#linuxcnc-devel Logs

Jul 25 2022

#linuxcnc-devel Calendar

02:03 AM rmu: most PRs seem pretty benign, except 581 from 2019
02:17 AM pere: rene-dev5: are you the one announcing the next video meeting on linuxcnc?
02:22 AM rs[m]: video meetings?
02:23 AM pere: rs[m]: we have had a couple to coordinate the po4a migration. I hope we can continue with those, and believe rene-dev5 might be the one that volunteered to call in the next one.
02:23 AM rmu: just FYI rs[m] and /me are the same, rs[m] is the matrix irc gateway. should probably fix the nick names.
02:24 AM pere: so sadly it is only the two of us. :(
02:24 AM rmu: where are those announced?
02:45 AM pere: not sure. I believe I either saw it here or in private email, or perhaps both. I am not subscribed to the list, so no idea if it was there too.
02:51 AM rmu: found it. there was a "doodle" on the list.
02:51 AM rmu: meeting minutes https://docs.google.com/document/d/1aWMxBY8IbCYXFLFvjgqUXZ09inZp5u0r-pZcjeb2bWk/edit#heading=h.n2ljcgb3niwg
02:52 AM pere: rmu: so, what is your role in the project?
02:52 AM rmu: user
02:53 AM pere: I guess there is a lack of active developers in the project. the ones that used to be active seem to be occupied elsewhere these days.
02:53 AM rmu: tried a concept of jerk limited TP modification
02:57 AM rmu: ran into a bunch of strange corner cases. https://github.com/rmu75/linuxcnc/tree/rs_6otp
03:00 AM pere: look like it need to re rebased on the current master.
03:02 AM rmu: yes that branch is stale. need to revisit it sometime
04:23 AM silopolis[m]: Yo!
04:24 AM silopolis[m]: <rmu> "ran into a bunch of strange..." <- I think we talked about those unknown branches during last meeting!
04:26 AM silopolis[m]: Could you link it to an issue with description of your work and the issues you're facing
04:27 AM silopolis[m]: Or maybe rebase and make a draft PR?
04:30 AM silopolis[m]: <pere> "not sure. I believe I either..." <- It was announced on devel ML and I relayed it at least two times here and on #qtpyvcp which brought @TurBoss to the party
04:30 AM rmu: silopolis[m]: it is too early for a PR. discussion is in the forum https://forum.linuxcnc.org/38-general-linuxcnc-questions/34666-c2-smooth-velocity-profile
04:32 AM rmu: silopolis[m]: also this whole approach is kind of a "hack"
04:33 AM rmu: silopolis[m]: so i decided i need more visibility into the TP and then got stuck/distracted there...
05:22 AM silopolis[m]: That's a pity, we cannot express more support than for TP work as we agreed this was one of the very features keeping us outside of the A league
05:36 AM rmu: it is not an easy problem. look at PR 518.
05:36 AM rmu: 581
05:37 AM silopolis[m]: rmu: It surely isn't
05:47 AM silopolis[m]: rmu: seen this https://forum.linuxcnc.org/38-general-linuxcnc-questions/46433-move-trajectory-planner-from-real-time-to-non-real-time ?
05:50 AM rmu: no. interesting.
05:59 AM silopolis[m]: just read the forum discussion, had you the chance to collaborate with @automata which offered help ?
06:08 AM rmu: not really. back then i was somewhat blocked by the python3 transition, that seems to be done. but, to paraphrase fox mulder, "the code is out there". Back in 2018 somebody immediately did a bunch of tests and discovered a bunch of acceleration violations. some of them went away after some bugfixes, but some I could not explain.
06:12 AM rmu: and, as i said, this is only a "hack" insofar as only tangential movement is jerk limited, individual axis are not. That would need some modification of the path, transition from linear movement to circular involves a jump in acceleration of axis.
06:15 AM rmu: i will try to find some time next week and rebase the code
06:26 AM pere: rmu: sadly, there is a lot of code out there, and only the code with a community trive. :)
06:31 AM rmu: I'm going to rebase it in the near future to current master, and then we will see.
06:44 AM -!- #linuxcnc-devel mode set to +v by ChanServ
06:44 AM -!- #linuxcnc-devel mode set to +v by ChanServ
07:39 AM pere: rmu: perhaps also reach out to the other person interested, to see if the interest is still there?
07:52 AM silopolis[m]: <rmu> "I'm going to rebase it in the..." <- Please open an issue and link to forum post once done. I'll make a post on devel ML in the mean time to dump community memory about past and current work and resources on TP... I'd like to have this put in the agenda for next meeting.
08:44 AM pere: wow, I got response on my email to the openhub maintainers. perhaps the emc/linuxcnc/linuxcnc-mirror entries can be merged now...
09:30 AM silopolis[m]: would be nice !
11:54 AM silopolis[m]: Just created a LinuxCNC Matrix Space https://matrix.to/#/#linuxcnc-space:matrix.org
11:55 AM * pere is working on bookworm live iso build
11:56 AM silopolis[m]: pere is doing good work 👍️
12:08 PM Tom_L: yes he is
04:18 PM pere: but it is hard to keep motivation going with so few active developers present her on IRC. Will see how long I can keep motivation going until I move on to a different project. :)
04:24 PM pere: bookworm live ISO build rules are now available from <URL: https://github.com/petterreinholdtsen/buster-live-build/ > branch bookworm. Not sure why the resulting ISO fail to boot with my qemu test machine.
04:29 PM pere: andypugh: hi. :)
04:30 PM andypugh: Hello.
04:30 PM andypugh: Was actually looking for syyl, though I haven’t seen him for a long time.
04:31 PM andypugh: I was watching his YouTube channel.
04:31 PM pere: not sure who that is. :)
04:31 PM pere: andypugh: btw, did you see my question regarding the live build git repos?
04:31 PM andypugh: Hmm, no.
04:32 PM pere: andypugh: <URL: https://github.com/LinuxCNC/buster-live-build/issues/3 >
04:34 PM andypugh: OK, yes, I support making the various debian (and other?) versions branches rather than repos.
04:34 PM pere: who decides, and who can execute?
04:35 PM andypugh: I guess I could, but it’s a bit late tonight.
04:36 PM pere: aha
04:37 PM pere: andypugh: I am currently testing my freshly built bookworm live iso. something is wrong, fail to find init. :)
04:38 PM pere: guess I need to build on something newer than buster. :)
04:38 PM andypugh: Are you on the live-build list?
04:38 PM pere: andypugh: btw, what is the url to syyls youtube channel?
04:38 PM pere: andypugh: nope. anything interesting going on there?
04:39 PM andypugh: No, it’s very, very quiet in general, but that’s where the experts are.
04:40 PM pere: I did send your live-build patch upstream, and it seem like we can drop it in bookworm.
04:41 PM andypugh: Yes, I saw that. I don’t recall why I didn;t try that myself.
04:41 PM pere: I am currently building without it, as it is purely cosmetic.
04:44 PM andypugh: Though not difficult to implement privately
04:45 PM pere: sure. but I plan to wait for it to show up in bookworm instead. :)
04:48 PM pere: andypugh: you should know that I very much appreciate your effort to merge my linuxcnc pull requests into master. :)
04:49 PM andypugh: It’s almost no effort, I just need a bit of a prod to notice them. Feel free to email me if I am slow.
04:49 PM pere: my hope is that the number of pull requests get down to zero, so I can look at issues instead of tracking pull requests. :)
05:49 PM pere: bookworm live ISO work, after giving my VM enough memory. :)
05:52 PM andypugh: Good news.
05:57 PM pere: andypugh: yeah. once the git repos for stretch/buster is cleaned up, I'll create a pull request for a bookworm branch. :)
06:05 PM pere: andypugh: btw, is there some automated job building live images regularly?
06:07 PM andypugh: No, it is done manually for each release. (actually, not necessarily even then, I think that there was only one 2.7 ISO and then the update process was expected to do the rest(