#linuxcnc-devel Logs

Feb 27 2020

#linuxcnc-devel Calendar

12:12 AM linuxcnc-build: build #3592 of 1903.clang-wheezy-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1903.clang-wheezy-amd64/builds/3592 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
12:16 AM linuxcnc-build: build #4601 of 1403.rip-wheezy-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1403.rip-wheezy-amd64/builds/4601 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
12:16 AM linuxcnc-build: build #4600 of 1400.rip-wheezy-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1400.rip-wheezy-i386/builds/4600 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
12:18 AM linuxcnc-build: build #6442 of 1306.rip-precise-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1306.rip-precise-amd64/builds/6442 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
12:19 AM linuxcnc-build: build #6439 of 1300.rip-precise-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1300.rip-precise-i386/builds/6439 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
12:20 AM linuxcnc-build: build #3591 of 1902.clang-wheezy-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1902.clang-wheezy-rtai-i386/builds/3591 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
12:21 AM linuxcnc-build: build #5660 of 1301.rip-precise-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1301.rip-precise-rtai-i386/builds/5660 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
12:47 AM linuxcnc-build: build #4804 of 1404.rip-wheezy-rtpreempt-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1404.rip-wheezy-rtpreempt-amd64/builds/4804 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
12:47 AM linuxcnc-build: build #4120 of 1402.rip-wheezy-rtpreempt-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1402.rip-wheezy-rtpreempt-i386/builds/4120 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
01:08 AM linuxcnc-build: build #4271 of 1401.rip-wheezy-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1401.rip-wheezy-rtai-i386/builds/4271 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
01:14 AM linuxcnc-build: build #1457 of 1630.rip-stretch-rtpreempt-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1630.rip-stretch-rtpreempt-amd64/builds/1457 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
01:15 AM linuxcnc-build: build #1457 of 1610.rip-stretch-rtpreempt-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1610.rip-stretch-rtpreempt-i386/builds/1457 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
01:41 AM linuxcnc-build: build #6460 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/6460 blamelist: Chris Morley <chrisinnanaimo@hotmail.com>
12:14 PM andypugh: Hmm,
12:14 PM andypugh: G71, what to do?
12:15 PM andypugh: The author (for reasons that I absiloutely do not understand) is using features not present in the Precise version of C++
12:16 PM -!- #linuxcnc-devel mode set to +v by ChanServ
12:16 PM cradek_ is now known as cradek
12:17 PM cradek: hi!
12:17 PM cradek: do you know how hard it is to instead not use those features?
12:17 PM andypugh: I have no idea. C++ is not my thing. I only really understand C and a bt of Python. (And VBA)
12:27 PM JT-Shop: I wonder if he is using the "features" of the newer compiler in G71? or is it just what he uses?
12:32 PM jepler: It's easy to use them accidentally, if you are just a "modern C++ coder"
12:32 PM andypugh: I don’t know, nor do I really know how to tell.
12:33 PM andypugh: He seems to be using complex numbers. Which never even ocurred to me as useful when I made a Python remap version of G71
12:33 PM JT-Shop: that makes sense
12:33 PM jepler: how do complex numbers relate to G71? I thought that was subroutines.
12:34 PM andypugh: I don’t know, I have not looked at the code in detail.
12:34 PM andypugh: But the G71 implementation uses a subroutine to hold the profile definition.
01:48 PM linuxcnc-build: build #6441 of 1300.rip-precise-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1300.rip-precise-i386/builds/6441 blamelist: Mark <mark.vandoesburg@hetnet.nl>
01:49 PM linuxcnc-build: build #6444 of 1306.rip-precise-amd64 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1306.rip-precise-amd64/builds/6444 blamelist: Mark <mark.vandoesburg@hetnet.nl>
01:50 PM linuxcnc-build: build #5662 of 1301.rip-precise-rtai-i386 is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1301.rip-precise-rtai-i386/builds/5662 blamelist: Mark <mark.vandoesburg@hetnet.nl>
01:55 PM andypugh: jepler: Is it a good idea to try to revert G71 back out of master?
02:10 PM jepler: andypugh: It would be nice to see it included. I should budget some time and see if I can untangle it. I wonder if I can get precise installed for testing purposes...
02:10 PM andypugh: I was just trying to find an old Precise ISO
02:10 PM andypugh: There are older ones here, and newer: http://www.linuxcnc.org/iso/
02:11 PM andypugh: (Hardy, Lucid, Wheezy)
02:11 PM andypugh: Even Ye Olde Dapper Drake
02:16 PM cerna: I admire your effort, guys, but I think you are little crazy.
02:17 PM jepler: is prceise the only one with the problem? We could also just get the RM's permission to officially drop it.
02:17 PM jepler: we still would have one supported rtai-based distro (wheezy)
02:18 PM jepler: so yeah at that point maybe it does make the most sense. I can pull request to disable building precise / master branch while leaving 2.7 and 2.8.
02:18 PM andypugh: We have a reasonably good RTAI kernel for Buster (etc)
02:20 PM jepler: not packaged as deb, afaik
02:20 PM andypugh: Though I don’t think that one programmer with new-fangled ideas is necessarily a good reason to drop Precise.
02:20 PM andypugh: jepler: www.linuxcnc.org/temp
02:20 PM jepler: packaging rtai kernels as debs is a prerequisite to building linuxcnc packages that use it
02:20 PM jepler: I stand corrected
02:20 PM jepler: (yay)
02:23 PM jepler: there's no source package, which is a barrier too. We can't add a binary package without adding the source package it was built from.
02:37 PM jepler: [I don't think I ever really used precise and man I just don't like their ux #opinions]
02:54 PM jepler: are there tests for this new functionality?
02:55 PM andypugh: I should be able to make a source package.
02:55 PM andypugh: (for the RTAI .debs)
02:57 PM jepler: andypugh: that would help, should be possible to add it to the package server then. (is there no development package for rtai, just the headers? I assume linuxcnc builds though? are the headers in the one monolithic package?)
02:57 PM andypugh: RTAI kernel or RTAI application?
02:58 PM jepler: somewhere the headers for building stuff that uses rtai functionality have to be installed. Maybe it's in the rtai-modules package, I didn't look.
02:58 PM andypugh: There are LinuxCNC debs in that directory too. Built on that kernel.
02:59 PM andypugh: (I did it as an experiment, but then didn’t have the signing keys etc to test it properly in /dists/
03:01 PM jepler: OK, well, I made a pull request that built on precise/uspace for me. Nothing is too scary, but it is old fashioned and of course there's the chance I introduced errors...
03:02 PM jepler: It's a long shot, but I'll be attending to OSHWA 2020 convention in a couple weeks in NYC. Anybody else going to be there?
03:03 PM andypugh: I wasn’t planning on being.
03:04 PM andypugh: jepler: How about I pull your stuff in to the g71_test branch and let the builbot play with it?
03:05 PM jepler: That would be great, thanks!
03:05 PM jepler: @and
03:05 PM jepler: argh new IRC user
03:06 PM jepler: tempted to start the "we need a post-IRC chat" flamewar but really don't want to. matrix and discord are both .. pretty not bad? and I've grown so used to using them for my other open source work...
03:07 PM andypugh: gitter.im is OK too.
03:07 PM jepler: speaking of "why do we stay fixated on old things like ubuntu precise"
03:08 PM jepler: I don't think I've used that particular one. Naturally I'd prefer it to be a service I already use :)
03:08 PM andypugh: But experience seems to show that adding more platforms just splits the support and user base into smaller chunks.
03:11 PM jepler: you mean communications platforms? yeah I'd also be on the opposite side of another debate than I was back in the day, which was whether to close the mailing lists when we created the forum..
03:11 PM andypugh: jepler: Did you start with (broken) master or with the g71_test branch?
03:11 PM jepler: with broken master, do I need to start over?
03:11 PM andypugh: Maybe not
03:12 PM jepler: too bad I didn't realize that some fixing had maybe already occurred :-/
03:12 PM andypugh: Yes, Mark V has already pushed a few fixes, and then declared no interest in anything older than C++11
03:13 PM jepler: I hope that wouldn't stop him fixing bugs after someone else adapted it to older systems.
03:14 PM andypugh: See #686
03:14 PM andypugh: https://github.com/LinuxCNC/linuxcnc/pull/686
03:15 PM andypugh: I am not sure that 8746c7b has the desired effect.
03:18 PM jepler: it might, depending on the changes to configure which I didn't look at. they'd have to be considered together.
03:18 PM jepler: but, we set -std=c++11 so that we had SOME ability of the compiler to catch constructs that were only valid in newer compilers...
03:18 PM jepler: .. I think
03:19 PM andypugh: New comment from the author on gitub
03:19 PM andypugh: Using words I don’t understand :-)
03:19 PM andypugh: (I just do C, me)
03:19 PM jepler: oh they added and then removed it, I see
03:22 PM andypugh: I am getting something of a merge conflict in my test branch now, trying to combine his work and yours.
03:22 PM andypugh: I am feeling a bit out of my depth.
03:23 PM jepler: I may know more about what's going on, and some of the stuff in 686 looks helpful for working with older compilers
03:23 PM jepler: .. so let me make a fresh patch on top of what's in 686
03:37 PM jepler: andypugh: https://github.com/LinuxCNC/linuxcnc/pull/689
03:38 PM jepler: and now I'll stop meddling again
03:53 PM CaptHindsight: someone was just asking about a compile error with G71 in the other channel
03:54 PM andypugh: papagano?
03:54 PM CaptHindsight: yeah
03:54 PM andypugh: We had a private chat
03:54 PM CaptHindsight: I'm working with your SVG HAL stuff
03:55 PM CaptHindsight: I'm writing a parser for SVG files that Slic3r outputs
03:56 PM CaptHindsight: and a QTPYVCP GUI to control it all
03:56 PM andypugh: Ah, OK. Not using my Inventor slicer code then?
03:57 PM CaptHindsight: I saw the binary
03:58 PM CaptHindsight: everyone tends to output slices differently
03:58 PM CaptHindsight: does yours work with inkscape to convert to SVG's?
03:59 PM CaptHindsight: i saw all the inkscape references in your example SVG's
03:59 PM andypugh: No, I was just ising inkscape as a viewer
03:59 PM CaptHindsight: ah
04:00 PM CaptHindsight: https://raw.githubusercontent.com/andypugh/SVG_Slicer/master/test.svg
04:00 PM andypugh: I needed a namespace (I think) so use the Inkscape one
04:00 PM CaptHindsight: I'm using Slic3r since #1 it is still open and it also handles adding supports to the prints
04:02 PM andypugh: I need to see if I can convert that .ivb file into programme text.
04:02 PM CaptHindsight: your SVG_Slicer doesn't like Slic3rs SVG's
04:03 PM CaptHindsight: I'll get it all sorted
04:04 PM andypugh: To be honest, it was all a long time ago and only really an intellectual exercise
04:04 PM CaptHindsight: yeah no problem
04:05 PM CaptHindsight: I've been needing to master all of this for a while anyway
04:05 PM CaptHindsight: most of the software for resin printer control is all going closed source
04:05 PM CaptHindsight: even the files are binary
04:06 PM CaptHindsight: so you can't even preview their output in Inkscape
04:06 PM CaptHindsight: so this will keep it open
04:07 PM CaptHindsight: I've wanted to get all the control into LCNC for resin printers as well as inkjet for some time
04:11 PM CaptHindsight: I also wanted a way to work with openCV again for machine vision
04:12 PM CaptHindsight: https://ibin.co/2u5W3RNmUMtN.jpg 50um gap of a dusty caliper
04:12 PM CaptHindsight: ^^ using $20 webcam
04:13 PM CaptHindsight: https://ibin.co/2u5XXEUdLjMj.jpg same 50um gap with Sobel filter to detect edges
04:14 PM CaptHindsight: so we can align parts using a cheap camera
04:15 PM andypugh: I keep meaning to take a look at OpenCV, but you know how it goes…
04:16 PM CaptHindsight: :)
04:16 PM CaptHindsight: I used to work with imaging back in the 80's
04:16 PM CaptHindsight: we were happy top grab a single B&W 512 x 512 frame in real time
04:17 PM CaptHindsight: top/to
04:18 PM andypugh: I did a course using a Keyence system. Rather neat, programmed with an Excel-like interface that appeared in a web browser on the device.
04:18 PM CaptHindsight: I can easily see having a camera inspect cuts and also listen for chatter to automatically tune feeds and speeds
04:18 PM andypugh: That would be early-noughties
04:19 PM CaptHindsight: i was working on old film restoration and also remote radiology
04:19 PM CaptHindsight: mid-80's
04:20 PM CaptHindsight: 20MHz DSP's and array processors
04:20 PM andypugh: no transputers?
04:21 PM andypugh: (We have some dynamometers at work with transputers)
04:22 PM CaptHindsight: pre-transputer
04:22 PM CaptHindsight: in a PC over ISA bus
04:22 PM CaptHindsight: then we moved to multibus
04:22 PM CaptHindsight: with 2MB of video memory :)
04:23 PM CaptHindsight: https://books.google.com/books?id=IGkuDrOYjOcC&ppis=_c&lpg=PA167&ots=jQ4l_rIgw_&dq=512%20frame%20grabber%20imaging%20technologies&pg=PA167#v=onepage&q=512%20frame%20grabber%20imaging%20technologies&f=false
04:26 PM andypugh: Whoo! 64 shades of grey!
04:27 PM CaptHindsight: yeah was closer to 6bit depth but grabbed at 8bit
04:27 PM CaptHindsight: 256 levels was hard back then at few MHz
04:29 PM andypugh: But, it’s still bette than 50 Shades of Grey :-)
04:34 PM jepler: remembers writing a quick cam userspace driver on linux in the 90s and I probably had it easy....
05:40 PM andypugh: I wonder if the Buildbot has got stuck?
06:14 PM CMorley: Andy there is a compile error in 2.8 missing bracket and an extra bracket....
06:15 PM andypugh: Eek!
06:15 PM andypugh: But I tested that!
06:16 PM CMorley: I just pulled and it would not compile without changes
06:16 PM andypugh: in cpuinfo.c ?
06:16 PM CMorley: yes
06:17 PM CMorley: it's fairly obvious when you look with gitk
06:18 PM andypugh: Yes. I see it now.
06:19 PM CMorley: after it compiled but had a warning
06:19 PM CMorley: Thanks Andy
06:27 PM andypugh: I had a few goes at that edit, as initially I didn’t have write access from this machine to the LinuxCNC drive. I think I got lazy on attempt 3 to make the same trivial edit.
06:39 PM andypugh: Hopefully fixed now. See you all tomorrow.
06:43 PM jthornton: night
11:02 PM fjungclaus1 is now known as fjungclaus