[06:43] morning [07:22] good morning [07:34] Goood morning! [07:45] * zyga got a power plug adapter for south africa [07:51] PR snapd#5845 closed: interface: add new `{personal,system}-files` interface [08:05] mornings === pstolowski|afk is now known as pstolowski [08:07] zyga: pstolowski: hey === JanC is now known as Guest41406 === JanC_ is now known as JanC [09:17] hello [09:17] is this the place for questions relating to snap packaging problems? [09:17] t1mp: hey, for various kinds of topics related to snaps [09:22] I created a snap package, which has a python+qt app inside (using PySide2, and PyInstaller). It works on most systems, but when the host is running Wayland, it fails as follows: [09:22] https://pastebin.ubuntu.com/p/whWXk2drjW/ [09:22] greyback is already helping me with it :) [09:27] for reference, my snapcraft.yaml is here: https://pastebin.ubuntu.com/p/KN9qsVRdJ8/ [09:33] morning all [09:33] Good day [09:36] t1mp: hey! don't you also need the 'wayland' plug? [09:36] I have seen some options where you use x11 and wayland plug but disable wayland in the environment which triggers it to use xwayland [09:37] popey: hello :) [09:37] I have no idea. Is there documentation about the wayland plug somewhere? [09:38] I thought all was working, but now when testing on different systems I found that the snap doesn't work on systems with wayland. [09:38] PR snapd#5915 closed: interfaces/network-setup-control: allow calling netplan generate/apply [09:38] https://forum.snapcraft.io/t/wayland-crash-with-electron-2/5752/4?u=popey [09:39] t1mp: theres a similar thread, where they use DISABLE_WAYLAND=1 in the environment to force xwayland [09:39] (I *think* that's how it works) [09:39] The documentation for the wayland interface is somewhat lean https://docs.snapcraft.io/the-wayland-interface/7784 [09:40] https://github.com/notepadqq/notepadqq/blob/master/snap/snapcraft.yaml#L16 [09:40] someone else using that method [09:42] shouldn't those kind of things be included in the desktop-helpers? [09:42] I could try the DISABLE_WAYLAND env var [09:42] possibly [09:43] plugs are irrelevant as long as I'm using --devmode right? [09:43] yes [09:43] (I think) [09:44] you can use "snappy-debug.security scanlog" to learn what's failing apparmor [09:44] leave that running in another terminal when you run the app [09:44] might be enlightening [09:49] brb [09:58] greyback: interesting, if I run the app (packaged with PyInstaller), but outside of the snap, it works fine. But if I set QT_QPA_PLATFORM=wayland, then it starts but without window decoration or images in the app. [09:59] Issue core18#93 closed: i386 support missing on core18-amd64 [10:02] The missing images looks familiar. I think I had the same on X when I missed some files to have GLSL support [10:10] quick coffee to raise the pressure [10:10] it's so sleepy today === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [11:06] PR snapd#6333 opened: daemon: intrduce /v2/connections snapd API endpoint [12:01] greyback, popey: https://github.com/notepadqq/notepadqq/blob/master/snap/snapcraft.yaml#L16 made the snap work for me [12:01] I don't know if it is the most elegant solution, but the package works now, which is what I need [12:01] thanks for the help :) [12:01] I still need to test it on some PCs though [12:02] t1mp: ok, glad you got it working [12:02] it appears wayland support still needs work in desktop-helpers [12:08] (bug reports welcome) :D [12:10] bug reports? https://www.youtube.com/watch?v=S-s4lEk91ng [12:11] brb again [12:14] diddledan: I'm doing my part! [12:14] mborzecki: hey, #6332 can land right? [12:14] PR #6332: spread: make Fedora 29 auto again [12:15] pstolowski: let me run the spread job again, sample of 2 is better than 1 ;) [12:18] PR snapd#6318 closed: release-tools: display self-help [12:18] ok [12:21] thanks pawel! [12:21] running some spread tests, I will go for a walk with the dog [12:21] it's snowing like crazy here [12:22] I don't want to miss this ;0 [12:22] :-) [12:22] Where is $here? I want to come! [12:22] (wait, that sounded creepy) [12:23] it's all bound to melt away, it will be +2 soon [12:23] but for now it's close to -1 so the view is very nice [12:23] (in Warsaw) [12:32] mborzecki: hi, are you mostly blocked waiting on various reviews at this point? [12:34] pedronis: i'm advancing respective pieces in private branches and working on a small fontconfig fix for fedora, but yeah, opening new PR is blocked until some of the current ones land === ricab is now known as ricab|lunch [12:35] mborzecki: ok, I will get to your PRs but there are even older PRs that needs some attention too, I can point at some of the "small" cards I mentioned yesterday in the standup [12:56] PR snapd#6334 opened: dirs, interfaces/builtin/desktop: system fontconfig cache path is different on Fedora [12:57] ^^ maybe a bit late for .37 but we could squeeze it in .37.1 [13:00] PR snapd#6332 closed: spread: make Fedora 29 auto again [13:03] off to pick up the kids === chihchun is now known as chihchun_afk [13:15] greyback: yeah. Is any OpenGL stuff not working with Wayland now? That seems quite important ;) [13:15] or is it just Qt? Or something specific that I'm using [13:16] t1mp: it has worked for me in the past. I don't know what broke for you, sorry [13:16] * Chipaca ⇝ lunch [13:26] re [13:26] t1mp: it probably depends [13:45] okay, thanks. :) For me it just matters that it works for my snap now ;) I'm sure the snap is optimal, but I'll see if I can work on it again in the near future. [13:46] *I'm sure my snap is NOT optimal :) [13:56] Chipaca: hi, could you review #6121 when you have a bit of time? [13:56] PR #6121: tests: new test for snapshots with more than 1 user [13:57] pedronis: yes [13:57] I guess the current test that creates snapshots for root + testuser isn't enough? [13:58] Chipaca: I suppose, it's question for cachio [13:59] mborzecki: btw, did you see the request for tweaks to #6317 ? [13:59] PR #6317: overlord/snapstate/backend: call fontconfig helpers from the new 'current' [14:02] pedronis: yup, saw that, it's in my queue [14:02] ok [14:31] Bug #1810948 opened: Internal Watcher Error [14:43] Bug #1810948 changed: Internal Watcher Error [15:08] zyga, https://travis-ci.org/snapcore/spread-cron/builds/476263240#L9952 [15:08] looking [15:08] that should be fixed with the update in the repo? [15:08] so [15:08] you need to update & reboot [15:08] let me check that locally [15:09] zyga, I did that [15:09] in that case the fix is not out yet [15:09] let's wait and see, I'll check the bug report in suse [15:09] thanks for trying! [15:09] np === chihchun_afk is now known as chihchun [15:54] * cachio lunch [16:12] jdstrand: thanks for adding the docs about personal/system-files, shouldn't the doc mention somehow that even just installation requires the snap-declaration ? [16:12] (Not sure how we did this for other interfaces with the same requirement) [16:18] feels like we should parse /etc/fonts/fonts.conf to figure out fontconfig locations rather than use hardcoded ones [16:19] fedora/rhel/centos all use /usr/lib/fontconfig/cache, but amzn2 is behind with updates and uses /var/cache/fontconfig and obviously has ID_LIKE=".. fedora.." [16:27] mborzecki: how hard is it to parse? [16:27] zyga: rather easy, it's xml, encoding/xml should do just fine [16:27] oh, that's good [16:28] is the location easy to extract? [16:29] zyga: yes, it's quite trivial :) [16:30] zyga: btw. wonder if we shuld also use ~/.local/share/fonts and ~/.fonts, we're only using system locations now [16:30] it depends if the code runs in user contetxt [16:31] zyga: we setup those mounts in the desktop interface, i'd guess it does [16:31] but who would run the parser? [16:33] zyga: snapd could parse /etc/fonts/fonts.conf in destkop iface mount connected plug [16:36] pedronis: probably, yes. we don't do that elsewhere that I can see but that is no reason not to do it here [16:36] pedronis: I'll adjust [16:44] mborzecki: can we discuss this tomorrow morning? [16:44] I'm adding some tests still [16:44] pedronis: done. fyi, I also updated snapd-control and kernel-module-control similarly. maybe that will reduce some questions [16:50] jdstrand: thanks [16:53] pstolowski: I left a comment in the hotplug-connect PR, we can talk about it tomorrow [16:54] pedronis: ok, thanks [16:55] zyga: sure === pstolowski is now known as pstolowski|afk [17:45] kenvandine: hello :) [17:45] I just saw your comment on https://github.com/ubuntu/snapcraft-desktop-helpers/issues/172 [17:46] hey t1mp [17:46] the way I understand it, plugs don't make a difference if the snap runs with --devmode (or as classic). Is that correct? [17:46] oh [17:46] right [17:46] hmm. I'm rebuilding my snap and (automatically, with jenkins) pushing it to the snap store, but I keep getting a 502 error [17:47] Snap Store encountered an error while processing your request: bad gateway (code 502). [17:47] The operational status of the Snap Store can be checked at https://status.snapcraft.io/ [17:47] script returned exit code 2 [17:49] I can see the revision though when I log in on snapcraft.io [17:50] ah no, I cannot. It is a different revision number [17:50] (just two digits swapped that's why I thought it is the same) [17:53] t1mp: try added QT_QPA_PLATFORM=wayland [17:53] i haven't trying anything qt/qml with wayland myself [17:54] kenvandine: I tried that. Still gives the same error. [17:54] I will add a comment to the issue [17:54] greyback: ^^ do you have any ideas? [17:54] greyback: https://github.com/ubuntu/snapcraft-desktop-helpers/issues/172#issuecomment-452388662 [17:55] greyback: was already helping me with it :) So he is aware. For me the problem is solved with the DISABLE_WAYLAND: 1. [17:55] That's enough for me, I have a working snap. But I guess it is still a workaround. [17:55] that's not a solution though :) [17:55] that makes it fall back to XWayland [17:59] true [17:59] but I need to release my snap this week :) [18:03] yeah... so probably good enough :) [18:06] Hey pedronis, you still around? [18:08] kenvandine: hey, yes I looked into it. I noticed the glvnd config was missing - $SNAP/usr/share/glvdn/egl_vendor.d/50_mesa.json - (stracing the binary, it was not opening any DRI drivers). When I hacked that file into the snap, the app worked for me on my wayland. So I recommended adding "libegl0-mesa0" to the stage packages, but that didn't help t1mp [18:09] hmm [18:09] kenvandine: I didn't try further tbh [18:09] understand === ricab|lunch is now known as ricab [20:45] cjwatson: do we have an eta for core18 builds in LP via build.snapcraft.io ? [20:47] popey: not yet [20:55] is that the eta is "not yet" or the eta on the eta is "not yet"? [20:56] kenvandine: I've done a lot of work yesterday and today on the gnome-extension POC that I created: https://github.com/snapcore/snapcraft/pull/2398 [20:56] PR snapcraft#2398: gnome extension [21:07] PR snapd#6335 opened: NOT-REVIEW: tests: fix "No space left on device" issue on amazon-linux [21:09] diddledan: I noticed! [21:09] it finally works [21:09] oSoMoN and I will look at it tomorrow [21:09] :-p [21:09] Woot [21:10] I've done a test build locally with my gnome-twitch snap which successfully built the snap and when installed it runs correctly [21:10] Excellent [21:10] including libwebkitgtk! [21:11] (I put a layout definition to fix that) [21:11] diddledan: thanks for that [21:12] it also uses the gnome-3.xx-1x04 platforms [21:14] I couldn't publish my test build of gnome-twitch to the store, though, because it rejected the `command-chain:` block because it doesn't believe such a config exists :-) [22:17] diddledan: yes [22:18] :-)