[08:02] <oSoMoN> good morning desktoppers
[08:02] <oSoMoN> and happy Friday!
[08:07] <duflu> Morning oSoMoN
[08:08] <oSoMoN> good afternoon duflu
[08:10] <jibel> tsimonq2, canary images are not published anywhere yet.
[08:10] <jibel> good morning everyone
[08:15] <duflu> Hi jibel
[08:16] <jibel> cpaelzer__, hi, I still have bug 1761497 which is pretty annoying for manual iso testing. Is there a way to listen to events like key strokes received by the guest?
[08:16] <ubot5`> bug 1761497 in virt-manager (Ubuntu) "syslinux menu is displayed when a desktop ISO is powered from virt-manager" [Low,Confirmed] https://launchpad.net/bugs/1761497
[08:18] <cpaelzer> jibel: re-reading old bug ...
[08:19] <oSoMoN> salut jibel
[08:20] <jibel> cpaelzer, it's like the VM receives a keyboard event on boot
[08:20] <cpaelzer> jibel: is still the different host HW the only difference between good&bad systems?
[08:21] <cpaelzer> jibel: also if on an affected machine you start it via virt-manager but without a view of the guest graphics open - does it work then
[08:22] <cpaelzer> i.e. does the UI hve to be attached to trigger the issue?
[08:23] <jibel> cpaelzer, it's only on one machine, the HW is the only difference between good vs bad systems.
[08:24] <jibel> cpaelzer, and yes, the UI seems to have a role in triggering the bug
[08:25] <jibel> and it happens with any release booted on the guest (i tried with trusty)
[08:25] <cpaelzer> for the sake of trying to be a friendly helper I tried to reproduce this once more now - but it doesn't trigger
[08:26] <cpaelzer> let me think about potential debugging options that I might have forgotten
[08:27] <jibel> yeah, I think it's really a problem with the host difficult/impossible to reproduce on other hw
[08:28] <jibel> cpaelzer, if I start the VM directly from the manager without the view it works as expected
[08:28] <cpaelzer> you can send-key extra keystrokes, but I don't know of monitoring them
[08:28] <cpaelzer> might be an unwanted big-brother feature after all
[08:30] <cpaelzer> jibel: for the sake of trying almost random things - switching the video backend between spice/vnc maybe?
[08:30] <cpaelzer> not sure how your guest looks like atm
[08:32] <jibel> cpaelzer, ah, actually switching to vnc solves the problem
[08:32] <cpaelzer> jibel: are by any accident the working systems already on vnc only instead of spice?
[08:33] <cpaelzer> jibel: and/or is this a sufficient workaround for you for now?
[08:36] <jibel> cpaelzer, the working systems are using spice.
[08:37] <jibel> cpaelzer, The workaround is sufficient, thanks for your help.
[08:37] <jibel> I'll see if I can dig further
[08:49] <willcooke> morning
[08:58] <duflu> Morning willcooke
[08:59] <willcooke> hi duflu
[08:59] <willcooke> re: https://bugs.launchpad.net/ubuntu/+source/gnome-shell-extension-desktop-icons/+bug/1821933
[08:59] <ubot5`> Ubuntu bug 1821933 in gnome-shell-extension-desktop-icons (Ubuntu) "Touch input is offset with two screens (even appearing on the wrong screen)" [Undecided,New]
[09:00] <willcooke> I dont have any scaling set, do you want me to set it and retest?
[09:00] <duflu> willcooke, no if you can reproduce it with 100% then that's plenty
[09:00] <willcooke> roger
[09:02] <willcooke> comment posted
[09:03]  * duflu stops typing and looks
[09:04] <duflu> willcooke, thanks. Back to being a pure gnome-shell bug. I did search for prior reports yesterday but found none
[09:04] <duflu> I can imagine the cause though
[09:04] <willcooke> jibel can reproduce, so I'm happy its not just me
[09:04] <duflu> The stage covers all monitors and possibly touch input is being scaled to that
[09:05] <duflu> So the rubber band would appear at the proportion of the full stage that's the same as the proportion across the touch screen
[09:05] <duflu> Or any touch would
[09:05] <duflu> I happened to order a tiny external monitor yesterday for such bugs but it is not touch
[09:13] <willcooke> duflu, do we need T_revinhos eyes on it?
[09:14] <duflu> willcooke, I can probably find the offending code myself but couldn't test any fix right now
[09:15] <willcooke> if you can stick it in a ppa, I'm happy to test
[09:15] <willcooke> as and why, like
[09:15] <willcooke> *when
[09:17] <duflu> willcooke, actually jibel beat you to it (bug 1723039)
[09:17] <ubot5`> bug 1723039 in gnome-shell (Ubuntu) "With 2 monitors, touches to the touch screen appear across both (and in the wrong place)" [Undecided,New] https://launchpad.net/bugs/1723039
[09:17] <duflu> which to use?
[09:18] <willcooke> I think that's different
[09:18] <willcooke> it talks about turning off one screen
[09:18] <duflu> willcooke, no he's only suggesting a solution I think
[09:19] <willcooke> "With an external monitor attached and the internal display being a touch screen, set the output to the external display only."
[09:19] <duflu> willcooke, oh, no. Same bug occurs even when the touch screen is off
[09:19] <willcooke> that's not the same thing
[09:19] <duflu> Still the same bug
[09:19] <willcooke> oki
[09:19] <willcooke> That bug is from 2017, but my machine has been find on cosmic and bionic
[09:20] <willcooke> so there's a regression there
[09:20] <willcooke> my bug has a video too, so..... ;D
[09:23] <jibel> duflu, different bugs
[09:23] <duflu> jibel, different situations yeah but the same fix should do it
[09:23] <duflu> Also there might have been a separate fix already to stop touches responding when a screen is off
[09:23] <jibel> no, my case was about disabling the touch screen when the display is off
[09:24] <duflu> Yes
[09:24] <jibel> the other one is about the offset when 2 screens are on. How is it the same thing?
[09:25] <duflu> jibel, oh, yes I see. When you say actions are visible you mean actions without a touch location.
[09:25] <jibel> yes
[09:25] <duflu> jibel, OK no problem. Sounds familiar so I think that might have been fixed in recent years
[09:25] <duflu> Now checking
[09:27] <willcooke> morning jibel
[09:29] <willcooke> oops
[09:33] <duflu> jibel, your bug is magically fixed. Apparently
[09:37] <willcooke> ah, now then....
[09:37] <willcooke> If I turn off the laptop screen, which is the touch screen
[09:37] <willcooke> and only have video out on the external monitor
[09:37] <willcooke> then touch events still work
[09:37] <willcooke> and
[09:38] <willcooke> they are positioned on the external screen, more or less in the right place
[09:38] <willcooke> like, if I touch in the middle of the (disabled) laptop screen, the event happens more or less in the middle of the external screen
[09:38] <willcooke> I assume "more or less" because the screens are different sizes and resolutions
[09:45] <duflu> willcooke, hmm. I just closed that because I found the fix in mutter 3.31.4
[09:46] <duflu> I guess it's not fully off?
[09:53] <willcooke> I guess so, the laptop lid is up but it looks like the backlight is off
[10:15] <duflu> willcooke, it took me longer to remember how to PPA than to actually write a patch
[10:15] <duflu> Give it some minutes
[10:16] <willcooke> duflu, this can wait if it's already EOW for you
[10:16] <duflu> willcooke, fair point. But wait for the build and let me know: https://launchpad.net/~vanvugt/+archive/ubuntu/lp1821933
[10:16] <willcooke> duflu, merci
[10:19] <duflu> willcooke, ugh. Actually the math is wrong and there is missing info I can't just fix :(
[10:19] <willcooke> nw
[10:19] <willcooke> Sounds like I should open a bug uptream
[10:21] <willcooke> I'll test on Fedora too
[10:22] <duflu> willcooke, yeah that can't hurt to log an upstream bug. The problematic function is the one in the patch I mentioned, but a fix would need to actually cover more code than that. If it's possible at all right now
[10:23] <duflu> willcooke, I guess my faulty patch might work only for cases where the touch screen is at the top left
[10:25] <willcooke> I'll test it anyway, see what happens
[10:27] <duflu> I think if the touchscreen and the LCD below it are the same resolution AND placed at the top left of all monitors, THEN it might work
[10:29] <duflu> So it may have the illusion of working in the end, but should not be proposed as a fix
[10:29] <willcooke> kk
[10:30] <willcooke> duflu, so for an upstream bug, do you think it should go against gnome shell?
[10:30] <duflu> willcooke, no it's purely mutter
[10:30] <willcooke> kk
[10:30] <duflu> I'd just leave a gnome-shell task in Launchpad to help others find it. That minimizes duplicates
[10:30] <duflu> Also minimises
[10:31] <willcooke> :)
[10:31] <duflu> Night
[10:31] <willcooke> see you!
[10:31] <willcooke> have a good one
[15:47] <tsimonq2> jibel: ack, but they are meant to be published at some point?
[15:47] <tsimonq2> jibel: I ask because Lubuntu would like to jump on the train, so to speak. :)
[16:00] <jibel> tsimonq2, yes, once it's installable.
[16:00] <jibel> I'll tell you
[16:02] <tsimonq2> jibel: Fair enough. Are the remaining issues in the tooling or the packages that are in the PPA?
[16:06] <jibel> tsimonq2, just lack of time
[16:07] <jibel> tsimonq2, actually it should work with ubiquity, I'll try to push a build somewhere for you next week if you want to have a look
[16:13] <tsimonq2> jibel: Oh, is this the one with the stacked squashfses?
[16:13] <tsimonq2> If that's the case, it's a two-liner in Calamares, I just need to verify and test it.
[17:53] <willcooke> night all, have a nice weekend