR snapd#3245: many: aliases v2 cleanups [07:11] pedronis: thanks, will do [07:24] mvo: any reason not to merge #3244, I don't see anything controversial in there [07:27] pedronis: I tihnk its fine to merge, gustavo asked for those changes so he might want to review but afaict its all addressed in this pr [07:38] mvo: no, now that I look at those commetns, there's a couple of things missing [07:42] Hi there. Has anyone experimented with making a build server for snaps? [07:45] beruic: yes [07:45] beruic: build.snapcraft.io [07:45] beruic: :-) [07:45] beruic: using snapcraft itself it is not that hard [07:48] @zyga Awesome. I'm considering trying to distribute my companys Django project on Core using snaps. [07:48] beruic: No such command! [07:49] zyga Awesome. I'm considering trying to distribute my companys Django project on Core using snaps. [07:49] Sorry for double posting [07:50] Does anyone know of a Freenode client for Ubuntu? Preferably one that comes as a snap package ;) [07:50] mvo: there's just a couple minor things there now that I looked more carefully [07:51] beruic: I don't know of any, I'd love to see polari IRC from gnome but that may require some more work on gnome/gui side; irssi should be easy to snap [07:54] zyga: I think there's a script to update the board, but gustavo needs to run it [07:54] it's semi-automated, not fully afaiu [07:56] pedronis: aha, good to know [07:59] mvo: seems there's a few things for 2.25 that we didn't get to do (https://forum.snapcraft.io/t/next-snapd-2-25/197), right? I suppose we didn't consider there was Easter [08:02] I'm having trouble figuring out the following about build.snapcraft.io: 1. I'm pretty sure it only connects to GitHub (we use Bitbucket :( ) 2. Is it possible to publish snaps that are only internally available, or does build.snapcraft.io always publish for public access? [08:07] beruic: I think it's meant for public snaps, you can build stuff yourself with snapcraft (from bitbucket) and as long as you don't push it to any channel you should be OK [08:07] beruic: you can get more answers by asking on forum.snapcraft.io [08:08] zyga: Thank you :) === JanC is now known as Guest35192 === JanC_ is now known as JanC [08:25] mvo: I see more sleep tweaks on https://github.com/snapcore/snapd/pull/3242 [08:25] PR snapd#3242: tests: tweak time for econnreset test a bit more [08:25] mvo: is there a way to make that test reliable without that? [08:26] tests and sleep, almost always tears [08:35] mvo: I made a suggestion on #3220 let me know what you think there? [08:37] mvo: what's the plan for 2.25? [08:39] error: cannot get data for "complexion": [0xc82016e300 0xc82016e480 0xc82016e600] [08:39] ^^^ WTF [08:39] zyga: maybe, I think it will need a muchbigger snap [08:40] Chipaca: hmm [08:40] zyga: the problem is that there is a bit of a race, when the print happens it has not connected yet, I could make another debug print when its connected. we want to only retry on rst [08:40] zyga: when rst happens in the middle not at the start [08:40] Chipaca: that's the snap you added in your tests, right? [08:40] zyga— it's a snap, yes [08:40] mvo: could we do something, anything, via a helper proxy that we control, to make the test easier to write [08:41] zyga— i did 'snap try complexion' [08:41] it printed that [08:41] Chipaca: what are those [] numbers? [08:41] but that was apparently a success message? because the snap was installed after that [08:41] zyga— I have absolutely no idea [08:41] zyga— that was the output of 'snap try' [08:41] pedronis: checking your suggestion now. the plan for 2.25 is to push it out once the aliases are ready. it looks like in the standup we can discuss what (if anything) else is a must for 2.25 [08:41] pedronis: definitly want to release today [08:41] Chipaca: ./cmd/snap/cmd_snap_op.go [08:42] $ sudo snap try complexion/ [08:42] error: cannot get data for "complexion": [0xc82016e300 0xc82016e480 0xc82016e600] [08:42] mvo: ok, aliases should be definitely in soon, mostly waiting on your feedback there ;) ? [08:42] Chipaca: the list is %v of `snaps` which came out of cli.List() [08:42] Chipaca: looks like wonky String method [08:42] that doesn't take pointer receiver [08:42] zyga: sending RST is tricky, the best way I found was via the iptables, however something like trickle might help to ensure the download is not too quickly over. or just a huge snpa, maybe I just go with that [08:43] looks like snap.List returned three items despite being given a snap name [08:43] Chipaca: yep [08:43] wonky indeed D: [08:43] why [08:43] Chipaca: let's patch the string methodfirst [08:43] all i wanted to do was manual testing of complexion [08:43] pedronis: ha! on it :) [08:44] http://paste.ubuntu.com/24465628/ [08:44] Chipaca: try with this [08:44] so this is something that's broken on master [08:44] PR snapd#3220 closed: many: implement snap prefer (aliases v2) [08:46] zyga— it's a %v, I don't think String is called on it [08:47] Chipaca: mmh [08:47] Chipaca: aha, that's also possible [08:49] ah, look at that, it would be called [08:50] ogra_: how can I watch your webinar? [08:58] hi mvo, this is the test i mentioned for the new local assertions capabilities of prepare-image https://github.com/mvo5/snappy/pull/16 it passes locally on your branch, let me know if i should change anything about it [08:58] PR mvo5/snappy#16: tests: add spread test for assertions in prepare-image's extra-snaps [08:58] mvo: updated https://github.com/snapcore/snapd/pull/3244 for pedronis comments [08:58] PR snapd#3244: many: fix review comments from PR #3177 [09:02] fgimenez_: nice work, thats some magic in there :) I have a look [09:02] morphis: thanks, also special thanks to pedronis for double checking that one [09:04] mvo: snapd#3245 is ready for review, it contains some of the follow ups I promised in previous reviews [09:04] PR snapd#3245: many: aliases v2 cleanups [09:04] plus other cleanups [09:06] Chipaca: maybe you could give it a look as well? ^ (it's smallish, repetitive) [09:10] Son_Goku: ping [09:14] gah, I was debugging something that kept failing because I forgot to merge mater [09:14] master [09:15] fgimenez_: could you please merge master and push again in the prepare-image-assert-test branch? there is some unrelated diff and I updated my branch but it looks GH is not updating the diff until you push something too [09:16] mvo: sure on it [09:16] ta [09:19] morphis: static checks are failing in your x11 pr, should be a trivial fix .) [09:19] mvo: ok [09:20] mvo: fixed [09:21] zyga: I will pull in 3243, jamie was asking for a test, maybe we can get one in a followup? [09:21] mvo: yes, I'll do one in a follow-up [09:21] mvo: I was looking at that already but it will need some extra mocking [09:21] mvo: and this will make CE happy [09:21] 3238 needs a second review [09:22] Chipaca: maybe you -^ ? branch is very small [09:22] PR snapd#3243 closed: cmd/snap-confine: don't use apparmor if it is disabled on boot [09:22] zyga: wasn't there a test we disabled about that? [09:23] pedronis: yes but the test depends on an unreleased kernel, CE has a kernel that they can use this on already [09:23] ah, I see [09:31] mvo: fixed, now it shows only the test files https://github.com/mvo5/snappy/pull/16 [09:31] PR mvo5/snappy#16: tests: add spread test for assertions in prepare-image's extra-snaps [09:38] morphis: pong [09:38] Son_Goku: you saw my patch recently for the rpm to enable unit tests for the vendorized build [09:38] yes [09:38] I wanted to ask you something [09:39] shoot :-) [09:39] does the unit tests require internet access? [09:39] fgimenez_: thanks, reading now [09:39] because if they do, then I'm not merging this [09:39] Son_Goku: they shouldn't as far as I know [09:39] Son_Goku: no, they don't [09:39] Son_Goku: regular unit tests [09:39] Son_Goku: spread requires internet access but that's not tested at package build time [09:39] okay, then [09:40] morphis, could you link me the patch again? I seem to have downloaded it on a computer that's not this one [09:40] Son_Goku: sure [09:40] zyga, you got to ask thibautr__ [09:42] PR snapcraft#1219 closed: kernel plugin: learn how to assemble the ubuntu config using kconfigflavour [09:53] morphis, nvm, I found it [09:54] Son_Goku: ok, was finishing something else but would have pasted it afterwards [09:54] zyga: what do you think about https://github.com/morphis/snapd/commit/75888e9198aedda1a862b7db496733e308b38a49 ? [09:55] morphis: that won't work [09:56] morphis: /etc/ssl is a symlink [09:56] no way to bind mount over one [09:56] (sorry) [09:56] sure but we have the target in /usr/share too [09:56] zyga: you mean /etc/ssl is a symlink on suse? [09:56] yes [09:56] no cheap way out other than to stop sharing /etc [09:57] ok that makes thing more complicate [09:57] zyga: " What we will then see is a symlink from /etc/ssl./certs to /var/lib/ca-certificates/pem that is just broken (the symlink)." [09:57] if I read that correct /etc/ssl/certs is the symlink [09:57] not /etc/ssl [09:58] morphis: aha, hmm [09:58] morphis: maybe that will work then [09:58] zyga: let me verify that [09:58] morphis: I still don't like it but it looks like a smaller change [09:58] morphis: please do [09:58] zyga: it's a cheap fix for the meantime [09:59] instead of taking quite some time to rework the rest [09:59] morphis, I don't have the patch you added to the spec, though [09:59] Son_Goku: ah [10:02] morphis, which sort of makes this pointless :) [10:03] mvo— you're cutting the release, yes? [10:04] mvo: landed #3245 [10:04] PR snapd#3245 closed: many: aliases v2 cleanups [10:07] Son_Goku: you mean 0001-many-fix-test-cases-to-work-with-different-DistroLib.patch ? [10:07] yes [10:07] that one is merged into 2.25 [10:07] I didn't even see that PR [10:07] sorry, is up for merge [10:07] Son_Goku: https://github.com/snapcore/snapd/pull/3222 [10:07] PR snapd#3222: many: fix test cases to work with different DistroLibExecDir [10:08] zyga here you go for the webinar... [10:08] https://www.brighttalk.com/webcast/6793/248379 [10:08] Son_Goku: 2.26 is more likely [10:09] thibautr__: thanks! [10:11] zyga: ok, /etc/ssl isn't a symlink on suse [10:11] morphis: ha [10:11] morphis: ok, let's do this, add this to suse build [10:11] morphis: as a distro patch [10:11] morphis: I'd love to see it work [10:12] zyga: :-) [10:12] this patch makes no sense [10:13] Son_Goku: why not? [10:14] not the suse one, the PR [10:14] I don't even know what you're changing in the suse one :) [10:14] ah, sorry [10:15] Son_Goku: ? [10:15] morphis: I left a comment in the PR, but I'm confused because it looks like you're testing the wrong thing now [10:15] Son_Goku: no you don't [10:16] so why is everything changing from distrolibexec to corelibexec? [10:16] Son_Goku: there are many reexec cases which were properly handleed [10:16] but those are never used on fedora/suse [10:16] so they were wrong for quite some time [10:16] and rexec never would have worked [10:18] ah [10:19] zyga: didn't you say you merged https://build.opensuse.org/request/show/490989 ? [10:20] morphis: I thought I did [10:20] wtf is rcsnapd? [10:20] Son_Goku: https://en.opensuse.org/openSUSE:Packaging_checks#suse-missing-rclink [10:20] it's not really of any use anymore but rpmlint complains about it .. [10:21] that's a bug, actually [10:21] morphis: aha, did that again [10:21] Son_Goku: maybe [10:21] because they just decided to get rid of that stuff a couple of weeks ago [10:21] Son_Goku: then we can drop this in the near future again [10:21] (yes, I'm involved in openSUSE, too :) ) [10:21] the only distribution family I'm not all that involved in is Debian :) [10:22] I even have a somewhat tenuous link to Yocto [10:22] Son_Goku: :-) [10:23] the reasons for not getting involved in Debian/Ubuntu are sadly long and terrible [10:23] I told mvo once, quite a while ago [10:24] oh, and ximion knows too :P [10:24] I guess if I get the time, I'll help Debian complete the transition to dracut [10:24] I already told ximion I'd help him on it, at least :) [10:28] Son_Goku: any progress on the postrm script? [10:29] it needs reworking, but I think I'm nearly there :) [10:29] great! [10:35] Chipaca: release> yes, very soon. was thinking right after the standup so that we can discuss as a group if there is anything missing that must go in but I think we are good [10:37] PR snapd#3238 closed: cmd/snap-confine: re-enable re-assciate fix for CE [10:42] PR snapcraft#1285 opened: kernel plugin: learn how to assemble the ubuntu config using kconfigflavour [10:45] mvo, my branch failed on econnreset test yesterday - https://s3.amazonaws.com/archive.travis-ci.org/jobs/225939208/log.txt?X-Amz-Expires=29&X-Amz-Date=20170427T104419Z&X-Amz-Algorithm=AWS4-HMAC-SHA256&X-Amz-Credential=AKIAJRYRXRSVGNKPKO5A/20170427/us-east-1/s3/aws4_request&X-Amz-SignedHeaders=host&X-Amz-Signature=014a53a5f0d9f89ee092ab7db9df6258e5f7cc3fa15bd68841c80719c2148446 - did you land any fixes since? shall I just retry? [10:48] pstolowski: yeah, I'm working on making them more reliable [10:48] pstolowski: sorry for that, just retry for now [10:48] zyga: do you think you could look at 3232? [10:49] zyga: a (nice) suggestion from samuele in there [10:49] zyga: I think this can go in then [10:49] (assuming it gets a second review) [10:50] mvo: looking [10:51] mvo: the "applied"? [10:51] mvo: I'd rather not change that in this branch alone, it's much cleaner to change it in one patch across the whole tree (there are other places) [10:52] zyga: ok, do you want to do a followup? works for me [10:52] mvo, sure, np. thanks! [10:52] mvo: yes, when other things land [10:52] mvo: this is in 3 or 4 places (the wording) [10:52] morphis: alright, so I'm trying this now... https://koji.fedoraproject.org/koji/taskinfo?taskID=19231693 [10:53] Son_Goku: it's a vendorized build? [10:53] nope [10:53] then it wont work [10:54] Son_Goku: I still need to file bugs for updates to a few golang-* packages and we need https://bugzilla.redhat.com/show_bug.cgi?id=1444819 [10:54] so only the vendorized build works for now [10:55] then I'll test that locally [10:57] Son_Goku: I will update https://bugzilla.redhat.com/show_bug.cgi?id=1444819 later today [10:57] okay === chihchun is now known as chihchun_afk [11:20] mvo: this fell through the cracks for 2.25 https://forum.snapcraft.io/t/gadget-snap-config-defaults-dont-work/409 I created the topic now, and marked upcoming (2.26) [11:20] mvo: don't know what else we missed in 2.25 [11:24] mvo: is 2.25 tagged now? [11:26] zyga: it's not tagged yet [11:26] zyga: he said after the standup [11:26] aha [11:27] I missed that , sorry [11:36] PR snapcraft#1280 closed: Added support for 'branches' in Store responses (release, close, and status) [11:37] yay [11:45] raah rahh [11:45] moo? [11:45] ogra_— bash is a bundle of sticks made entirely of poo [11:45] Trevinho: sorry I didn't get a chance to test that sooner [11:45] Trevinho: and thanks you for the fix! :) [11:46] jdstrand: no worries.. [11:46] ooh, a jdstrand! [11:46] Chipaca, fully agreed :) [11:46] hey Chipaca :) [11:46] jdstrand— in etelpmoc, change the ()s around the definition of comopt to {}s [11:46] compopt* [11:46] Chipaca: I couldn't stop thinking about bash completion after you left (sorry) [11:48] jdstrand— still debugging why it doesn't get it quite right, but that'll get the slashes and such [11:48] Chipaca: does that give the 'ls' behavior [11:48] ? [11:48] jdstrand— closer. it depends. yes? sorta. [11:48] heh [11:48] jdstrand— it trips up and returns nothing as soon as it's quoted or there is a space in there [11:49] jdstrand— and i'm tracing that and very puzzled by it [11:49] but, yeah, it wasn't setting compopt -o filenames, meaning you didn't get ls-like behaviour ever [11:49] because, subshell /o\ [11:49] my bad there [11:51] interesting [11:51] Chipaca: does that mean the tab-completion feature is out of 2.25?? [11:52] I'm torn [11:52] I think we should land it, and then iterate [11:52] but I'm open to other people saying otherwise [11:52] Chipaca: if we land it, please add '&' to the grep [11:52] jdstrand— will do [11:53] however, if we land it as is, I'm quite sure someone will file a bug [11:53] that's ok. we could pre-file it :) [11:53] and even reference it in the code [11:54] * zyga breaks for lunch :/ [12:01] Mornings [12:08] niemeyer: good morning :) [12:11] stgraber: hey, I just noticed that my lxd snap is not upgrading. I'm at 2.5 (r241). have you seen this? [12:18] stgraber: fyi, https://forum.snapcraft.io/t/lxd-snap-not-refreshing-says-it-is-local-but-it-isnt/411 [12:19] mvo: sorry to bother you. this seems like something potentially bad ^ [12:19] pstolowski: 3242 is ready for a second review, should fix the econnreset tstuff [12:19] jdstrand: looking [12:20] mvo, thanks, approved [12:20] jdstrand: that looks alarming indeed - last refresh was month and month ago [12:21] jdstrand: I assume you installed this a long time ago? I suspect there is somehting missing in the state, would you mail the state file to me (privately) ? it may contain the macaroon so either edit it out or I can write something that just extracts the lxd information [12:22] mvo: I was one of the first users of the lxd snap, so, 'yes' [12:22] mvo: /var/lib/snapd/state.json ? [12:23] mvo: should I stop snapd first or anything? [12:23] jdstrand: yes, this one. should be fine [12:24] mvo: jdstrand: that error can happen only if the state has no snapid for that snap [12:25] we had some bugs around preserving side info on some ops, might be because of one of those [12:25] pedronis: yeah, I hope the state give some more clues [12:26] zyga— do i have to do anything to have ./run-checks --static not complain about things in configure? [12:26] Chipaca: try distclean in cmd but other than that, no :/ [12:26] Chipaca: I'm working on a way to switch away from automake [12:26] Chipaca: sorry about that [12:29] mvo: sent [12:31] Chipaca: you might have to remove some stuff that distclean doesn't I don't know if that got cleaned up, but I have in my notes: cd cmd ; make clean ; make distclean ; rm -f aclocal.m4 config.sub config.guess depcomp configure snap-confine/snap-confine.apparmor snap-confine/Makefile snap-confine/Makefile.in snap-confine/tests/Makefile snap-confine/tests/Makefile.in ; rm -rf autom4te.cache [12:32] jdstrand, pedronis: look like side-info is totally busted for this particular revision :/ http://paste.ubuntu.com/24466403/ [12:32] jdstrand: as a workaround you can snap rmeove lxd/snap install --candidate lxd [12:32] jdstrand: the interessting question of course is what happend [12:33] jdstrand— I was a lazy bag of lazy stuff and moved ahead by putting my static check before the spellcheck [12:33] jdstrand— pushed a branch full of comments, and a & [12:33] mvo: yea, that seems the effect of disable/enable at some point I think [12:33] for example [12:33] pedronis: aha, indeed - most likely that bug [12:34] mvo: I'm kinda worried to remove it cause I don't want to lose my container. it has my whole build environment [12:35] it is plausible I did enable/disable. I've done that before, but I don't know if I ever did on this snap. what if I disable/enable now? [12:35] jdstrand: it won't fix it [12:35] it lost the information [12:36] jdstrand: what you can do is stop snapd, edit that bit of state to put back the snapid, restart snapd and run refresh [12:38] I'd much rather do that than remove/install [12:38] jdstrand: yeah, was about to suggest the same, copy the snap-id from the pastebin [12:38] jdstrand: with that you are hopefully back [12:39] PR snapd#3242 closed: tests: tweak time for econnreset test a bit more [12:41] mvo: I'm looking at a wall of text and I have the paste in front of me. you are saying replace '{"name":"lxd","snap-id":"","revision":"241"}' with '{"name":"lxd","snap-id":"id from paste","revision":"241"}'? [12:42] jdstrand: correct [12:42] jdstrand: and make a backup of the state file first :) [12:42] jdstrand: (I guess you did by mailing it to me) [12:49] mvo: ok, that worked. thanks! I updated the forum for some of this, but you guys might want to add more detail [12:49] pedronis: thank you too :) ^ [12:49] jdstrand: thank you and sorry for the trouble [12:50] jdstrand: I think we know which bug it caused, the issue is long fixed since but of course there is fallout :/ [12:51] mvo: do you have a json linter/checker? it seems like this could be detected programmatically. I'm not sure how many state.json issues there are that could benefit from a linter/checker, but it's an idea [12:52] jdstrand: the issue here is that no snap-id is a legal state, it happens if you sideload without any assertions [12:52] jdstrand: this makes it hard to cleanup [12:53] mvo: yes, but the revision didn't start with an 'x' [12:53] jdstrand: but I think we could be smarter actually - if it has a normal revision but no snap-id and there is a snapid available in the previous sequence its probably that bug [12:53] jdstrand: yeah, was just thinking this too :) [12:53] exactly [12:59] hah! tests failed because i can't type [13:00] standup ? [13:24] PR snapcraft#1263 closed: lxd: pass through commands into the container [13:29] PR snapd#3233 closed: packaging: remove current mount profiles when purging [13:30] PR snapcraft#1278 closed: recording: save the snapcraft.yaml in the resulting snap [13:46] zyga: where is /var/lib/snapd/mount coming from? [13:47] morphis: snapd writes it [13:48] morphis: it's where mount profiles are stored [13:48] zyga: getting: cannot perform operation: mount --bind /snap/core/current//var/lib/snapd mount /tmp/snap.rootfs_5lYRhE/var/lib/snapd/mount: No such file or directory [13:48] none of the core snaps (stable, candidate or edge) has that directory [13:49] morphis: it should be created by snapd [13:49]