#linuxcnc-devel Logs

Jan 29 2023

11:33 AM linuxcnc-build: build #3258 of 1640.rip-buster-rtpreempt-amd64 is complete: Failure [4failed compile runtests] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1640.rip-buster-rtpreempt-amd64/builds/3258 blamelist: Sebastian Kuzminsky <seb@highlab.com>
12:02 PM linuxcnc-build: build #9667 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/9667 blamelist: Sebastian Kuzminsky <seb@highlab.com>
12:07 PM seb_kuzminsky: oops :-(
12:07 PM Tom_L: still running the old buildbot?
12:08 PM seb_kuzminsky: yes :-(
12:09 PM pere: hi
12:09 PM pere: seb_kuzminsky: what is the status of the opengl stuff now? I still have to edit out the glBitmap calls to get my machine running. :/
12:10 PM seb_kuzminsky: pere: linuxcnc's opengl is all broken in bookworm and sid
12:10 PM pere: I figured out the spindle rotary encoder protocol. It is a incremental ABZ encoder.
12:10 PM pere: Started looking for a replacement VFD in case I am unable to get the old one working.
12:11 PM seb_kuzminsky: i've got a PR open that works around part of the problem, and jepler's got one that fixes part of the underlying problem (he's the opengl wizard)
12:11 PM pere: yours is the =x11 env?
12:11 PM seb_kuzminsky: yeah
12:12 PM seb_kuzminsky: that's pyopengl-speak for "use GLX, not EGL"
12:12 PM pere: did not change anything for me, but I doubt xfce uses wayland.
12:12 PM seb_kuzminsky: yeah, if you're not on wayland you're already using pyopengl glx
12:12 PM seb_kuzminsky: try jepler's PR
12:12 PM seb_kuzminsky: see if that helps any
12:12 PM seb_kuzminsky: now, what did i break this time...
12:13 PM pere: but <URL: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1029936 > is a fatal problem for me. have to downgrade python3-opengl.
12:13 PM seb_kuzminsky: oh, the buildbot breakage isn't my fault, that's the broken mb2hal test
12:14 PM seb_kuzminsky: ... someone should fix that
12:16 PM pere: did the github ci tests not discover the mb2hal test in time?
12:17 PM seb_kuzminsky: maybe that, or maybe it's intermittent, idk
12:18 PM seb_kuzminsky: Tom_L: we're running both the old and the new buildbots in parallel, but the new one only has a very few workers/platforms in it
12:19 PM seb_kuzminsky: one of these weekends i'll move all the workers from the old to the new buildbot
12:19 PM seb_kuzminsky: and decommission the old one, and put the new one in its place
12:20 PM pere: seb_kuzminsky: do you want a buildbot master outside your basement?
12:20 PM seb_kuzminsky: pere: i think jeff's PR will fix the bug you reported
12:20 PM pere: building as we speek to test. :)
12:21 PM pere: seb_kuzminsky: I doubt it will fix the bug, which is in python3-opengl, but at least linuxcnc will not be bit by it. I notice the glBitmap calls are not touched by jepler.
12:21 PM seb_kuzminsky: pere: eh, that's not important to me. i doubt anyone else would work on it, so i'd rather not spend energy refactoring the buildbot to run the master elsewhere
12:22 PM pere: hard to predict, the future is. if no-one else can work on it, I am sure no-one else will work on it.
12:23 PM seb_kuzminsky: pere: you're right, it won't fix the bug in pyopengl. try opening an issue in github.com/mcfletch/pyopengl, maybe @swt2c will come and fix that one too
12:23 PM seb_kuzminsky: pere: fair enough. how about this: let's hold off on that for now, and after the current buildbot transition settles we can talk about it
12:25 PM pere: hm, build failure in configure:
12:25 PM pere: checking for GTK 2.4.0 or above... no
12:25 PM pere: configure: error: GTK2 missing. Install with "sudo apt install libgtk2.0-dev" or specify --disable-gtk2 to skip building classicladder
12:25 PM seb_kuzminsky: is that jeff's branch? what platform are you building on?
12:25 PM pere: seb_kuzminsky: I have not done anything to prepare the VM we could get at nuug.no, been waiting to see if there was any interest before starting any work.
12:26 PM pere: this is jeffs branch rebased on 2.9, and building on Bookworm.
12:26 PM pere: I get this sequence:
12:26 PM pere: checking for glib... yes - 2.74.5
12:26 PM pere: checking for GTK 3.22.4 or above... yes - 3.24.36
12:26 PM pere: checking for GTK 2.4.0 or above... no
12:26 PM pere: Why is it looking for both 2 and 3?
12:27 PM seb_kuzminsky: probably some of our programs need 2 (classicladder?) and some need 3 (the guis that rene et all updated)
12:28 PM seb_kuzminsky: try `debian/configure; dpkg-checkbuilddeps`
12:28 PM pere: I did run apt build-den . to get the build dependencies. look like gtk2 failed to install.
12:28 PM seb_kuzminsky: but jeff's branch didn't introduce any new build deps, so mabye something else is wrong
12:29 PM pere: libgtk2.0-dev is listed as a build depend in .top.in, but not in d/control. configured using no-docs, perhaps it pick the wrong set?
12:29 PM pere: or perhaps I ran configure before rebasing? not sure. Reran it and now it showed up.
12:29 PM seb_kuzminsky: remember you have to run debian/configure to update debian/control
12:30 PM pere: I thought I did, but might have been using the wrong time to do so.
12:30 PM pere: second try. :)
02:32 PM pere: the patch from jepler did not fix my runtime problem. still fail on glBitmap().
06:50 PM linuxcnc-build2: Build [#319](http://buildbot2.highlab.com/buildbot/#builders/13/builds/319) of `00-checkin` 4failed.
08:24 PM linuxcnc-build2: Build [#321](http://buildbot2.highlab.com/buildbot/#builders/13/builds/321) of `00-checkin` 3completed successfully.
08:26 PM linuxcnc-build: build #9670 of 0000.checkin is complete: Failure [4failed fetch branch to local git repo] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/9670 blamelist: Sebastian Kuzminsky <seb@highlab.com>