[01:42] kenvandine: yeah I noticed it at least needs a newer PyGObject for Gtk.Template() but I didn't have much luck figuring out how to add that appropriately as a part (was messing around with the python plugin to try and use that but wasn't successful) [01:59] amurray: i'll get it in the store soon [01:59] kenvandine: nice - cheers [04:26] PR snapcraft#2652 closed: meta: transparently support command-chain === pstolowski|afk is now known as pstolowski [07:02] mornings [07:29] PR snapd#7196 opened: packaging: use passthrough for type:snapd [10:08] Hello [10:09] hey zyga [10:48] #1838654 looks like fun [10:48] Bug #1838654: Snap completely broke [10:50] amazing [10:50] IKR [11:58] PR snapd#7206 opened: overlord/snapstate: use reflect to improve snapstate.Get [12:00] PR snapd#7207 opened: snap: prevent duplicated snap name and snap files when parsing seed.yaml [12:17] ^^^ a bit of friday fun [12:18] sigh, what's up with spread _now_ [12:21] pstolowski: did you check how your dupe checker worked from 'snap debug validate-seed'? [12:22] pstolowski: also note image/validate_seed_test.go [12:26] Chipaca: hmm, nope, will do, ty [12:37] morning folks [12:39] hey ijohnson! [12:40] o/ pstolowski [12:41] Chipaca: ValidateSeed calls ReadSnapYaml, so it's covered by the fix (and so is snap debug validate-seed) [12:44] Hi guys, I have a quick question regarding access to /proc. Under flatpak apps such as Discord and Spotify have no access to /proc whatsover, however having made the switch to snap recently I realized that the same apps have access to /proc. As far as I know both Discord and Spotify have not been installed with the --classic flag so I was wondering [12:44] if this is intended? (If not then) Correct me if I am wrong but shouldn't access to /proc be limited because sensitive data could be read from there? [12:44] Chipaca: but surely an extra test for cmd_debug_validate will not harm [12:46] snapuser9: some snaps declare to use interfaces to access certain things inside /proc, so discord and spotify use some of those interfaces [12:47] what interface depends on what exactly in /proc is desired to access [12:48] snapuser9: access to proc is limited; is there anything the snap has access to that is sensitive? [12:53] PR snapd#7208 opened: cmd/snap, data/completion: improve completion for 'snap debug' [12:53] snapuser9: (it's not an all-or-nothing thing) [12:54] PR snapcraft#2651 closed: Extension confinement [12:54] roadmr, cjwatson, i just built a new gadget on build.s.io ... this rightly went into manual review ... yet ... build.s.io tels me "Built and released" and also tells me to "sudo snap install --edge " ... i'm pretty sure former gadgets properly showed the store status on build.s.io [12:57] ogra: there's a bug report link at the bottom of the page :) [12:57] heh, k [12:58] it says "we're hiring" though :) [12:59] I'm going to be a minute late for the standup [12:59] k [13:06] * ogra files https://github.com/canonical-web-and-design/build.snapcraft.io/issues/1231 [13:11] ogra: while you're there you could re-file https://bugs.launchpad.net/bugs/1838718 [13:11] Bug #1838718: Downloaded file is called "snaprcaft.yaml" instead of "snapcraft.yaml" [13:12] Chipaca, ?? [13:12] ENOTMYBUG [13:14] aw [13:14] adopt a bug ❤ 🕷 ❤ [13:14] hah [13:28] EWHOCARES [13:29] EGOAWAY [13:29] :-p [13:29] no [13:29] oops [13:29] EHIDANHOWAREYOUDOINGTODAY [13:30] EERRORING [13:31] :-D [13:31] I'm good ta [13:32] ETHOSEARENTERRORS [13:32] EHOWDOYOUKNOW ? [13:34] EERRORNOTANERROR [13:34] EIGAVEUPTHINKINGOFMEANINGFULERRORCONSTANTNAMESANDIDONTKNOWWHATTHISERRORACTUALLYIS [13:34] I love those errors, ijohnson_ [13:35] "Unknown Error" is a great one, too [13:35] that's where the developer didn't have a clue what they were doing [13:35] i still love "Error: success" [13:35] "Unexpected Error" is also a hit [13:35] yes! [13:35] ogra: what about openSSL with their goto fail; on success?! [13:36] haha, yeah ! [13:41] we need more emojis in in error messages E😊😊😊😊😊 [13:44] E 🤮 [13:53] Anyone know of a good way to convince squid to cache downloads from the snap store? I've set up a simple squid server to try this, and done 'snap set system proxy.http(s)=...' to point it at my proxy, but when I download anything through the store, it doesn't seem to get cached. [13:53] no growth of the cache dir, no speedup on subsequent downloads. squid access log occasionally only logs something like: 1564751506.978 51166 192.168.2.96 TCP_TUNNEL/200 14198 CONNECT api.snapcraft.io:443 - HIER_DIRECT/91.189.92.20 - === jdstrand_ is now known as jdstrand [14:02] plars: it might be uncached because it's an HTTPS request [14:02] you need to be intercepting requests for https://fastly.cdn.snapcraft.io [14:03] diddledan: that's what I was thinking, just curious if anyone had any luck getting squid to play nice with that sort of use case. I'm reading up on the ssl_bump stuff on squid which is kind of a terrifying mitm attack I think, and it could be workable if I can figure it out, but would probably take convincing snapd that the cert for squid is valid [14:04] plars: snapd will detect mitm attacks [14:04] plars: I think you could try setting up the enterprise proxy instead [14:05] that cost money tho [14:05] plars: it's not transparent as you need to opt in but that's one way out [14:05] diddledan: not for home use [14:05] aah ok [14:05] zyga: that's my backup plan, just trying to see if there's something more basic since I don't need any of the revision control stuff [14:05] plars: I think that's the only thing that really works [14:05] ..."opt in to way out" ... [14:06] zyga: kinda what I figured, thanks for confirmation [14:06] curiously, the fastly cdn is sending this header as part of its response: x-bzr-revision-number: 7464 [14:06] I wonder what that's about [14:06] dunno :) [14:08] diddledan: it's the revision number of the backing service on our side ;) [14:09] yeah I figured it was the django app [14:09] diddledan: think the commit id of the tree from which the code serving the snap was taken [14:10] plars: shit, sorry :( looks like my squid suggestion is bogus and you'll need the store proxy which is more aware of the content [14:16] roadmr: no worries, it was an interesting experiment, and it really is a *good* thing that snapd doesn't want to play nicely with that sort of thing. It would be troubling if it were easy to trick it into downloading some random thing from a proxy pretending to be the store :) [14:23] plars: we'd still verify the assertions on the response so it'd be hard to trick anyway [14:23] yeah [14:23] but at least snapd replies in clear and simple terms that MITM is in progress [14:24] plars: snap-proxy is a thing [14:24] Sorry for the late reply, I was thinking about something like retrieving window names (e.g. when in incognito mode of firefox) but apparently no such info is stored inside /proc but rather somewhere else? Chipaca ijohnson_ [14:24] plars: snap-store-proxy* [14:24] Chipaca: he knows; we were just wondering if a squid and a simple snap set system.http= would achieve the purpose of faster downloads for frequently-used snaps (he installs a lot) [14:24] ah ok [14:24] atleast I wasn't able to find anything inside /proc [14:24] Chipaca: we're going the store-proxy now :) [14:41] * cachio afk [15:24] snapuser9: yeah I don't know where window names are stored but snaps shouldn't affect where that is stored, once you figure out what you need access to, feel free to start a forum post and we can try to help you out more [15:44] * cachio lunch === pstolowski is now known as pstolowski|afk [18:31] amurray: the drawing snap has been uploaded to the store awaiting manual review :) [18:33] PR snapd#7074 closed: interfaces: add gpio-control interface [18:39] kenvandine: awesome! [21:08] hi everyone, i'm new to snap and trying to package a simple gtk application but falling short and seeking some help. [21:08] I've looked at several forum posts, existing gtk snapcraft.yaml files for examples, and tried lots of things. [21:08] would it be best to put all the hairy details into a forum post? [21:09] hellsworth: might be worth a forum post. being that we've got the weekend just appearing in a few hours most of the canonical folk will be elsewhere till monday [21:10] ah yes good point. a forum post it is then! [21:31] kenvandine: approved (but perhaps too soon - I am seeing the following error: https://paste.ubuntu.com/p/wkhbcXKhzz/ ) [21:32] kenvandine: I'm about to board my flight home so will catch you next week - cheers [21:32] Thanks [21:33] amurray: safe travels [23:01] PR snapcraft#2653 opened: scriplets: run override-pull on update_pull