#linuxcnc-devel Logs

Jul 15 2022

#linuxcnc-devel Calendar

12:38 AM -!- #linuxcnc-devel mode set to +v by ChanServ
04:10 AM pere: silopolis[m]: perhaps it is time to address the open documentation issues, <URL: https://github.com/LinuxCNC/linuxcnc/issues?q=is%3Aissue+is%3Aopen+label%3Adocumentation >?
04:15 AM silopolis[m]: pere: that'd surely be nice indeed
04:16 AM silopolis[m]: My work on qtvcp docs is so long overdue 😔
04:16 AM pere: just had a look at your <URL: https://github.com/LinuxCNC/linuxcnc/pull/1742 >, and it is hard to understand the rationale behind all the changes by looking at the complete patch.
04:25 AM silopolis[m]: Rationale is improving formating
04:44 AM pere: silopolis[m]: the .git* and debian/ changes is less clear in that context.
04:45 AM pere: and removals and renames likewise.
04:46 AM pere: some of the changes seem to be inherited from some of my patches, already merged into master.
05:48 AM silopolis[m]: pere: it's been a couple of month since I've last worked on this and have to get my head back around it.
05:48 AM silopolis[m]: Like told in email just sent, will do my best to at least send a status update by the end of the WE
05:50 AM pere: ack. I hope the amount of activity can lead to more people looking at the docs.
06:02 AM silopolis[m]: I really hope too...
06:02 AM silopolis[m]: Transition and beginnings at new job have been much harder than expected.
06:02 AM silopolis[m]: Started to slowly cool down this week, so I hope things will settle very soon
06:05 AM pere: I normally assume around 6 months to settle into a new job.
06:08 AM pere: looking at the various hal components, it occur to me that the .c files should probably be converted to .comp to ensure all components behave consistently.
06:09 AM pere: 24 .c files, 115 .comp in src/hal/components/. If I am not mistaken, this conversion was already done in machinekit.
06:11 AM pere: silopolis[m]: I do find it a bid sad that so few are active here on irc. :)
06:11 AM pere: silopolis[m]: I am glad you are here, at least. :)
06:19 AM silopolis[m]: <pere> "silopolis: I do find it a bid..." <- That is indeed 😢
06:26 AM silopolis[m]: Would make the dev community much smoother and lively
06:26 AM pere: yeah
09:03 AM mrec: Hi, does anyone know is G71 fixed for the lathe? eg. when diving into the material there should be another feed rate possible, also some dwelling
09:04 AM mrec: I'm going to look at the code now here, but just in case someone fixed it already
09:04 AM mrec: I'm cutting some thinner stainless parts and it doesn't look good with the current G71 implementation
10:24 AM mrec: is there a function to dump the content of a block_pointer?
10:44 AM mrec: seems like it's not so difficult to fix G71
10:53 AM mrec: if((*p)->dive(location,x,out,p==begin()))
10:54 AM mrec: p == begin() ... it's doing a rapid cut into the material it seems
10:55 AM mrec: -4 2.3
10:55 AM mrec: -4 2.12 (straight_rapid)
10:55 AM mrec: -8.23 2.12 (straight move)
10:58 AM mrec: luckily my X is slow.
11:04 AM mrec: does anyone know how to set the feed rate with c++?
11:04 AM mrec: maybe something like set_feed_rate
11:14 AM silopolis[m]: smoe63 pere would you be willing to mentor TurBoss for QtPyVCP packaging and Debian upload?
11:22 AM pere: silopolis[m]: sound likely, if I understood what the question was. What is QtPyVCP?
11:27 AM -!- #linuxcnc-devel mode set to +v by ChanServ
11:27 AM -!- #linuxcnc-devel mode set to +v by ChanServ
11:44 AM silopolis[m]: https://www.qtpyvcp.com/
11:49 AM silopolis[m]: pere: a nice and featureful lcnc VCP with a nice community 🥰
11:52 AM pere: so a new source deb and binary. sure, I can probably give advice.
12:12 PM TurBoss: Hello pere,
12:12 PM TurBoss: we are currently using a python setup script and relay on pypi packages for deps,
12:12 PM TurBoss: many of them are on debian now, but i need to figure 2 of them
12:12 PM TurBoss: https://github.com/zerwes/hiyapyco and https://github.com/wimglenn/oyaml
12:18 PM pere: TurBoss: are you aware of the python group in Debian? They some times hang out on irc.debian.org, #debian-python.
12:19 PM TurBoss: no didn't know
12:19 PM TurBoss: thanks
12:22 PM pere: TurBoss: are you building debs today?
12:23 PM TurBoss: not yet
12:27 PM pere: you could find some help in the dh-make package, it can provide a scaffold to build the package on.
12:31 PM TurBoss: thanks will take a look
12:39 PM pere: one of the most important files to get correct to get the package accepted into debian is the debian/copyright file, listing licensing and copyright status. it is a lot easier to handle if all the program code have license info in the file headers.
01:09 PM pere: Am I on the right track with <URL: https://github.com/LinuxCNC/linuxcnc/pull/1839 >?
01:13 PM mrec: nice my G71 is now diving in slowly into the material
01:14 PM * pere look up G71
01:30 PM mrec: the result is better, I'm going to add dwelling support now
01:30 PM mrec: https://github.com/LinuxCNC/linuxcnc/issues/1146#issuecomment-1185727887
01:30 PM mrec: I will take some pictures from the part which I'm doing for comparison
01:31 PM mrec: the current implementation seems to be dangerous
01:44 PM pere: mrec: what kind of memory use is top reporting for your linuxcnc instance?
01:47 PM mrec: I think around 110 mb linuxcnc usually runs for days here
01:47 PM mrec: but that 110 mb is just a fresh start
01:47 PM mrec: dwelling works with G71
01:47 PM pere: my process was touching 3GB when I killed it, after loading the g-code program at the start of issue #1146.
01:47 PM mrec: time to cut the stick again
01:48 PM mrec: my linuxcnc version is 2.9 from last year
01:48 PM mrec: last commit 12. Jun 2021
01:49 PM mrec: I have modified the XHC pendant, and wrote a plugin for controlling the old spindle of my CNC mill
01:50 PM mrec: oooh you're on another issue here
01:50 PM mrec: yes there's an endless loop I think
01:50 PM mrec: in case you do something wrong with the G71 command it can end up like that
01:50 PM mrec: I did not have a look at that yet
01:51 PM mrec: it's an absolute trap for beginners
02:13 PM pere: mrec: but I expected the hang to show up when running the program, not when loading it?
02:16 PM mrec: it should hang when loading it
02:25 PM pere: aha. that was not quite clear from my reading.
02:27 PM mrec: I'm into the code now should it happen again I'll fix it.
02:28 PM mrec: I did not experiene that problem for a long time because I'm used to the correct syntax now
03:38 PM -!- #linuxcnc-devel mode set to +v by ChanServ
03:38 PM -!- #linuxcnc-devel mode set to +v by ChanServ
05:07 PM pere: is it possible to write a test for <URL: https://github.com/LinuxCNC/linuxcnc/issues/1813 >? If so, where can I find an exapmle?
05:10 PM Tom_L: i agree it shouldn't do it but that's a pretty crappy cam that would output code like that
05:11 PM pere: should we care about that?
05:11 PM Tom_L: if it's going to grind linuxcnc to a snail's pace yes
05:12 PM Tom_L: it's still going to have to read that line and process it somehow
05:12 PM pere: I mean, if you want to send a bug report to the cam system, please do. I just found it to be a rather obvious thing that 'keep the same speed as before' should not behave as it it was a 'change speed' command.
05:12 PM Tom_L: i don't use that cam so i likely won't
05:12 PM Tom_L: i agree with you
05:13 PM pere: I would like to add a test to linuxcnc to detect the problem, and ensure my patch fixes it, but do not know how to do that.
05:14 PM Tom_L: i'm not much on writing testing code either
05:22 PM pere: Are the tests in tests not run automatically? <URL: https://github.com/LinuxCNC/linuxcnc/pull/1840 > must have been wrong for a while...
05:23 PM pere: I hope someone look at the pull requests soon. Once most of those are processed, we can focus on the issues.
11:50 PM mrec: I have some blending phenomen here:
11:50 PM mrec: I'm setting F100
11:50 PM mrec: run G1 Z100
11:50 PM mrec: run G1 X0
11:50 PM mrec: sorry G1 X0 F3
11:51 PM mrec: it will run into X with F100 (the original speed) until the curve is over
11:51 PM mrec: G64 P.. fixes that problem
11:51 PM mrec: but X should at any time be F3 in that case no?
11:52 PM mrec: phenomenon*