#emc-devel | Logs for 2010-04-08

[01:05:53] <skunkworks> heh - when you mdi a tool prep (say t1) emc gets very sluggish until the tool-prepared is true.
[07:57:52] <alex_joni> interesting reading material: http://www.stardust.webpages.pl/files/handbook/handbook-en-0.3-rc1.pdf
[08:00:14] <aystarik> right... Rafael is professor in Poland, very smart :)
[08:03:47] <alex_joni> not related to NURBS TP though
[08:17:05] <aystarik> he is my co-worker at SUSE... :)
[08:32:39] <alex_joni> aystarik: cool
[08:34:12] <aystarik> * aystarik met Linus once or twice... And broke his laptop with the kernel update once :)
[08:38:39] <alex_joni> heh
[08:38:45] <alex_joni> sounds fun
[08:39:40] <aystarik> becomes boring after a while as well... :)
[11:20:49] <alex_joni> aystarik: everything does
[11:21:34] <aystarik> and your big steel "toys" too? :)
[11:21:55] <alex_joni> yup
[12:03:22] <CIA-2> EMC: 03jthornton 07v2.4_branch * rf642d5da3f57 10/docs/src/Master_User.lyx: Change order of chapters
[12:03:29] <CIA-2> EMC: 03jthornton 07v2.4_branch * rd12728e00d40 10/docs/src/common/User_Concepts.lyx: Fix wrong discription for G61
[17:15:16] <mozmck_work> another paper: http://www.escholarship.org/uc/item/1x12k114?display=all
[17:42:04] <skunkworks> huh - I really didn't know matt and ray actually worked for smithy https://www.smithy.com/about_us.php?id=14
[17:46:12] <awallin> easy-troll ?? lol.
[17:48:30] <SWPadnos> they have both done contract work for Smithy over the years
[18:14:17] <cradek> a combined total of 160 years of experience
[18:14:25] <cradek> what a silly number
[18:15:25] <SWPadnos> I bet that EMC2 contributors have more than that :)
[18:16:25] <cradek> is this where I make fun of matt for being the only one wearing a tie?
[18:17:06] <SWPadnos> heh
[18:25:58] <skunkworks> heh.. 'I now fully understand why the developers don't hop onto each release of Ubuntu... getting it right is a lot of work!'
[18:41:41] <mozmck> skunkworks: where is that quote from?
[18:42:53] <SWPadnos> latest email on users list, from "dave"
[18:46:40] <mozmck> I see it. :) it's true.
[18:47:07] <mozmck> not too bad to get it running on one machine, but making it generic enough is harder for sure.
[19:14:47] <cradek> interp_check.cc:177: CHKS((block->m_count > MAX_EMS), NCE_TOO_MANY_M_CODES_ON_LINE);
[19:14:50] <cradek> ^ weird
[19:17:06] <skunkworks> now many can you have? ;)
[19:17:21] <cradek> 4
[19:17:29] <cradek> I don't know why. I can't see any reason for the limit.
[19:19:05] <skunkworks> our old comtroller could only have one of each letter on each line.
[19:19:09] <skunkworks> :)
[19:19:29] <cradek> emc is that way for some letters, but not g or m
[19:19:44] <skunkworks> right
[19:20:22] <cradek> the boss8 does things like G789 X123. X234. X345. Y1. Y2. F3. F4. F5.
[19:20:38] <cradek> where G789 is some crazy cycle that has three different feeds and a bunch of coordinates
[19:20:46] <skunkworks> yiked
[19:20:52] <skunkworks> yiked?
[19:20:55] <cradek> yiked!
[19:21:11] <skunkworks> that is the past tense of yike
[19:21:11] <cradek> it's actually neat. it pretty much solves the problem of running out of letters.
[19:21:33] <skunkworks> interesting if not confusing. ;)
[19:21:45] <cradek> to specify a unit rectangle, why not use X0.Y0.X1.Y1.
[19:21:59] <cradek> yeah you have to get them in the right order...
[19:22:31] <cradek> I never figured out - if you use a drill cycle with several Z in it, then on the next line you specify fewer Zs, which ones are sticky?
[19:22:48] <skunkworks> heh
[19:22:50] <cradek> the rules weren't spelled out very well, and the control usually just crashed when I left out too many words
[19:23:19] <cradek> it was pretty delicate
[19:23:27] <cradek> poor seb isn't going to get to use it. he has to use emc instead.
[19:23:37] <skunkworks> darn. ;)
[19:23:55] <skunkworks> crashed as in - restart machine?
[19:24:21] <cradek> yeah there was a reboot button on the side (factory)
[19:24:35] <cradek> reboot and then rehome
[19:25:25] <SWPadnos> cradek, they don't have to be in the right order - you only have to specify one point, since the current point is the "other" one
[19:25:47] <SWPadnos> but that only gives you axis-aligned rectangles, and you need a CW/CCW code (like G2/G3)
[19:25:52] <cradek> SWPadnos: some of their cycles used several points in addition to the current point
[19:26:10] <SWPadnos> yeah. I have seen the same thing, on a big EDM machine
[19:26:21] <cradek> like the peck drills had depth, first peck increment, subsequent peck increments (all being Z words)
[19:26:31] <SWPadnos> eek
[19:26:31] <cradek> ... which is actually a better cycle than emc has
[19:27:03] <SWPadnos> I think what I saw was something like M27 I3 Q12 M45 I17 P24 etc
[19:27:04] <cradek> you very frequently want a deeper first peck
[19:27:17] <SWPadnos> multiple M codes with their own sets of parameters, on the same line
[19:27:47] <cradek> what did we do to deserve gcode?
[19:27:59] <SWPadnos> we let someone else design it
[19:28:00] <cradek> we must've kicked puppies in previous lives or something
[19:28:03] <SWPadnos> in the 1970's :)
[19:28:06] <skunkworks> it seemed like a good idea at the time...
[19:28:11] <cradek> 60s surely
[19:28:32] <SWPadnos> I think the "standard" was actually in the '70s-'80s
[19:28:38] <SWPadnos> NGC anyway
[19:28:40] <skunkworks> atleast - the k&t was designed in 62
[19:28:41] <cradek> haha standard
[19:28:45] <SWPadnos> yeah
[19:29:05] <SWPadnos> it's standard like XML, only without the benefit of ignoring things you don't understand
[19:30:11] <skunkworks> the only reall odd thing about the k&t was you could only do arcs in the 4 quadrants.
[19:31:08] <skunkworks> so a complete circle was 4 lines of gcode.
[19:33:44] <cradek> wtf kl? interp_read.cc:1652...
[19:35:42] <SWPadnos> 1652-1679, it seems
[19:36:20] <SWPadnos> and of course some other lines above, and the entire idea of not using a switch
[19:36:49] <SWPadnos> err
[19:37:06] <SWPadnos> or should I say not including all the cases in the case above?
[20:00:50] <cradek> just all the if(a==b) a=b; stuff seemed particularly strange to me. It leaves me wondering if that's what he even really meant to do.
[20:05:15] <SWPadnos> yeah. it seemed as though there should have been some sort of an assignment to a different variable (though even that would have been an unconditional c=b)
[20:11:54] <JT-Hardinge> even I understand that is a do nothing statement I think
[20:12:52] <SWPadnos> heh
[20:56:02] <alex_joni> anyone around with a 2.4 checkout?
[20:56:25] <alex_joni> emc2 v2_4_branch that is...
[20:56:56] <mozmck_work> yes
[20:57:51] <alex_joni> I'd appreciate a pastebin of the latest pot files
[20:58:06] <alex_joni> saves me a reboot + checkout
[20:58:47] <mozmck_work> bleh, my machine is off right now and I have to leave - sorry.
[20:59:02] <alex_joni> ah, no problem
[21:20:32] <alex_joni> micges: hi
[21:20:50] <micges> hi
[21:21:28] <alex_joni> got a git checkout around?
[21:21:52] <micges> yes
[21:22:05] <alex_joni> can you email me the pot files?
[21:22:35] <alex_joni> there's a guy who wants to do a .sk translation
[21:22:50] <micges> jas
[21:26:44] <alex_joni> great
[21:29:41] <micges> sended
[21:39:07] <alex_joni> micges: thanks a lot
[21:39:22] <micges> welcome
[21:39:29] <alex_joni> is that all in one file?
[21:39:39] <alex_joni> axis, tcl and rs274ngc ?
[21:42:57] <micges> yes
[21:43:16] <micges> and hal tools
[21:44:37] <alex_joni> cool, thanks again