#emc-devel | Logs for 2008-10-21

[01:11:16] <fenn_> fenn_ is now known as fenn
[08:20:00] <CIA-37> EMC: 03tissf 07TRUNK * 10emc2/docs/src/config/ini_config_fr.lyx: french translation update
[08:20:10] <CIA-37> EMC: 03tissf 07TRUNK * 10emc2/docs/src/index_fr.tmpl: french translation update
[10:39:38] <BigJohnT> I just noticed on my 8.04 install that the nc_files folder did not get put in and the default for file open is /configs/myconfig instead of /nc_files as in 6.06 Is this a change?
[11:23:13] <jepler> no; 8.04 and 6.06 should behave the same in this respect
[11:23:31] <jepler> how did you copy/create your configuration?
[11:23:48] <BigJohnT> livecd
[11:24:24] <BigJohnT> then I did a stepconf to create a sample config
[11:27:05] <CIA-37> EMC: 03tissf 07TRUNK * 10emc2/docs/src/common/Diagnostics_fr.lyx: french translation fix some lyx errors and update
[11:27:07] <CIA-37> EMC: 03tissf 07TRUNK * 10emc2/docs/src/ (5 files): french translation fix some lyx errors and update
[11:27:08] <CIA-37> EMC: 03tissf 07TRUNK * 10emc2/docs/src/config/lathe_config_fr.lyx: french translation fix some lyx errors and update
[11:27:10] <jepler> it looks like stepconf doesn't know to create the nc_files directory or the symlink to the samples
[11:27:27] <BigJohnT> ok
[11:31:01] <micges> when I'm moving axis x in + direction I have different vel shown in axis preview and in pin axis.0.joint-vel-cmd
[11:31:30] <micges> current vel 5000 and joint vel 4100
[11:33:46] <jepler> it must be config specific
[11:34:02] <jepler> I see e.g., 72 (ipm) and 1.2 (ips) using configs/sim/axis.ini and adjusting jog speed slider to 72.
[11:34:43] <micges> ok
[11:36:57] <micges> traj: MAX_VELOCITY = 200
[11:37:10] <micges> axis_0: MAX_VELOCITY = 134
[11:37:36] <BigJohnT> If I create a config from the configuration selector it creates the nc_files directory and and symlink
[11:40:06] <CIA-37> EMC: 03bigjohnt 07v2_2_branch * 10emc2/docs/src/hal/pyvcp.lyx: replace single quotes with double quotes for font
[11:40:13] <CIA-37> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/stepconf/stepconf.py: create nc_files directory and examples symlink if necessary
[11:40:13] <jepler> BigJohnT: ok that helps confirm that it's stepconf, not 8.04.
[11:40:28] <jepler> thanks
[11:41:04] <BigJohnT> jepler: dang your fast :)
[11:41:32] <CIA-37> EMC: 03bigjohnt 07TRUNK * 10emc2/docs/src/hal/pyvcp.lyx: replace single quotes with double quotes for font
[11:43:54] <jepler> BigJohnT: in the e-mail you sent to the developers' list it looks like the error is no close paren on the <font> line
[11:44:18] <BigJohnT> that might have been a typo when I typed it
[11:44:33] <BigJohnT> when I change the single quotes to double all work well
[11:44:52] <jepler> did you cut & paste it to make your xml file, or did you type it?
[11:45:22] <BigJohnT> I cut & paste it from the integrators manual
[11:45:42] <jepler> from the html or from the pdf?
[11:45:51] <micges> jepler: when I'm sending mdi command while emc isn't in coord mode there is an error shown and after that emc is in undefined state (jog doesn't work, exec_state is fast change between error and done)
[11:45:55] <BigJohnT> I just looked at the e-mail and it has a closed paren
[11:46:02] <BigJohnT> pdf jepler
[11:48:57] <BigJohnT> When I copied from the pdf it was not a single quote but some other thing like an accent mark
[11:48:59] <jepler> BigJohnT: the problem is that lyx put a RIGHT SINGLE QUOTATION MARK character (a kind of smart quote character), while Python only accepts APOSTROPHE for strings
[11:49:22] <BigJohnT> yep, when I typed it in I saw the difference
[11:50:06] <jepler> there was also a problem like this for the double quotes; I had to find a special way to enter the actual QUOTATION MARK character and avoid getting LEFT DOUBLE QUOTATION MARK or RIGHT DOUBLE QUOTATION MARK characters
[11:50:22] <BigJohnT> it is Ctrl shift "
[11:50:48] <BigJohnT> shift " gives the left right thing
[11:51:37] <BigJohnT> so it was an lyx typo :)
[11:52:23] <jepler> yeah I can't immediately determine if there's something to type to get APOSTROPHE and not RIGHT SINGLE QUOTATION MARK
[11:52:39] <BigJohnT> I'm just looking for that too
[11:53:00] <jepler> but for most Python code, " and ' are interchangeable so what you did is probably OK
[12:09:22] <jepler> micges: I can also see that problem with mdi. I filed a bug.
[12:09:32] <jepler> bbl
[12:13:35] <micges> this state is also when I'm change soft limit so that machine is not in the table
[12:14:57] <jepler> don't do that, then
[12:15:21] <jepler> a patch to reject changing the soft limit to not include the current point might be a good idea
[12:19:48] <micges> indeed
[12:23:27] <jepler> anyway, like I said ten minutes ago, bbl
[12:23:28] <jepler> :-P
[12:23:59] <micges> heh
[13:32:56] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[16:00:52] <CIA-37> EMC: 03cradek 07TRUNK * 10emc2/src/emc/task/emctaskmain.cc:
[16:00:52] <CIA-37> EMC: fix bug 2184280 and probably also 1734318 and 1701792. This was a bug with
[16:00:52] <CIA-37> EMC: motion not resetting its error flag if in teleop or free (trivkins manual)
[16:00:52] <CIA-37> EMC: modes. Long ago this caused "abort storms" and someone suppressed the "storm"
[16:00:52] <CIA-37> EMC: in task but did not fix the original problem where motion was still left in an
[16:00:54] <CIA-37> EMC: error state (and therefore is triggering error handling behavior in task over
[16:00:56] <CIA-37> EMC: and over).
[16:00:58] <CIA-37> EMC: 03cradek 07TRUNK * 10emc2/src/emc/motion/command.c:
[16:01:00] <CIA-37> EMC: fix bug 2184280 and probably also 1734318 and 1701792. This was a bug with
[16:01:02] <CIA-37> EMC: motion not resetting its error flag if in teleop or free (trivkins manual)
[16:01:04] <CIA-37> EMC: modes. Long ago this caused "abort storms" and someone suppressed the "storm"
[16:01:08] <CIA-37> EMC: in task but did not fix the original problem where motion was still left in an
[16:01:10] <CIA-37> EMC: error state (and therefore is triggering error handling behavior in task over
[16:01:12] <CIA-37> EMC: and over).
[16:02:24] <jepler> cradek: ooh thanks for looking into that
[16:02:43] <jepler> cradek: I noticed it's possible to MDI "M3 S100" while in manual mode...
[16:02:50] <jepler> this might be considered a feature, I dunno
[16:09:21] <alex_joni> whee.. cradek is messing around emctaskmain :)
[16:13:15] <jepler> Revision 1.11: download - view: text, markup, annotated - select for diffs
[16:13:18] <jepler> Mon Aug 16 00:26:08 2004 UTC (4 years, 2 months ago) by jmkasunich
[16:13:19] <jepler> first partly working version of EMC2 - runs 3D_Chips.ngc correctly with steppers
[16:13:30] <jepler> whee, that's been awhile
[16:13:41] <alex_joni> right
[16:13:52] <alex_joni> and 2.0.0 1-2 years later?
[16:14:37] <alex_joni> 2006.05.08 - RELEASE_2_0_0
[16:16:50] <alex_joni> jepler: is CVS high on cpu usage?
[16:17:31] <alex_joni> load averages: 0.64, 0.38, 0.25
[16:17:45] <cradek> hm, the taskAborted stuff I took out was in emc1 cvs version 1.1
[16:17:47] <alex_joni> hmm, not extensively (was waitng for a tarball download)
[16:18:01] <alex_joni> cradek: now that's a bit dated :)
[16:18:10] <cradek> I guess we'll eventually see if that is problematic
[16:18:18] <cradek> I figured it was much later
[16:18:26] <jepler> alex_joni: no, but my dsl is sluggish for interactive stuff
[16:18:36] <cradek> * cradek curses the use of indent by people far in the past
[16:18:39] <alex_joni> remember there were a couple years of devel before 1.1 in cvs
[16:19:04] <SWPadnos> several
[16:19:13] <alex_joni> several couples?
[16:19:14] <SWPadnos> like maybe 5 years
[16:19:42] <alex_joni> depends how you count them, might be even more
[16:19:53] <SWPadnos> yep
[16:20:06] <SWPadnos> I think the first public exposure of the code was in 1997 or 1998
[16:20:15] <SWPadnos> but it had been worked on at NIST for some time before that
[16:22:11] <alex_joni> Proctor, F. M., and Michaloski, J., "Enhanced Machine Controller Architecture Overview," NIST Internal Report 5331, December 1993.
[16:23:06] <SWPadnos> yep. I think that was a description of what they wanted to accomplish though, the proof-of-concept (EMC) wasn't made until later
[16:24:04] <alex_joni> counts as a start though :)
[16:25:25] <alex_joni> http://www.isd.mel.nist.gov/documents/shackleford/4191_05.pdf
[16:26:07] <alex_joni> "In 1992, following the NML work, NIST began the Enhanced Machine Controller (EMC) program under funding from the U.S. Navy to investigate standards issues in open architecture controller for machien tools."
[16:41:17] <cradek> jepler: I doubt that's a feature...
[16:43:51] <jepler> cradek: if it's not, then refusing earlier to execute the mdi command seems like the winning strategy
[16:44:08] <cradek> I agree
[16:47:08] <CIA-37> EMC: 03cradek 07TRUNK * 10emc2/src/emc/task/emctaskmain.cc: refuse MDI commands if not in MDI mode
[16:48:19] <cradek> this fix would have covered up the other bug...
[17:02:10] <alex_joni> http://eneas.juve.ro/~juve/stats2/
[17:02:19] <alex_joni> fresh stats, get them while they're hot
[17:03:00] <alex_joni> BigJohnT made it in the top 10 :)
[17:05:24] <jepler> yay I'm #2
[17:06:01] <jepler> those must be number of surviving lines, not number of committed lines
[17:06:57] <jepler> or maybe not -- I guess the sum of lines of code by author is more than "total lines of code"
[17:07:03] <alex_joni> jepler: the number of commits is more appropriate
[17:07:08] <alex_joni> and there you're #1 by far ;)
[17:07:32] <alex_joni> jepler 4350
[17:07:37] <alex_joni> jmkasunich 2168
[17:07:46] <alex_joni> alex_joni 2044
[17:07:50] <alex_joni> cradek 1421
[17:07:57] <alex_joni> paul_c 1409
[17:08:48] <alex_joni> for jmk and fenn it counts dxf's and such as loc's
[17:09:52] <SWPadnos> so if you commit single-line or single character changes, you get better stats
[17:10:07] <cradek> haha http://eneas.juve.ro/~juve/stats2/directory_sizes.png
[17:10:25] <SWPadnos> code seems a little underrepresented, doesn't it? :)
[17:10:50] <jepler> cradek: that graph is not terribly useful
[17:11:21] <SWPadnos> hmmm. how did we end up with an extra 150000 lines of code in late 2006/early 2007?
[17:11:28] <SWPadnos> is that docs or something?
[17:11:35] <jepler> SWPadnos: yes I think it was the docs
[17:11:47] <alex_joni> yup
[17:11:51] <SWPadnos> there's also a 100k bump in August 06
[17:11:57] <alex_joni> that's when I started moving stuff from docs/ to emc2/docs/src
[17:12:03] <SWPadnos> ah
[17:12:46] <alex_joni> the 150k bump
[17:13:00] <alex_joni> the first bump might have been some indent flurry or such iirc
[17:13:13] <SWPadnos> well, that's LOC, not commits (isn't it?)
[17:14:14] <alex_joni> right
[17:14:42] <SWPadnos> it looks like those may be Eric Johnson
[17:14:54] <SWPadnos> but I can't really see the colors well in the LOC graph by contributor
[17:17:09] <alex_joni> same here :)
[17:17:24] <alex_joni> in aug. 2006 the main loc's were: docs, AXIS part of emc23
[17:17:30] <alex_joni> and more docs
[17:18:14] <SWPadnos> right - I'm just noticing that AXIS was added then (for a 23565 LOC increase)
[17:18:28] <alex_joni> the second bump was jmk adding dxf's to docs
[17:18:37] <alex_joni> 120k loc's
[17:19:09] <SWPadnos> and 50k for docs in August - that more or less explains the 100k bump there :)
[17:19:41] <alex_joni> yup
[17:21:25] <alex_joni> I wonder how it counts lines inside png's :)
[17:21:35] <alex_joni> Section_Display_Build01.png (new 144), Section_Display_Build02.png (new 158), Section_Display_Build03.png (new 178), Section_Display_Grid.png (new 126)
[17:40:27] <BigJohnT> mmm bbq
[17:53:36] <skunkworks> so - what kind of time constant should the filter be for the current limit? I am thinking around 100khz?
[17:55:08] <BigJohnT> doing a mold rotation is fun with the anilam :)
[17:55:40] <skunkworks> right now it is 1mhz.
[17:55:51] <skunkworks> (1k and .001uf)
[17:56:37] <skunkworks> I am getting a little different current trip + to -. Seems to smooth out with a .01uf across it (100k)
[17:56:47] <skunkworks> 100khz)
[18:21:14] <BigJohnT> http://www.cnczone.com/forums/showthread.php?t=66530
[18:23:47] <alex_joni> nice
[18:24:49] <jepler> earlier I wondered how long a cvs checkout takes over the internet .. I timed it as 3:25
[18:27:20] <alex_joni> something like that..
[18:27:51] <SWPadnos> I think if we all check at the same time, we may skew the results ;)
[18:27:52] <jepler> here's a different view of the size of directories & files in emc2 cvs: http://emergent.unpy.net/files/sandbox/directory_sizes.png
[18:27:56] <jepler> (fresh checkout)
[18:28:38] <BigJohnT> cool nesting software
[18:29:05] <jepler> old project of mine: http://unpy.net/~jepler/tkdu/
[18:29:44] <BigJohnT> cool, bbl
[18:29:47] <SWPadnos> there's a similar one included with Ubuntu 8.04
[18:30:21] <jepler> oh really?
[18:30:34] <SWPadnos> I think so - somewhere in the admin apps
[18:30:54] <SWPadnos> it may be something I installed though - I'll let you know when I next fire up my laptop
[18:31:09] <alex_joni> SWPadnos: http://www.ghielectronics.com/products.php
[18:31:11] <SWPadnos> there are nearly identical programs available for Windows and Linux also
[18:31:37] <SWPadnos> alex_joni, ewwww. .net :P
[18:32:03] <SWPadnos> oh, they sell the eye-fi
[18:32:06] <SWPadnos> cool
[18:32:20] <SWPadnos> hmmm - or it's based on one of their chips
[18:45:44] <alex_joni> jepler: very interesting to look at
[18:46:15] <alex_joni> the "actual" emc source is quite small
[18:46:21] <alex_joni> compared to drivers and docs
[18:46:51] <alex_joni> and infrastructure (configs, nc-files, debian stuff, tests)
[18:47:45] <SWPadnos> I was surprised at the size of debian/
[18:47:57] <SWPadnos> 800k seems like a lot for that (though I don't know exactly what "that" is)
[18:48:05] <alex_joni> well.. it contains lots of icons and links and whatnot
[18:48:29] <SWPadnos> ok, icons I can imagine. links/menu entries are pretty small
[18:48:45] <SWPadnos> or should be anyway
[18:49:45] <alex_joni> 300K extras-sim-Ubuntu-5.10
[18:50:05] <alex_joni> same for the rest (7 in total)
[18:51:00] <alex_joni> hmm.. that seems odd
[18:51:21] <alex_joni> I think there's a problem with du
[18:51:39] <alex_joni> I get about 100k for debian/
[18:52:29] <alex_joni> but running du:
[18:52:29] <alex_joni> juve@proxy:~/emc2$ du -s -h debian
[18:52:30] <alex_joni> 2.1M debian
[18:54:34] <alex_joni> ah, du counts 4k for each file and dir
[18:54:38] <alex_joni> even if < 4k
[18:55:28] <alex_joni> juve@proxy:~/emc2$ du -s --apparent-size -h debian
[18:55:29] <alex_joni> 804K debian
[18:57:24] <SWPadnos> hey LawrenceG, can you easily add a checkbox to engrave.py to suppress output of "common code"? (like subroutines and preamble/postamble)
[19:01:05] <LawrenceG> SMOP....possible.... the pre and post amble are easy to clear during normal use..... advanced users have no problem with edits to gcode.... new users could just be more confused with another checkbox...
[19:03:21] <alex_joni> make it hidden :D
[19:03:39] <alex_joni> ctrl+alt+make the checkbox visible
[19:03:48] <alex_joni> to make it visible :)
[19:09:35] <LawrenceG> thanks alex.... the secret 1 pixel high checkbox.... send $19.95 to enable!
[19:11:09] <alex_joni> reminds me of minesweeper
[19:12:08] <LawrenceG> ahh... now I got it
[19:12:50] <alex_joni> heh, seems you know what I'm talking about :P
[19:12:52] <alex_joni> cheater :P
[19:13:25] <LawrenceG> I hate getting blown up!
[19:16:04] <alex_joni> I managed 3s on beginner :)
[19:16:12] <alex_joni> but that was quite a while ago
[19:26:58] <BigJohnT> hi LawrenceG
[19:27:41] <LawrenceG> Hi John... what is cooking in the USA? venison burgers?
[19:30:53] <BigJohnT> BBQ today
[19:31:57] <LawrenceG> http://pastebin.ca/1233097
[19:32:01] <BigJohnT> we need to invent software that knows what you want to do via esp
[19:33:38] <LawrenceG> yea.... its too bad that one still has to have a basic understanding of what you want it to do to make it work
[19:33:54] <alex_joni> esp?
[19:33:57] <BigJohnT> too bad you can't have pistols in
[19:34:00] <BigJohnT> Canada
[19:34:25] <BigJohnT> extra sensory perception
[19:35:12] <LawrenceG> thats only a problem if you try and register them :}.... pretty soon it will be OK to do the crime as long as you are certified
[19:38:40] <BigJohnT> well time to go see you guys later
[19:38:52] <jepler> see you BigJohnT
[19:39:41] <LawrenceG> good bbq
[19:40:45] <alex_joni> LawrenceG: sounds like the assassin's guild :P
[19:42:20] <LawrenceG> yes, the job world has changed a lot here in the last 10 years...... if you can pass an internet based multiple choice open book exam, you can be "certified" as an expert
[21:04:16] <skunkworks> hmm - some more odd issues
[21:14:23] <skunkworks> heh - my reference ground to the comparator is movine over 50mv..
[21:14:26] <skunkworks> moving
[21:14:39] <skunkworks> that can't be good ;)
[21:14:56] <skunkworks> only when I move the shaft 1 direction
[21:29:00] <skunkworks> http://www.electronicsam.com/images/KandT/servostart/boardagain.png
[21:30:26] <skunkworks> the large trace - 2 up from pad 4. I can measure a fluctuation in the ground between where the 7 watt resistor is soldered in to logic ground and pin 4 of the lm393 as much as 100mv.
[21:30:46] <skunkworks> which suprises the hell out of me.
[21:30:49] <skunkworks> :)
[21:32:00] <SWPadnos> using difference mode on the scope?
[21:33:38] <SWPadnos> uou may need plated through holes for this - you should have some ties between the shield planes on the two sides
[21:33:57] <SWPadnos> assuming they're all ground *or* Vsupply :)