#linuxcnc-devel Logs

Jan 02 2023

06:09 AM pere: hi
08:25 AM linuxcnc-build2: Build [#77](http://buildbot2.highlab.com/buildbot/#builders/7/builds/77) of `10-rip.debian-11-bullseye-rtpreempt-amd64` 4failed.
08:25 AM linuxcnc-build2: Build [#80](http://buildbot2.highlab.com/buildbot/#builders/11/builds/80) of `10-rip.debian-12-bookworm-rtpreempt-amd64` 4failed.
08:26 AM linuxcnc-build2: Build [#76](http://buildbot2.highlab.com/buildbot/#builders/14/builds/76) of `10-rip.debian-10-buster-rtpreempt-amd64` 4failed.
08:26 AM linuxcnc-build2: Build [#87](http://buildbot2.highlab.com/buildbot/#builders/13/builds/87) of `00-checkin` 4failed.
08:39 AM linuxcnc-build2: Build [#77](http://buildbot2.highlab.com/buildbot/#builders/14/builds/77) of `10-rip.debian-10-buster-rtpreempt-amd64` 8completed with warnings.
08:40 AM linuxcnc-build2: Build [#78](http://buildbot2.highlab.com/buildbot/#builders/7/builds/78) of `10-rip.debian-11-bullseye-rtpreempt-amd64` 8completed with warnings.
08:44 AM linuxcnc-build2: Build [#81](http://buildbot2.highlab.com/buildbot/#builders/11/builds/81) of `10-rip.debian-12-bookworm-rtpreempt-amd64` 8completed with warnings.
08:56 AM linuxcnc-build2: Build [#88](http://buildbot2.highlab.com/buildbot/#builders/13/builds/88) of `00-checkin` 3completed successfully.
10:58 AM seb_kuzminsky: happy new year y'all! 2023 is going to be great for linuxcnc :-)
10:59 AM roguish[m]: seb_kuzminsky: happy new year to you. your buldbot2 seems to be working pretty good. I'm running one of its bullseye debs......
11:02 AM seb_kuzminsky: me too
11:03 AM seb_kuzminsky: i'm hoping to turn off the old buildbot and replace it with the new one soon
11:05 AM roguish[m]: seb_kuzminsky: quick question. hope you can give a bit of guidance. QTvcp (cmorley's thing) does not get installed with any debs. gotta run an install script that's buried pretty deep and nondocumented. Waz up with this? It is in all the docs and has lots of examples in sim, so shouldn't it get installed? I don't want you to take that on yourself. Morley should.... yes, no, ????
11:05 AM seb_kuzminsky: hmm, i was not aware of that
11:06 AM seb_kuzminsky: yeah, if we have stuff in our repo it should get built and installed along with everything else as part of the debs
11:06 AM seb_kuzminsky: cmorley is super sharp and competent, i bet he's got a good reason for why it's not like that already, but i don't know what it is
11:06 AM roguish[m]: yes, it is in the github repo
11:07 AM seb_kuzminsky: i've seen the qtvcp code in our repo (i used some icons from it for a pywebio UI i was hacking on the other day)
11:09 AM seb_kuzminsky: hmm, i've got the 1:2.9.0~pre1.160.g3ccda74cc deb installed from buildbot2, and i have /usr/bin/qtvcp and /usr/lib/python3/dist-packages/qtvcp
11:09 AM seb_kuzminsky: what install script do you have to run?
11:11 AM roguish[m]: copy and pasted from the forum....: /usr/lib/python3/dist-packages/qtvcp/designer/install_script
11:11 AM seb_kuzminsky: the sim/axis/qtvcp/qtvcp_vismach config seems to run fine here, i get a vismach tab in axis
11:11 AM roguish[m]: https://forum.linuxcnc.org/qtvcp/38414-installing-qtvcp
11:13 AM seb_kuzminsky: that script seems to install qtvcp runtime dependencies (which the deb also does), and install the qt designer development environment (not sure what that does)
11:13 AM seb_kuzminsky: so for a user (as opposed to a developer), i think it's not required to run that
11:13 AM roguish[m]: the qt designer let's one customize or create a gui via qtvcp.
11:14 AM seb_kuzminsky: having a developer's helper script in /usr/lib/python3 is probably not right, it should be in /usr/share/doc or something
11:14 AM seb_kuzminsky: right, so that script is a developer's helper
11:14 AM seb_kuzminsky: it just looks slightly misfiled to me
11:15 AM roguish[m]: I installed bullseye, then one of your debs for 2.10, and could not get qtdragon (one of the pre-made guis) to run. had to run that install script. check this: https://forum.linuxcnc.org/qtvcp/47756-up-to-dateness
11:17 AM roguish[m]: principally what I proposed there is that if it's in the repo and docs, it should run 'out of the box' and not have to do a bunch of messing around.
11:18 AM seb_kuzminsky: roguish[m]: i totally agree with the "run out of the box" thing, it looks to me like you've found a packaging bug
11:19 AM roguish[m]: seb_kuzminsky: I think this should be cmorley's chore. he is typically very nice and helpful, but lately lagging on updates caused by changes in linux, qt, and other libraries.
11:20 AM roguish[m]: any encouragement that direction would be super helpful, for all.
11:22 AM seb_kuzminsky: it might help if you opened an issue on github for this, at least that's where i prefer to keep track of bugs
11:23 AM roguish[m]: seb_kuzminsky: ok, i'll try to. thanks for the ear....
11:28 AM seb_kuzminsky: we have automated tests that verify that some parts of linuxcnc run in a fresh, minimal install, but unfortunately none of the GUIs are covered by that testing
11:29 AM roguish[m]: I have seen Pere chat about that.
11:29 AM seb_kuzminsky: yeah, he proposed a system for testing the guis, but it's pretty fiddly to set up, a lot of infrastructure needed
11:29 AM roguish[m]: maybe there's a way to at a minimum, they start up without error.....
11:30 AM seb_kuzminsky: yeah, that would be super great
11:30 AM seb_kuzminsky: much easier than one of us making a new netinst VM and installing our deb and testing it by hand, which is what i'm doing right now...
11:31 AM roguish[m]: running could be difficult... like qtvcp is really intended for touch screens....
11:31 AM seb_kuzminsky: but it still works with a regular screen and a mouse, i would think?
11:32 AM roguish[m]: yes
11:32 AM seb_kuzminsky: cool
12:10 PM seb_kuzminsky: roguish[m]: i'm not seeing the problem you're describing
12:10 PM seb_kuzminsky: http://highlab.com/~seb/linuxcnc/qtvcp-on-bullseye.png
12:11 PM seb_kuzminsky: that's a fresh, minimal install of bullseye, then i installed linuxcnc-uspace from buildbot2, and started the sim/axis/qtvcp/qtvcp_tab.ini config, and it started
12:21 PM roguish[m]: sure didn't work for me.....
12:25 PM seb_kuzminsky: can you elaborate?
12:30 PM roguish[m]: I made a debian 11 install from a download from debian, added a few repositories so I could get a 4.xx amd-rt kernel, and installed 4.19 rt. then installed linuxcnc from your buildbot2 for bullseye. the latest on at the time (about a week ago). added my config for qtdragon and removed from it any of my customizations, and it did not run.
12:35 PM roguish[m]: the qtdragon (and I think all qtvcp) requires a 'handler' python file. that's what I cleaned up. I did not mess with the as-installed widget python files, or hal_glib.py, which I had to modify before to get the customizations I wanted.
12:36 PM roguish[m]: to get it to run, I have to run that installer script. it's pretty long, and looks like it does a lot
12:41 PM seb_kuzminsky: ah, so this is a config that's not in our git repo? can you send it to me?
12:44 PM roguish[m]: My config is specific to my mill. uses a Mesa 7i80hd, has a vfd for the spindle via modbus comp. I can send it if you really want to check it out. it's a bunch of files.... the .ini, .hal's, and several required for the qt gui.
12:44 PM roguish[m]: but I can zip 'em all up....
12:56 PM seb_kuzminsky: a sim version would be easier for me to test
01:02 PM seb_kuzminsky: i wonder if the included qtdragon sim config runs...
01:03 PM roguish[m]: seb_kuzminsky: I have a fair bit of time. and a legit copy of Vmware Workstation.
01:03 PM seb_kuzminsky: it does not, but not because of any dependency problem
01:05 PM roguish[m]: I have a vm of bullseye with 2.10 installed. new as of a few days ago. but honestly, I don't recall if I ran that install script.
01:05 PM seb_kuzminsky: so the sim/qtvcp_screens/qtdragon/qtdragon.ini config starts fine too on a minimal machine without that script ran
01:05 PM roguish[m]: ya know how you can use the tab key in a command terminal window to scroll back thru commands? where is that history list stored?
01:06 PM roguish[m]: if ya happen to know
01:07 PM seb_kuzminsky: ~/.bash_history
01:07 PM seb_kuzminsky: (assuming your shell is bash, otherwise i do not know)
01:08 PM roguish[m]: that's it. and i did run the install script.....
01:10 PM roguish[m]: I'll create another D11 + 2.10 and see what happens.
01:10 PM roguish[m]: vm
01:11 PM roguish[m]: I have an older Dell dual zeon cad machine...... lots of memory and disk....
01:24 PM roguish[m]: and there's a debian mirror right over the hill at UC Berkeley
01:35 PM roguish[m]: seb_kuzminsky: would you happen to have a decent installation preconfiguration file? I usually install with the graphical installed, but it's a drag. i would like to try an 'automatic' install....
01:37 PM seb_kuzminsky: no, i install my VMs by hand, and yeah it's a bit of a pain
01:37 PM roguish[m]: ok. i'll ask pere next time he's around... i'm creating a new vm now.
01:39 PM roguish[m]: oh, is it possible to install software with a script, rather than synaptic? must be....
01:41 PM pere: ?
01:41 PM roguish[m]: I have a fair number of programs that I like to use, and synaptic is a pain....
01:41 PM roguish[m]: hey Pere...
01:42 PM seb_kuzminsky: roguish[m]: yeah, `apt-get install`
01:42 PM pere: or the newer apt install
01:42 PM roguish[m]: how do ya do that in a script? doesn't is need sudo and all....
01:43 PM pere: installing system packages require root access, yes. sound like you want to talk to pkgmanager.
01:45 PM roguish[m]: pere: I don't want to deal with complete packages, just little things like grub-customizer, and mozo, and search-monkey, and samba, ......
01:45 PM roguish[m]: with a script, rather than one by one in synaptic, or my bad typing....
01:46 PM pere: roguish[m]: perhaps <URL: https://www.linux.com/training-tutorials/streamlining-software-installation-packagekit/ > can help you? do you have pkcon installed?
01:47 PM pere: pkcon install packagename might work.
01:47 PM roguish[m]: i'll check it out...thanks.
01:47 PM pere: I always use apt and sudo/run as root, so I am not very familiar with the details of pkcon.
01:51 PM seb_kuzminsky: pere: i've been playing around a bit with the "run linuxcnc guis displaying to vnc" thing you suggested and it seems promising
01:52 PM seb_kuzminsky: i'm running in linuxcnc in docker and running the vnc viewer on the host
01:52 PM seb_kuzminsky: so far i'm not able to reproduce the qtdragon error roguish[m] is describing
01:52 PM seb_kuzminsky: this is with bullseye in the container
01:54 PM pere: did you also check out my gui test sketch using xnee?
01:54 PM seb_kuzminsky: no, got a link?
01:57 PM pere: sure, <URL: https://github.com/petterreinholdtsen/linuxcnc/tree/test-gui >
01:58 PM seb_kuzminsky: cool
01:58 PM seb_kuzminsky: back in a bit
01:59 PM pere: xnee seem to depend a lot of specific screen coordinates, which I am not too happy with. there is also vlcdo, <URL: https://pypi.org/project/vncdotool/ >, which can be more position agnostic, but harder to time.
02:00 PM pere: sadly no-one packaged vncdotool for debian yet.
02:42 PM pere: I just upgraded my bookworm install and AXIS refuse to starte over some OpenGL issue. Got Gmoccapy limping along.
03:28 PM pere: <URL: https://bugs.debian.org/1027756 > is the state of vncdo in debian.
03:34 PM seb_kuzminsky: i wish that if you ran `linuxcnc foo.ini`, then hit ^C, the linuxcnc runner would clean up the linuxcnc session
03:35 PM seb_kuzminsky: it currently just dies and leaves all the linuxcnc stuff still running
03:35 PM seb_kuzminsky: as is, i don't know of a way to reliably kill a linuxcnc session
04:31 PM roguish[m]: seb_kuzminsky: hey. Ok, created a new D11 install vm. added an amd-rt kernel. installed linuxcnc from one of your bullseye debs (doc, debug, uspace, and dev) and even restarted. started linuxcnc from the menu, selected sim, qtdragon, and FAILED. here is the Report File.
04:31 PM * roguish[m] posted a file: (6KiB) < https://libera.ems.host/_matrix/media/v3/download/matrix.org/eqTYgRvGgwpEvYkJUmdwvNQI/qtdragon%20sim%20fail.txt >
04:44 PM seb_kuzminsky: roguish[m]: that's helpful, thank you
04:44 PM pere: seb_kuzminsky: why is the trap in linuxcnc.in enough to clean up the stuff left behind?
04:44 PM seb_kuzminsky: looks like you don't have the Recommended packages installed
04:44 PM pere: s/is/is not/
04:45 PM seb_kuzminsky: pere: it's some interaction with `su`, not sure of the details yet, still debugging it
04:45 PM linuxcnc-build2: Build [#78](http://buildbot2.highlab.com/buildbot/#builders/14/builds/78) of `10-rip.debian-10-buster-rtpreempt-amd64` 4failed.
04:46 PM seb_kuzminsky: roguish[m]: how did you install linuxcnc-uspace, with synaptic?
04:46 PM linuxcnc-build2: Build [#79](http://buildbot2.highlab.com/buildbot/#builders/7/builds/79) of `10-rip.debian-11-bullseye-rtpreempt-amd64` 4failed.
04:46 PM roguish[m]: I use gdebi
04:47 PM seb_kuzminsky: maybe we should promote our Recommends to Depends...
04:47 PM seb_kuzminsky: or maybe you can convince gdebi to install Recommended packages by default (and then we teach all our users how to do that...)
04:47 PM pere: seb_kuzminsky: what about having a way to tell linuxcnc to exit, for example a hal command or kill -USR2 or a dbus message or something.
04:48 PM pere: what is the rationale behind which packages are depends and recommends for linuxcnc?
04:49 PM seb_kuzminsky: at one point we got scared that "apt install linuxcnc-uspace" was getting too big, because we have N different GUIs each using a different toolkit, so we made "some" of the GUIs dependencies Recommends, on the theory that they'd be installed by default but people could removed the ones they didn't need
04:49 PM seb_kuzminsky: obviously the correct answer it to make each GUI its own deb, with its own Dependencies, but that sounds like effort
04:49 PM pere: aha
04:50 PM Tom_L: what would the default startup one be then?
04:50 PM pere: same as today?
04:51 PM seb_kuzminsky: Tom_L: yeah, it's tricky
04:51 PM Tom_L: or treat it like a desktop for debian ie xfce mate etc
04:51 PM Tom_L: pick it on the install
04:51 PM seb_kuzminsky: because all the configs depend on some gui, and if linuxcnc-core.deb only Reommends linuxcnc-gui-{axis,gmoccapy,...}.deb, then the same problem roguish[m] is having would mean *no* guis get installed
04:52 PM pere: it could depend on gui1 | gui2 | gui3, this way at least one of them would be installed.
04:52 PM seb_kuzminsky: maybe there's one blessed GUI (axis?) in linuxcnc-core.deb, so at least *something* works on all platforms?
04:52 PM seb_kuzminsky: pere: yeah, that's good!
04:53 PM seb_kuzminsky: or it depends on linuxcnc-gui, and linuxcnc-gui-* Provides linuxcnc-gui?
04:53 PM roguish[m]: axis has always worked. out of the box. always
04:53 PM seb_kuzminsky: so we don't have to keep updating the Depends list as we add and remove guis
04:54 PM seb_kuzminsky: roguish[m]: yeah, that one is well tested and widely used, and hasn't seen much change in a long time
04:54 PM Tom_L: axis seems to be the 'goto'
04:54 PM roguish[m]: exactly
04:55 PM Tom_L: i don't see why ppl stray from it honelstly. it just works TM.
04:55 PM seb_kuzminsky: and i guess we would move all the configs for gui "X" from linuxcnc-core.deb to linuxcnc-gui-x.deb
04:55 PM seb_kuzminsky: Tom_L: same! :-)
04:55 PM pere: seb_kuzminsky: yeah, a provides would work too, but should list the prefered one as well, ie depends: prefered-gui | linuxcnc-gui
04:55 PM seb_kuzminsky: i've been told it's a gui only a programmer could love
04:55 PM pere: new provides fields should be discussed on debian-devel@.
04:55 PM seb_kuzminsky: pere: aha, good to know
04:55 PM Tom_L: i can see the desire for touch screen
04:55 PM seb_kuzminsky: eyah, axis is no good on a touchscreen
04:56 PM Tom_L: honestly, i've never used one
04:56 PM seb_kuzminsky: touchy is, though
04:56 PM pere: seb_kuzminsky: do you experience the opengl issue in bookworm?
04:56 PM seb_kuzminsky: i don't run bookworm, so no
04:56 PM seb_kuzminsky: i'm no bullseye still, like an old fogey
04:56 PM Tom_L: pere has me hesitant to update bookworm now
04:56 PM pere: the mazak machine did not work after an upgrade today. did not debug, just saw it was broken. I guess some library...
04:57 PM Tom_L: i was doing steady updates for a while
04:57 PM seb_kuzminsky: pere: :-(
04:57 PM seb_kuzminsky: pere: are you running axis on the mazak still?
04:57 PM pere: got gmoccapy limping along, while axis failed to start.
04:57 PM seb_kuzminsky: boo
04:57 PM pere: seb_kuzminsky: yes, still axis. will look at gui options once the hw is working.
04:58 PM seb_kuzminsky: eh, axis is fine, like Tom says it's all you need
04:58 PM pere: gmoccapy also had opengl problems, but only show some error popups and continued, while axis failed to start.
04:58 PM Tom_L: pere, write your own in qtpyvcp
04:58 PM Tom_L: jt did
04:59 PM seb_kuzminsky: lol
04:59 PM Tom_L: seems you have _plenty_ of time :)
04:59 PM Tom_L: i can't use it as i'm still on wheezy on the mill
05:00 PM pere: Tom_L: If I had to make a nice gui to survive, I would die. I am esthetically challenged.
05:00 PM Tom_L: i think probe basic is nice but maybe 'too much'
05:00 PM roguish[m]: now see qtpyvcp is more up-to-date. turboss works on it a lot, and he is a gaming programmer.....
05:01 PM CaptHindsight[m]: https://forum.linuxcnc.org/21-axis/40659-axis-2-touchscreen-version-of-old-friend also this
05:01 PM Tom_L: wondered what he did
05:01 PM linuxcnc-build2: Build [#89](http://buildbot2.highlab.com/buildbot/#builders/13/builds/89) of `00-checkin` 4failed.
05:02 PM Tom_L: never seen axis2
05:02 PM pere: seb_kuzminsky: are you aware of the upcoming package freeze in debian?
05:03 PM seb_kuzminsky: vaguely...
05:03 PM pere: seb_kuzminsky: <URL: https://release.debian.org/ >
05:04 PM CaptHindsight[m]: somebody in the LCNC channel was just asking about gmocappy on Deb12 or 13, they had to build gtk4 since it required gtk2 from before 2011, let me find the info
05:04 PM roguish[m]: jeeze seems like I stirred the pot a bit
05:05 PM seb_kuzminsky: pere: soft freeze feb 12
05:06 PM seb_kuzminsky: i guess we should update our package before then
05:06 PM CaptHindsight[m]: " Need to recompile glade to get version >3.8 without deprecated widgets." to get gmoccapy on deb13
05:06 PM pere: as we got working autopkgtest scripts, we can even be lucky after that.
05:07 PM pere: seb_kuzminsky: we should update the package well ahead of the freeze, to have time to handle any problems.
05:07 PM pere: for example like the opengl issue I saw.
05:08 PM pere: seb_kuzminsky: did you see my mqtt hal component?
05:08 PM seb_kuzminsky: pere: su --pty makes it pass SIGINT (^C) to the child process
05:08 PM seb_kuzminsky: touchy (for example) handles SIGINT by exiting cleaning, which makes the linuxcnc runner kill the rest of the stuff right
05:09 PM seb_kuzminsky: axis (for example) ignores SIGINT and things hang until manual cleanup
05:09 PM pere: sound like axis is broken.
05:09 PM seb_kuzminsky: yep, for some value of "broken"
05:11 PM pere: I tried to write a script testing all configs/, but could not find a way around the unability to stop linuxcnc in a predictable way.
05:11 PM pere: inability?
05:11 PM seb_kuzminsky: inability
05:12 PM seb_kuzminsky: i'm fetching up against the same issue
05:12 PM seb_kuzminsky: my current plan is to run each config in a separate container, and destroy the container at the end of the test :-/
05:12 PM seb_kuzminsky: and then fix each one that doesn't clean up correctly (or at least file a bug)
05:13 PM -!- #linuxcnc-devel mode set to +v by ChanServ
05:13 PM -!- #linuxcnc-devel mode set to +v by ChanServ
05:14 PM pere: seb_kuzminsky: I also tried to reproduce and debug the arm classicladder issue, but could not trigger the crash. seem to only trigger some race condition where parts of linuxcnc took too long to start.
05:14 PM pere: tried running linuxcnc within valgrind, but could not get any crash, just timeouts.
05:15 PM -!- #linuxcnc-devel mode set to +v by ChanServ
05:17 PM -!- #linuxcnc-devel mode set to +v by ChanServ
05:17 PM pere: seb_kuzminsky: <URL: https://www.openhub.net/p/emc > show linuxcnc doing quite well the last year. :)
05:18 PM seb_kuzminsky: yeah, looking good!
05:18 PM seb_kuzminsky: all those translations really boost our stats :-)
05:19 PM pere: yeah, but number of contributors are all time high. :)
05:26 PM pere: <URL: https://devhumor.com/media/monkeyuser-test-drive-development > :)
05:36 PM linuxcnc-build2: Build [#79](http://buildbot2.highlab.com/buildbot/#builders/14/builds/79) of `10-rip.debian-10-buster-rtpreempt-amd64` 8completed with warnings.
05:37 PM linuxcnc-build2: Build [#80](http://buildbot2.highlab.com/buildbot/#builders/7/builds/80) of `10-rip.debian-11-bullseye-rtpreempt-amd64` 8completed with warnings.
05:39 PM pere: <URL: https://www.turnoff.us/geek/tdd-vs-ftt/ > :)
05:41 PM seb_kuzminsky: it sure is nicer finding bugs before your users do
05:47 PM pere: would be nice to have a axis test in debian/tests/
06:04 PM linuxcnc-build2: Build [#90](http://buildbot2.highlab.com/buildbot/#builders/13/builds/90) of `00-checkin` 3completed successfully.
06:19 PM roguish[m]: seb_kuzminsky: well, gentlemen. I think I just found one of my main problems now. qtvcp requires a python handler file. That file ties the gui and linuxcnc together. my modified gui and handler file are from a year ago. I have got the gui working, only checking the minor changes that I made in the handler py file. well, cmorley has made lots of update changes in the supplied sim handler files......including a fix for
06:19 PM roguish[m]: units. just did a diff on the files. big diff found. I hope if fixes my units problem in the tools and probing.......
06:21 PM roguish[m]: seb, you probably recall it took years to get axis working really reliably. I fear the same for qtvcp. it's way more complicated than axis.
06:37 PM roguish[m]: well, just tried my slightly modified new version on the stock handler file. no luck. there is sufficient differences in the old gui that I modified and the new stock gui to throw it all off........ guess I'll have to modify a new stock gui file..... this is the difficulty developing a new, fancy gui system. I recall cradek going through this with axis....... he perservered.
06:42 PM pere: hm, interesting. just tested ./scripts/runtests tests/classicladder/ on arm64, and it worked.
07:45 PM -!- #linuxcnc-devel mode set to +v by ChanServ
09:17 PM seb_kuzminsky: pere: i don't know if i should be happy or sad about that
10:27 PM seb_kuzminsky: the sim/axis/lathe-fanucy/lathe-fanucy.ini config fails to start
11:00 PM pere: seb_kuzminsky: most likely sad. can you reproduce it still?
11:29 PM pere: seb_kuzminsky: I tested on amdahl.debian.org, which is not a rpi4.
11:42 PM seb_kuzminsky: several other configs fail to start:
11:42 PM seb_kuzminsky: axis/profile_axis.ini
11:42 PM seb_kuzminsky: axis/remap/getting-started/demo.ini
11:42 PM seb_kuzminsky: axis/remap/manual-toolchange-with-tool-length-switch/manualtoolchange.ini
11:42 PM seb_kuzminsky: axis/remap/rack-toolchange/racktoolchange.ini
11:42 PM seb_kuzminsky: craftsman/craftsmancnc.ini
11:42 PM seb_kuzminsky: gscreen/gscreen_custom/9axis.ini
11:42 PM seb_kuzminsky: gscreen/gscreen_custom/gaxis_no_plot.ini
11:43 PM pere: seb_kuzminsky: similar to the failures in <URL: https://github.com/LinuxCNC/linuxcnc/pull/2236 >? Are you talking about master on any arch, or just rpi4?
11:43 PM seb_kuzminsky: gscreen/gscreen_custom/gscreen_gaxis.ini
11:43 PM seb_kuzminsky: gscreen/gscreen_custom/industrial.ini
11:43 PM seb_kuzminsky: gscreen/gscreen_custom/tester.ini
11:43 PM seb_kuzminsky: gscreen/gscreen_custom/tester_metric.ini
11:43 PM seb_kuzminsky: gscreen/industrial_lathe_wear/industrial_lathe_wear.ini
11:43 PM seb_kuzminsky: gscreen/silverdragon/silverdragon.ini
11:43 PM seb_kuzminsky: gscreen/spartan/spartan.ini
11:43 PM seb_kuzminsky: qtvcp_screens/qtaxis_mpg.ini
11:43 PM seb_kuzminsky: qtvcp_screens/qtdragon/qtdragon_xyyz.ini
11:43 PM seb_kuzminsky: qtvcp_screens/qtvcp_custom/qtdefault_ngc.ini
11:44 PM seb_kuzminsky: qtvcp_screens/qtvcp_experimental/qtdefault_ngc.ini
11:44 PM seb_kuzminsky: qtvcp_screens/woodpecker/woodpecker_xyzya.ini
11:47 PM seb_kuzminsky: pere: no, not like failure
11:48 PM seb_kuzminsky: these failures are all with 2.9 on amd64
11:58 PM seb_kuzminsky: https://github.com/SebKuzminsky/linuxcnc-config-test
11:59 PM pere: seb_kuzminsky: is this the same bookworm opengl issue I saw yesterday?
11:59 PM seb_kuzminsky: these failures are mostly different, and none of them that i've looked at have anything to do with opengl