#linuxcnc-devel Logs

May 10 2017

#linuxcnc-devel Calendar

01:02 AM KGB-linuxcnc: 03Sebastian Kuzminsky 052.7 f7ce94a 06linuxcnc 10src/hal/drivers/mesa-hostmot2/hm2_eth.c hm2_eth: don't segfault on interfaces without addresses * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f7ce94a
01:27 AM linuxcnc-build_: build #4925 of 1200.rip-lucid-i386 is complete: Failure [4failed compile runtests] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1200.rip-lucid-i386/builds/4925 blamelist: Sebastian Kuzminsky <seb@highlab.com>
01:53 AM linuxcnc-build_: build #4942 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4942 blamelist: Sebastian Kuzminsky <seb@highlab.com>
02:15 AM seb_kuzminsky: task: 227 cycles, min=0.000011, max=14.907989, avg=0.067226, 2 latency excursions (> 10x expected cycle time of 0.001000s)
02:15 AM seb_kuzminsky: linuxcnc-build_: force build --branch=2.7 0000.checkin
02:15 AM linuxcnc-build_: build forced [ETA 5h21m57s]
02:15 AM linuxcnc-build_: I'll give a shout when the build finishes
03:51 AM linuxcnc-build_: Hey! build 0000.checkin #4943 is complete: Success [3build successful]
03:51 AM linuxcnc-build_: Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4943
04:55 AM emcPT: I am stuck at a midle of a job and I require some help. Should be simple.
04:55 AM emcPT: I have upload a image to https://ibb.co/hShyqk
04:55 AM emcPT: The problem is:
04:56 AM emcPT: I am making a internal thread
04:56 AM emcPT: and the tool is almost the size of the inside hole
04:56 AM emcPT: but it can machine the thread
04:56 AM emcPT: The problem is that the G76 cycle is retracting too much the tool
04:57 AM emcPT: and it does not respect the "I" value of the G76 cycle as described here: http://linuxcnc.org/docs/2.6/html/gcode/gcode.html#sec:G76-Threading-Canned
04:58 AM emcPT: For me, it makes sense that the retraction is always done to the same X value
04:58 AM emcPT: (X=diameter)
04:59 AM emcPT: The G76 used is on the image
04:59 AM emcPT: Any ideias?
05:21 AM emcPT: well I am making it with G33, but I still think that something is wrong with the G76
07:46 AM skunkworks: jepler, changing the video driver fixed my laptop - Would that fix the forum?
07:46 AM skunkworks: ;)
07:46 AM jepler: skunkworks: I'll do that right away
07:54 AM jepler: emcPT: as shown below, the tool doesn't return to the original "drive line" each time. Instead, it returns to different locations so that the entry move (green vertical move at right) always has the same length. I believe this helps ensure that the start of the thread is uniform; if the line changed lengths, then the uniformity at the start of the thread suffers.
07:54 AM jepler: emcPT: as you have identified, you have a need that is not served by G76, so you are right to choose G33.
07:55 AM jepler: (also I think the exit move is also made to be the same length every time)
07:55 AM jepler: The documentation could be improved, but based on the information at hand I think G76 is behaving as designed.
08:22 AM jepler: all that said, it's possible that there is a bug. If you still think there is, you should make a simple standalone gcode program like the one in our source tree at tests/interp/g76/g76only.ngc that you believe shows the problem, and open a bug report on github.
08:23 AM jepler: the existing test uses a negative i value for outside threads, so the positive i value for inside threads is not tested.
08:41 AM emcPT: the i positive and negative works. I use it very often and it works fine.I think that for obvious reasons that the X of the retract should be always to the same spot (to avoid internal collisions). But it can also be possible that it is behaving like it should.
08:56 AM kwallace2: emcPT, I rewrote the G76 code for PathPilot to fix the retract problem that you mentioned. My documentation is here: http://www.wallacecompany.com/t_tmp/G76_doc/G76.html . Hmm, it looks like this may not be totally up to date. My last change was to not only retract to the same line but the line follows the main taper if called for. So in Figure 7 segment 4 would be parallel to 2. Figure 8 shows that end tapers still might interfere with tight ba
08:57 AM kwallace2: ... tight back clearance.
09:01 AM kwallace2: The slightly out of date source: http://www.wallacecompany.com/t_tmp/G76_doc/g76kw_interp_convert.cc
09:22 AM emcPT: Yes, it seams much better. Any reason why this is not on the linuxcnc dev?
09:47 AM seb_kuzminsky: i can not ssh to forum.linuxcnc.org, and i can not connect to its DO console, so i'm power cycling it
09:47 AM jepler: seb_kuzminsky: ugh thanks
09:49 AM seb_kuzminsky: geez DO, how long does it take to yank the power cord out of a VM?
09:49 AM jepler: it's usually pretty quick
09:50 AM seb_kuzminsky: it's coming back now
09:51 AM jepler: OFFS
09:51 AM seb_kuzminsky: the last entry in the memory monitor log shows it 3.7 GB into swap
09:51 AM jepler: with their new monitoring, the info doesn't survive a power cycle !?
09:51 AM jepler: the "graphs" page is just empty
09:51 AM seb_kuzminsky: there are many apaches
09:52 AM jepler: my change was supposed to limit it to 25 apache processes, if I understood things right
09:52 AM seb_kuzminsky: 1 gig of mysqld
09:52 AM jepler: 75 apache2 processes
09:53 AM seb_kuzminsky: yeah, that agrees with the ps listing
09:53 AM jepler: no I mean that's how many I counted in the list
09:53 AM seb_kuzminsky: oh
09:53 AM jepler: but I thought I set the limit to 25
09:53 AM seb_kuzminsky: yeah me too
09:53 AM jepler: .. fresh after reboot mysql is VSZ 880772 RSS 90700
09:53 AM jepler: so I'm not sure how much attention to pay to its VSZ having grown modestly to 901520
09:55 AM jepler: ./conf-enabled/local.conf:MaxRequestWorkers 75
09:55 AM jepler: OK this 75 does match something I've configured apparently
09:56 AM jepler: the great thing about apache is... you can configure it so much
09:57 AM seb_kuzminsky: heh yeah
09:57 AM jepler: OK maybe it's actually restarted with a limit of 25 workers
09:57 AM seb_kuzminsky: 75 * 400MB >> 1 GB + 6 GB
09:59 AM jepler: I don't know how much to believe VSZ numbers
10:02 AM seb_kuzminsky: yeah, much of that should be shared, and at least some of it not paged in
10:03 AM seb_kuzminsky: between 19:25:30 and 19:26:30 the system added a ton of apache processes, and that's also when the last entry was written to access.log
10:04 AM seb_kuzminsky: the machine stayed up until 19:51, but then "locked up"
10:05 AM jepler: - - [09/May/2017:19:25:36 -0400] "GET /media/kunena/attachments/22936/C10-C41-KBIC120.jpg HTTP/1.0" 200 620202 "-" "-"
10:06 AM jepler: I wonder why the machine requests its own images
10:06 AM seb_kuzminsky: yeah, many of the connections during that time are internal, to attachment urls
10:07 AM seb_kuzminsky: i dont know if that's normal or not
10:07 AM jepler: - - [07/May/2017:06:44:54 -0400] "GET /media/kunena/attachments/1749/Selection_002_2013-11-17.png HTTP/1.0" 200 74907 "-" "-"
10:08 AM seb_kuzminsky: the number of connections during the partial 19:2* 10-minute window doesn't look unusually high, compared to some other 10-minute windows leading up to there
10:13 AM jepler: it looks like because stupid, php/joomla will fetch an image via URL to determine its size
10:13 AM seb_kuzminsky: the number of connections from forum.linuxcnc.org and from also seem about average just ebfore the crash
10:14 AM seb_kuzminsky: you reduced max workers and bounced apache? I'm tempted to restart the memory logger and walking away from this until the next crash
10:14 AM jepler: yes please start the memory logger
10:15 AM seb_kuzminsky: done
10:15 AM seb_kuzminsky: the old log i renamed to .0, the new log is accumulating now
10:15 AM seb_kuzminsky: bbl
10:17 AM jepler: digitalocean's alerts are useless, digitalocean's new stats reporting is useless
10:18 AM jepler: the new stats reporting requires userspace to be alive, so it never reached the sustained CPU usage threshold I set
10:32 AM KGB-linuxcnc: 03Jeff Epler 05master 699f6b0 06linuxcnc 10src/emc/rs274ngc/interp_convert.cc 10src/emc/rs274ngc/interp_python.cc 10src/emc/rs274ngc/rs274ngc_interp.hh 10src/emc/rs274ngc/rs274ngc_pre.cc Merge branch 'remap-io-lcnc-2.6' of https://github.com/zultron/machinekit * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=699f6b0
10:32 AM KGB-linuxcnc: 03Jeff Epler 05master 9e4641a 06linuxcnc 10tests/remap/remap-io/core_sim.hal 10tests/remap/remap-io/test-ngc.ini 10tests/remap/remap-io/test-py.ini update test configs for master branch * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9e4641a
10:37 AM kwallace2: emcPT, there hasn't been quite enough roundtoit to get the G76 update pushed over the finish line (polish the code, testing, docs).
10:50 AM emcPT: Thank you. I will take a look into it and ask you if I have questions during testing.
10:51 AM emcPT: The part that I was making is here: https://youtu.be/70HlJlqCvSA
10:59 AM cradek: square threads? or acme?
11:00 AM cradek: sounded terrible but it worked...
11:01 AM KGB-linuxcnc: 03Jeff Epler 05master 1a1d958 06linuxcnc 10src/emc/usr_intf/pncconf/pncconf.py Merge branch 'issue-159' of https://github.com/jepler/linuxcnc * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=1a1d958
11:14 AM emcPT: The sound is not from the thread, but the axis moving.
11:15 AM emcPT: Too much stif
11:15 AM emcPT: Following error very,very low
11:15 AM emcPT: Like 0.02mm in rapids of 20000m/min
11:16 AM emcPT: " The sound is not from the thread, but the axis moving." this is, if I machine air, it will make the same noise
11:19 AM emcPT: square thread
11:24 AM emcPT_ is now known as emcPT
12:52 PM emcPT_ is now known as emcPT
03:27 PM jepler: sigh, javascript style creeping into AXIS
03:27 PM jepler: if type(i) != type(0): return
03:32 PM seb_kuzminsky: it's what the kids today use instead of tcl/tk style :-P
03:33 PM Tom_L: what do they know?
03:34 PM pcw_mesa: LinuxCNC seem to work fine on Bunsenlabs Linux...
08:19 PM andypugh: I am trying to find a thread from emc-dev on the soucreceforge/mailman archive to reference in a forum answer. And I just can’t find it.
08:19 PM andypugh: Post title “[Emc-developers] serious PCIe parallel port problem"
08:19 PM andypugh: Any search-fu clues?
08:21 PM jepler: > Your search - "serious PCIe parallel port problem" site:*.sourceforge.net - did not match any documents. Reset search tools but google finds other sites with archives such as http://www.mail-archive.com/emc-developers@lists.sourceforge.net/msg16095.html
08:22 PM andypugh: Thanks.
08:22 PM andypugh: And here is the post: http://www.mail-archive.com/emc-developers@lists.sourceforge.net/msg16090.html
08:23 PM andypugh: I wonder what is going one there with the archive search?
08:36 PM Roguish: does this look a little familiar? https://stepcraft.us/product/uccnc-control-software-oem/
08:43 PM KGB-linuxcnc: 03Chris Edwards 05master af98ca0 06linuxcnc 10docs/src/remap/remap.txt Fix minor spelling and punctuation errors * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=af98ca0
08:43 PM KGB-linuxcnc: 03Chris Edwards 05master 8654564 06linuxcnc 10docs/src/remap/remap.txt Edit existing link to FEATURES section to use NOTE admonition and FEATURES (not FEATURE) label * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=8654564
08:43 PM KGB-linuxcnc: 03Chris Edwards 05master 5df27d3 06linuxcnc 10docs/src/config/ini-config.txt Explain how to use custom vars from G-code * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5df27d3
08:43 PM KGB-linuxcnc: 03Chris Edwards 05master 531e09b 06linuxcnc 10docs/src/config/ini-config.txt Add FEATURES to list of RS274NGC section variables; harmonise bullet list format * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=531e09b
08:43 PM KGB-linuxcnc: 03Chris Edwards 05master 145dbd2 06linuxcnc 10docs/src/config/ini-config.txt Correct cross-reference markup for Remap:FEATURES link * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=145dbd2
08:43 PM KGB-linuxcnc: 03Chris Edwards 05master 774890f 06linuxcnc 10docs/src/config/ini-config.txt Reword cross-ref text for better readability if in print form * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=774890f
08:43 PM KGB-linuxcnc: 03Chris Edwards 05master a56f4a3 06linuxcnc 10docs/src/config/ini-config.txt In INI-file variable docs, mention Z axis and move feedrate to G38 block; amend/add cross-references * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a56f4a3
08:51 PM KGB-linuxcnc: 03Jeff Epler 05master dcf77d4 06linuxcnc Merge branch 'master' of https://github.com/LearningLinuxCNC/linuxcnc * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=dcf77d4
09:10 PM KGB-linuxcnc: 03Jeff Epler 05master caab9ae 06linuxcnc 10(15 files in 7 dirs) Merge branch 'tseufl-fix-docs/2.7-add-img-alt' of https://github.com/tseufl/linuxcnc-mirror * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=caab9ae
09:10 PM KGB-linuxcnc: 03Jeff Epler 05master bdc4ec0 06linuxcnc 10docs/src/drivers/gm.txt 10docs/src/gui/tklinuxcnc.txt docs: don't use ampersands in alt= text * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=bdc4ec0
09:10 PM KGB-linuxcnc: 03Jeff Epler 05master 7adb6dc 06linuxcnc 10docs/src/drivers/gm.txt docs: add missing = * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=7adb6dc
09:10 PM KGB-linuxcnc: 03Jeff Epler 05master c8747e4 06linuxcnc 10docs/src/hal/tools.txt docs: fix build error when image alt text is multiline * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c8747e4
09:21 PM KGB-linuxcnc: 03Jeff Epler 05master e68ab24 06linuxcnc 10docs/src/checkref doc build: don't print "***" when all is well * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=e68ab24
09:29 PM KGB-linuxcnc: 03Jeff Epler 05master 5286664 06linuxcnc 10src/emc/rs274ngc/gcodemodule.cc gcodemodule: arbitrary number of preamble godes * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5286664
09:29 PM KGB-linuxcnc: 03Jeff Epler 05master 596363b 06linuxcnc 10lib/python/rs274/glcanon.py glcanon: allow old and new forms of arguments for gcode.parse * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=596363b
09:29 PM KGB-linuxcnc: 03Jeff Epler 05master 001eb00 06linuxcnc 10src/emc/usr_intf/axis/scripts/axis.py axis: set up most parts of interpreter state based on info from task * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=001eb00
09:40 PM cradek: wow
09:43 PM Tom_L: busy bot
09:44 PM Tom_L: i should come up with a way to ignore it in the logs or is it a helpful thing to have in them?
09:50 PM jepler: Tom_L: I don't have a strong feeling that KGB- needs to be logged
09:50 PM jepler: or if you mean -linuxcnc-github- I likewise don't have a strong feeling
09:50 PM Tom_L: both
09:51 PM Tom_L: the information is there to see anyway isn't it?
09:51 PM jepler: yeah
09:54 PM Tom_L: is -linuxcnc-github: a channel message or user?
10:08 PM jepler: 21:38:27 [freenode] -!- linuxcnc-github [~linuxcnc-@]
10:08 PM jepler: 21:38:27 [freenode] -!- was : GitHub IRCBot - LinuxCNC/linuxcnc
10:08 PM jepler: 21:38:27 [freenode] -!- server : moon.freenode.net [Thu May 11 02:00:20 2017]
10:08 PM jepler: 21:38:27 [freenode] -!- End of WHOWAS
10:09 PM jepler: linuxcnc-github is a user and I think it's sending a NOTICE in IRC parlance but that lingo may also be 10 years out of date
10:33 PM jepler: hoo-boy I do not know how to make gscreen do things
10:37 PM KGB-linuxcnc: 03Jeff Epler 05master 71566e2 06linuxcnc Merge branch 'tseufl-bugfix-glade/2.7-9_axis' of https://github.com/tseufl/linuxcnc-mirror * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=71566e2