#linuxcnc-devel Logs

Sep 05 2022

#linuxcnc-devel Calendar

01:11 AM silopolis[m]: Hello friends o/
01:18 AM silopolis[m]: <seb_kuzminsky> "pere: interesting, thanks - i..." <- But it's still unclear to me why we need it?
01:21 AM silopolis[m]: <roguish[m]> "seb_kuzminsky: i'll figure out..." <- I would NOT recommend adding the unstable/sid repo to your sources.
01:21 AM silopolis[m]: Even with package pinning, this puts your install at risks.
01:21 AM silopolis[m]: Considering the great work seb_kuzminsky is doing, I'm confident the upload chore is going to lighten son. Then we can hope for a more regular update flow in sid and consecutively testing.
01:21 AM silopolis[m]: If you leave on the cutting edge because of your work with @cmorley I can't see a better option that building locally
01:22 AM silopolis[m]: pere: In my mind smart + useful + wanted = needed :)
01:25 AM silopolis[m]: pere: so first what would it be useful to?
01:27 AM pere: the idea is to make the packages from linux.org and debian.org compatible, so those using an old version of packages from either will upgrade seamlessly to a package from the other if it is newer. at the moment those using debian packages will upgrade to linux.org packages, even if they are older.
02:25 AM silopolis[m]: OK, thanks for clarification 👍🙏
03:14 AM pere: silopolis[m]: note, it is perfectly ok to decide that the linuxcnc.org packages should take preference over the debian packages, no matter their version number. I assumed that was the decision taken when the version schema was decided, but later talks with seb_kuzminsky make me believe this was not the intention.
03:36 AM silopolis[m]: pere: absolutely, needs the release and publishing of packages to be somehow synced in a way expected by users...
03:37 AM silopolis[m]: AFAIK, I would consider simply dropping distribution on lwo
03:38 AM pere: silopolis[m]: I got the impression that they also provide master snapshots, which debian is never going to do.
03:38 AM pere: as in every time master is updated, a new deb is generated.
03:38 AM silopolis[m]: As soon as lcnc stable is available in debian stable, there's is no point in duplicating
03:40 AM silopolis[m]: pere: ... And for master as well as those darn old debian versions it seems we can't get rid off, use buildbot
03:43 AM silopolis[m]: Speaking of which, it should also be added and prioritized for next discussions!
03:43 AM pere: any idea when the next videochat will take place?
03:44 AM pere: I suspect we should do a talk regarding weblate setup independent of the linuxcnc developer chat, as it is a lot fewer people with knowledge about the weblate setup.
03:46 AM silopolis[m]: It is to much of a chore and burden for seb_kuzminsky alone, and a risk for the project to rely on this self hosted (I believe) black box!
03:46 AM silopolis[m]: Buildbot should have it's GH repository with devops piping handy
03:46 AM silopolis[m]: pere: Nope
03:47 AM pere: silopolis[m]: well, I believe it is a very bad idea to hand more information than required to Microsoft (aka github), and very much appreciate the self hosting done by seb_kuzminsky.
03:47 AM pere: silopolis[m]: what make you believe the buildbot setup is not in git and on github already?
03:49 AM silopolis[m]: Words from andypugh in our meetings saying knowledge of how it was setup and running was known by few and sometime forgotten
03:50 AM silopolis[m]: But I may have misunderstood, again!
03:51 AM pere: the same has been said of stuff I have set up, where everything is on github already...
03:51 AM silopolis[m]: True
03:52 AM silopolis[m]: Hopefully everything is in farm-scripts directory of the infrastructure repository
03:53 AM silopolis[m]: But skimming over I suspect it's only for build slaves
05:49 AM -!- #linuxcnc-devel mode set to +vv by platinum.libera.chat
10:57 AM seb_kuzminsky: the buildbot config is in a private git, because it contains the passwords that the workers use to connect to the master
10:59 AM seb_kuzminsky: this may be an overabundance of caution on my part, since the buildbot runs on a mostly isolated local network (the only connection to the internet is via the apache that runs on the buildmaster to expose the deb archive, and the buildbot webserver itself that serves the waterfall status pages, err and the kgb instance that connects the github webhooks to the buildmaster)
11:00 AM seb_kuzminsky: linuxcnc-build: hi
11:14 AM seb_kuzminsky: linuxcnc-build: hello
11:14 AM linuxcnc-build: yes?
11:14 AM seb_kuzminsky: ah
01:35 PM -!- #linuxcnc-devel mode set to +v by ChanServ
01:37 PM hansu: Just a short question: I thought the buildbot is also building a deb package for every commit on the 2.8 branch. Maybe my assumption is wrong, because I couldn't find a recent packe on http://buildbot.linuxcnc.org/dists/buster/2.8-rtpreempt/binary-amd64/ ?
01:41 PM pere: I thought so too, but have not used them nor looked for them.
02:01 PM hansu: For master there are a lot, but not for 2.8
02:06 PM pere: perhapd the 2.8 builds failed on the buildbots?
03:17 PM seb_kuzminsky: the buildbot builds all branches
03:18 PM pere: seb_kuzminsky: hi. :)
03:19 PM seb_kuzminsky: the most recent commit on 2.8 was 92cfdf71955 on Aug 20, matching the most recent deb in the buildbot deb archive: http://buildbot.linuxcnc.org/dists/buster/2.8-rtpreempt/binary-amd64/?C=M;O=D
03:19 PM seb_kuzminsky: hi pere :-)
03:20 PM seb_kuzminsky: the buildbot prunes old debs, it only keeps the most recent month or something like that (i'd have to look up the exact criterion)
03:43 PM pere: good
03:58 PM seb_kuzminsky: pere: i've made a new dsc that fixes https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1019212, testing it now, will have an upload for you in an hour or so if this goes well
04:01 PM pere: seb_kuzminsky: if I am awake when it is done, I'll upload swiftly, if not it will be done tomorrow morning.
04:04 PM seb_kuzminsky: thank you :-)
04:04 PM seb_kuzminsky: i'll work on getting my keys in order and getting my new-maintainer stuff done so i dont have to bug you every time
04:05 PM pere: sound good. perhaps time to also visit <ULR: https://nm.debian.org/ > to become a full member? :)
04:06 PM seb_kuzminsky: will do :-)
04:34 PM pere: seb_kuzminsky: note, you might want to include 9e2785801ea78d7745dc5677c4b29ab1e54c0c74 in the next upload, yet another try to get the debian/tests/ script to work.
04:34 PM pere: just pushed to master.
04:40 PM seb_kuzminsky: hrm, i update the gbp repo already and am halfway through testing, how about next time
04:42 PM pere: no problem.
04:42 PM pere: it will reduce the migration time from 5 to 2 days, if successful.
04:58 PM seb_kuzminsky: did you run it locally? does it fix the problem?
05:05 PM pere: the tests have always worked locally, I just have failed to get them to work in the automatic test environment in Debian so far.
05:12 PM seb_kuzminsky: running `autopkgtest` locally works? but it fails on debian's machines?
05:14 PM pere: running the test scripts locally.
05:16 PM seb_kuzminsky: ah, ok, yeah that's well tested
06:32 PM seb_kuzminsky: pere: if you're still awake, http://highlab.com/~seb/linuxcnc/debian.org/2.9.0~pre0+git20220905.b2e28a88c3-1/
06:44 PM seb_kuzminsky: pere: oops, i forgot to mark it as released-to-unstable in d/changelog, files updated, should be good to go now
10:18 PM andypugh_ is now known as andypugh
11:20 PM pere: seb_kuzminsky: unable to connect to highlab.com port 443.