#linuxcnc-devel Logs

May 30 2019

#linuxcnc-devel Calendar

09:08 AM mozmck: jthornton: did you mean to respond to the mailing list?
09:11 AM jthornton: yep, I wondered what happened but now I know I pressed the wrong button
09:11 AM mozmck: Yeah, I hate when mailing lists set the Reply-To to the sender.
09:11 AM mozmck: Seems to me to defeat the purpose of a mailing list
09:11 AM mozmck: jthornton: I'll respond back to the list then unless you want to send it again
09:13 AM jthornton: I thought our list used to do that
09:37 AM mozmck: cradek?
09:58 AM cradek: which list?
10:01 AM cradek: [assuming emc-users] it is set for replies to go to the list, but to leave alone any Reply-To that the individual sender might add.
10:02 AM cradek: so I think if there's a Reply-To it's because you yourself put it there
10:03 AM cradek: oh wait I'm wrong, it interacts with another setting: Setting this value to either This list or Explicit address causes Mailman to insert a specific Reply-To: header in all messages, overriding the header in the original message if necessary
10:03 AM cradek: and we have it set to "This list"
10:08 AM cradek: users and devel are set the same
10:08 AM cradek: I am getting Reply-To headers pointing to the list, and my mailer honors them
10:09 AM cradek: actually what mutt does is when you hit reply it shows you what it will do, offering the reply-to first, and asks y/n
10:10 AM cradek: if you hit n, it offers the sender instead
10:13 AM mozmck: Ah, I use thunderbird. It has a button for reply-to-list now, but it used to not have that.
10:14 AM mozmck: cradek: while you are here, can someone else be added as an owner on Github so they can add developers?
10:15 AM mozmck: cradek: or are you going to continue to be around more? Congrats on doing the clocks/watches full time now by the way!
10:25 AM mozmck: huh, I guess I don't really know how all the email headers work. I don't even see a reply-to if I view all headers, and I also don't in some yahoo groups mailing lists. The yahoo groups lists only have a "Reply" button on the emails, and it replies to the list. In the linuxcnc lists there is a "Reply" and a "Reply List" button.
12:05 PM cradek: mozmck: hm, I don't know what thunderbird does, but it sounds too smart by half
12:06 PM cradek: I regret to admit that I'm not sure what powers I have on github
12:06 PM cradek: thanks for your congratulations - it does leave me a bit more time for other things
12:08 PM JT-Shop: cradek: your one of the owners along with Jeff and Seb
12:09 PM cradek: ok I see "maintainer" and "admin" in various places, but not "owner"
12:09 PM JT-Shop: https://github.com/orgs/LinuxCNC/people
12:10 PM JT-Shop: now I wonder why my github does not have a People tab...
12:11 PM cradek: haha it does say "owner" there, cool
12:11 PM cradek: why use just one when 3 is better
12:11 PM cradek: I assume they mean the same thing but who can know
12:13 PM cradek: mozmck: you're already a release manager - are you comfortable being a github admin/maintainer/owner?
12:19 PM cradek: brb
12:56 PM mozmck: cradek: sure! I would nominate andypugh for that as well if he doesn't mind.
12:58 PM andypugh: No, that’s fine.
02:06 PM cradek: you should both be admins now
02:06 PM cradek: I confirmed that "adding collaborators" is one of the powers of admins
02:10 PM mozmck: hmm, at least for me it's still showing us as "Member": https://github.com/orgs/LinuxCNC/people
02:10 PM CaptHindsight: using 4.19.0 rt kernel on Sid as a test system currently
02:12 PM cradek: can you see this? https://github.com/orgs/LinuxCNC/people/mozmck
02:12 PM mozmck: cradek: no, 404 error
02:12 PM CaptHindsight: should have LCNC running later with multiple 4K monitors for SLA/DLP printing
02:13 PM cradek: turns out "admin" and "owner" don't mean the same thing (I haven't figured out what maintainer means yet)
02:13 PM cradek: you are an admin of LinuxCNC/linuxcnc
02:14 PM cradek: ("Can read, clone, and push to this repository. Can also manage issues, pull requests, and repository settings, including adding collaborators.")
02:14 PM mozmck: Interesting. I do see a "Settings" option which I don't think was there before.
02:16 PM cradek: I don't think it's actually true that you can "add" collaborators, I think you can invite them.
02:19 PM mozmck: Yes, looks that way. I sent Rob an invite.
02:21 PM cradek: looks like zultron and jmk also have active but unaccepted invitations
02:22 PM mozmck: yes, I haven't seen much from jmk in a long time.
02:22 PM cradek: I talked with him last July
02:43 PM rene_dev_: andypugh stuttgart andy doesnt care which date we pick in august
02:43 PM rene_dev_: andypugh which one do you prefer?
02:53 PM JT-Shop: when you pick a date I can add that announcement to the main page
03:00 PM rene_dev_: ok
03:17 PM andypugh: rene_dev_: Three of my trips have been cancelled, so I am much more flexible.
03:18 PM andypugh: What dates are in play?
03:23 PM CaptHindsight: now running 5.0.19 rt11 on Sid, 4K 3d accell works great
03:33 PM rmu: andypugh: rene_dev_: are you talking about another "integrator" meeting?
03:52 PM andypugh: rene_dev_: I can’t do the 31st of August any more. So the 3rd suite me better.
03:52 PM andypugh: (Though I need to go to Stuttgart direct from Spain)
04:02 PM rmu: aug 21-28 is chaos camp
06:40 PM memfrob: Hi, I'm trying to build LinuxCNC on debian sid and I ran into this: http://dpaste.com/0BSH1NS
06:41 PM memfrob: PREEMPT_RT / uspace configuration
06:42 PM memfrob: python --version returns 2.7.16
06:42 PM memfrob: git master branch btw.
06:45 PM memfrob: Well this is a bit confusing.. from __future__ import print_function is the first line of src/objects/hal/utils/halcompile.py if you don't include the copyright header
06:47 PM memfrob: I'm no python expert, but from __future__ import print_function is actually written twice
06:48 PM memfrob: But it's only written once here: https://github.com/LinuxCNC/linuxcnc/blob/master/src/hal/utils/halcompile.g
06:49 PM jthornton: I wonder why someone did that python 2.7 can use either version of print
06:49 PM memfrob: import sys is twice as well.. my halcompile.py file is corrupt I think
06:50 PM memfrob: http://dpaste.com/0NPGYZR -- what went wrong here?
06:51 PM jthornton: yea 2006 you might need to from future import print if you use print()
06:52 PM jthornton: that's weird for sure, that ain't the complete file
06:52 PM memfrob: No I only grabbed the start to show that the first ~30 lines are all messed up
06:52 PM jthornton: # Begin -- grammar generated by Yapps dunno what that is
06:53 PM memfrob: Everything else after the paste is irrelevant.. At least I'm not there yet.
06:53 PM memfrob: yeah the Yapps thing isn't in halcompile.g
06:58 PM memfrob: It's all because it's importing REEEEEEEEEEE
09:58 PM memfrob: I wanted to switch to python 3 and try that, but debian doesn't provide a boost library for python version 3, only 2, which is stupid.
09:59 PM memfrob: I'm supposed to build boost myself, on debian? I expect this from gentoo.
10:04 PM memfrob: Ah, find / -name "libboost_python37.so" -> /usr/lib/x86_64-linux-gnu/libboost_python37.so
10:05 PM memfrob: --with-boost-python=boost_python37 -- Yay!
10:07 PM memfrob: Same problem.. SyntaxError: from __future__ imports must occur at the beginning of the file
10:08 PM memfrob: I edited halcompile.py by hand to fix the problem, but now there's this: http://dpaste.com/1NA1A18
10:15 PM memfrob: Downgrading to gcc 7 (using update-alternatives) didn't fix it either.
10:17 PM memfrob: Forgot about g++!
10:26 PM memfrob: So the closest I get is by building with g++-7 and python3.7 -- without modifying anything, I get this: http://dpaste.com/2Q58TNX
10:30 PM mozmck_lp: I don't think python3 support is finished yet.
10:31 PM memfrob: I got to build but this isn't right. Current combination of hacks and tricks: Set GCC and G++ to 7, use Python 2.7, wait for the error: SyntaxError: from __future__ imports must occur at the beginning of the file
10:31 PM memfrob: After that error occurs, modify halcompile.py, remove the 2nd instance of "from __future__ import print_function" and install a package called python-yapps
10:31 PM memfrob: Run make again, and it works.
10:32 PM mozmck_lp: Not sure why you would be getting that. I've built with no problems on ubuntu 16.04 and 18.04. Is debian sid the unstable branch?
10:32 PM memfrob: Indeed!
10:34 PM mozmck_lp: Did you make sure all the right build dependencies are installed?
10:34 PM memfrob: Yes.
10:35 PM CaptHindsight: the info should go up on the forums or the mail list
10:35 PM CaptHindsight: no more magic builds
10:36 PM mozmck_lp: What info?
10:36 PM memfrob: I think this is a linuxcnc bug than rather proper instructions on how to build it.
10:37 PM memfrob: *rather than
10:38 PM mozmck_lp: Well, I suspect it has something to do with changes in sid. I know it builds fine in other versions of debian - I think including buster.
10:38 PM memfrob: LinuxCNC shouldn't be distro-dependent.
10:39 PM memfrob: Debian sid exposed a bug in LinuxCNC -- Of course it won't work.
10:42 PM memfrob: Because I can document a semi-working fix, it should help a developer solve this the right way.
11:03 PM mozmck_lp: Hmm, looks like the yapps2 package in buster and sid no longer installs the python 2.7 version of yapps, but depends on python3 and python3-yapps. That is why the python-yapps package was needed. This is the kind of thing that often happens in a distro upgrade and has to be worked around.
11:07 PM memfrob: You still need to wait for the "SyntaxError: from __future__ imports must occur at the beginning of the file" to come up and remove the 2nd instance.