#linuxcnc-devel Logs

Nov 03 2021

#linuxcnc-devel Calendar

02:07 AM pere: seb_kuzminsky: any idea when the txt->adoc change will happen?
09:32 AM seb_kuzminsky: pere: i started the review last night, hope to finish today
04:21 PM seb_kuzminsky: pere: does the Rename_Documentation_txt_to_adoc-rebased-2021-11-03 branch build for you? It does not build for me on bullseye:
04:21 PM seb_kuzminsky: make[1]: *** No rule to make target 'depends/Master_Getting_Started_fr.d', needed by 'Makefile'. Stop.
04:21 PM pere: seb_kuzminsky: did not test. will test and fix.
04:21 PM seb_kuzminsky: thx...
04:21 PM pere: ran out of time and had to just publish and run.
04:29 PM seb_kuzminsky: i know how that goes - thanks for working on it
04:31 PM pere: found a few files where the txt->adoc move had been lost.
04:32 PM pere: seb_kuzminsky: btw, is there a plan to get rid of the compiler warnings?
04:33 PM seb_kuzminsky: there's a desire but not really a plan
04:34 PM seb_kuzminsky: jepler's got an open PR that fixes a bunch of warnings, i gave some comments but i think it's not quite ready to merge yet
04:34 PM seb_kuzminsky: https://github.com/LinuxCNC/linuxcnc/pull/1278
04:34 PM pere: seb_kuzminsky: I looked at it, and suspect it is trying to gap over a too large bite.
04:35 PM seb_kuzminsky: yeah, big PRs are sure harder to accept than small ones...
04:36 PM seb_kuzminsky: jepler is a master coder and a long-time core developer in linuxcnc, i'm confident he'll fix it up when he gets the chance
04:39 PM jepler[m]: yeah I bit off a lot and haven't had the motivation to get back to it. sorry.
04:41 PM seb_kuzminsky: a wild jepler appears!
04:41 PM jepler[m]: the plague of modern "notifications" :)
04:41 PM seb_kuzminsky: heh
04:41 PM seb_kuzminsky: you should turn off all notifications except linuxcnc irc
04:43 PM seb_kuzminsky: i'm heading out and leaving my phone behind, back in an hour or two
04:45 PM jepler[m]: good idea
04:48 PM pere: seb_kuzminsky: branch updated and now builds.
04:50 PM pere: jepler[m]: what about splitting up the pull request into smaller parts, and fix some of the classes of warnings?
04:56 PM jepler[m]: That would have been wise and could still be done
04:59 PM pere: for example the strncpy issues might be security issues, and could be fixed independent of the other issues.
06:44 PM pere: seb_kuzminsky: seem like steffen updated several pull requests, hope they are ready now.
06:55 PM seb_kuzminsky: pere: i'm still getting build failures on bullseye at 2092877ebd:
06:56 PM seb_kuzminsky: make[1]: *** No rule to make target 'depends/getting-started/index_fr.d', needed by 'Makefile'. Stop.
07:14 PM seb_kuzminsky: maybe you ran configure without --enable-build-documentation?
07:24 PM seb_kuzminsky: maybe it's counterintuitive that we don't build docs by default...
07:33 PM seb_kuzminsky: that index_fr thing seems to be the last issue, now it builds
07:34 PM Tom_L: fixing linuxcnc errors keeps you outta the bars!
07:34 PM Tom_L: thanks.....
08:01 PM seb_kuzminsky: this is a new feature, sort of, that pere and smoe are working on
08:02 PM seb_kuzminsky: by renaming the asciidoc source files of our documentation from ".txt" to ".adoc", they'll get rendered properly in github, and people can edit them there and submit PRs with the updates easily
08:02 PM seb_kuzminsky: pretty neat, and hopefully it'll lower the threshold of nerdiness needed to contribute to the docs
08:03 PM Tom_L: nice
08:04 PM seb_kuzminsky: pere: uh-oh, the docs are different before and after that change
08:04 PM seb_kuzminsky: looks like a bunch of stuff didn't make it through the transition
08:05 PM seb_kuzminsky: (i build linuxcnc-doc-*.deb before and after that branch, extracted them with dpkg-deb -x, and ran diffpdf on the pdfs)
08:39 PM seb_kuzminsky: the difference has to do with the generation of svgs from dot files, i think i have a fix
08:44 PM seb_kuzminsky: compare this in our current code: https://github.com/linuxcnc/linuxcnc/blob/master/docs/src/getting-started/about-linuxcnc.txt
08:44 PM seb_kuzminsky: to pere & smoe's new stuff: https://github.com/petterreinholdtsen/linuxcnc/blob/Rename_Documentation_txt_to_adoc-rebased-2021-11-03/docs/src/getting-started/about-linuxcnc.adoc
08:52 PM seb_kuzminsky: (that reminds me, generated images, like the homing diagrams made from dot files, will be missing in the github view of the docs, but oh well)
09:13 PM Tom_L: maybe i can ask here..
09:14 PM Tom_L: rotary axis needs to stop before linear move. any idea what code for that? G81 G61.1 G04 don't work for that
09:14 PM Tom_L: s/G81/G61
09:37 PM seb_kuzminsky: hmm, sounds like you're talking about a "LOCKING_INDEXER_JOINT" maybe? http://linuxcnc.org/docs/2.8/html/config/ini-config.html#_axis__lt_letter_gt_section
09:38 PM Tom_L: i've been looking at that
09:39 PM Tom_L: but that limits it to an indexer only
09:39 PM Tom_L: which might be ok for some things
09:39 PM Tom_L: i don't mind running more than one config...
09:40 PM seb_kuzminsky: i don't understand what hardware you have, and what behavior you want
09:41 PM Tom_L: i will share a yt moment with you ...
09:41 PM Tom_L: https://www.youtube.com/watch?v=BZMAnER9WFo
09:42 PM Tom_L: working to get that set as either an indexer or rotary that can profile with X
09:48 PM seb_kuzminsky: i've never used a rotary or an indexer, but can't you just write your gcode to do "software indexing"? g0 in a, then profile in xyz, g0 in a to another place, profile again, and so on?
09:48 PM Tom_L: the problem is it doesn't wait for G0 A to finish
09:49 PM Tom_L: but i don't have that LOCKING_INDEXER_JOINT enabled yet
09:49 PM seb_kuzminsky: that's surprising to me
09:49 PM seb_kuzminsky: what's your G64 set to? maybe it's blending the g0 with the following g1?
09:49 PM Tom_L: i'm also writing a post for this as i go
09:50 PM Tom_L: not using G64
09:51 PM seb_kuzminsky: does that mean you're in G61 exact path mode? or both G61 and G64 are unspecified?
09:51 PM Tom_L: i tested G61 and G61.1 and neither seemed to have an effect on it
09:51 PM Tom_L: haven't tried G64 yet
09:52 PM seb_kuzminsky: what happens if you insert a G4 P0 after you move A?
09:52 PM Tom_L: it pauses a little but continues on with linear moves before A finishes
09:53 PM Tom_L: jt seemed to think that might break up the que so it would stop and wait
09:53 PM seb_kuzminsky: do you have position feedback from A, or is it open loop?
09:53 PM Tom_L: open
09:54 PM seb_kuzminsky: are you sure the stepper is tuned properly? maybe it's lagging behind motion's commanded position
09:54 PM Tom_L: i'm in the first testing stages so i don't expect perfection
09:54 PM Tom_L: just got it moving yesterday
09:54 PM Tom_L: no i'm sure the stepper isn't fully tuned
09:55 PM seb_kuzminsky: it'd be interesting to set up halcsope to look at commanded position for A, and see if motion thinks A is done while the axis is still moving
09:55 PM Tom_L: but right now, linear motion begins as soon as angular does
09:55 PM Tom_L: no pause
09:55 PM Tom_L: once i can see a pause i'll know i'm headed in the right direction
09:56 PM Tom_L: and this is new territory for me :)
10:00 PM seb_kuzminsky: i'm not able to replicate this behavior in sim
10:01 PM Tom_L: i'll try to get the axis calibrated better then come back to this
10:01 PM seb_kuzminsky: sounds good
10:01 PM Tom_L: not alot of docs on tuning rotary
10:02 PM seb_kuzminsky: it should be very similar to tuning a linear axis
10:02 PM seb_kuzminsky: unless i'm missing something...
10:02 PM Tom_L: yeah, the numbers are alot different
10:02 PM seb_kuzminsky: yeah, better docs would be good
10:03 PM Tom_L: units for alot of it are angular
10:03 PM Tom_L: this is more a learning exercise for me anyway
10:03 PM Tom_L: nothing pressing
10:03 PM seb_kuzminsky: looks like a neat project
10:04 PM Tom_L: got jt's interest.. he ordered one :)