#linuxcnc-devel Logs

Dec 10 2022

01:07 AM -!- #linuxcnc-devel mode set to +v by ChanServ
01:34 AM -!- #linuxcnc-devel mode set to +v by ChanServ
03:52 AM rmu: this copyright pedantry exists because of SCO lawsuit and is somewhat necessary
04:20 AM pere: I would take it a step further, and say it exist because copyright law has been expanded way past its break point, and we have to live in this dysfunctional world.
09:01 AM unterhaus: I don't know how much nist code is left, but the tp code was awful last time I tried to decipher it. Reusing variable names to save memory
09:04 AM unterhaus: I wonder if this is the only file that guy found and if so that's pretty good
09:39 AM JT-Cave: is there an option to not build the docs when building a RIP? I can't find it anymore
09:42 AM Tom_L: yes i think it's --no-docs or something very similar
09:45 AM JT-Shop_ is now known as JT-Shop
09:45 AM JT-Shop: --no-docs is not in configure anymore for 2.9
09:46 AM * JT-Shop starts a fire to warm up the shop
09:46 AM Tom_L: i see alot of constants in configure BUILD_DOCS=
09:46 AM Tom_L: yes/no
09:47 AM -!- #linuxcnc-devel mode set to +v by ChanServ
09:48 AM Tom_dev: JT-Cave, look for --enable-build-documentation
09:48 AM Tom_dev: it's in configure
09:48 AM Tom_dev: not sure where or how you set it
09:49 AM Tom_dev: read optional features line 1531 in configure
09:53 AM JT-Shop: ./configure --help will list all options
09:53 AM JT-Shop: maybe set it to no or false?
09:54 AM Tom_L: doesn't look that way
09:54 AM JT-Shop: ./configure --enable-build-documentation=no
09:55 AM JT-Shop: that seemed to work
09:55 AM Tom_L: good
10:01 AM JT-Shop: off to start another fire
10:08 AM roguish[m]: pyro....
10:12 AM -!- #linuxcnc-devel mode set to +v by ChanServ
10:13 AM Tom_dev: JT-Cave, also in the install.adoc:
10:13 AM Tom_dev: If you don't want to generate the documentation from source use
10:13 AM Tom_dev: `--disable-build-documentation`
10:14 AM JT-Shop22: did you try that?
10:14 AM Tom_L: no
10:14 AM Tom_L: but i read that :)
10:14 AM Tom_L: in 2.10 pre
10:15 AM JT-Shop22: should be the same as 2.9
10:15 AM JT-Shop22: but I didn't see that in configure
10:15 AM JT-Shop22 is now known as JT-Shop2
10:15 AM Tom_L: i didn't either
10:19 AM * JT-Shop2 swims back to the mahcine shop to test it out
10:21 AM Tom_L: i'll build some debs sometime today
10:22 AM JT-Shop: ./configure --disable-build-documentation seems to work as well
10:22 AM Tom_L: cool
11:48 AM JT-Cave: checking whether readline license is compatible with GPL-2... no
11:48 AM JT-Cave: configure: WARNING: The LinuxCNC binary you are building may not be
11:48 AM JT-Cave: distributable due to a license incompatibility with LinuxCNC (some portions
11:48 AM JT-Cave: GPL-2 only) and Readline version 6 and greater (GPL-3 or later).
11:48 AM JT-Cave: configure: error: To configure LinuxCNC in this way, you must
11:48 AM JT-Cave: invoke configure with "--enable-non-distributable=yes". Note that on
11:48 AM JT-Cave: Debian-based systems, you may be able to use libreadline-gplv2-dev instead.
11:49 AM JT-Cave: libreadlne-dev is 8.2 so something is broken maybe with make?
11:49 AM JT-Cave: I mean configure
11:50 AM JT-Cave: so a gpl license issue with configure?
11:51 AM JT-Cave: { printf "%s\n" "$as_me:${as_lineno-$LINENO}: checking whether readline license is compatible with GPL-2" >&5
11:51 AM JT-Cave: printf %s "checking whether readline license is compatible with GPL-2... " >&6; }
11:51 AM JT-Cave: cat confdefs.h - <<_ACEOF >conftest.$ac_ext
11:52 AM JT-Cave: configure line 1198
11:52 AM JT-Cave: 11978
12:27 PM cradek: readline changed from gpl2 to gpl3, so if you combine readline with gpl2 code, you can't distribute the result. configure is warning you about that
12:42 PM JT-Shop2: are we gpl2?
01:30 PM cradek: yes
05:13 PM andypugh: unterhaus: No, I think this is his opening salvo in a war against LinuxCNC.
05:14 PM andypugh: unterhaus: https://www.mail-archive.com/debian-legal@lists.debian.org/msg45679.html
05:16 PM JT-Cave: yep sounds like it might be Paul... on a vendetta
05:22 PM unterhaus: He's wrong about the one file, it seems like the rest is a batch of hoo as well
05:23 PM CaptHindsight[m]: calmly write up a response using the name Secret Squirrel
05:26 PM Tom_L: JT-Cave, 3 failed attempts pushing the armhf deb
05:32 PM CaptHindsight[m]: context https://hanna-barbera.fandom.com/wiki/The_Atom_Ant/Secret_Squirrel_Show
05:33 PM JT-Cave: why not use your real name?
05:34 PM unterhaus: After what Andy found, it seems to me that Paolo probably needs to update his file
05:34 PM unterhaus: I imagine he's a bit careless about that sort of thing
05:36 PM JT-Cave: the two files are not even close to the same thing
05:41 PM JT-Cave: Tom_L, are you trying to push to my repo again?
05:41 PM Tom_L: turned out to be wifi
05:42 PM Tom_L: and yes
05:42 PM Tom_L: working on the arm64 now
05:42 PM JT-Cave: cool
05:42 PM Tom_L: tell the chickens not to get their feathers ruffled.. i'm trying to hurry :)
05:43 PM Tom_L: andypugh, were you planning to add that helix fix to 2.8 as well?
05:44 PM JT-Cave: is the helix bug in 2.8?
05:44 PM Tom_L: it's possible
05:44 PM Tom_L: you'll have to verify that with andy but maybe introduced in a late 2.8 release
05:45 PM JT-Cave: I should have 2.8 on the BP mill
05:45 PM Tom_L: i doubt it would show it. it might
05:45 PM Tom_L: do an axis sim...
05:46 PM Tom_L: pr 61 up
05:47 PM JT-Cave: merged and pulled
05:47 PM Tom_L: noticed
05:47 PM andypugh: The fix that cused the bug was not put in 2.8, as I thought it needed more testing at the time.
05:47 PM andypugh: Which looks like it was a good decision.
05:48 PM Tom_L: ok
05:48 PM andypugh: But according to the original reporter the bug is still not fixed
05:48 PM Tom_L: so it appears in 2.10 pre only?
05:48 PM Tom_L: it changed behavior here. i think it is fixed
05:49 PM JT-Cave: I could not reproduce the bug in 2.9 today
05:49 PM Tom_L: i was able to produce the bug and noted a fix
05:49 PM Tom_L: in 2.9.0 pre
05:49 PM JT-Cave: in 2.9 built from a clone?
05:49 PM Tom_L: i'm not sure about that on the original install
05:50 PM Tom_L: it may have been from the package manager
05:50 PM Tom_L: pretty sure.. it is available there
05:50 PM Tom_L: i did build 2.10 from clone
05:52 PM JT-Cave: yep in 2.9 from debian it's still a bug
05:52 PM Tom_L: i dont't think bullseye has it in package manager. i'm using bookworm
05:52 PM JT-Cave: debian 12 here
05:52 PM JT-Cave: I think
05:53 PM JT-Cave: I built a rip but could not get the bug to do a z only move
05:53 PM andypugh: I can’t reproduce it either. Which is making it hard to fix.
05:53 PM Tom_L: is he sure he's running your fix?
05:59 PM JT-Cave: this pc is debian 12 with linuxcnc installed from debian and the bug is there... estop, power on, home. then G2 X0 Y0 Z-2 R4 F1000
06:00 PM Tom_L: crap. i can't run 2 instances of lcnc.. i was going to post both axis screens 2.9.0 & 2.10
06:00 PM JT-Cave: jambalaya time
06:00 PM CaptHindsight[m]: is this the test? : estop, power on, home. then G2 X0 Y0 Z-2 R4 F1000
06:01 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/arc_test/arc_test_G91.png
06:01 PM Tom_L: if you're in G90 increment Z
06:03 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/arc_test/axis_test_fix.png
06:03 PM Tom_L: with andy's patch
06:08 PM -!- #linuxcnc-devel mode set to +v by ChanServ
06:11 PM andypugh: Tom_L: I think that I fixed a different (and more important) bug than was originally reported.
06:11 PM Tom_L: ok
06:12 PM Tom_L: but tell me this. why doesn't it show up in G3?
06:13 PM Tom_L: or is it just best not to ask :)
06:29 PM andypugh: Why did it work some of the time on spirals before?
06:30 PM andypugh: /* now angle is in range 0..PI . Check if cross is antiparallel to
06:30 PM andypugh: normal. If so, true angle is between PI..2PI. Need to subtract from
06:30 PM andypugh: 2PI. */
06:30 PM andypugh: pmCartCartCross(&circle->rTan, &rEnd, &v);
06:30 PM andypugh: pmCartCartDot(&v, &circle->normal, &d);
06:30 PM andypugh: if (d < DOUBLE_FUZZ) {
06:30 PM andypugh: circle->angle = PM_2_PI - circle->angle;
06:30 PM andypugh: }
06:31 PM pere: andypugh: can these examples be turned into a tests/ script?
06:31 PM andypugh: Because some of the time d < DOUBLE_FUZZ and the angle is zero, so you get a PM_2_PI circle anyway.
06:32 PM andypugh: pere: Possibly, but only bu timing how long the path takes to run, I think.
08:51 PM CaptHindsight[m]: https://sourceforge.net/p/emc/mailman/message/37746740/
08:52 PM CaptHindsight[m]: if anyone else wants to test, here are the steps
09:21 PM Tom_L: CaptHindsight[m], did you test it on master with the fix?
11:18 PM CaptHindsight[m]: Tom_L: no, I just provided the link with the instructions ^^