[04:50] <RAOF> duflu: Are you looking at the pulseaudio autopkgtest failure?
[04:58] <duflu> RAOF: I did not see a failure when the first email arrived today...
[04:58] <duflu> RAOF: And now I do.
[04:58] <duflu> Will get it sorted today
[04:58] <RAOF> duflu: http://people.canonical.com/~ubuntu-archive/proposed-migration/bionic/update_excuses.html#pulseaudio suggests a regression in gsequencer?
[04:58] <RAOF> Ah. Heh.
[06:14] <willcooke> jibel, morning!  Could you run the weekly meeting today?  Ken is still on hols and me & Seb are sprinting
[06:23] <flocculant> willcooke: up early :)
[06:38] <willcooke> morning flocculant, travelling this week, UTC-2
[06:39] <willcooke> flocculant, whats your excuse?
[06:40] <flocculant> willcooke: leave for work at 6:45 every day :(
[06:40] <flocculant> in the rain on a motorbike - on our roads - awesome :p
[06:40] <flocculant> and with that - have a good day :)
[06:41] <jibel> willcooke, okay, I'll do it
[06:42] <jibel> and good morning everyone
[06:56] <duflu> Morning willcooke, thumper, seb128, flocculant, jibel
[06:56] <duflu> And welcome to summer for some
[06:56] <seb128> hey duflu
[06:58] <jibel> summer ...
[06:58] <jibel> hi duflu
[06:58] <jibel> and seb128
[06:58] <seb128> hey
[07:09] <duflu> RAOF (or anyone), the gsequencer build failure looks like its own fault. It uses fakeroot without having a build dependency on it.
[07:09] <duflu> Not pulseaudio's fault
[07:10] <RAOF> duflu: care to fix gsequencer, then? 😉
[07:10] <duflu> RAOF: It has a whole new version waiting in proposed
[07:10] <duflu> Might not need to
[07:11] <duflu> Hmm looks like proposed has the same mistake
[07:27] <oSoMoN> good morning desktoppers
[07:27] <didrocks> good morning
[07:28] <duflu> Morning oSoMoN
[07:28] <oSoMoN> salut didrocks, ça va?
[07:28] <oSoMoN> hey duflu
[07:28] <duflu> Hey didrocks
[07:29] <didrocks> oSoMoN: ça va, et toi ?
[07:29] <didrocks> hey duflu
[07:29] <oSoMoN> didrocks, bien!
[07:42] <duflu> RAOF: I think a fix in Debian would be preferable to us forking it just for this. So... https://bugs.launchpad.net/debian/+source/gsequencer/+bug/1743511
[07:47] <Nafallo> morning
[07:48] <seb128> hey Nafallo
[08:04] <RAOF> duflu: you'll need to find someone to mark gsequencer as badtest. I can do that for SRUs, but not bionic.
[08:05] <duflu> seb128, ^ ?
[08:10] <seb128> duflu, I don't know how to do that offhand, I'm in meetings this morning but I can try to have a look later
[08:10] <seb128> maybe Laney can help you when he gets online though
[08:10] <duflu> seb128, no worries. Maybe didrocks?
[08:10] <seb128> could be
[08:10] <duflu> P.S. Go have dinner RAOF
[08:10] <duflu> it's probably time
[08:12] <RAOF> duflu: I've had dinner 😛
[08:13] <zyga> o/
[08:13] <duflu>         \o
[08:14] <RAOF> duflu: needs an archive admin to tweak the Britney hints.
[08:15] <RAOF> I think that requires AA
[08:15] <duflu> OK, ta
[08:20] <didrocks> duflu: the branch for the hints are in ~ubuntu-release
[08:20] <didrocks> so, not an AA, but a release team member
[08:21] <didrocks> sil2100, Laney and others ;)
[08:22]  * duflu smiles and rolls eyes
[09:02] <andyrock> hey all
[09:03] <Laney> hi
[09:05] <duflu> Hey hi andyrock, Laney
[09:06] <didrocks> hey andyrock, Laney
[09:07] <duflu> Laney, do you know how to mark packages as "badtest"?
[09:07] <Laney> sure
[09:07] <Laney> moin didrocks
[09:07] <duflu> Laney, could you help please? gsequencer from bug 1743511
[09:08] <oSoMoN> hey andyrock, Laney
[09:10] <Laney> ahoy oSoMoN!
[09:11] <Laney> duflu: this is broken in some way that is not force-badtest
[09:11] <Laney> build-deps should get you dpkg-dev which depends on fakeroot
[09:12]  * didrocks added ubuntu support to travis.debian.net (derivatives support in general ;))
[09:13] <duflu> Laney: That's a recommends
[09:13] <duflu> Not depends
[09:14] <seb128> hey Laney oSoMoN didrocks
[09:15] <Laney> oh yeah
[09:15] <Laney> ok, so why did it work on the previous run?
[09:15] <oSoMoN> salut seb128, ça va?
[09:17] <duflu> Laney, can't be sure but it appears likely previous runs had fakeroot already installed, for other reasons
[09:17] <seb128> oSoMoN, ouais, et toi?
[09:18] <Laney> you can see https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-bionic/bionic/amd64/g/gsequencer/20180102_123735_6177c@/log.gz installing fakeroot
[09:18] <Laney> hey duflu
[09:18] <Laney> hey seb128!
[09:18] <Laney> erm
[09:19] <oSoMoN> seb128, bien!
[09:20] <ricotz> good morning desktopers
[09:21] <duflu> Laney, oh, because it's not a build, maybe. Just a re-test: http://people.canonical.com/~ubuntu-archive/proposed-migration/bionic/update_excuses.html#pulseaudio
[09:23] <duflu> Argh. Or maybe because "Maximum number of ports exceeded (HTTP 403)"
[09:25] <oSoMoN> good morning ricotz
[10:13] <duflu> Long overdue for cooking something healthy
[10:13] <duflu> later
[10:16] <Laney> dang it
[10:16] <Laney> I found out what his problem was
[10:17] <Laney> new version of gsequencer has the depends, old doesn't
[11:41] <GunnarHj> seb128: Hi Sebastien! I'd need some help to make ibus-libpinyin ready to be MIR'ed (bug #1738582). One thing is bugsubscriber to ibus-libpinyin and libpinyin (comment #4). Then the dependency on lua5.1 should be replaced with lua5.3. To achieve that, and at the same time upgrade to a more recent version, I put my suggestions in comment #16 and #17 of bug #1735362.
[11:45] <seb128> GunnarHj, hey, I'm travelling and in meetings this week, I can do the bug subscription but I can't look at the lua problem atm
[11:46] <seb128> GunnarHj, desktop-bugs subscribed
[11:47] <GunnarHj> seb128: Thanks! Is the rest of the team traveling too? (It's a bit quiet here today.)
[11:47] <seb128> no, only will_cooke & me, others are probably just busy with work
[11:47] <didrocks> correct :)
[11:47] <seb128> or slacking away while we are not watching? :p
[11:48] <GunnarHj> seb128: ;)
[11:48] <didrocks> corre…oh wait!
[11:48] <seb128> :p
[12:31] <fossfreedom_> hi all - anybody got a bit of spare bandwidth to have a look at my patch update to gnome-screensaver please? https://bugs.launchpad.net/ubuntu/+source/gnome-screensaver/+bug/1740150
[13:47] <GunnarHj> jbicha: Hi Jeremy! Do you have time to help with the ibus-libpinyin MIR? You find my ideas of what to do next in comment #16 and #17 of bug #1735362.
[14:07] <GunnarHj> Hi Laney! When we discussed bug #1720250, you indicated that systemd might recognize graphical-session-pre.target services in 18.04. Can you tell if that will happen? Asking because it would be desirable to get rid of the /etc/profile.d hack.
[14:24] <Laney> GunnarHj: I'm working on it.
[14:25] <GunnarHj> Laney: Ok, thanks for letting me know.
[14:25] <Laney> It's a bit of a delicate task
[14:25] <GunnarHj> Laney: Don't try to explain to me - I wouldn't understand anyway. ;)
[14:27] <jibel> desktop meeting in 3min
[14:27] <didrocks> waow, countdown! That's the efficient french way :p
[14:29] <jibel> not a countdown, a reminder ;)
[14:29] <oSoMoN> SUBTLE REMINDER: 1 MINUTE
[14:30] <jibel> #startmeeting Desktop Team Weekly Meeting - 2018-01-16
[14:30] <meetingology> Meeting started Tue Jan 16 14:30:21 2018 UTC.  The chair is jibel. Information about MeetBot at http://wiki.ubuntu.com/meetingology.
[14:30] <meetingology> Available commands: action commands idea info link nick
[14:30] <oSoMoN> 🐵/
[14:30] <didrocks> hey
[14:30] <jibel> Roll call: andyrock, dgadomski, didrocks, duflu (out), jbicha, jamesh (out), jibel/heber, kenvandine (out), laney, oSoMoN, seb128 (out), tkamppeter, trevinho, robert_ancell (out)
[14:30] <andyrock> o/
[14:30] <heber> o/
[14:30] <Trevinho> \o
[14:31] <jibel> #topic andyrock
[14:31] <andyrock> 1. Updated udisks patch to hide snap squashFS in gnome-disk-utilities
[14:31] <andyrock> 2. Debugging a update-notifier issue with systemd path units
[14:31] <andyrock> 3. working on Ubiquity to enable canonical-livepath from the installer
[14:31] <andyrock> 4. eof
[14:31] <jibel> thanks andyrock
[14:32] <jibel> #topic dgadomski
[14:32] <dgadomski> hey
[14:32] <dgadomski> one issue has be brought to my attention: bug #1743422, is that something already known?
[14:32] <jibel> Trevinho, ^ do you know?
[14:33] <Trevinho> mh, no... I didn't notice that
[14:33] <Trevinho> let me see it
[14:33] <jibel> thanks dgadomski , anything else?
[14:33] <dgadomski> jibel: nope, that's it, thanks
[14:33] <jibel> #topic didrocks
[14:33] <didrocks> * Worked and commited in our branch date centering relative to workarea on the Shell (bug #1716432) and sent the patch upstream (no answer yet).
[14:33] <didrocks> * Implemented new sound option in Gnome Control Center (will be a minimal ubuntu patch) following Allan's design and GNOME Tweaks.
[14:34] <didrocks> * Mentored GCI. We got 2 large contributions: new getting started blender videos adapted to our patch. Needs packaging (video generation when building) though. The second one is notifications on Terminal. Not fully working yet but vte & g-t Fedora's patches are applied. Helped the students with pointers, tutoring, packaging 1o1…
[14:34] <didrocks> * Caught up on every PR/issues for the communitheme, did some comments, proposed some fixes.
[14:34] <didrocks> * Integrated some CI on both communitheme repos, ensuring that package is building via Travis on every PR (and no typo on the sass syntax thus!).
[14:34] <didrocks> * Follow up discussions about our ambiance theme refresh with Trevhino & Will
[14:34] <didrocks> * Some sponsoring and reviews
[14:34] <didrocks> .
[14:34] <jibel> didrocks, does this dot mean your done ?
[14:34] <jibel> you're*
[14:34] <didrocks> jibel: yes, as always :)
[14:34] <jibel> thanks didrocks
[14:35] <jibel> #topic duflu
[14:35] <jibel> * Pure Wayland apps (including mpv) not responding to the mouse (https://launchpad.net/bugs/1739625)
[14:35] <jibel>   - This week the fix landed in both upstream master and 3.26 branches.
[14:35] <jibel>   - Now definitely scheduled for release in mutter 3.36.3.
[14:35] <jibel> * Unresponsive touchpads (https://launchpad.net/bugs/1696929)
[14:35] <jibel>   - Conducted more hardware testing of the current and proposed patches: https://bugs.freedesktop.org/show_bug.cgi?id=98839#c85
[14:35] <jibel>   - Looking at the upstream bug list, this seems to be the most reported/duplicated complaint against libinput, accounting for over 11% of open libinput bugs. So it's not just me
[14:35] <jibel> * PulseAudio: Revised 11.1 for bionic, with additional fixes:
[14:35] <jibel>   - https://launchpad.net/bugs/1718824
[14:35] <jibel>   - https://launchpad.net/bugs/1732629
[14:35] <jibel>   - HELP: pulseaudio 1:11.1-1ubuntu4 stuck in proposed for no fault of its own, but: https://bugs.launchpad.net/debian/+source/gsequencer/+bug/1743511
[14:35] <jibel> * BlueZ 5.48: The testing PPA is finally built now:
[14:35] <jibel>   - https://launchpad.net/~bluetooth/+archive/ubuntu/bluez
[14:36] <jibel> * Gnome Shell performance work:
[14:36] <jibel>   - Good news: Found that clutter's frame clock logic needs the same kind of fixes as have recently landed in GTK.
[14:36] <jibel>   - Bad news: Not that much other than some shell animations rely on the precision of the clutter frame clock so experimental fixes to its precision so far don't yield any major visual improvement to everyday usage. I might propose changes to upstream clutter later. Not a shell priority any more, probably.
[14:36] <jibel>   - Presently: I am now working on generating some real-time profiles of gnome-shell to see where it blocks (or wastes CPU).
[14:36] <jibel> * Daily bug management across gnome-shell, mutter, gdm3, ubuntu-themes, bluez, pulseaudio, dkms, mir, wayland, totem, mpv, libinput.
[14:36] <jibel> * HELP: Fixes still awaiting sponsorship:
[14:36] <jibel>   - https://code.launchpad.net/~vanvugt/gtk/fix-1698270/+merge/331846
[14:36] <jibel>   - https://code.launchpad.net/~vanvugt/totem/fix-1502476/+merge/333195
[14:36] <jibel> thanks duflu
[14:36] <jibel> #topic jbicha
[14:37] <jibel> no jbicha, next
[14:38] <jibel> #topic jibel / heber
[14:38] <jibel> 17.10.1 tested and released
[14:38] <jibel> Continued the work on bootspeed and fixed bug 1742403
[14:38] <jibel> Added new KPI to the desktop dashboard.
[14:38] <jibel> Update grafana dashboards with correct releases
[14:38] <jibel> Continue adding gnome-software automated tests
[14:38] <jibel> ..
[14:38] <jibel> #topic kenvandine
[14:38] <jibel> on holidays
[14:38] <jibel> #topic Laney
[14:39] <jibel> Laney, around?
[14:39] <Laney> yeah I was commenting on a bug
[14:39] <Laney> here I am
[14:39] <Laney> o Reviews / fixes / testing / sponsoring for andyrock's SRUs
[14:39] <Laney> o GLib updates in Debian, debugged & fixed a GDateTime problem on i386
[14:39] <Laney> o More work on systemd user session stuff, copying my local setup in the VM to the actual trees including autotools work :(
[14:39] <Laney> ?
[14:40] <Laney> oh I helped a GCI guy a bit too
[14:40] <Laney> thanks to didrocks for pointing him at me
[14:40] <jibel> thanks Laney
[14:40] <jibel> #topic oSoMoN
[14:40] <oSoMoN> hey
[14:40] <oSoMoN> • firefox
[14:40] <oSoMoN>   ∘ helped with snap issues, and asked if upstream would be interested in patches to use xdg-open and xdg-settings
[14:40] <oSoMoN>   ∘ started https://forum.snapcraft.io/t/opening-a-local-file-with-its-default-application/3493
[14:40] <oSoMoN>   ∘ tested the patch attached to https://bugzilla.mozilla.org/show_bug.cgi?id=1430073 and provided feedback
[14:40] <oSoMoN> • chromium
[14:40] <oSoMoN>   ∘ moved ownership of my chromium-snap branches and packages to ~chromium-team
[14:40] <oSoMoN>   ∘ working on a patch to allow searching for the widevine CDM libs in $HOME/.local/lib, to make this mechanism snap-friendly
[14:40] <oSoMoN>   ∘ working on bug #1742653
[14:40] <oSoMoN> • libreoffice
[14:40] <oSoMoN>   ∘ successfully built and tested 5.4.4 snap, issued call for testing (https://forum.snapcraft.io/t/call-for-testing-libreoffice-5-4-4/3542)
[14:41] <oSoMoN>   ∘ excluded third-party fonts from the snap (used only at build time to run unit tests), saved 16MB
[14:41] <oSoMoN>   ∘ updating bionic to 5.4.4
[14:41] <oSoMoN> that's it from me
[14:41] <jibel> thanks oSoMoN
[14:41] <jibel> #topic seb128
[14:41] <jibel> seb128, is sprinting
[14:41] <jibel> #topic tkamppeter
[14:42] <tkamppeter> Did not prepare the list yet, if we are not yet at the end of the line, will post it at the end.
[14:42] <jibel> tkamppeter, that's fine, we'll come back to you in 5 min
[14:42] <jibel> #topic Trevinho
[14:42] <Trevinho> · Some more digging into fprintd and pam finger print
[14:42] <Trevinho> · Some improvements to the gnome-shell / mutter patches for scaling
[14:42] <Trevinho> · Even more JS fixes to gnome-shell and updates to old patches
[14:42] <Trevinho> · Review of some CSS changes to Ambiance, and some look at Ambiance-RW
[14:42] <Trevinho>  ...
[14:43] <jibel> thanks Trevinho
[14:43] <jibel> #topic robert_ancell
[14:44] <jibel> - Guest session work
[14:44] <jibel> - Working on upstreaming gnome-software channel selection
[14:44] <jibel> - LightDM XDMCP fixes
[14:44] <jibel> thanks robert_ancell
[14:44] <jibel> #topic AOB
[14:44] <jibel> anyone wants the mic?
[14:44] <didrocks> nothing for me
[14:45] <jibel> okay then
[14:45] <jibel> #topic tkamppeter
[14:45] <tkamppeter> - cups-filters: Released 1.19.0: Overtaken all new functionality from the driverless PPD generator of CUPS to the one in cups-filters, fixed color space/color depth determination in PPD generator, improved human-readable strings in the PPDs.
[14:45] <tkamppeter> - cups: Fixed bug of 16-bit high color depth not working in Apple Raster (AirPrint printers, most common driverless printing type).
[14:45] <tkamppeter> - Hardware: Searching for a new printer to purchase, investigated models.
[14:45] <tkamppeter> - Google Summer of Code 2018:
[14:46] <tkamppeter> Interviewed first group of students
[14:46] <tkamppeter> Put up application for Linux Foundation
[14:46] <tkamppeter> Put up web pages with project ideas.
[14:46] <tkamppeter> - Bugs.
[14:47] <jibel> thanks tkamppeter, anything else?
[14:47] <jibel> I take your silence as a no.
[14:47] <tkamppeter> OK, that's all.
[14:48] <jibel> Thanks everyone and see you next week!
[14:48] <jibel> #endmeeting
[14:48] <meetingology> Meeting ended Tue Jan 16 14:48:11 2018 UTC.
[14:48] <meetingology> Minutes:        http://ubottu.com/meetingology/logs/ubuntu-desktop/2018/ubuntu-desktop.2018-01-16-14.30.moin.txt
[14:48] <didrocks> thanks!
[14:48] <tkamppeter> One problem: Need xnox to check the systemd problem.
[14:48] <didrocks> Trevinho: a question, I'm curious about your js GNOME Shell changes, do you have pointers?
[14:48] <tkamppeter> And need someone to sponsor my NEW packages.
[14:48] <oSoMoN> thanks jibel
[14:49] <Trevinho> didrocks: they're related to https://bugzilla.gnome.org/show_bug.cgi?id=791233 but I didn't push the final patches yet
[14:49] <Trevinho> didrocks: as I've been discussing about the best design for them
[14:50] <seb128> tkamppeter, we talked about systemd with Steve, it would help if you could figure out a way to reproduce easily the bug that doesn't involve owning a specific printer or such ... do you think it's possible?
[14:50] <seb128> tkamppeter, foundations is going to look at it but it would make easier to debug if they could reproduce easily
[14:50] <didrocks> Trevinho: lucky you, you get reviews! :)
[14:50] <Trevinho> didrocks: well, after pinging people a lot.....
[14:51] <Trevinho> didrocks: and even after that not many really, well not specific as Jonas does
[14:51] <didrocks> Trevinho: btw, if you have an idea to replace the idle_add, I'm all ears
[14:51] <tkamppeter> seb128, probably any USB printer would serve, not only the concrete model I am using.
[14:51] <Trevinho> I don't think a "I don't like it" is enough for a review :)
[14:51] <didrocks> Trevinho: yeah, it's really hard to get Shell reviews for the js part :/
[14:51] <Trevinho> didrocks: I've been trying to think another way, let me go back to that
[14:52] <didrocks> Trevinho: would love to :)
[14:52] <seb128> tkamppeter, right, I'm not sure they own a printer, Steve doesn't travek with one for sure at least so having a script or something to simulate it would be nice
[14:53] <seb128> tkamppeter, could umockdev allows to do that?
[14:58] <tkamppeter> seb128, I think this is difficult, to have some software emulating the appearance of a USB device.
[14:59] <tkamppeter> seb128, where is Steve currently? In the Bluefin office?
[14:59] <seb128> tkamppeter, no, in south africa for the same sprint will and I are at
[15:00] <tkamppeter> In South Africa, interesting.
[15:00] <seb128> tkamppeter, umockdev could do what you need to record/replay the event maybe?
[15:01] <seb128> tkamppeter, well I'm going to ask them if they can access an usb printer to debug
[15:01] <seb128> tkamppeter, did you try if the bug happens on Debian as well?
[15:01] <tkamppeter> What one can do is, right after the Sprint, buy a cheapo HP, like the DeskJet 2540, expensed to Canonical.
[15:01] <seb128> I wonder if other distributions are having the same issue and if not why they don't
[15:02] <tkamppeter> What is the venue in South Africa? A hotel?
[15:02] <seb128> yes
[15:03] <tkamppeter> Any printer which is around has a USB connector, you could reproduce the bug even if there is no working Linux driver for the printer.
[15:04] <tkamppeter> So some printer in the hotel's business center or office could even help.
[15:04] <tkamppeter> Simply go with a laptop and a USB cable into the business center turn on and connect the printer and see what the computer does.
[15:05] <seb128> k
[15:05] <seb128> tkamppeter, umockdev allows you to record udev events and reply them to simulate devices like that
[15:06] <seb128> tkamppeter, https://github.com/martinpitt/umockdev
[15:06] <seb128> tkamppeter, but that's maybe not required in that case, should be easy enough to access a printer
[15:06] <tkamppeter> pitti has the solution for everything ...
[15:07] <seb128> indeed
[15:07] <tkamppeter> I want to have pitt back at Canonical ...
[15:07] <seb128> please do :)
[15:08] <tkamppeter> But probably Canonical would have to big very high, like for a great socker player.
[15:09] <tkamppeter> s/big/bid/
[15:13] <pitti> tkamppeter: indeed, mocking the *presence* of a USB printer is fairly straightforward with umockdev
[15:13] <pitti> mocking its *behaviour* is another topic, and that's currently rather brittle
[15:13] <pitti> I did some R&R on this a few years ago, and it kind of works, but maybe don't bet on it
[15:14] <pitti> but udev doesn't care about that, so the "static" udev properties are fine
[15:14] <Laney> yeah it's just adding it
[15:15] <Nafallo> hey! pitti didn't leave. he's here!
[15:15] <pitti> sure! I can't leave my old desktop friends! :-)
[15:17] <pitti> tkamppeter: I did a demo of this at this part: https://media.ccc.de/v/ASG2017-121-simulate_hardware_for_integration_testing#t=697
[15:17] <seb128> salut pitti
[15:17] <pitti> tkamppeter: (< 10 minutes)
[15:17] <pitti> ça va seb128 !
[15:17] <seb128> pitti, ça va, j'ai suis en Afrique du Sud mais j'avais la grippe, ça va mieux maintenant !
[15:18] <seb128> pitti, tkamppeter just needs the usb printer detection, it's for https://github.com/systemd/systemd/issues/7109
[15:18] <pitti> seb128: ouch à nouveau ? quel dommage !
[15:18] <seb128> " Services asked for by UDEV do not get triggered #7109"
[15:18] <seb128> seems an escaping issue with paths or something like that
[15:19] <seb128> lennart added patches to that bug but they don't seem to work and it stalled since
[15:19] <seb128> pitti, et toi, comment ça va ?
[15:19] <seb128> pitti, oh et bonne année au fait :)
[15:20] <pitti> seb128: bonne année à toi aussi ! (mais je crois que nous à parlè déjà cette année :) )
[15:21] <seb128> c'est bien possible, je ne me rappelle de plus rien à cause de la grippe :p
[15:22] <tkamppeter> ubot5 is not answering to GitHub bugs and GitHub is so common now.
[15:24] <tkamppeter> pitti, por favor, fale alemão, português, ou Inglês.
[15:25] <Laney> ????
[15:25] <Laney> neu Gymraeg?
[15:26] <tkamppeter> Laney, was sagst Du da?
[15:27] <Nafallo> hohum. ska vi byta till svenska istället kanske? :-)
[15:27] <tkamppeter> Laney, which language is that?
[15:29] <xnox> tkamppeter, i'm not sure how i can help you. The upstream bug report looks valid, and it is correct that you do not get the expected units started.
[15:29] <Laney> tkamppeter: japanese then welsh
[15:30] <tkamppeter> Laney, then Japanese is easy: ??? ?? ?????? ??? ??  ???? ...
[15:30] <seb128> xnox, we have been told you maintain systemd and such are the right person for looking at getting such issues resolved?
[15:30] <xnox> tkamppeter, the only solution i can offer you is as follows - make all usb-printers trigger "setup-usb-printer.service" which correctly scans all printers and effectively does `systemctl start usb-printer@the-correct-path.service`
[15:30] <Laney> tkamppeter: you know that if you have ENV{SYSTEMD_WANTS}="foo@" then the path is added automatically?
[15:31] <Laney> +=
[15:31] <xnox> Laney, the issue here, is that tkamppeter is adding SYSTEMD_WANTS... but the systemd escaping sysfs-path and unescaping it is not round trip save, and in the end the unit is triggered with a non-existant sysfs.
[15:31] <xnox> due to loss of too many or not enough backslashes.
[15:31] <Laney> xnox: yeah, but you can remove some manual handling of the path by using this facility
[15:32] <Laney> 70-printers.rules computes the path using a PROGRAM thing
[15:32] <Laney> but AIUI that shouldn't be necessary
[15:32] <xnox> hm, true.
[15:32] <Laney> if you just want to start a template unit with the path
[15:33] <Laney> https://www.freedesktop.org/software/systemd/man/systemd.device.html -> SYSTEMD_WANTS section
[15:33] <tkamppeter> xnox, Laney, and there are two types of USB printers affected: Conventional ones which need to have the auto-setup tool of system-config-printer started and IPP-over-USB ones which need to have ippusbxd started.
[15:34] <Laney> I just made that change
[15:35] <Laney> and it resulted in https://paste.ubuntu.com/26398450/
[15:35] <Laney> tkamppeter: the sysfs path there looks correct?
[15:35] <xnox> Laney, i wanted to try to want an "ls" unit which would /bin/ls the whatever was passed to it
[15:36] <xnox> unable to access /sys/sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.1 -> is the path right?
[15:36] <tkamppeter> Laney, you get /sys/sys/devices/...
[15:36] <xnox> actually I guess i can do that on _all_ devices, to see where the paths break
[15:36] <Laney> it's got an extra /sys
[15:36] <Laney> but line 2 doesn't have that
[15:36] <xnox> tkamppeter, does udev-configure-printer adds on?
[15:36] <xnox> Jan 16 15:34:09 raleigh.local udev-configure-printer[18659]: add /sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.1
[15:36] <xnox> Jan 16 15:34:09 raleigh.local udev-configure-printer[18659]: unable to access /sys/sys/devices/pci0000:00/0000:00:1a.0/usb1/1-1/1-1.1
[15:37] <Laney> and that is a print of argv[n]
[15:37] <Laney> so I'm guessing u-c-p is adding it
[15:37] <xnox> cause yeah, it is called with /sys, and then it changes to /sys/sys?
[15:37] <tkamppeter> Laney, originally, the s-c-p tool was called with /devices/...., your new method seemas to have called it with /sys/devices/...
[15:37] <Laney> yeah that must be a difference
[15:37] <Laney> so I guess this is a fix/workaround, if you remove that prepending?
[15:38] <tkamppeter> Laney, looks like so.
[15:38] <xnox> maybe the tool should check if starts with /sys do not add an extra /sys -> to keep it compatible with both old and new world orders
[15:39] <tkamppeter> Laney, what exactly did you change?
[15:39] <Laney> tkamppeter: ACTION=="add", SUBSYSTEM=="usb", ENV{DEVTYPE}=="usb_device", ENV{ID_USB_INTERFACES}=="*:0701??:*", TAG+="udev-configure-printer", TAG+="systemd", ENV{SYSTEMD_WANTS}+="udev-configure-printer@"
[15:42] <Laney> I would guess the old way has some delicate problems with quoting
[15:42] <tkamppeter> Laney, will this still use the file /lib/systemd/system/udev-configure-printer@.service?
[15:43] <tkamppeter> Laney, or does it call the udev-configure-printer utility directly?
[15:43] <Laney> it's starting that service
[15:44] <tkamppeter> OK. The systemd service is needed that the tool can spawn a daemon without systemd killing it.
[15:44] <Laney> it should be equivalent to what you had
[15:44] <Laney> except apparently you were getting the sysfs path without /sys before
[15:47] <tkamppeter> Laney, can you also check whether something similar is possible with /lib/systemd/system/ippusbxd@.service
[15:49] <tkamppeter> Laney, and the ippusbxd UDEV rule:
[15:49] <tkamppeter> ACTION=="add", SUBSYSTEM=="usb", ENV{DEVTYPE}=="usb_device" ENV{ID_USB_INTERFACES}=="*:070104:*", OWNER="root", GROUP="lp", MODE="0664", TAG+="systemd", PROGRAM="/bin/systemd-escape --template=ippusbxd@.service $env{BUSNUM}:$env{DEVNUM}", ENV{SYSTEMD_WANTS}+="%c"
[15:50] <gQuigs> if you need a real printer tester - I have a USB printer -  epson stylus cx 7450 if that helps...
[15:52] <Laney> tkamppeter: is that broken?
[15:54] <Laney> seems okay to me
[15:56] <tkamppeter> Laney, checking now.
[16:07] <tkamppeter> Laney, I have the folowin in /lib/systemd/system/ippusbxd@.service:
[16:08] <tkamppeter> [Unit]
[16:08] <tkamppeter> Description=Daemon to make IPP-over-USB printers available as network printers (%i)
[16:08] <tkamppeter> [Service]
[16:08] <tkamppeter> Type=forking
[16:08] <tkamppeter> GuessMainPID=true
[16:08] <tkamppeter> ExecStart=/usr/sbin/ippusbxd --bus-device %i --from-port 60000 --logging
[16:08] <tkamppeter> # ExecStop= Not needed, ippusbxd stops by itself on shutdown of the printer
[16:08] <tkamppeter> and above-mentioned rule.
[16:08] <Laney> does it work?
[16:09] <tkamppeter> I have started "udevadm monitor --environment" and plugged in the printer.
[16:09] <tkamppeter> I get a lot of output with
[16:09] <tkamppeter> SYSTEMD_WANTS=ippusbxd@002:008.service 'udev-configure-printer@devices-pci0000:00-0000:00:14.0-usb2-2\x2d1.service' printer.target
[16:10] <tkamppeter> in it, but ippusbxd does not get started.
[16:12] <Laney> well that service name looks right to me
[16:12] <tkamppeter> But if I enter "sudo systemctl start ippusbxd@002:008.service" ippusbxd gets correctly started. This should usually go automatically.
[16:12] <Laney> take a look in the journal
[16:13] <tkamppeter> Laney, Which command do I use for that?
[16:13] <Laney> journalctl -e
[16:13] <Laney> look for stuff that happened when you plugged it in
[16:13] <Laney> probably Starting Daemon to make IPP-...
[16:17] <tkamppeter> It says "Jan 16 14:06:02 till-x1carbon kernel: usb 2-1: usbfs: process 16465 (ippusbxd) did not claim interface 0 before use"
[16:17] <tkamppeter> Laney, ^^ but ippusbxd can be started normally from the command line or when manually starting the systemd service.
[16:17] <Laney> don't know what that means but it looks like systemd did start the service for you
[16:18] <Laney> is that a race condition?
[16:18] <tkamppeter> Laney, perhaps. I will remove the 70-.. UDEV rule and see whether this sokves it.
[16:20] <Laney> try adding ExecStartPre=/bin/sleep 5 or something?
[16:20] <Laney> don't forget to systemctl daemon-reload after editing units
[16:21] <tkamppeter> Laney, now it works, but strangely I get again
[16:22] <tkamppeter> Jan 16 14:19:44 till-x1carbon kernel: usb 2-1: usbfs: process 16931 (ippusbxd) did not claim interface 0 before use
[16:22] <tkamppeter> and this time the daemon is running though, but as PID 16938.
[16:23] <tkamppeter> Will try the  ExecStartPre=/bin/sleep 5
[16:23] <Laney> do you see that when you start it yourself?
[16:24] <Laney> sounds to me like it is not on systemd's side anyway (i.e. if you see that message, systemd has started the service for you)
[16:38] <tkamppeter> Laney, there is some hidden race condition.
[16:39] <tkamppeter> I have removed 70-printers.udev, so there is only 55-ippusbxd.udev.
[16:39] <tkamppeter> Laney, without the sleep the ippusbxd starts sometimes sometimes not, also for sleep 1 or sleep 2.
[16:40] <tkamppeter> Laney, With spleep 3 it never starts.
[16:41] <tkamppeter> In any case when ippusbxd starts, there is this "usbfs: process 16931 (ippusbxd) did not claim interface 0 before use" and ippusbxd starts with another process number.
[16:44] <Laney> Sounds like ippusbxd is using libusb wrong
[16:46] <tkamppeter> Laney, ippusbxd directly started from the command line starts reliably, also when starting the service manually, ippusbxd starts reliably.
[16:48] <tkamppeter> Laney, also, in the cases when ippusbxd does not start after plugging the printer, journalctl -e shows no hint that the service gets started. The "Starting Daemon to make IPP-over-USB printers available as network printers (001:001)" line does not appear.
[16:48] <Laney> ????????????????????????????????????????????????????????????????
[16:49] <Laney> why are you telling me about this random message if it's not starting?
[16:49] <Laney> how can it be printing a message in that case :/
[16:55] <tkamppeter> Laney, no I have confused you. Let us start over:
[16:55] <tkamppeter> 1. Starting ippusbxd from the command line with the printer plugged always works.
[16:57] <tkamppeter> 2. Plugging the printer gives "SYSTEMD_WANTS=ippusbxd@002:031.service printer.target" in the output of "udevadm monitor --environment", see full output in https://paste.ubuntu.com/26398883/
[17:00] <tkamppeter> 2a. In most cases usbippxd does not start. Then in journalctl -e there is no message of the ippusbxd unit, not even "Starting Daemon to make IPP-over-USB printers available as network printers (001:001)".
[17:02] <tkamppeter> 2b. In some cases ippusbxd actually gets started. Then journalctl -e actually shows "Starting Daemon to make IPP-over-USB printers available as network printers (001:001)" and even "Started Daemon to make IPP-over-USB printers available as network printers (001:001)", but it also shows "usbfs: process 16931 (ippusbxd) did not claim interface 0 before use" and ippusbxd starts with another process number. Looks like that ippusbxd got started twice (
[17:02] <tkamppeter> and only one of the two stays running).
[17:02] <tkamppeter> Laney, I hope it is easier to understand now for you.
[17:03] <ogra> did you try 8for laughs) to set Type=simple in your systemd unit ?
[17:03] <Laney> Sure, but now I'm afraid I have no idea why the unit isn't getting started when it's in SYSTEMD_WANTS
[17:03] <Laney> What does systemctl status <that service> show after it fails to start?
[17:06] <tkamppeter> ogra, tried it and no change.
[17:07] <ogra> well, was worth a try :)
[17:08] <tkamppeter> Laney, udo systemctl status ippusbxd@002:033.service
[17:08] <tkamppeter> ● ippusbxd@002:033.service - Daemon to make IPP-over-USB printers available as n
[17:08] <tkamppeter>    Loaded: loaded (/lib/systemd/system/ippusbxd@.service; static; vendor preset:
[17:08] <tkamppeter>    Active: inactive (dead)
[17:10] <Laney> no log output?
[17:10] <Laney> ok, I'm out of ideas, sorry
[17:11] <Laney> if it's in WANTS and it's not being started and it's not failed or anything, I have no idea
[17:13] <tkamppeter> Laney, is there another log than only  journalctl -e
[17:14] <tkamppeter> xnox, pitti, are you still here?
[17:15] <tkamppeter> Laney, and I still can start the service manually and get
[17:15] <tkamppeter> till@till-x1carbon:~$ sudo systemctl start ippusbxd@002:033.service
[17:15] <tkamppeter> till@till-x1carbon:~$ sudo systemctl status ippusbxd@002:033.service
[17:15] <tkamppeter> ● ippusbxd@002:033.service - Daemon to make IPP-over-USB printers available as network printers (002:033)
[17:15] <tkamppeter>    Loaded: loaded (/lib/systemd/system/ippusbxd@.service; static; vendor preset: enabled)
[17:15] <tkamppeter>    Active: inactive (dead) since Tue 2018-01-16 15:14:54 -02; 2s ago
[17:15] <tkamppeter>   Process: 19766 ExecStart=/usr/sbin/ippusbxd --bus-device 002:033 --from-port 60000 --logging (code=exited, status=0/SUCCESS)
[17:15] <tkamppeter>  Main PID: 19766 (code=exited, status=0/SUCCESS)
[17:15] <tkamppeter> Jan 16 15:14:54 till-x1carbon systemd[1]: Started Daemon to make IPP-over-USB printers available as network printers (002:033).
[17:15] <tkamppeter> Jan 16 15:14:54 till-x1carbon ippusbxd[19766]: 60000|19768|
[17:15] <tkamppeter> till@till-x1carbon:~$
[17:16] <Laney> Did the process exit?
[17:16] <Laney> or did it get the PID wrong?
[17:18] <tkamppeter> Laney, sorry, this was of ogra's suggestion with "Type=simple", which does not work.
[17:18] <tkamppeter> Laney, here is the correct one:
[17:18] <tkamppeter> till@till-x1carbon:~$ sudo systemctl start ippusbxd@002:034.service
[17:18] <tkamppeter> till@till-x1carbon:~$ sudo systemctl status ippusbxd@002:034.service
[17:18] <tkamppeter> ● ippusbxd@002:034.service - Daemon to make IPP-over-USB printers available as network printers (002:034)
[17:18] <tkamppeter>    Loaded: loaded (/lib/systemd/system/ippusbxd@.service; static; vendor preset: enabled)
[17:18] <tkamppeter>    Active: active (running) since Tue 2018-01-16 15:17:22 -02; 9s ago
[17:18] <tkamppeter>   Process: 19892 ExecStart=/usr/sbin/ippusbxd --bus-device 002:034 --from-port 60000 --logging (code=exited, status=0/SUCCESS)
[17:18] <tkamppeter>  Main PID: 19894 (ippusbxd)
[17:18] <tkamppeter>     Tasks: 3 (limit: 4915)
[17:18] <tkamppeter>    CGroup: /system.slice/system-ippusbxd.slice/ippusbxd@002:034.service
[17:18] <tkamppeter>            └─19894 /usr/sbin/ippusbxd --bus-device 002:034 --from-port 60000 --logging
[17:18] <tkamppeter> Jan 16 15:17:22 till-x1carbon systemd[1]: Starting Daemon to make IPP-over-USB printers available as network printers (002:034)...
[17:19] <tkamppeter> Jan 16 15:17:22 till-x1carbon systemd[1]: Started Daemon to make IPP-over-USB printers available as network printers (002:034).
[17:19] <tkamppeter> till@till-x1carbon:~$
[17:19] <Laney> seems ok then
[17:20] <xnox> tkamppeter, why is the number different? it got incremented due to plugging in / unplugging? 002:031 vs 002:034?
[17:21] <tkamppeter> Laney, journalctl -e still shows "kernel: usb 2-1: usbfs: process 19892 (ippusbxd) did not claim interface 0 before use", but the process number in there is of the ExecStart call and ippusbxd has actually another one, compare the status output.
[17:22] <tkamppeter> Laney problem is still that the service does not start automatically, when I plug in the printer, but starting manually works perfectly.
[17:22] <tkamppeter> xnox, yes. after each plugging the printer gets a new number, increased by 1.
[17:23] <tkamppeter> Laney, seems that we can ignore the "kernel: usb 2-1: usbfs: process 19892 (ippusbxd) did not claim interface 0 before use" message. It seems not to affect the problem.
[17:23] <Laney> so it did something before forking
[17:23] <Laney> I don't think that matters
[17:24] <Laney> What is more interesting is why systemd doesn't start it *sometimes*
[17:24] <Laney> no ideas on that though
[17:24] <tkamppeter> Laney, xnox, the problem is that the service appears in SYSTEMD_WANTS but does not get started.
[17:24] <xnox> tkamppeter, i'd ratehr use --no-fork int he ExecStart and remove both Type=forking and GuessMainPID
[17:24] <Laney> maybe show us the whole journal from before you plug it in
[17:25] <tkamppeter> Laney, xnox, and here in the ippusbxd case there seems also not to be any quoting/unquoting problem.
[17:25] <Laney> yes
[17:26] <xnox> tkamppeter, not sure why there is need to invoke systemd-escape in this case either.
[17:29] <tkamppeter> xnox, Laney: --no-fork (with Type=simple) behaves the same:
[17:29] <tkamppeter> till@till-x1carbon:~$ sudo systemctl status ippusbxd@002:035.service
[17:29] <tkamppeter> ● ippusbxd@002:035.service - Daemon to make IPP-over-USB printers available as network printers (002:035)
[17:29] <tkamppeter>    Loaded: loaded (/lib/systemd/system/ippusbxd@.service; static; vendor preset: enabled)
[17:29] <tkamppeter>    Active: inactive (dead)
[17:29] <tkamppeter> till@till-x1carbon:~$ sudo systemctl start ippusbxd@002:035.service
[17:29] <tkamppeter> till@till-x1carbon:~$ sudo systemctl status ippusbxd@002:035.service
[17:29] <tkamppeter> ● ippusbxd@002:035.service - Daemon to make IPP-over-USB printers available as network printers (002:035)
[17:29] <tkamppeter>    Loaded: loaded (/lib/systemd/system/ippusbxd@.service; static; vendor preset: enabled)
[17:29] <tkamppeter>    Active: active (running) since Tue 2018-01-16 15:28:25 -02; 2s ago
[17:29] <tkamppeter>  Main PID: 20221 (ippusbxd)
[17:29] <tkamppeter>     Tasks: 4 (limit: 4915)
[17:29] <tkamppeter>    CGroup: /system.slice/system-ippusbxd.slice/ippusbxd@002:035.service
[17:29] <tkamppeter>            └─20221 /usr/sbin/ippusbxd --no-fork --bus-device 002:035 --from-port 60000 --logging
[17:30] <tkamppeter> Jan 16 15:28:25 till-x1carbon systemd[1]: Started Daemon to make IPP-over-USB printers available as network printers (002:035).
[17:30] <tkamppeter> till@till-x1carbon:~$ ps auxwww | grep ippusbxd
[17:30] <tkamppeter> root     20116  0.0  0.0  80192  4804 pts/1    S    15:26   0:00 sudo emacs /lib/systemd/system/ippusbxd@.service
[17:30] <tkamppeter> root     20117  0.7  0.6 395956 55380 pts/1    Sl   15:26   0:00 emacs /lib/systemd/system/ippusbxd@.service
[17:30] <tkamppeter> root     20221  0.2  0.0 270272  3180 ?        Ssl  15:28   0:00 /usr/sbin/ippusbxd --no-fork --bus-device 002:035 --from-port 60000 --logging
[17:30] <tkamppeter> till     20234  0.0  0.0  14356  1044 pts/1    S+   15:28   0:00 grep --color=auto ippusbxd
[17:30] <tkamppeter> till@till-x1carbon:~$
[17:31] <tkamppeter> Laney, xnox, the problem seems not to be affected by what the service unit does but by the service not started automatically.
[17:31] <Laney> yeah that's the bit I don't understand
[17:32] <Laney> please just share the whole journal from just before you plug it in when it doesn't work
[17:32] <Laney> maybe we can see something
[17:33] <tkamppeter> xnox, the /bin/systemd-escape call does not seem to cause any problem, as the SYSTEMD_WANTS shows the correct service name, the one which I can manually start.
[17:33] <tkamppeter> Laney, how to obtain the full journal in its maximum verbosity?
[17:34] <Laney> journalctl | tee till.log
[17:34] <Laney> journalctl -f sorry
[17:39] <tkamppeter> Laney, xnox, here is the full log without "-f": https://paste.ubuntu.com/26399230/
[17:41] <tkamppeter> Laney, xnox, here is the journalctl -f of only plugging the printer:
[17:41] <tkamppeter> Jan 16 15:40:13 till-x1carbon kernel: usb 2-1: new high-speed USB device number 37 using xhci_hcd
[17:41] <tkamppeter> Jan 16 15:40:13 till-x1carbon kernel: usb 2-1: New USB device found, idVendor=03f0, idProduct=c211
[17:41] <tkamppeter> Jan 16 15:40:13 till-x1carbon kernel: usb 2-1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[17:41] <tkamppeter> Jan 16 15:40:13 till-x1carbon kernel: usb 2-1: Product: Deskjet 2540 series
[17:41] <tkamppeter> Jan 16 15:40:13 till-x1carbon kernel: usb 2-1: Manufacturer: HP
[17:41] <tkamppeter> Jan 16 15:40:13 till-x1carbon kernel: usb 2-1: SerialNumber: BR54BFB02C05XK
[17:41] <tkamppeter> Jan 16 15:40:13 till-x1carbon kernel: usblp 2-1:1.1: usblp1: USB Bidirectional printer dev 37 if 1 alt 0 proto 2 vid 0x03F0 pid 0xC211
[17:41] <tkamppeter> Jan 16 15:40:13 till-x1carbon mtp-probe[20729]: checking bus 2, device 37: "/sys/devices/pci0000:00/0000:00:14.0/usb2/2-1"
[17:41] <tkamppeter> Jan 16 15:40:13 till-x1carbon mtp-probe[20729]: bus: 2, device: 37 was not an MTP device
[17:41] <tkamppeter> Jan 16 15:40:18 till-x1carbon colord-sane[20731]: io/hpmud/musb.c 2101: Invalid usb_open: Permission denied
[17:41] <tkamppeter> Jan 16 15:40:18 till-x1carbon colord-sane[20731]: io/hpmud/musb.c 2101: Invalid usb_open: Permission denied
[17:41] <tkamppeter> Jan 16 15:40:18 till-x1carbon colord-sane[20731]: io/hpmud/musb.c 2101: Invalid usb_open: Permission denied
[17:41] <tkamppeter> Jan 16 15:40:18 till-x1carbon colord-sane[20731]: io/hpmud/musb.c 2101: Invalid usb_open: Permission denied
[17:41] <tkamppeter> Jan 16 15:40:18 till-x1carbon colord-sane[20731]: io/hpmud/musb.c 2101: Invalid usb_open: Permission denied
[17:41] <tkamppeter> Jan 16 15:40:18 till-x1carbon colord-sane[20731]: io/hpmud/musb.c 2101: Invalid usb_open: Permission denied
[17:43] <tkamppeter> Laney, xnox, this is the log of manually starting the service:
[17:43] <tkamppeter> Jan 16 15:42:24 till-x1carbon systemd[1]: Started Daemon to make IPP-over-USB printers available as network printers (002:037).
[17:43] <tkamppeter> Jan 16 15:42:24 till-x1carbon sudo[20876]: pam_unix(sudo:session): session closed for user root
[17:43] <tkamppeter> Jan 16 15:42:24 till-x1carbon kernel: usb 2-1: usbfs: process 20879 (ippusbxd) did not claim interface 0 before use
[17:43] <tkamppeter> Jan 16 15:42:25 till-x1carbon kernel: TCP: request_sock_TCP: Possible SYN flooding on port 60000. Sending cookies.  Check SNMP counters.
[17:43] <tkamppeter> Jan 16 15:42:26 till-x1carbon colord[1598]: failed to get session [pid 6984]: No data available
[17:43] <xnox> tkamppeter, i need the status of the device
[17:43] <xnox> unit
[17:44] <xnox> tkamppeter, cause if device unit is not "considered plugged" the wants are not started, as far as i understand....
[17:44] <tkamppeter> Laney, xnox: The manual start was successful.
[17:44] <Laney> Manually starting things isn't really interesting for this problem
[17:44] <tkamppeter> xnox, how do I get the log of the device unit?
[17:45] <xnox> tkamppeter, can you plugin the thing; and find it in $ systemctl list-units ? and check if it is there or not?
[17:46] <xnox> and then check status, if it is "loaded active plugged"?
[17:52] <tkamppeter> xnox, Laney: It contains
[17:52] <tkamppeter>   sys-devices-pci0000:00-0000:00:14.0-usb2-2\x2d1.device                                     loaded active plugged   Deskjet_2540_series
[17:52] <tkamppeter> regardless of whether the printer is plugged or not.
[17:53] <tkamppeter> So it seems that the printer is not correctly unregistered from systemd on unplug, and so it does not get considered newly plugged.
[17:53] <tkamppeter> xnox, Laney: ^^
[17:54] <tkamppeter> xnox, Laney: First step would be to somehow manually reset it and see whether the service gets started automatically when plugging right after the reset.
[17:55] <Laney> did you experiment with umockdev or something?
[17:55] <xnox> tkamppeter, reboot? =)
[17:55] <tkamppeter> xnox, Laney, next step is to find a way that the status of the device gets reset/removed on unplug automatically.
[17:56] <xnox> tkamppeter, run udevadm monitor to figure things out. but it usually monitors only things it knows about pre-plug. possibly a kernel driver bug.
[17:56] <tkamppeter> xnox, Laney: sudo systemctl daemon-reload removes the status.
[17:56] <xnox> good
[17:57] <Laney> for me it goes active / inactive corresponding to whether the printer is plugged in or not
[17:57] <tkamppeter> xnox, Laney: And plugging right after that starts the service automatically!!
[17:58] <tkamppeter> xnox, Laney: I have never "inactive".
[17:59] <Laney> suggest restarting
[18:01] <xnox> tkamppeter, maybe your printer is broken; or something else starts that holds up the device; i see other things starting e.g. colord-sane? what is up with that?
[18:02] <tkamppeter> Laney, xnox: turning off the printer has the same problem, no reset of loaded active plugged.
[18:02] <Laney> I mean restarting your computer :P
[18:02] <Laney> but now I'm off, can carry this on tomorrow if required
[18:02] <Laney> happy evening o/
[18:04] <tkamppeter> Laney, thanks, see you tomorrow.
[18:04] <seb128> Laney, have a nice evening
[18:05] <tkamppeter> xnox, do not know how colord-sane is working. It does not come fro me, but it seems to react on the presence of the printer, too.
[18:05] <tkamppeter> xnox, how can I stop that.
[18:08] <tkamppeter> xnox, in "udevadm monitor --environment" is nothing output at all during unplug of the printer. Is this correct?
[18:12] <tkamppeter> xnox, is there some way to see whether a device like sys-devices-pci0000:00-0000:00:14.0-usb2-2\x2d1.device in systemd is held up by something else?
[18:27] <xnox> tkamppeter, no idea. Laney's printer unplugs just fine....
[18:27] <xnox> most devices do too....
[18:35] <tkamppeter> xnox, I have reduced the ippusbxd problem onto the unplug problem now. If I plug the printer after a daemon-reload, it reliably starts the ippusbxd service, even without ExecStartPre delay, independent of forking or non-forking ippusbxd and also with presence of the 70-printers.udev rule.
[18:37] <tkamppeter> xnox, perhaps it has to do with the fact that it is an MF device with scanner.
[18:37] <tkamppeter> xnox, or a device supported by HPLIP.
[18:38] <tkamppeter> xnox, is there actually some way to see whether a device like sys-devices-pci0000:00-0000:00:14.0-usb2-2\x2d1.device in systemd is held up by something else?
[18:41] <xnox> tkamppeter, it's not a systemd question. it's a kernel question. try #ubuntu-kernel
[18:47] <tkamppeter> xnox, the removal of the device also does not take place when removing HPLIP, so HPLIP is also not the cause.
[19:06] <tkamppeter> xnox, sorry, I had some problems to get into the #ubuntu-kernel channel, but now I am in and asking there.