#emc | Logs for 2008-03-03

[00:57:19] <JymmmEMC> So, are there going to be any other emc releases in the next 90 days, or is 2.2.2 what should be installed?
[01:04:47] <jmkasunich> I think we're at 2.2.3
[01:05:07] <SWPadnos> I think the liveCD may be 2.2.2, so an update is needed after install
[01:05:15] <SWPadnos> I could be wrong about that though
[01:05:17] <jmkasunich> anyway, if you install 2.2.something, upgrades in the 2.2 series are automatic and won't break your configs
[01:05:33] <jmkasunich> 2.3 is the next major revision, and will involve some config changes, etc
[01:05:45] <jmkasunich> it will probably be some time in the summer
[01:06:06] <SWPadnos> that won't be until the end of April at the soonest, if current ideas/plans come to fruition
[01:06:21] <SWPadnos> mainly that the new liveCD will be based on Ubuntu 8.04
[01:06:38] <SWPadnos> which should be released April 24 (Ubuntu, not EMC/8.04)
[01:07:37] <jmkasunich> if you are thinking of setting up a PC for EMC and aren't in a hurry, its tempting to wait for ubuntu 8.04
[01:28:59] <cradek> I disagree
[01:29:27] <cradek> doubt I will trust 8.04 like I trust 6.06 before summer
[02:01:56] <jepler> barring the unforeseen, the whole 2.2.x release series of emc2 will get official packages for ubuntu 6.06. You don't have to worry that you can't get emc2 bugfixes without reinstalling your OS.
[02:45:58] <steves_logging> steves_logging is now known as steve_Stallings
[03:12:01] <tomp> alex_joni: thanks for the info on M66, i was looking for 'current' value of #1. M66 doesnt know about # variables, from what i read in the docs. but maybe the code for M66 is a framework to look at.
[03:17:23] <tomp> how do i find the code for M66? everytime i look for something, i end up wading thru everything!
[03:20:53] <eric_U> I always find things in the last place I look
[03:22:36] <tomp> grep only shows m66 in interp_array.cc, and there it only is identified in a group, not the code executed or the message created by it.
[03:27:32] <jepler> I'd search for '66' in the rs274ngc directory; it gets a few false hits but something there should lead you to the relevant code
[03:36:30] <tomp> thx, i find some in interp_convert, but stymied by some ?macros? CHK(blah), convoluted macros defined in hdrs.... been here before , almost unreadable
[03:36:47] <cradek> it used to be much worse :-)
[03:38:12] <cradek> *left SET_TO (((*left IS 0.0) AND (*right ISNT 0.0)) OR
[03:38:13] <cradek> ((*left ISNT 0.0) AND (*right IS 0.0))) ? 1.0 : 0.0;
[03:38:17] <cradek> hahahaha
[03:38:54] <jepler> hm, why didn't they have a define for the ternary operator ... say, #define IFELSE_EXPR(a,b,c) ((a)?(b):(c))
[03:39:22] <cradek> and ADDITION, SUBTRACTION
[03:39:49] <SWPadnos> hmmm. don't make me find that web page again
[03:40:01] <jepler> yes, judicious use of macros could have settled whether it was pronounced TIMES or MULTIPLIED_BY
[03:40:03] <SWPadnos> the one that describes all the mathematical / logical operations
[03:40:12] <SWPadnos> ASTERISK
[03:40:14] <SWPadnos> :)
[03:40:23] <tomp> yes, i remember fighting thru those 'aids'
[03:40:26] <cradek> I wonder how you could differentiate between NEGATION and SUBRACTION
[03:40:55] <tomp> lotsa CAPS
[03:42:36] <cradek> I have to say the formatting was pretty before someone porked it all up with indent
[03:42:53] <cradek> aside from SET_TO it's quite easy to read
[03:46:10] <tomp> " CHK(((block->p_flag == ON) && (block->e_flag == ON)),NCE_BOTH_DIGITAL_AND_ANALOG_INPUT_SELECTED);"
[03:46:10] <tomp> no i dont get it, i see condition , comma, msgthatis#defsassomeinteger, but no action! wheres the code?
[03:46:43] <cradek> if condition is true, throw that error
[03:46:59] <cradek> the code is somewhere following #define CHK
[03:47:25] <cradek> if you have an editor with tags, just go to the CHK tag to find the definition
[03:48:32] <cradek> emacs: put the cursor over CHK, M-., enter
[03:50:40] <jepler> you probably have to run "make tags" in the emc2/src directory before that works, and that requires extra program(s) to be installed (ctags-exuberant for vi-style tags, and etags for emacs-style tags)
[03:51:31] <tomp> no editor like that, using kate, but i do see the code finally. you're right its just past. big mental block on those macros for many years now. thx
[03:52:14] <tomp> those macros actually stopped me for many years, never got to next line.
[03:52:38] <cradek> I'm sure kate can use tags if it's a programmer's editor
[03:53:00] <tomp> did the authors HATE c? or being masonic?
[03:53:38] <eric_U> they're special
[03:53:45] <cradek> well that's C even if you think it's ugly C
[03:53:46] <jepler> "Your search - site:kate-editor.org ctags OR etags - did not match any documents." -- google
[03:53:55] <cradek> jepler: pfft
[03:54:59] <jepler> http://www.kde-apps.org/content/show.php/Kate+CTags+Plugin?content=47743
[03:55:48] <tomp> thx
[03:55:48] <cradek> jepler: vim doesn't seem to do that upward search like emacs does
[03:56:07] <jepler> cradek: oh, you mean to find a tag in the same file?
[03:56:27] <cradek> no, to find the tags file in parent dirs
[03:56:29] <jepler> oh
[03:57:13] <jepler> :tag works when you start from emc2/src which I must do by habit
[03:57:44] <cradek> ah
[04:04:17] <tomp> i used to write to Borland telling them their program was notc ( an audio pun )
[04:08:19] <steve_Stallings> steve_Stallings is now known as steves_logging
[04:21:45] <SWPadnos> does anyone else remember the sound example in the Borland C manual? (or it could have been Turbo Pascal - I forgot)
[04:22:15] <SWPadnos> they generate a 3Hz tone, and there's a long comment explaining that this is the resonant frequency of a chicken's head
[04:31:40] <SkinnYPuppY> 3HZ, no wonder chicken heads don't chime
[05:16:18] <JymmmEMC> SWPadnos: I probably have both around here somewhere
[05:16:49] <JymmmEMC> SWPadnos: manuals and software for Borland C v3or4 and Turbo Pascal 7
[06:23:42] <scutsxg> still quiet here?
[16:25:18] <micges> hello all
[16:44:30] <jymm> any suggestion on a DWG to DXF converter?
[16:45:14] <cradek> autocad
[16:50:52] <jymm> anything else?
[16:53:47] <cradek> not that I know of
[16:57:03] <jymm> thanks. I found a utility, but it's converting with errors.
[16:57:54] <SWPadnos> it's probably better to convert without errors
[16:57:56] <cradek> do you know what autocad version your dwg is? if it's a one-time deal (and mine can read it) I can convert it for you
[16:58:38] <cradek> anything other than autocad that reads dwgs is "guessing" since it's autocad's proprietary format. dxf is their equivalent open format.
[16:59:01] <archivist> and dxf is broken enough
[16:59:30] <cradek> dxf is fine (especially compared to exchanging stuff in dwg)
[16:59:54] <archivist> heh I had an unreadable dxf the other day
[17:00:03] <SWPadnos> dxf is like G-code - no two programs/people use it quite the same way
[17:00:13] <archivist> turbocad output
[17:00:29] <jymm> Oooops I take that back, it did convert it
[17:06:24] <micges> archivist: what mean unreadable ?
[17:06:30] <archivist> solidworks could not import
[17:07:00] <micges> do you have that file?
[17:07:19] <archivist> I also had a jpeg sent so no need
[17:07:20] <micges> will be very helpfull to my dxf->ngc converter
[17:08:01] <archivist> was a gear assembly drawing for a quote
[17:08:41] <micges> ok
[17:09:31] <archivist> I will be interested in dxf->ngc soon whe I finish my mill
[17:09:37] <archivist> when
[17:09:46] <jymm> ***** http://opendwg.org/ has a (M$) DWG viewer and converter utilities
[17:10:28] <jymm> My mistake on the converter, it defaults to auditing the file, which you have to uncheck to convert
[17:51:19] <ChanServ> [#emc] "This is the #emc channel - talk related to the Enhanced Machine Controller and general machining. Website: http://www.linuxcnc.org/, wiki at http://wiki.linuxcnc.org/"
[17:51:31] <cradek> I think jummm has turbocad, I don't
[17:51:35] <cradek> y
[17:52:39] <micges> jymm: could you send me some dxf's ?
[17:55:10] <micges> cradek: have any testing files like yours ARC-TEST.DXF ?
[17:56:31] <cradek> I'll make you something, one sec
[17:58:27] <cradek> here is something I have been working on: http://timeguy.com/cradek-files/emc/nums.dxf
[18:00:41] <cradek> bbl
[18:00:52] <micges> cool thanks
[18:35:03] <alex_joni> http://images.wikia.com/wikitex/images/0/06/06e/efa80cee74d633f6a0f3fe9aef8f21.png
[18:38:06] <archivist> alex_joni, how does one get beans on cheesa on toast out of that
[19:08:13] <alex_joni> you need to pray to the FSM for that
[19:08:30] <cradek> micges: are there any particular types of file you want for troubleshooting?
[19:14:52] <micges> polylines
[19:15:06] <cradek> I think that was a good one then
[19:15:12] <cradek> it is all polylines with lots of arcs
[19:15:32] <alex_joni> archivist: and don't forget to end your prayer with ramen
[19:17:14] <cradek> micges: REALIZE does not support splined polylines. I don't know if you intend to or not. [I think they are very rarely used]
[19:23:55] <micges> splined ?
[19:24:15] <cradek> yes you can tell autocad that the polyline points are spline control points
[19:24:54] <cradek> I doubt anything but autocad would handle that properly
[19:37:16] <micges> Full support of normal POLYLINE LWPOLYLINE interesting me
[19:37:28] <micges> nothing more
[19:38:25] <cradek> I understand
[19:38:27] <micges> my import is quite complicated now and I wanted to strip code to pure dxf, and I didn't make it
[19:38:39] <cradek> I can not make LWPOLYLINE
[19:38:44] <cradek> my autocad is too old
[19:38:52] <micges> cool
[19:39:18] <micges> when we have LW in some file then I can test it
[19:39:35] <cradek> yes maybe someone else can help with that
[19:42:18] <micges> cradek: can you send me circle divided to 8 arcs, 45 deg each, for testing
[19:42:53] <micges> I haven't now any cad that could generate polylines
[19:43:29] <cradek> micges: do you mean 8 separate arcs, or a polyline made up of 8 arcs?
[19:43:53] <cradek> (that is probably not a good test because they will all have the same bulge)
[19:44:12] <micges> you right
[19:44:38] <cradek> did you get my file with the six numbers? that should be a very good test of polyline arcs
[19:45:35] <micges> yes
[19:46:50] <micges> cradek: do you know sth about bulge?
[19:47:05] <cradek> micges: sure
[19:50:38] <micges> I attempt to program in bulge 3rd time
[19:52:26] <cradek> micges: your program is GPL right?
[19:52:39] <micges> yep
[19:52:47] <cradek> http://timeguy.com/cradek-files/autocad/realize.lsp
[19:53:08] <cradek> see defun cut-to
[19:53:18] <cradek> it is correct and very well tested
[19:53:35] <cradek> bulge -> I,J format arc
[19:54:20] <micges> (dig into code)
[19:55:33] <cradek> I see rad (radius) is calculated but not used; ignore that
[19:56:42] <micges> ok
[19:58:01] <micges> ok I see interesting part
[20:00:13] <micges> bbl
[20:19:20] <micges> cradek: can you translate to c ?
[20:19:51] <micges> cot (* 0.5 (- (/ 1.0 bulge) bulge))
[20:19:51] <micges> cx (/ (- (+ prev-x x) (* (- y prev-y) cot)) 2.0)
[20:19:51] <micges> cy (/ (+ (+ prev-y y) (* (- x prev-x) cot)) 2.0)
[20:20:59] <cradek> cot = 0.5 * ((1.0 / bulge) - bulge)
[20:22:02] <cradek> cx = ((prev_x + x) - (cot * (y - prev_y))) / 2.0
[20:22:09] <cradek> etc. :-)
[20:22:10] <micges> ohhh I get It
[20:22:18] <micges> inverse notation
[20:22:25] <cradek> yes
[20:23:05] <micges> much more clear now :)
[20:34:39] <jepler> double angle formula?
[20:45:45] <micges> hmm
[20:46:02] <micges> it works much better than my code :P
[20:55:59] <cradek> yay :-)
[21:31:11] <micges> no t working..
[21:33:33] <cradek> what's not working?
[21:35:15] <micges> bulge in my test files has values from -5 to 5
[21:36:54] <jepler> in the reference I found, bulge may only go from -1 to 1. http://www.autodesk.com/techpubs/autocad/acad2000/dxf/vertex_dxf_06.htm
[21:37:06] <cradek> 42
[21:37:06] <cradek> -1.022045
[21:37:07] <micges> g2 g3 distinguish expression not working at all
[21:37:08] <jepler> 42 Bulge (optional; default is 0). The bulge is the tangent of one fourth the included angle for an arc segment, made negative if the arc goes clockwise from the start point to the endpoint. A bulge of 0 indicates a straight segment, and a bulge of 1 is a semicircle.
[21:37:23] <cradek> jepler: that's surely wrong. you could not have more than a semicircle.
[21:37:25] <jepler> er, maybe I misinterpreted it
[21:37:34] <cradek> yes I think you did
[21:37:55] <jepler> OK, I take it back
[21:38:21] <cradek> the file I sent you nums.dxf has bulges from about -1.02 to 1.02
[21:38:32] <cradek> if you are getting -5 maybe the file is not reading properly
[21:39:54] <micges> ARC-TEST:
[21:39:54] <micges> 42
[21:39:54] <micges> -6.941783
[21:40:15] <cradek> yes that could be right, I thought you were talking about nums.dxf
[21:41:02] <micges> ups sorry
[21:44:26] <cradek> g2 vs g3 is just bulge<0.0
[21:44:30] <cradek> that should be the easiest part
[21:47:07] <micges> could you send me number zero and one in separated files from nums.dxf? (will be faster than manual editing)
[21:49:24] <cradek> http://timeguy.com/cradek-files/emc/one.dxf
[21:49:24] <cradek> http://timeguy.com/cradek-files/emc/zero.dxf
[21:51:12] <micges> thanks
[21:51:21] <cradek> welcome
[21:52:37] <alex_joni> good night all
[21:53:14] <micges> night alex
[21:53:50] <cradek> bye
[22:16:28] <micges> polylines works :)
[22:17:11] <micges> testing..
[22:23:01] <micges> about 40 polylines, all correct
[22:23:12] <cradek> yay
[22:23:14] <micges> 40 polyline files
[22:23:29] <cradek> so you handled the "closed" polylines too
[22:23:48] <micges> yes
[22:24:50] <micges> have some question about generating gcode
[22:25:48] <micges> what (M, G) function should include ?
[22:26:37] <cradek> I don't know. that is a hard question.
[22:28:32] <micges> ok
[22:31:15] <micges> you want to translate interface ?
[22:31:46] <micges> translated*
[22:32:33] <micges> yes I think :)
[22:33:06] <cradek> if you can do a basic translation I can help make it better
[22:33:07] <micges> bbl
[22:33:23] <cradek> I don't know any polish but maybe I can help anyway
[22:33:31] <cradek> bbl, me too
[23:02:09] <micges> good night
[23:15:04] <BigJohnT> jepler: you around?