#linuxcnc-devel Logs

Jul 24 2017

#linuxcnc-devel Calendar

07:33 AM skunkworks: jepler, :) https://forum.linuxcnc.org/9-installing-linuxcnc/32941-linuxcnc-on-debian-stretch?start=30#96339
08:32 AM jepler: so we can make some of the people happy some of the time
10:16 AM skunkworks: jepler, exactly
10:26 AM cradek: I hate to ask, but are we going to roll another 2.7 soon to fix the problem with gcode error messages being wrong?
10:27 AM cradek: with a new cd we might get a flurry of people installing, and I'd hate for them to get new breakage
10:29 AM jepler: yeah it's a bad bug I introduced :(
10:39 AM skunkworks: we have all done that.. (well - I guess I haven't - but cradek has ;) )
10:44 AM cradek: :-P
10:49 AM seb_kuzminsky: i'm happy to make 2.7.11 today or tomorrow
10:50 AM seb_kuzminsky: i will look at PR #309 today
10:50 AM seb_kuzminsky: i'd also like to propose that I'll build pycam debs and put them on wlo, and ask that they be included in -r8
10:54 AM seb_kuzminsky: it's disturbing to me that none of our tests caught #308
10:58 AM seb_kuzminsky: but then i feel that way about all bugs, so...
11:10 AM skunkworks: still not out..
11:10 AM skunkworks: http://www.centroidcnc.com/centroid_diy/acorn_cnc_controller.html
11:10 AM skunkworks: look at the limitation of the 'free cnc'
11:10 AM skunkworks: 10 tools - you can only automeasure the first 3 tools?
11:11 AM skunkworks: only 1 coordinate system
11:12 AM skunkworks: only 3 axis simultaneous motion.
11:13 AM skunkworks: I guess - only $99 gets you more stuff
11:13 AM jepler: clearly we need to revise the pricing structure of *our* product upwards, based on what the market will bear.
11:13 AM cradek: we should at least double the price
11:13 AM skunkworks: jepler, exactly - we could be rich I tell you! rich!
11:16 AM jepler: should my e-mail announce about the actual github transition today be terse (here are git commands to type to point your existing clone at github) or long (here are some github documentations, etc etc)
11:20 AM jepler: on stretch, 'dd status=progress' is a thing. it's not in the version on wheezy, boo
11:21 AM cradek: jepler: since I have various clones, I'll depend on the push error message to tell me the commands to change origin.
11:22 AM kwallace2: I saw this show yesterday and thought "all that work to invent something that already exists?". I don't want to be "mean" so maybe there is something I'm missing? https://kamlashow.com/2016/10/27/video-a-converstation-with-danielle-applestone-of-other-machine-co/
11:23 AM kwallace2: https://othermachine.co/
11:24 AM * cradek warily scans the page for a link to the transcript
11:24 AM cradek: ah pcb routing
11:24 AM seb_kuzminsky: i think there's a natural tendency of programmers to want to reinvent everything
11:24 AM kwallace2: Sort of, it seem to be just a mini mill.
11:25 AM cradek: the little fan on the tool is neat!
11:25 AM seb_kuzminsky: at least partially because designing and writing new programs is more fun than (ahem) trying to understand someone else's 20 year old broken undocumented design
11:25 AM cradek: routed pcbs can't have more than two layers, or through-plated vias
11:25 AM cradek: ... which is pretty limiting with today's packages
11:26 AM cradek: however I still find it a useful technology for my hobby projects
11:26 AM seb_kuzminsky: there's also a tendency i think to want to "control" stuff by having it be in-house software
11:29 AM kwallace2: I suppose an upgrade to LinuxCNC would be fairly easy.
11:52 AM jepler: they claim the machine makes double-sided 6/6 boards and then they offer to sell you double sided scotch tape for fixturing. at least there's not much markup on the tape(?)
11:55 AM cradek: I don't think that's very likely on a machine made out of HDPE
11:55 AM cradek: you need really good flatness and depth control to get those small cuts
12:07 PM archivist: 3200 for a plastic machine \o/
12:10 PM archivist: considering how plastic behaves with temperature, that needs an hour warm up if you machined the base flat on it's self
01:03 PM seb_kuzminsky: i don't know how to repro #308, but if you're confident #309 fixes it i say go for it
01:04 PM seb_kuzminsky: jepler: ^^^
01:58 PM -!- #linuxcnc-devel mode set to +v by ChanServ
01:58 PM jepler: cradek: can you test #308/#309?
01:59 PM seb_kuzm1nsky: or better yet, teach the buildbot to test it? ;-)
01:59 PM jepler: it involves testing AXIS, that's .. not going to happen right this second
01:59 PM seb_kuzm1nsky: yeah, agreed
01:59 PM seb_kuzm1nsky is now known as seb_kuzminsky
02:00 PM seb_kuzminsky: i was hopeful that a test could import gcodemodule like axis does, but i totally understsand that that's not always easy either
02:02 PM jepler: I suppose it could
04:34 PM jepler: http://media.unpythonic.net/github-reconfigure-draft.txt comments?
04:34 PM jepler: also I'm asking cradek to go ahead and mark glo read-only
04:34 PM jepler: cradek: can you go ahead and mark glo read-only?
04:35 PM cradek: sure
04:36 PM jepler: .. I'm ready to test the readonlyness, let me know
04:37 PM cradek: ok
04:38 PM jepler: remote: Since we've moved to github, this repository is now read-only.
04:38 PM jepler: error: failed to push some refs to 'git.linuxcnc.org:/git/linuxcnc.git'
04:38 PM cradek: I'll update the url when you're done
04:38 PM jepler: OK
04:38 PM jepler: one last mirror to github, then shut that off?
04:38 PM cradek: you got the please see [url], right?
04:38 PM jepler: yes
04:38 PM jepler: remote: Please see
04:38 PM jepler: remote: http://media.unpythonic.net/github-reconfigure-draft.txt
04:39 PM cradek: ok mirroring
04:39 PM cradek: last mirror done, and disabled
04:40 PM cradek: [there is also a hostmot2-firmware.git mirror we need to deal with]
04:40 PM jepler: OK
04:40 PM jepler: OK, I'm giving github push permissions now based on the list of active developers
04:46 PM jepler: cradek: thanks, I think that's what I need from you right now
04:47 PM cradek: yay!
05:22 PM linuxcnc-build: build #250 of 1540.rip-jessie-armhf is complete: Failure [4failed compile] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1540.rip-jessie-armhf/builds/250 blamelist: Jeff Epler <jepler@unpythonic.net>
05:25 PM jepler: seb_kuzminsky: so did the problem follow the SD card or the mainboard?
05:25 PM * jepler has lost track
05:29 PM mozmck: thanks jepler!
05:32 PM jepler: 10:29 <linuxcnc-build_> build #164 of 4025.deb-jessie-armhf is complete: Failure [failed shell_3]
05:32 PM jepler: Build details are at http://buildbot.linuxcnc.org/buildbot/builders/4025.deb-jessie-armhf/builds/16
05:32 PM jepler: 4 blamelist: Jeff Epler <jepler@unpythonic.net>
05:32 PM jepler: looks like the error has gone with jessie (the SD card)
05:32 PM jepler: 11:06 <seb_kuzminsky> asciidoc: include: /tmp/buildd/linuxcnc-2.8.0~pre1.3265.ge200479/docs/src/com
05:32 PM jepler: mon/gpld-copyrighSegmentation fault
05:32 PM jepler: pardon the terrible paste :-/
05:32 PM jepler: 11:07 <seb_kuzminsky> when i get home i'll swap the wheezy-armhf and the jessie-armhf SD cards and see if the problem follows the SD card or stays with the U3 board
05:37 PM linuxcnc-build: build #5066 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/5066 blamelist: Jeff Epler <jepler@unpythonic.net>
05:39 PM jepler: hm but on july 3 there was a wheezy-armhf weird crash build failure too
05:39 PM jepler: Segmentation fault
05:39 PM jepler: make: *** [.html-images-stamp] Error 139
05:58 PM seb_kuzminsky: yay github! thanks jepler
05:59 PM seb_kuzminsky: so yeah, the problem followed the SD card (mostly)
05:59 PM seb_kuzminsky: it's up to date, but maybe that odroid kernel is just broken?
05:59 PM jepler: sigh who can say
05:59 PM jepler: the systems have swap, right?
06:00 PM Tom_L: so the source is in a different github location now?
06:00 PM Tom_L: just reading the scrollback here...
06:00 PM jepler: Tom_L: the canonical location for linuxcnc source is now github. git.linuxcnc.org is now readonly and will not be updated by new pushes.
06:00 PM jepler: we have a couple more items to move off git.linuxcnc.org and then its lights will be turned out
06:01 PM Tom_L: ok
06:05 PM jepler: seb_kuzminsky: I wonder what all these messages in dmesg ar
06:05 PM jepler: [9582047.300000] CPU2: Booted secondary processor
06:05 PM jepler: [9582049.405000] CPU3: Booted secondary processor
06:05 PM jepler: [9582065.800000] CPU3: shutdown
06:05 PM jepler: [9582066.410000] CPU2: shutdown
06:05 PM jepler: http://buildbot.linuxcnc.org/buildbot/builders/1405.rip-wheezy-armhf/builds/3176/steps/clear-dmesg/logs/stdio
06:06 PM Tom_L: is this it's new home? https://github.com/LinuxCNC/linuxcnc
06:06 PM jepler: Tom_L: right
06:06 PM Tom_L: thanks
06:07 PM jepler: as far as I can see, we don't mention git on our jekyll website, so nothing to change there
06:08 PM jepler: (we do in the docs but those automatically rebuild and will be updated soon if they haven't been already)
06:10 PM jepler: https://forum.odroid.com/viewtopic.php?f=73&t=9573
06:10 PM jepler: after killing App2 kernel kills App1 with : Illegal instruction,
06:10 PM jepler: and immediately after killing App2 i'm getting message from kernel:
06:10 PM jepler: [ 8868.964804] IRQ155 no longer affine to CPU3
06:10 PM jepler: [ 8868.966017] CPU3: shutdown
06:20 PM jepler: though maybe the "shutdown" messages are part of "normal" power management on those things
06:22 PM jepler: it occurs many times in a successful build too http://buildbot.linuxcnc.org/buildbot/builders/1405.rip-wheezy-armhf/builds/3207/steps/get-dmesg/logs/stdio
06:22 PM jepler: it's probably not good for latency, disabling the CPU that rtapi_app is supposed to be scheduled on!
06:40 PM seb_kuzminsky: jepler: yeah, i think it dynamically shuts down and boots processors under power management
06:41 PM seb_kuzminsky: it started showing up in dmesg (or got much more frequent) when i got more aggressive about slowing down the jessie armhf machine, to try to cool it
06:44 PM seb_kuzminsky: uildslave has 2 GB swap, and the oom killer never kills
06:49 PM seb_kuzminsky: i wonder if the debian.org linux-image-armmp will run on the u3...
06:55 PM jepler: seems optimistic
06:59 PM seb_kuzminsky: looks like no
06:59 PM seb_kuzminsky: the d.o kernel supports exynos 5 but not 4 like the u3 has
07:58 PM Roguish: hey, can anyone comment on this latency. http://i.imgur.com/YgXbsBZ.png
08:36 PM MrHindsight is now known as captHindsight
09:41 PM jepler: Roguish: fine for running with FPGA
10:57 PM Roguish: jepler. thanks.