[07:27] good morning desktoppers [08:24] good morning [08:51] good morning desktopers [08:51] lut oSoMoN, re didrocks [08:51] re seb128 [08:59] morning all [09:00] change of plan, not in London today [09:01] hey willcooke, Didier was saying so [09:04] salut didrocks, seb128, willcooke [09:04] hey willcooke, morning oSoMoN [09:05] hey didrocks , how are you feeling? [09:05] willcooke: a little bit better, still fever and headaches (but pills start to have an effect) and need toilets not really far… [09:06] :/ [09:06] Shouldn't you take a day off? [09:06] hint: you should [09:06] willcooke: I have some emails, especially with the GNOME guys to answer, and that's good enough to WFH [09:06] kk [09:06] take it easy [09:06] thanks! Will do [09:06] but yeah, today is email/community hub day, no coding [09:19] oSoMoN: I've put together another desktop-launch patch that should help with the user fonts issue: https://github.com/ubuntu/snapcraft-desktop-helpers/pull/87 [09:27] jamesh, excellent! I'll take a look right away [10:08] #ubuntu [12:22] seb128: hey I was backporting py-macaroonbakery to xenial [12:22] and it depends on python3-protobuf [12:22] that it's not available on xenial [12:23] so we should backport python3-protobuf too but this could break packages that already depends on the old protobuf [12:31] andyrock, hey, why would it if it's another package? do they conflict? [12:31] there cannot be two versions of protobuf in the same system [12:31] they conflict [12:32] so e.g. mir will not start anymore [12:32] well, then it's not an issue having the binding [12:32] but the binding require a newer protobuf/different version than xenial has? [12:32] indeed [12:32] can you build a python3-protof for/from the xenial version? [12:33] there is no support for python3 in protobuf 2.6 [12:33] that was my initial idea [12:34] hum, normally, you don't ned to install python3-protobuf or anything protobuf. You are supposed to generate the client bindings you are using in the tarball with protobuf [12:34] isn't what py-macaroonbakery is doing? [12:35] nope on bionic it just depends on python3-protobuf [12:35] let me check [12:37] well you still need the bindings [12:40] I guess maybe a bug upstream to ask them to generate the binding they are using as per protobuf doc? [12:40] didrocks, is that standard practice in debian/ubuntu packages? (/me doesn't know much about that stack) [12:40] (doesn't help you on the backport, sorry, but that's really what they should do) [12:41] seb128: basically, it's how protobuf upstream is telling how you should use it [12:41] the service exposing the protocol has the .proto file + the server-side binding generated [12:41] didrocks: they're already generated [12:41] which Debian disagrees with I guess if they now have a deb and make things depends on it? [12:41] also upstream [12:41] https://github.com/go-macaroon-bakery/py-macaroon-bakery/blob/master/macaroonbakery/internal/id_pb2.py [12:41] then, any client should sync the proto files and generate their own client code [12:42] ah :) so upstream has it and Debian does differently? [12:42] debian does the same [12:42] but you still need the bindings [12:42] seb128: yeah, could be, but really, this is code that has to be in sync with your repo, and there is different version of protobufs, that's why it's the upstream practice [12:42] wwhy? [12:42] https://github.com/go-macaroon-bakery/py-macaroon-bakery/blob/master/macaroonbakery/internal/id_pb2.py is what is generated by protobuf [12:42] (well, protoc) [12:42] otherwise all the "from google.protobuf import" will fail [12:43] we don't use the compiler but we use bindings [12:43] * didrocks knows more the Go generate bindings which don't depend on protobuf [12:43] weird the python version is different [12:43] (meaning: not self-sufficient) [12:43] * didrocks looks at some experiment I did in the past in python [12:46] hum, that's right, the python bindings are behaving differently: https://github.com/ubuntu/christmas-music-carousel-snap/blob/master/music-grpc-events/musicevents/piglow_pb2.py [12:46] I can try to add python3 support to protobu 2.6 [12:47] and edit the debian to build the python3 too [12:48] when static linking is an option you can workaround this issue [12:48] I did something like that when working on chromium on mir [12:48] yeah, it imports proto1 at build time: https://github.com/didrocks/grpc-piglow/blob/master/proto/piglow.pb.go [12:50] mmm adding python3 supports looks like a huge task [12:50] yeah, I bet… [12:53] andyrock, yeah, I wouldn't try to go this way [12:54] need another approach for xenial imho [12:54] https://github.com/google/protobuf/issues/7 [12:54] it says that 2to3 should do the job [12:56] could you just copy the content in an 3rd-party module dir? [12:58] oSoMoN, hey, LO 6.0beta1 builds fine so far, but there are some more packaging fixes/tweaks needed [13:02] didrocks: I can try [13:03] vendoring sounds like the best approach to me [13:08] hi. fwiw, I'm back on IRC for the time being. LTNS :-) [13:48] ricotz, are those changes things that are already in the debian packaging branch? [14:04] oSoMoN, no [14:06] oSoMoN, "minor" fallouts due ubuntu specific changes [14:23] ricotz, ack, I assume you're looking into them already? [14:27] oSoMoN, yeah, will do so soon [14:28] thanks! [14:33] Trevinho, not sure where to file this issue. I've found an issue with the shell when a snap is updated for a snap that's already running [14:33] Trevinho, for example if the chromium snap is running when the snap refreshes the shell will no longer associate the dock icon with the running process [14:34] no mark showing it's running and clicking on the icon launchers a new process [14:34] Trevinho, so far i've only reproduced it with chromium though, the gnome snaps behave. So perhaps the gnome apps are more resilient because they register as a gapplication [14:55] I've found an issue with snaps. People see them on the software center, install them, and then never know that they've installed a snap until they have an issue with it and try to get help and none of the troubleshooting tips work for some reason. [14:57] Trying to troubleshoot for something like that is not fun. Can't blame the user because no where at all is it obvious that they've installed a snap. [15:52] simonizor, how is that an issue with snaps? why do the user need to care how was the thing they use installed? it's rather an issue for the bug reporting process if anything [16:02] hi europe.... (from medellin) [16:13] hey Trevinho, how are you? back in shape? [16:13] seb128: hey [16:14] seb128: well.... I've been feeling better I've to say, I'm still very tired physically.... Not really energetic. [16:14] But, I can live with that... [16:15] No problems for working, thank god [16:15] no crazy parties this week, just sleep at night :) [16:21] btw, is ubufox coming back in firefox 57+, or is it dropped now? :-P [16:22] hey Nafallo, it has been a while, how are you? [16:22] what is ubufox doing? [16:22] seb128: it has indeed :-). I'm alright. back in Sweden these days, working on Ubuntu projects ;-) [16:22] nothing at the moment because it's an empty package [16:22] seb128: how's you? :-) [16:23] seb128 ^ ;) [16:23] chrisccoulson: heh, it's a legacy extension and not even showing up :-P [16:24] chrisccoulson, :) [16:24] Nafallo, I'm goog thanks [16:24] you are back to Ubuntu, nice ;) [16:24] I never really left :-P [16:25] I think I've done Ubuntu somewhere or the other since September 2004 now ;-) [16:25] seb128: If the user has no idea what type of package they've installed, they have no idea how to ask for help with it [16:25] It should be extremely obvious in the software center that you are installing a snap and not a regular deb package [16:26] Have had multiple users who have installed Discord lately through the software center trying to get help in a Linux Discord server that I'm in, and they had no idea they were running a snap [16:27] So these users go reporting bugs with the snap directly to the developers not knowing that they're even using a snap and that the problems were caused by the snap [16:28] you would have the same issue if they had t he deb [16:28] and the bug was in the deb [16:28] No. [16:28] simonizor: "extremely obvious in the software center" > the details page for the app already has a Source field [16:36] bug #1729656 [16:36] bug 1729656 in ubufox (Ubuntu) "changes to distribution.ini gets overwritten on upgrade" [Undecided,New] https://launchpad.net/bugs/1729656 [16:36] chrisccoulson: ^ reason I was asking about ubufox btw ;-) [16:37] might be deprecated now [18:00] night gang [18:24] good night al [18:24] all [21:16] RAOF: what was the name of that space board/card game you brought to one of the sprints? One of the US sprints, can't remember which one. We played it in someones hotel room. [21:17] robert_ancell: You're probably thinking of `Race For The Galaxy` [21:17] RAOF: bingo! [21:17] robert_ancell: The other one which would be a candidate is Eclipse, but I'm *pretty* sure I never brought that; it takes upwards of 3 hours. [21:18] RAOF: you on boardgamegeek.com? [21:18] bingo's a different kind of game entirely.. [21:18] I have browsed there; I don't think I've got an account theer. [21:19] sarnold: 🔥 [21:19] :D [21:20] sarnold: that's why everyone keeps looking at me weird when I should it out mid-game... [21:21] hehehe [21:22] Been playing lots of new games at a local group, trying to remember all the random ones I've played previously. [23:16] jbicha: http://paste.ubuntu.com/26061112/ [23:17] jbicha: while I don't want to be a DD is it possible to get push access to pkg-ayatana? [23:20] robert_ancell: done, please add yourself to Uploaders for snapd-glib to silence the lintian warning [23:21] jbicha: should I set the changelog to UNRELEASED or unstable? [23:23] it doesn't matter today since I'll go ahead and upload for you, but UNRELEASED is probably better in the future [23:24] btw, sunweaver usually uses gbp dch --auto (for easier cherry-picking to other releases) [23:25] acronym acronym abbreviation :) I'll have to look that one up [23:25] jbicha: still getting pubkey failure - where does debian check the ssh key from? [23:26] are you able to push to pkg-gnome repos (svn or git) ? [23:26] I was able to push to pkg-gnome ages ago. [23:26] check https://alioth.debian.org/account/ [23:27] if that looks ok, you might just have to wait a bit for Alioth to pick up your new group membership [23:27] yeah, key is OK. I'll try again later [23:28] your username is robert_ancell-guest [23:29] yes [23:30] becoming a Debian Maintainer should be pretty easy (it's not like Debian Developer) and it would let you do uploads yourself for packages you have been whitelisted for [23:31] jbicha: hmm, I can't pull from gnome-calculator either. [23:31] Are there any other permissions I should need? [23:33] make sure your url is set to ssh://robert_ancell-guest@git.debian.org/git/pkg-ayatana/snapd-glib [23:34] gnome-calculator and most GNOME pkgs still uses svn [23:34] jbicha: ah, that fixed it! I thought most git servers don't care about the username when using ssh keys [23:35] jbicha: thanks! [23:36] if you want yet another IRC channel, #debian-ayatana on irc.debian.org is set up to get git push notifications via a kgb bot [23:37] the username might not matter if you set in ~/.ssh/config like this [23:37] Host *.debian.org [23:37] User robert_ancell-guest [23:38] jbicha: you can never have enough IRC channels, right ;) [23:38] ok, me -> lunch