ijohnson | jdstrand: thanks! | 00:09 |
---|---|---|
mborzecki | morning | 06:14 |
mborzecki | brb, new kernel | 06:22 |
mborzecki | and back, 5.4.1 | 06:25 |
mup | PR snapd#7834 closed: tests: move the watchdog timeout to 2s to make the tests work in rpi <Created by sergiocazzolato> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/7834> | 06:25 |
mup | PR snapd#7833 closed: HACKING.md: add nvidia options to configure example <Documentation> <Simple π> <Created by anonymouse64> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/7833> | 06:27 |
mborzecki | not sure what the purpose of tests/main/install-store-laaaarge is | 06:44 |
mup | PR snapd#7743 closed: snap-bootstrap: force partition table operations <UC20> <Created by cmatsuoka> <Merged by bboozzoo> <https://github.com/snapcore/snapd/pull/7743> | 06:55 |
zyga | mborzecki: that we donβt buffer snaps in memory perhaps? | 07:06 |
zyga | Good morning | 07:06 |
zyga | Still sleepy | 07:06 |
zyga | Fought mould in the office late las night | 07:07 |
mborzecki | zyga: ayy, not good, make sure to remove all of it | 07:12 |
zyga | good morning mvo | 07:30 |
mvo | hey zyga | 07:31 |
mborzecki | mvo: hey | 07:36 |
mvo | mborzecki: good morning! | 07:37 |
mborzecki | ok, off to the bank, back ~12 hopefully | 07:42 |
* zyga jumps into https://github.com/snapcore/snapd/pull/7825 | 07:43 | |
mup | PR #7825: many: use transient scope for tracking apps and hooks <Created by zyga> <https://github.com/snapcore/snapd/pull/7825> | 07:43 |
pstolowski | morning | 08:02 |
sdhd-sascha | Hello, | 09:03 |
zyga | hey | 09:03 |
sdhd-sascha | :-) | 09:03 |
sdhd-sascha | on 19.04, when i run the test-suit, i get: | 09:03 |
zyga | so, snap run --explain, there's a small twist in I/O sync across all the interacting programs | 09:03 |
sdhd-sascha | - {"", time.Time{}}, // zero time default | 09:03 |
sdhd-sascha | + {"", time.Time{}}, // zero time default | 09:03 |
zyga | need to think how to make that all work cleanly | 09:03 |
sdhd-sascha | in the formattting test | 09:03 |
zyga | sdhd-sascha: ignore it, it's busted | 09:04 |
zyga | sdhd-sascha: we format with a fixed gofmt version | 09:04 |
zyga | sdhd-sascha: unless you happen to have that it will always differ | 09:04 |
sdhd-sascha | well, on 18.04 it works and go's on... | 09:04 |
sdhd-sascha | is there a workaround | 09:04 |
zyga | format with that version | 09:06 |
sdhd-sascha | On 18.04, it ./run-checks stops, with "Crushing failure and despair." | 09:06 |
zyga | you can install it as a snap :) | 09:06 |
zyga | yeah, it's silly | 09:06 |
zyga | mvo: ^ | 09:06 |
sdhd-sascha | i installed go as snap | 09:06 |
Chipaca | sdhd-sascha: just run the unit tests? ./run-checks --unit | 09:07 |
sdhd-sascha | :-) i will | 09:08 |
sdhd-sascha | What gofmt version i should install ? | 09:08 |
zyga | brb, I'm starving | 09:11 |
Chipaca | sdhd-sascha: try SKIP_GOFMT=1 ./run-checks | 09:20 |
Chipaca | sdhd-sascha: i think it's the gofmt from go 1.10 but i might be wrong on this | 09:21 |
Chipaca | sdhd-sascha: if you're running go from a snap, snap refresh --channel=1.10 go | 09:21 |
sdhd-sascha | Chipaca: SKIP_GOFMT works. I need this, because in the night i have to shutdown my noisy server ;-) thanks | 09:22 |
Chipaca | noisy stuff in the home needs to die | 09:27 |
sdhd-sascha | Chipaca: right | 09:39 |
pedronis | mvo: what UC20 bits need reviewing? | 09:45 |
sdhd-sascha | What is best pratice? Build and install master branch? Is master mostly compatible with snapcraft store, or test this seperatly ? | 09:45 |
zyga | sdhd-sascha: master is compatible with the store | 09:46 |
zyga | sdhd-sascha: we take backwards compatibility seriously | 09:46 |
* zyga breaks branch iteration to review stuff for 2.34 | 09:47 | |
zyga | 43 even | 09:47 |
mvo | pedronis: 7831 should be ready, 7832 too, I'm just adding a spread test | 09:47 |
pedronis | mvo: 7831 has two +1 | 09:56 |
mup | PR snapd#7831 closed: gadget: extract and export new DiskFromPartition() helper <UC20> <Created by mvo5> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7831> | 10:00 |
mvo | pedronis: thanks, merged and I updated 7832 against master plus it has a spread test now | 10:10 |
mvo | pedronis: that should unblock 7762 :) | 10:10 |
zyga | pedronis: reviewed https://github.com/snapcore/snapd/pull/7228#pullrequestreview-326719099 | 10:36 |
mup | PR #7228: interfaces: add audio-playback/record and pulseaudio spread tests <Created by jdstrand> <https://github.com/snapcore/snapd/pull/7228> | 10:36 |
zyga | pstolowski: about your work on the label bug | 10:36 |
zyga | pstolowski: do you think we can chop it into parts so that the several bugs this uncovered can be addressed in isolation | 10:37 |
pedronis | mvo: it's Ubuntu Core not Ubuntucore, so UC not Uc | 10:37 |
pstolowski | zyga: i think so yes, nb i think i see the issue, just trying a tentative fix | 10:37 |
zyga | oh, while are are at naming, why do we have Grubenv and not GrubEnv | 10:37 |
zyga | it's rubbing salt into my eyes | 10:37 |
mvo | pedronis: thank you, I keep this in mind | 10:37 |
zyga | pstolowski: super, thank you for taking this | 10:37 |
pedronis | mvo: thanks, not the most important thing in the world but well | 10:38 |
mvo | pedronis: yeah, we should be consistent! | 10:38 |
pedronis | mvo: is that right that cmd/snap-bootstrap/bootstrap is mostly tested via spread? | 10:41 |
zyga | they are shooting an episode next door again | 10:52 |
zyga | mvo: https://github.com/snapcore/snapd/pull/7832#pullrequestreview-326757177 | 10:54 |
mup | PR #7832: snap-bootstrap: support auto-detect device in create-partitions <UC20> <Created by mvo5> <https://github.com/snapcore/snapd/pull/7832> | 10:54 |
mvo | zyga: thank you, looking | 11:05 |
mvo | (was in a meeting) | 11:05 |
zyga | no worries | 11:05 |
pstolowski | zyga: ok, wrt to the label issue.. another problem confirmed and fix works, it's implicit stuff biting again | 11:08 |
pstolowski | zyga: we bind implicit hooks when reading the yaml, but implicit slots arrive later to the picture :/ | 11:09 |
pstolowski | we should probably tear all that apart and redo at some point, but this needs a good plan | 11:10 |
zyga | pstolowski: as expected the, no disagreement there | 11:22 |
zyga | short coffee break | 11:23 |
zyga | it's not so cold today somehow | 11:24 |
zyga | back | 11:40 |
mup | PR snapd#7835 opened: gadget: add missing test for duplicate detection of roles <Simple π> <Created by mvo5> <https://github.com/snapcore/snapd/pull/7835> | 11:44 |
cachio | mvo, hey | 11:55 |
mvo | hey cachio | 11:55 |
cachio | did you fix the test uc20-snap-recovery ? | 11:56 |
cachio | in the last PR merged? | 11:56 |
cachio | mvo, I see the tests passed during last execution on travis | 11:59 |
cachio | but I see it is failing on master when I execute it | 11:59 |
mup | PR snapd#7836 opened: o/ifacestate: fix binding of implicit hooks to implicit slots on core snap <Bug> <Created by stolowski> <https://github.com/snapcore/snapd/pull/7836> | 11:59 |
pstolowski | zyga: ^ | 12:01 |
zyga | nnn | 12:01 |
mvo | cachio: it seems racy | 12:02 |
mvo | cachio: it's strange, I wonder what broke it | 12:02 |
mvo | cachio: do you know when it started failing? | 12:03 |
cachio | mvo, yesterday the bionic image was updated | 12:03 |
cachio | perhaps something new is breaking it | 12:03 |
cachio | I ran the tests yesterday for the new image and that test passed | 12:04 |
mvo | cachio: there are also changes in the sfdisk code | 12:04 |
mvo | cachio: 7743 landed | 12:05 |
mvo | cachio: and after that I see an error on master and I also seen an error in my own stuff - but not in qemu so far (but only ran twice). runing on google now | 12:05 |
cachio | mvo, nice | 12:06 |
mvo | looks like we also need to improve the error, i.e. what exactly is not compatible | 12:06 |
cachio | mvo, hehe, I'll try to reproduce the error using the previous image | 12:06 |
cachio | so if I'm able to do it the problem should be in the code | 12:06 |
* Chipaca needs to lie down for a while | 12:07 | |
zyga | pstolowski: reviewed | 12:09 |
pstolowski | zyga: thank you; please also see my comment under 1st PR | 12:10 |
zyga | ack | 12:10 |
mvo | cachio: thanks! | 12:11 |
pstolowski | zyga: as touched briefly yesterday, remaining 3rd issue is peer=snap.core.configure now generated for e.g. modem-manager plug on classic | 12:12 |
sdhd-sascha | hey, can i ask a off-topic question? i want to forward a unix-domain socket via ssh & netcat. With "ssh <host> -- nc -q0 -U </path/to/socket>" i got a connection. | 12:14 |
sdhd-sascha | But now my application needs a socket-filename | 12:15 |
sdhd-sascha | i want with emacsclient securly connect to server (tramp plugin didn't work like expected...) | 12:15 |
sdhd-sascha | :-) | 12:16 |
pedronis | pstolowski: do we really need to bind hooks fore core, given that the only hook there is hijacked anyway? | 12:23 |
mvo | cmatsuoka: hey, good morning | 12:25 |
pedronis | pstolowski: I'm basically wondering whether we can cut/simplify somehow differently this area | 12:25 |
cachio | mvo, the test fails with the old image as well | 12:25 |
mvo | cmatsuoka: "fun" - right now master is unhappy, it looks like 7743 sometimes works and sometimes does not. | 12:26 |
mvo | cmatsuoka: it looks like adding "blockdev --rereadpt" fixes it | 12:26 |
mvo | cmatsuoka: well, "fixes" | 12:26 |
mvo | cmatsuoka: I wonder if we need our own code afterall :/ | 12:26 |
pstolowski | pedronis: if we don't then label generation would need some other logic / special casing for core i think (for the reference - #7830) | 12:26 |
mup | PR #7830: interfaces: include hooks in plug/slot apparmor label <Bug> <Created by stolowski> <https://github.com/snapcore/snapd/pull/7830> | 12:26 |
cmatsuoka | mvo: is it failing on the gadget vs device consistency checking? | 12:27 |
pedronis | pstolowski: I would like us to explore that | 12:27 |
pedronis | pstolowski: it's a bit silly to generate profiles for something we never run | 12:27 |
pedronis | core is anyway special cases in various ways | 12:27 |
pedronis | but I might be missing something | 12:28 |
mup | PR snapd#7837 opened: devicestate: implement creating partitions in "install" mode <Created by mvo5> <https://github.com/snapcore/snapd/pull/7837> | 12:29 |
mvo | cmatsuoka: yeah | 12:29 |
mvo | cmatsuoka: exactly the issue you describe in the gdoc | 12:29 |
mvo | cmatsuoka: seems to break master right now :( | 12:29 |
mvo | cmatsuoka: do you think you could try just adding "blockdev --rereadpt" as a workaround? | 12:30 |
mvo | cmatsuoka: need to get lunch in a wee bit but would love to have someone look at this while I'm away :) | 12:30 |
cmatsuoka | mvo: yes, at which point you tried this rereadpt? | 12:30 |
cmatsuoka | mvo: the error pattern is very odd, last night I couldn't reproduce it anymore after ~10PM UTC | 12:31 |
pstolowski | pedronis: okay, i need to take a step back, and clear my mind. but lunch first | 12:32 |
mvo | cmatsuoka: fun! I can reproduce it in gce | 12:33 |
pedronis | pstolowski: ok, we ignored the silly work because it didn't have consequences so far, but if it starts to give trouble we need to consider | 12:33 |
mvo | cmatsuoka: but not in qemu | 12:33 |
mvo | cmatsuoka: I see it in my PR everytime in gce and also when running manually I hit it right away | 12:33 |
mvo | cmatsuoka: I did run "blockdev --rereadpt" after the failure manually i nthe shell and suddently things went away | 12:34 |
mvo | cmatsuoka: thanks for looking into this, if you have further question please use tg, I will go to lunch now | 12:34 |
mvo | cmatsuoka: but happy to help as good as I can | 12:34 |
cmatsuoka | mvo: where did you insert the blockdev that fixes the problem? (I'm afraid it's more like it's disturbing the system enough to make a racy situation work again) | 12:35 |
mvo | cmatsuoka: after the test failed I just ran it manually in a shelll | 12:35 |
mvo | cmatsuoka: which is obviously not that great of an answer | 12:36 |
* mvo really is away now | 12:36 | |
* cmatsuoka earns the "breaking master" badge π± | 12:39 | |
mup | PR snapcraft#2829 closed: store cli: push title and license on push-metadata <Created by sergiusens> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2829> | 12:40 |
mup | PR snapd#7838 opened: cmd/snap-bootstrap: stub out snap.SanitizePlugsSlots for real <Bug> <Simple π> <UC20> <Created by chipaca> <https://github.com/snapcore/snapd/pull/7838> | 12:53 |
* Chipaca _really_ needs to lie down for a while, now | 12:54 | |
mup | PR snapd#7839 opened: tests: prevent partitioning test errors <Created by cmatsuoka> <https://github.com/snapcore/snapd/pull/7839> | 13:08 |
cmatsuoka | mvo: issued a quick workaround atempt, but also working on a more definitive solution that drops lsblk entirely and uses blkid instead | 13:24 |
pedronis | mvo: I removed the 2.43 milestone from snap run --explain, seems a bit unlikely to make it | 13:27 |
zyga | pedronis: I just pushed an update there | 13:37 |
zyga | with spread tests and more things that it does | 13:37 |
zyga | but it can be added in later if you feel like it's not ready | 13:37 |
pedronis | zyga: for one it needs a jdstrand review, 2nd it needs careful review of each output line | 13:38 |
zyga | pedronis: https://github.com/snapcore/snapd/pull/7728/files#diff-771e057220c48a44a34166289f5339f4R1 is the starting point | 13:39 |
mup | PR #7728: cmd: implement snap run --explain <Created by zyga> <https://github.com/snapcore/snapd/pull/7728> | 13:39 |
zyga | we can tweak it all the way we want but I think, since it's not meant to be parsed by machines, it should go in after jdstrand's review | 13:41 |
pedronis | zyga: ? | 13:42 |
zyga | I'm saying we should not go over the output with a fine comb and not ship it | 13:42 |
pedronis | zyga: machine parsing is not my worry here | 13:42 |
zyga | we should ship it and improve it after people use it | 13:42 |
pedronis | zyga: we should start from a good place, atm the are open questions and the output formatting not very consistent | 13:44 |
zyga | pedronis: open questions? | 13:44 |
zyga | pedronis: what is not consistent? | 13:44 |
pedronis | zyga: the usage of colons for once | 13:45 |
zyga | pedronis: what about it? | 13:45 |
zyga | honestly, that's the last thing anyone using this for real will care about | 13:45 |
zyga | but I don't think it's inconsistent now, maybe you are looking at the older version | 13:46 |
pedronis | zyga: maybe, it needs review either way | 13:46 |
zyga | that's true | 13:46 |
pedronis | from me | 13:46 |
zyga | I'd just hate to blow it out of proportion as to what it needs to be - it's a tool, not a designer hat; it's good to ship it if it's 80% there and can benefit someone | 13:47 |
pedronis | not more than any other snapd output | 13:49 |
pedronis | anyway jdstrand review is still the main blocker atm | 13:49 |
pedronis | zyga: the open questions is the remark Maciej made about separator, and also how that matches or not the original sketch plan | 13:51 |
zyga | pedronis: the separator probably cannot be used if we want wrappers and other layer to handle explain-like output | 13:52 |
zyga | pedronis: because while there's a clear start, there is no clear end | 13:53 |
pedronis | zyga: I agree but the last output had still missing bits vs the original sketch | 13:53 |
pedronis | not sure if your last pass fixed this or not | 13:53 |
zyga | pedronis: I didn't check it against the sketch but I think the sketch is just that, there are TODOs to add more things based on what snap run really does | 13:53 |
zyga | pedronis: I'll double check what's going on in the code and if there's something to add based on the original sketch | 13:54 |
pedronis | zyga: yes, but the sketch has a separator just before the start of the app command itself | 13:54 |
zyga | pedronis: but my point is that it's meant to be realistic, not idealistic | 13:54 |
pedronis | which partly addressed Maciej worry | 13:54 |
pedronis | addresses | 13:54 |
pedronis | anyway my point stand that still need a proper review after jdstrand one | 13:55 |
zyga | s/after// but yeah | 13:55 |
zyga | I think we don't need to wait for jamie | 13:55 |
jdstrand | you touch snap-confine | 13:55 |
jdstrand | I'd like to review at least those bits | 13:56 |
pedronis | zyga: well I plan to wait, given I have also other things | 13:56 |
zyga | not wait for the other review | 13:56 |
=== hunterk_ is now known as hunterk | ||
pedronis | zyga: I suppose I could to a quick pass over the output in the spread test when I have a moment | 13:58 |
zyga | thanks! | 13:58 |
zyga | I can quickly adjust it after initial output review | 13:58 |
pedronis | is the output up-to-date ATM in tha test? | 14:00 |
zyga | yes | 14:00 |
pedronis | ok, thx | 14:00 |
zyga | cannot join standup | 14:02 |
zyga | sorry | 14:02 |
zyga | I'll join when I xan | 14:02 |
=== ricab is now known as ricab|lunch | ||
zyga | oh my | 14:26 |
zyga | sorry guys | 14:26 |
zyga | chrome love | 14:26 |
pedronis | mvo: there is, snap find mumble | 14:30 |
pstolowski | heh | 14:30 |
pstolowski | jdstrand: hey, question about modem-manager interface | 14:30 |
jdstrand | ok | 14:31 |
pedronis | mvo: ijohnson: can we do it 30 mins before standup tomorrow? | 14:31 |
jdstrand | that interface is pretty ancient | 14:31 |
ijohnson | pedronis: mvo: yes that's fine for me | 14:31 |
pstolowski | jdstrand: is this intended that AppArmorConnectedPlug always adds modemManagerConnectedPlugAppArmor snippet, and on classic in addition appends modemManagerConnectedPlugAppArmorClassic ? should there be if classic - else core? | 14:33 |
jdstrand | pstolowski: I'm looking at it | 14:36 |
pstolowski | jdstrand: that extra snippet gets label=snap.core.* on classic in the generated profile | 14:38 |
jdstrand | pstolowski: so, like I said, this interface is ancient and doesn't benefit from the moderm thinking around app snaps, classic, etc. looking at it, if you install the snap on a classic system, you get a weird peer=(label="snap.core.*") rule | 14:40 |
jdstrand | pstolowski: iirc, on or the other was tacked on later with the understanding that "you're not expected to install the modem-manager snap on a core system" | 14:42 |
jdstrand | but, we've handled this better lately | 14:42 |
jdstrand | pstolowski: eg, in avahi-control | 14:42 |
jdstrand | pstolowski: PermanentSlot and ConnectedSlot could also be updated | 14:44 |
jdstrand | pstolowski: basically, it and most likely bluez and network-manager would likely benefit from being updated to model after avahi-control. though, still, "you're not expected to install the * snap on a core system" with all of these :) | 14:45 |
ijohnson | jdstrand: but you are definitely expected to install the network-manager and bluez snap on UC? | 14:46 |
jdstrand | erf | 14:46 |
mvo | pedronis, ijohnson sure, let me schedule something | 14:46 |
ijohnson | and modem-manager snap too actually | 14:46 |
jdstrand | s/core/slassic/ | 14:46 |
jdstrand | classic* | 14:46 |
ijohnson | ahhh yes that makes much more sense +1 | 14:46 |
jdstrand | ijohnson: I totally typoed that :) | 14:46 |
jdstrand | pstolowski: ^ | 14:46 |
ijohnson | it's still early :-) | 14:47 |
jdstrand | pstolowski: is this something you are thinking of fixing or should I add it to the list for the next batch of updates? | 14:47 |
jdstrand | ijohnson: not sure you saw, but nice job on the lib caching forum topic :) | 14:47 |
ijohnson | I think I did see, but thanks again :-) | 14:48 |
jdstrand | :) | 14:48 |
pstolowski | jdstrand: no, at least not now; it just popped because of this weird label (in one of the spread tests), which becomes even more weird with my #7830 | 14:49 |
mup | PR #7830: interfaces: include hooks in plug/slot apparmor label <Bug> <Created by stolowski> <https://github.com/snapcore/snapd/pull/7830> | 14:49 |
jdstrand | pstolowski: ack, I'll add it to my list then | 14:50 |
pstolowski | jdstrand: i though a simple if-classic-else-core in modem manager would do it, but sounds like you envison to streamline it more | 14:51 |
pstolowski | *thought | 14:51 |
pstolowski | jdstrand: thanks, and thanks for explaining | 14:51 |
jdstrand | np | 14:52 |
jdstrand | yeah, they just need updating. they're a bit crufty | 14:52 |
pedronis | mvo: thx, I see it in your cal but not mine | 14:56 |
mvo | pedronis: yeah, sorry, was wondering if we should just use the 4:30 slot I have with ian tomorrow anyway - this would mean that ian does not have to get quite so early | 14:58 |
mvo | pedronis: or the 5pm slot | 14:58 |
mup | PR snapd#7733 closed: tests: disable nova from install-snaps test <Created by sergiocazzolato> <Closed by sergiocazzolato> <https://github.com/snapcore/snapd/pull/7733> | 14:58 |
mvo | pedronis: after the certbot meeting | 14:58 |
pedronis | mvo: well the first slot overlaps with the certbot one, after the cerbot is fine with me if it works for you | 14:59 |
mvo | pedronis: I would have to skip/move one meeting | 15:00 |
ijohnson | mvo: it's ok, it's not too early | 15:00 |
pedronis | mvo: no strong preference, except I think I need to be at the certbot meeting | 15:00 |
zyga | brb, lunch | 15:01 |
zyga | (late) | 15:01 |
mvo | ijohnson: it could either be 30min before the meeting or 2h later | 15:18 |
ijohnson | mvo: whatever is most convenient for you, I don't have any other meetings except our 1:1 and the SU and it's only 7:30 so that's doable for me | 15:18 |
zyga | Lunch over but Iβm looking after Lucy now | 15:28 |
zyga | I will return in about an hour | 15:29 |
* cachio lunch | 15:39 | |
=== ricab|lunch is now known as ricab | ||
zyga | re | 15:58 |
mup | PR snapd#7838 closed: cmd/snap-bootstrap: stub out snap.SanitizePlugsSlots for real <Bug> <Simple π> <UC20> <β Critical> <Created by chipaca> <Merged by chipaca> <https://github.com/snapcore/snapd/pull/7838> | 16:03 |
Chipaca | xnox: ^ | 16:04 |
pedronis | ijohnson: thanks for the review | 16:18 |
jdstrand | mvo, pedronis: hey, fyi, zyga gave his approval on PR 7228 and I incorporated his feedback. He said to feel free to merge, but I wasn't sure if that meant after another +1 or right away. I think there is a question in my mind since this PRs only adds spread tests and doesn't touch code, and not sure if 2 approvals is required for that | 16:18 |
ijohnson | pedronis: np | 16:18 |
mup | PR #7228: interfaces: add audio-playback/record and pulseaudio spread tests <Created by jdstrand> <https://github.com/snapcore/snapd/pull/7228> | 16:18 |
pedronis | ijohnson: about the confusing code, it copied and just slightly modified from some code where the error can happen or not | 16:19 |
jdstrand | (in any case, obviously I'll what for the tests to pass) | 16:19 |
pedronis | but I agree is confusing if the error is expected | 16:19 |
pedronis | jdstrand: let me look | 16:19 |
ijohnson | pedronis: yeah that's kinda what I thought but still seems like either a comment explaining that or not doing it that way would be good, because I could see myself needing to write a similar test and taking this one as a template and then being very confused about why it was written this way | 16:20 |
pedronis | jdstrand: you probably want a 2nd review from some of the people that looked at it already, doesn't need to be me tough | 16:21 |
jdstrand | pedronis: ok | 16:22 |
pedronis | I actually didn't look at it, just did a meta-comment at some point | 16:23 |
zyga | jdstrand: I meant "from my pov it is ok" but I don't think I have the power to merge with one review | 16:24 |
jdstrand | that's fine | 16:25 |
* zyga fights snap security tag hydra | 16:25 | |
jdstrand | I didn't want to assign anyone to review it, so I asked in a comment if Ian or Maciej wants to do it | 16:25 |
ijohnson | jdstrand: I can take a look today probably | 16:29 |
jdstrand | ijohnson: thanks. it is only as urgent as cleaning out old PRs is | 16:30 |
ijohnson | ack | 16:30 |
zyga | jdstrand: merge master into it if you haven't done that recently please | 16:32 |
cmatsuoka | mvo: the empty filesystem bug seems to be a race involving node creation and the udev event queue, I'll test a queue flush after partitioning and before creating the filesystems | 16:34 |
jdstrand | zyga: yep, already done | 16:35 |
jdstrand | I learned from last time | 16:35 |
zyga | super, just wanted to double check after recent old-branch-post-merge-fixups | 16:35 |
zyga | :) | 16:35 |
zyga | I think we all learned, it's a shame CI is too costly to run on each master change | 16:36 |
mup | PR snapd#7826 closed: tests: use test-snapd-sh snap instead of test-snapd-tools - Part 1 <Created by sergiocazzolato> <Merged by sergiocazzolato> <https://github.com/snapcore/snapd/pull/7826> | 16:39 |
mup | PR snapd#7840 opened: tests: use test-snapd-sh snap instead of test-snapd-tools - Part 2 <Created by sergiocazzolato> <https://github.com/snapcore/snapd/pull/7840> | 16:41 |
mvo | cmatsuoka: cool, thanks for digging into this | 16:46 |
jdstrand | ijohnson and bloodearnest: fyi, https://github.com/sergiusens/snapcraft-preload/pull/39 | 17:22 |
mup | PR sergiusens/snapcraft-preload#39: preload.cpp: use setgroups()/initgroups() in sandbox-compliant manner <Created by jdstrand> <https://github.com/sergiusens/snapcraft-preload/pull/39> | 17:22 |
ijohnson | \o/ yay thanks jdstrand | 17:22 |
zyga | I feel cold getting me | 17:32 |
zyga | Need a break | 17:32 |
zyga | Sleepy | 17:32 |
mup | PR snapcraft#2830 opened: elf: properly handle corrupted ELF files <Created by kyrofa> <https://github.com/snapcore/snapcraft/pull/2830> | 17:56 |
mup | PR snapcraft#2828 closed: Appstream <Created by sergiusens> <Merged by sergiusens> <https://github.com/snapcore/snapcraft/pull/2828> | 18:29 |
mup | PR snapcraft#2831 opened: appstream extractor: add support for code <Created by sergiusens> <https://github.com/snapcore/snapcraft/pull/2831> | 18:44 |
zyga | cold meds are working so I'm back | 18:45 |
zyga | wondering how to proceed with snap security tag validation | 18:45 |
zyga | security tag are super diverse | 18:45 |
zyga | and it feels like something to make tidy now | 18:45 |
mup | PR snapd#7839 closed: tests: prevent partitioning test errors <Created by cmatsuoka> <Merged by mvo5> <https://github.com/snapcore/snapd/pull/7839> | 18:49 |
mup | PR snapd#7841 opened: tests: fix partitioning test debug message <Created by cmatsuoka> <https://github.com/snapcore/snapd/pull/7841> | 19:39 |
cmatsuoka | cachio: PR #7841 fixes a silly shell error in the previous PR | 20:02 |
mup | PR #7841: tests: fix partitioning test debug message <Created by cmatsuoka> <https://github.com/snapcore/snapd/pull/7841> | 20:02 |
cachio | cmatsuoka, taking a look | 20:04 |
cmatsuoka | cachio: when the udev fstype error happens, grep fails and the test was aborted | 20:05 |
cachio | cmatsuoka, ah, ok | 20:05 |
cachio | thanks for the explanation | 20:05 |
cachio | +1 | 20:05 |
cmatsuoka | cachio: I was expecting ID_FS_TYPE to be there with an empty value when it fails, but in fact the key is not listed | 20:07 |
cachio | cmatsuoka, nice, thanks for the fix | 20:08 |
ackk | mvo, hi around? | 20:08 |
cmatsuoka | cachio: sorry for the basic error in the first attempt | 20:11 |
mup | PR snapd#7842 opened: tests: cache snaps also for ubuntu core and add new snaps to cache <Created by sergiocazzolato> <https://github.com/snapcore/snapd/pull/7842> | 20:11 |
cachio | cmatsuoka, no problem, this happens often | 20:12 |
mvo | ackk: a bit, it's rather late already - what can I do for you :) ? | 20:13 |
ackk | mvo, sorry, just checking if https://bugs.launchpad.net/snapd/+bug/1817276 is supposed to be fixed in 2.42.1+18.04 (deb package in bionic) | 20:13 |
mup | Bug #1817276: snapfuse use a lot of CPU inside containers <maas> <snapd:Fix Released by mvo> <https://launchpad.net/bugs/1817276> | 20:14 |
mvo | ackk: that's correct, it should be much better with this version | 20:14 |
mvo | ackk: I think I got one positive report already | 20:14 |
ackk | mvo, I still see it use 100% cpu at times | 20:14 |
mvo | ackk: is it not helping for you :/ ? | 20:14 |
ackk | mvo, well it's better for sure | 20:14 |
ackk | but I guess under heavy I/O cpu usage is expected? | 20:15 |
mvo | ackk: yes - ish - I mean, if it's still not good enough we need to talk again. the performance should be in the order of 10x better than before | 20:15 |
ackk | mvo, yeah it's definitely much better. I wasn't sure if it was supposed to be entirely gone | 20:16 |
ackk | mvo, anyway, thanks for the info, don't let me keep you :) | 20:16 |
mvo | ackk: aha, right. it's still fuse currently so a certain degree of slowness is expected unfortunately | 20:17 |
mvo | ackk: but again - if it's a problem we could talk about options (which would be hackish at this point :( | 20:17 |
mvo | ackk: but let's do this tomorrow :) | 20:17 |
ackk | mvo, I haven't tested it extensively, yet | 20:17 |
ackk | mvo, yeah, good night, thanks again :) | 20:18 |
mup | PR snapd#7843 opened: tests/cmd/snapctl: unset SNAP_CONTEXT for the suite <Simple π> <Created by anonymouse64> <https://github.com/snapcore/snapd/pull/7843> | 20:19 |
mvo | ackk: my pleasure - good night to you too! | 20:19 |
mup | PR snapd#7841 closed: tests: fix partitioning test debug message <Created by cmatsuoka> <Merged by cmatsuoka> <https://github.com/snapcore/snapd/pull/7841> | 23:03 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!