[00:08] kyrofa: it's not that bad http://paste.ubuntu.com/24284541/ [00:13] zyga, elopio: LP: #1673568 and LP: #1626359 are linked from the changelog of the SRU for snapd 2.23.6, but neither bug has an SRU template. Will someone be adding test cases to these bugs, or should the source be reuploaded without the bug refs? [00:13] Bug #1673568: snapd 2.23.6 SRU tracking bug [00:13] Bug #1626359: Cannot authorise quotactl syscall for Q_GETQUOTA

[00:14] zyga, elopio: sorry, LP: #1673568 is clearly an SRU tracking bug from the title - so only LP: #1626359 is at issue === chihchun_afk is now known as chihchun [04:59] Son_Goku: hah, https://koji.fedoraproject.org/koji/buildinfo?buildID=874052 ; that is awesome! great work! [06:17] zyga: hey! can you give https://github.com/snapcore/snapd/pull/3096 a review today? [06:17] PR snapd#3096: many: abstract path to /bin/{true,false} [06:19] morphis: hey [06:19] morphis: looking [06:21] zyga: thanks [06:23] zyga: and not sure if you saw https://koji.fedoraproject.org/koji/buildinfo?buildID=874052 already :-) [06:25] morphis: done [06:25] very nice! [06:26] morphis: what was the problem with ppc64? I see it's green now [06:26] morphis: btw, why two binary packages? [06:26] morphis: snap-confine / snapd [06:26] zyga: good question for Son_Goku [06:26] zyga: ppc64le was green before too [06:27] ppc64le is green [06:27] ppc64 was failing and still is [06:27] so the be variant [06:27] aah [06:27] ok [06:27] Son_Goku: I'd rather have one binary package (snapd/snap-confine) [06:27] Son_Goku: any reason why I don't see snapd yet with dnf on my rawhide vm? still need to wait for the sync? [06:27] Son_Goku: it will always be tightly coupled anyway so why splitting it [06:28] thanks for working on this guys! awesome work [06:28] I need to help with kids to get them out to school [06:28] ttyl [06:34] zyga: because I needed the old RPM to go away in the F26 and Rawhide trees [06:34] I'll merge snap-confine into snapd for 2.24, probably [06:43] Son_Goku: so why do I not see snapd in rawhide yet? is there any special trick like adding kojii repos? [06:43] morphis: it depends on your local mirror [06:43] it is definitely in the Koji internal repos [06:43] Son_Goku: according to https://apps.fedoraproject.org/packages/snapd it should be there [06:43] it may not have been mashed out to the mirrors yet [06:43] Son_Goku: where do I check which mirror I am using? [06:44] dnf will tell you when it starts doing stuff [06:44] but you can also check with #fedora-releng about these things [06:44] morphis: anyway, I just updated snapd-glib as well [06:44] we need snapd-xdg-open for desktop apps to work [06:45] Son_Goku: is there a spec file for it already? [06:46] morphis, zyga had a review that he never responded to... https://bugzilla.redhat.com/show_bug.cgi?id=1369560 [06:47] hah [06:47] Son_Goku: then let me put that on my list [06:47] I'd like to have snapd, snapd-glib, and snapd-xdg-open all pushed in the same update [06:48] sounds good [06:49] Son_Goku: will clean the spec file today and integrate your review [06:49] looks like we don't do proper releases for snapd-xdg-open [06:51] yeah [06:51] anyway, must sleeps [06:51] I've been up this whole time... [07:21] Son_Goku: makes sense [07:22] ? [07:22] Son_Goku: I replied to your earlier comment about snap-confine [07:23] Son_Goku: sorry, daughter had damp moo [07:23] Son_Goku: didn't want to go to school; didn't want to stay at home [07:23] Son_Goku: decided to actually get dressed and eat breakfast as we were leaving [07:29] hey pstolowski [07:29] let me know if you have something to review [07:29] PR snapd#3117 opened: tests: parameterize gadget snap channel [07:29] zyga, hi! [07:30] zyga, yeah, i'll have 3070 in a moment, something landed and it has new conflicts [07:31] pstolowski: sounds good [07:36] pstolowski: sorry, my fault, I landed the locking changes we were discussing yesterday, https://forum.snapcraft.io/t/transactionality-locking-and-other-concurrency-coordination/50/8 [07:37] pedronis, no worries, nice change btw and conflicts were trivial [07:38] zyga, pedronis #3070 ready for re-review [07:46] PR snapd#3118 opened: cmd: import snapd-xdg-open and integrate with our infrastructure [07:48] Son_Goku: didn't you say you need some sleep? :-) [08:11] pstolowski: looking now [08:13] thx [08:15] hey zyga a new kernel snap was promoted to beta yesterday, we had this new test error https://travis-ci.org/snapcore/spread-cron/builds/216891204#L392 could you please take a look? [08:16] fgimenez: hey [08:16] fgimenez: looking [08:16] fgimenez: can you in turn tell me what you think about this [08:16] 2017/03/30 19:17:29 Discarding autopkgtest:ubuntu-17.04-amd64, cannot connect: cannot connect to autopkgtest:ubuntu-17.04-amd64: ssh: must specify HostKeyCallback [08:16] fgimenez: I got this on https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-zesty-snappy-dev-image/zesty/amd64/s/snapd/20170330_191740_5aeb7@/log.gz [08:17] fgimenez: aha, let me prepare a quick PR, the test needs to be disabled; it will not work without another kernel change [08:17] zyga: sure looking [08:17] zyga: great thank you [08:19] PR snapd#3119 opened: interfaces: API additions for interface hooks [08:20] zyga: i've never seen before the autopkgtest error, it seems to happen at a very early stage, maybe something changed in zesty recently that makes the ssh connection fail, let me try to reproduce locally [08:22] fgimenez: I just pushed a change to https://github.com/snapcore/snapd/pull/3076, once it passes we should merge this [08:22] PR snapd#3076: cmd: disable the re-associate fix as requested by jdstrand [08:22] PR snapd#3120 opened: [WIP] interfaces: expose attrs to the interface API, snapctl enhancements (step #4) [08:28] zyga: if you could execute the failing test locally using the beta kernel snap that would be great to be extra sure, "export SPREAD_KERNEL_CHANNEL=beta" and then run spread with the failing test filter [08:34] fgimenez: you mean 1644439 or the earlier one? [08:34] fgimenez: I mean that test is definitely off for the next few weeks, we can only release this after the _next_ kernel gets released [08:37] only for the brave: #3119, #3120 [08:39] haha, is that the hook code? [08:39] the attribute things [08:39] yes [08:40] the 2nd PR isnt really as big as it may seem (it includes the 1st one) [08:40] zyga: aha if it's going to be disabled it's ok then, thanks === carlolo_ is now known as carlolo [08:50] zyga: you had a chance to look at https://build.opensuse.org/request/show/483839 already? [08:52] morphis: no, still on my queue [08:53] pstolowski: I just reviewed 3070 [08:53] pstolowski: have a look, I think you want to clarify locking and maybe think how this will scale [08:53] pstolowski: we won't have *that* many revisions but we can have any number of snaps [08:53] morphis: looking now [08:54] morphis: nice, too easy :) [08:54] morphis: thank you! [08:54] :-) [08:54] morphis: can you look at https://github.com/snapcore/snapd/wiki/Distributions#support-matrix [08:54] morphis: I think it needs a small lifting :) [08:55] zyga, thanks, looking. yes, that's a valid point although this is a general problem of state now, i think we need to solve this globally. with config snapshots we will have ~3 configs per snaps [08:56] pstolowski: yes, though I wonder how this works now, does golang json marshaller support partial encoding/decoding or is it all just smoke and mirros? [08:56] pstolowski: there's this json.RawMessage thing but I didn't look at the implementation to see what happens there [08:59] zyga, dunno, but I'd be surprised if it was that smart [09:00] pstolowski: and locking? [09:04] zyga, locking is fine, done by the caller, i'm adding a comment to these methods [09:05] pstolowski: I assumed as much because Get/Set check locking AFAIR [09:05] pstolowski: but I wanted to be sure [09:05] yes [09:05] pstolowski: can you look at https://github.com/snapcore/snapd/pull/3103 [09:05] PR snapd#3103: interfaces/mount: add function for parsing fstab-like file [09:05] pstolowski: it's super short, 2nd review [09:06] zyga, sure [09:11] fgimenez: I think the issue affects everything now `2017/03/31 08:43:13 Discarding autopkgtest:ubuntu-16.04-amd64, cannot connect: cannot connect to autopkgtest:ubuntu-16.04-amd64: ssh: must specify HostKeyCallback` [09:12] * ogra curses ... so why does core fail to build today [09:13] ogra: must be Friday [09:13] heh, yeah [09:16] zyga: running now, let's see.. [09:16] fgimenez: maybe ssh changed recently? [09:18] pedronis: hi, i've uploaded the expect snap to staging, these are the results on the ubuntu-core system http://paste.ubuntu.com/24286896/ could you please take a look when you have a moment? i think you mentioned we needed to run security-devpts in there, and it is failing [09:20] PR snapd#3076 closed: cmd: disable the re-associate fix as requested by jdstrand [09:20] ok, I've merged the patch that disabled 1644439 regression test [09:21] pstolowski: quick shot: https://github.com/snapcore/snapd/pull/3092 [09:21] PR snapd#3092: interfaces: capitalize Udev... as UDev [09:21] pstolowski: let's land it [09:22] morphis: 2nd review ^: this is a 1 char rename across the tree [09:22] (capitalization change) [09:22] zyga: aye [09:24] niemeyer: I'd like to land https://github.com/snapcore/snapd/pull/3039 [09:24] PR snapd#3039: many: add support for partially static builds [09:24] fgimenez: mvo: hi, I was getting this on all my PRs from autopkgtests: Discarding autopkgtest:ubuntu-16.04-i386, cannot connect: cannot connect to autopkgtest:ubuntu-16.04-i386: ssh: must specify HostKeyCallback do you know what it is about ? [09:24] niemeyer: I think all the feedback is addressed now, please have a 2nd look [09:25] pedronis: I don't see mvo here today [09:25] pedronis: is he off/ [09:25] zyga: btw. is there a way to invoke snap-confine manually? [09:25] morphis: sure [09:25] morphis: `make hack` in the tree [09:25] morphis: then you can use it [09:25] morphis: alternatively just run it, you need two things: [09:25] morphis: export SNAP_NAME=... [09:25] pedronis: looking into it right now, zyga pointed it out too [09:25] morphis: and call it with a security tag (snap.$SNAP_NAME.$APP_NAME) [09:25] ah [09:25] morphis: make hack is nice as it just works [09:25] morphis: replaces system snap-confine [09:25] morphis: very useful for iteration [09:26] zyga: yeah, but SNAP_NAME=.. is enough for now [09:28] morphis: sure but you need to call it with the mandatory argument ^_^ [09:28] yeah [09:28] works already :-) [09:28] slangasek, looks like your live-build change scrwed up our core snap builds (it tries to switch the image over to upstart) [09:28] upstart ? [09:28] yes [09:28] wow, blast from the past :) [09:28] retro base :) [09:29] http://launchpadlibrarian.net/309249343/live-build_3.0~a57-1ubuntu25.1_3.0~a57-1ubuntu25.2.diff.gz [09:29] ++# remove all installed packages that are Prio: required in the release [09:29] ++# pocket but something else in the latest version [09:29] so we remove systemd and get upstart to satisfy some dependencies? [09:29] in the logs i see the builds remove systemd-sysv and try to install all upstart packages [09:29] right [09:29] https://launchpadlibrarian.net/313704330/buildlog_snap_ubuntu_xenial_amd64_core_BUILDING.txt.gz [09:31] zyga, and it seems you were right ... its friday (that looks like something to fill the day :/ ) [09:36] I could also use a review for https://github.com/snapcore/snapd/pull/3114 [09:36] PR snapd#3114: interfaces/mount: add function for saving fstab-like file [09:36] it's even smaller and would allow me to start doing real stuff (tm) in update-ns [09:41] pstolowski: 3119 reviewed [09:41] pstolowski: have a look [09:41] pstolowski: get a 2nd review and lets land it [09:43] zyga, thanks, looking [09:45] pstolowski: I'll review the 2nd one after this lands, I need to focus on a branch I want to land today [09:57] PR snapd#3092 closed: interfaces: capitalize Udev... as UDev [09:58] morphis: test seem to fail on https://travis-ci.org/snapcore/snapd/builds/217086502 [09:58] aha, timeout :/ [09:58] some trouble allocating VMs [09:58] morphis: maybe restart it when more tests are quiet [09:59] zyga: yeah let me do a force push to trigger another run .. [09:59] PR snapd#3097 closed: interfaces/seccomp: add bind as part of the default seccomp policy (backport) [10:01] morphis: maybe merge master [10:01] morphis: that should cure some failures [10:01] morphis: I'm going for a small walk now [10:02] morphis: I'll check back in half an hour :) [10:02] morphis: please pleaes please review the two fstab PRs, they are tiny and I need them today [10:02] pstolowski: tell me if you want me to the Validate helpers that cuts boilerplate [10:02] pstolowski: and get a 2nd review and land your PR [10:02] pstolowski: I'll do it on otp [10:02] top [10:03] ttyl [10:13] PR snapd#3121 opened: WIP: cmd/snap-confine: always pass xauth file through regardless where it is located [10:22] Bug #1678042 [10:22] Bug #1678042: removal of priority changed packages breaks builds of the core snap [10:22] sigh ... [10:34] pedronis: not sure what's going on with the autopkgtest failure, using locally a prebuilt xenial image, without updating spread and running autopkgtest with snapd from the archive i still get the same error, checking changes in the dependencies now.. [10:40] ok, back to work for me [10:40] not 100% yet but getting there [10:40] anything on fire? :-) [10:43] Chipaca: hey :) [10:43] Chipaca: autopkgtests are all broken, something changed in ssh [10:43] Chipaca: but nothing on fire, just faint smoke [10:43] Chipaca: how are you? [10:46] Son_Goku, Pharaoh_Atem, King_InuYasha: when you're back, https://github.com/snapcore/snapd/pull/3118 is something we should consider instead of creating another package we will drop real soon [10:46] PR snapd#3118: cmd: import snapd-xdg-open and integrate with our infrastructure [10:48] zyga— headache died down, stomach settling, bored out of my mind but until recently couldn't turn my head fast [10:48] morphis: reviewed 3121 [10:48] Chipaca: ouch :/ [10:48] Chipaca: I had a rough evening, that youghurt *was* expired but ... well [10:49] Chipaca: it was still tasty though ;-) [10:49] heh [10:49] Chipaca: I could use some reviews on the fstab PRs [10:49] me it's like a hangover, but without the fun [10:49] Chipaca: and later on if you have a sec, I could use some hints on how to test some tricky code [10:53] zyga— https://thisisagoodsign.files.wordpress.com/2014/08/do_not_feed_hallucinogens_to_the_alligators.jpg [10:53] * Chipaca looks at PRs [10:54] Chipaca: poor crocks :) [10:54] well, at least it was mushrooms and not nutmeg [10:59] zyga— what's the expected use case for `SaveFSTab`? [11:04] Chipaca: snap-update-ns will write the things it did to /run [11:04] Chipaca: it will be a file in /run/snapd/ns/$SNAP_NAME.fstab [11:04] so, /run doesn't need atomicity in the sense we mean there [11:04] Chipaca: it should match /var/lib/snapd/mount/ [11:04] but /var does [11:05] Chipaca: but if it crashes the code will pick up from the start [11:05] Chipaca: and see if it needs to mount anything [11:05] ah, if snapd crashes? yeah [11:05] Chipaca: not snapd, this will be in a separate process [11:05] k [11:05] i was thinking of the machine crashing, not the program crashing [11:05] Chipaca: if machine crashes it's all good [11:06] Chipaca: this is purely ephemeral [11:06] yes [11:06] zyga— in what sense should things in run match things in var? [11:06] if all is good yes [11:06] some changes can be rejected [11:06] so they will not be mounted [11:06] and such entries won't show up [11:07] some entries can fail for various reasons [11:07] this is just a helper for the algorithm we merged earlier [11:07] that descries 'current' state [11:07] the desired state is always in /var/ [11:07] written by snapd [11:07] using all the nice atomic helpers [11:07] and the current file will be only written by this thing [11:08] and maybe by snap-confine but this is really optional since it will cross-check with mountinfo to see what is really mounted [11:08] zyga— ok. Can you add a note about this to the method? [11:08] Chipaca: sure [11:08] zyga, thanks [11:12] oh my how quickly my branches fall into conflicts recently ;) [11:13] I guess this is what you get if you touch 10*n files (where n>1) [11:14] Chipaca: added [11:14] pstolowski: try touching n**10 files ;) [11:14] ;) [11:16] woow [11:17] snow in the pyreneeyes [11:18] zyga— +1 [11:18] now i need to person up and go get lunch going before the boys get home ravenous [11:18] so much snow :) [11:18] PR snapd#3103 closed: interfaces/mount: add function for parsing fstab-like file [11:18] (term's out) [11:21] morphis: I'm not a fan of this whole merging everything into snapd [11:25] Son_Goku: why not? [11:25] Son_Goku: it really belongs together [11:27] * zyga small break [11:30] Son_Goku: that is the future and just follows what we're already doing [11:37] Son_Goku: so what do you think about integrating that as a patch now into the 2.23 package instead of doing the migration with 2.24 of a then to be dropped snapd-xdg-open package? [11:38] PR snapd#3122 opened: packaging: do not compile spread for autopkgtests [11:43] Bug #1678076 opened: console-conf crashes with eth0 and wlan0 on Pi 3 === pstolowski is now known as pstolowski|schoo === pstolowski|schoo is now known as pstolowski-schoo [12:26] Chipaca: hmm, how do I pass a []byte to a C function that gets a char *? [12:27] Chipaca: I see C.CBytes [12:27] is that it? [12:27] NOTE: I need an out argument (I want the C function to populate the buffer) [12:36] watched the video - but didn't feel it completed well : https://youtu.be/GiRbILdIVnA?t=15m58s [12:37] morphis: no [12:37] garh [12:38] I don't know how to get this to work [12:39] zyga— ? [12:39] yay ... core builds fixed again ... phew [12:39] Chipaca: having C function; ssize_t read_cmdline(char* buf, size_t buf_size) [12:40] Hi, I am getting this apparmor error using dbus in strict: [ 816.152055] audit: type=1400 audit(1490957861.012:208): apparmor="DENIED" operation="connect" profile="snap.screenly-client.websocket" name="/run/dbus/system_bus_socket" pid=3893 comm="python3" requested_mask="wr" denied_mask="wr" fsuid=0 ouid=0 [12:40] any idea what can it be caused by? I connect to dbus plug. So it works in devmode at least. [12:41] zyga— unsafe.Pointer(&buf[0]) ? [12:42] Chipaca: I tried that, cgo says no [12:42] ./bootstrap.go:64: cannot use unsafe.Pointer(&buf[0]) (type unsafe.Pointer) as type *C.char in argument to _Cfunc_read_cmdline [12:42] zyga— show me da code? [12:42] Chipaca: pushed as 'WIP' to setns-on-startup branch [12:42] Chipaca: have a look [12:44] Chipaca: note that place is also a bit bogus as len != cap [12:44] Chipaca: and even if I get this to work via the pointer somehow it will not change len [12:45] sborovkov: hmmm [12:45] sborovkov, sounds like a question for jdstrand [12:45] sborovkov: yes (who is off today) [12:45] let me look [12:46] so I am running a daemon. And this happens when connecting to system bus [12:46] zyga— ah, convert them to the right type and you're set [12:47] zyga— https://play.golang.org/p/BLn0jYkDmX [12:47] sborovkov: that seems to be covered by dbus-strict abstraction [12:47] sborovkov: so no idea, please report a bug [12:48] d'oh :)) [12:49] zyga: is it supposed to be used manually somehow in the snapcraft.yaml's code? just for reference that's what I have https://hastebin.com/apaxuseyob.lua [12:49] zyga— type safety goes out the window, very obviously [12:49] Chipaca: who cares about safety ;-) [12:49] Chipaca: this is just for unit testing [12:49] o/ [12:49] sborovkov, you want "daemon: dbus" i belive [12:49] looking at https://github.com/snapcore/snapd/blob/4ef0a46ea8ee10f45e73a91925cbfa5fdfbbe2e5/snap/validate.go#L137 [12:50] Chipaca: trying, thanks [12:50] ogra: ahh, alright I will try converting all my services that use dbus to that type from simple then. Thanks. [12:51] wooot [12:51] Chipaca: thanks!! [12:51] Chipaca: two things [12:51] Chipaca: casting "backwards" (C habbit) [12:51] sborovkov, well, perhaps i'm wrong, try with one first :) [12:51] Chipaca: and mandatory () around *C.char [12:51] and all backwards type syntax ;) [12:51] zyga— it's not casting, it's converting [12:51] there's a copy involved [12:51] but yes, it gets awkward quickly [12:52] Chipaca: copy of a pointer is OK [12:52] Chipaca: let me see if this works [12:52] yes!! [12:52] zyga— but then you get a copy of a copy and it starts to degrade :-p [12:52] suddenly all the halftones are gone [12:53] i swear it was just ibuprophene [12:53] ibuprofen* [12:54] Chipaca: meds ? What are you telling the group ? [12:54] CoderEurope— ? [12:54] telling which group? [12:55] use small words [12:55] Chipaca: you mentioned ibuprofen on a packaging channel ? [12:56] CoderEurope— I am not well, today [12:57] CoderEurope— so I've got an excuse for my usual non-sense [12:57] Chipaca: have you tried the 4-7-8 rule ? [12:58] CoderEurope— that's for sleeping? [12:58] yeah, --okay well try this sometime, bye-ya https://redd.it/5zyykt === pstolowski-schoo is now known as pstolowski [13:05] ogra: hmm, no that does not seem to work. Issues while validating snapcraft.yaml: The 'apps/websocket/daemon' property does not match the required schema: 'dbus' is not one of ['simple', 'forking', 'oneshot', 'notify'] Or it's in new snapcraft.. I updated it like last week [13:06] sborovkov: that is not supported I think [13:07] zyga: alright I will try asking jdstrand when he is around, and if it's not something I am doing wrong then I will file a bug. [13:07] sborovkov: I think we had a PR to support dbus deamones [13:07] ah, ok [13:09] zyga, https://github.com/snapcore/snapd/blob/4ef0a46ea8ee10f45e73a91925cbfa5fdfbbe2e5/snap/validate.go#L137 [13:09] zyga, is that not released yet ? [13:09] (that is what i just pointed sborovkov to) [13:10] niemeyer: I think I'm a bit too much out in the nature today [13:10] niemeyer: my connection is not sufficient for the hangout because of latency [13:10] niemeyer: I can do my update here [13:11] zyga: Remember that topic in the cafe? :) [13:12]