#emc-devel | Logs for 2009-03-04

[00:00:18] <BigJohnT> * BigJohnT says good night Gracie
[16:43:33] <cradek> seb_kuzminsky: is there a way to turn off the smart quotes on the compiler output, or else somehow cause it to render correctly in the browser?
[16:43:44] <cradek> I know I've fought with this before but I don't remember the answer
[16:44:27] <cradek> oh hey, if I tell firefox to use unicode encoding, it looks right
[16:45:35] <cradek> (still, I think ascii would be better so we can cut/paste it etc.)
[16:53:05] <alex_joni> cradek: you mean in the buildbot?
[16:53:07] <jepler> $ gcc fnoodle.c
[16:53:07] <jepler> fnoodle.c:1: error: syntax error before ‘blooze’
[16:53:10] <jepler> $ LANG=C gcc fnoodle.c
[16:53:10] <jepler> fnoodle.c:1: error: syntax error before 'blooze'
[16:53:19] <jepler> maybe you can't see it, but the second variation is dumb quotes
[16:53:21] <cradek> yeah
[16:53:35] <cradek> ah, so it's easy
[16:55:20] <cradek> btw, I used 2.3.0~beta1 yesterday for some work - I didn't see any problems (not to anybody's surprise I suppose)
[16:58:13] <alex_joni> still nice to hear it :)
[16:59:23] <seb_kuzminsky> i'll try to teach the buildbot to run with LANG=C
[17:00:09] <cradek> the most exciting moment was when I thought "maybe it'll drill anyway" and went ahead and used the drill with the bogus-looking point
[17:00:17] <cradek> (thanks, harbor freight)
[17:00:38] <cradek> it was supposed to be a split point, but the split was splitted too far and it didn't really have a point
[17:01:00] <cradek> it did drill for a bit, but then sort of exploded
[17:01:11] <fenn> i like when they grind the angle on the drill face backwards
[17:01:25] <jepler> cradek: I'm glad you don't blame emc2 for that
[17:01:30] <cradek> heh, not sure I've seen that one
[17:01:53] <cradek> jepler: emc2 should have an explosion sensor so it knows to stop
[17:02:07] <jepler> well really that sensor should be in your external estop chain
[17:02:08] <jepler> duh
[17:02:39] <fenn> is that like a caffeinated beverage on monitor sensor?
[17:03:56] <jepler> does that one test for its presence or its absence
[17:03:57] <jepler> ?
[17:08:51] <skunkworks_> I want a N.C. explosion sensor.
[17:09:49] <cradek> even better would be a sensor that would hear the little "tink" of the cutting points coming off, shortly before the explosion
[17:12:08] <BJT-Work> try drilling at 56 IPM into steel with an F bit :/
[17:12:23] <cradek> well, 56 seems a little fast to me
[17:12:37] <cradek> how about 2 instead
[17:12:45] <BJT-Work> it goes about 0.050" deep before it explodes
[17:13:58] <cradek> at least there's no drill left in the work then
[17:14:36] <BJT-Work> somehow my tool table had three drills with over 30 IPM
[17:14:52] <BJT-Work> no bits that I could see were left in the work
[17:14:55] <cradek> that just does not seem right.
[17:15:23] <alex_joni> tool table and 30 IPM?
[17:15:36] <BJT-Work> MC tool table
[17:15:51] <BJT-Work> it does my g code
[17:16:23] <alex_joni> * alex_joni only knows MC hammer table
[17:19:19] <alex_joni> BJT-Work: I was just discussion tool compensation with a friend
[17:19:30] <BJT-Work> yes
[17:19:36] <alex_joni> and I was wondering if you could add an example for an inside square
[17:19:49] <alex_joni> or some figure where the "smart" algorithm against gouging is visible
[17:20:30] <BJT-Work> any ideas on what that would look like?
[17:20:51] <BJT-Work> ok I think I know
[17:21:08] <cradek> I'd plunge in the center somewhere and then go out and cut the square
[17:21:30] <cradek> if I was being fancy I'd use arcs to enter and exit the cut tangentially
[17:21:53] <BJT-Work> that would give a better finish
[17:22:04] <fenn> like this? http://heekscnc.blogspot.com/2009/03/profile-section-of-sketch.html
[17:22:07] <BJT-Work> at the entrance and exit
[17:22:12] <cradek> yes also a slanted plunge would be easier on the end mill
[17:22:35] <cradek> programming the square is easy (well, now) but knowing how to do a smart entry is harder
[17:23:07] <cradek> fenn: looks to me like that path will leave dimples on the sides
[17:23:13] <fenn> yeah :\
[17:23:22] <cradek> and why two cuts?
[17:23:33] <fenn> uh, i'll ask
[17:24:19] <BJT-Work> how about a spiral entry to a pocket
[17:24:42] <alex_joni> certainly not like this: http://4.bp.blogspot.com/_lpgMldGi1O4/Saxn-bSoLfI/AAAAAAAAAIw/bMmj_OLtr6M/s1600-h/adaptivesketch.png
[17:24:43] <fenn> like this? :) http://heekscnc.blogspot.com/2009/03/sketch-as-boundary-for-adaptive.html
[17:24:53] <alex_joni> fenn: ha
[17:24:59] <cradek> alex_joni: haha wtf
[17:25:15] <BJT-Work> time to start drinking :)
[17:25:21] <fenn> "red means OK right?"
[17:25:48] <alex_joni> BJT-Work: those lines only get more vurved from drinking :P
[17:25:54] <cradek> maybe I don't understand that picture
[17:26:17] <BJT-Work> I'm sure I don't understand it
[17:26:48] <alex_joni> I think it counts as "art"
[17:27:06] <fenn> it's high speed machining
[17:27:18] <alex_joni> I'd say too high speed ;)
[17:27:28] <BJT-Work> is it a Japanese dragon head?
[17:27:30] <cradek> slow down and color inside the lines!!
[17:27:39] <alex_joni> machining by numbers?
[17:27:41] <fenn> yeah the red lines being outside is wrong
[17:28:15] <alex_joni> "machining by numbers" isn't that funny really :/
[17:28:43] <fenn> is that like "coding by patterns"?
[17:29:02] <alex_joni> it was supposed to be like "painting by numbers" ;)
[17:29:13] <alex_joni> guess it should be "machining by feeling"
[17:30:13] <BJT-Work> maybe blind machining using braille
[17:34:22] <cradek> http://www.codinghorror.com/blog/archives/000922.html
[17:34:24] <cradek> interesting
[17:35:42] <fenn> cradek: the dimples were on purpose, as an example for tabs so the part doesn't fall out (or similar)
[17:36:00] <cradek> fenn: oh I see
[17:37:54] <BJT-Work> http://imagebin.ca/view/XuXltO.html
[17:38:29] <BJT-Work> * BJT-Work heads out of the shop
[17:38:30] <cradek> you and your fancyschmancy software
[17:38:43] <fenn> i dont get it]
[17:38:51] <BJT-Work> LOL
[17:39:05] <BJT-Work> talk to you guys later
[17:39:59] <alex_joni> this is interesting: http://bikeshed.com/
[17:41:17] <jepler> cradek: I agree with a lot of that blog entry
[17:41:38] <cradek> I do too
[17:41:46] <jepler> (I don't always agree with that guy)
[17:43:22] <cradek> I often say "don't worry about consensus, just do what seems right if you know what it is" and "you don't WANT any more opinions - so don't ask for them" -- both often poorly-received
[17:46:25] <fenn> fwiw the boundary wasn't the border of the pocket, it was just a "machine the surface in this area only" directive
[17:49:57] <cradek> alex_joni: that only makes me wonder who Brett Glass is...
[17:50:10] <alex_joni> it was linked from the page you pasted
[17:50:13] <jepler> hm, I'm worried about that heekscnc application. "HeeksCNC uses libarea for the "Pocket" operation" "HeeksCNC Code License: New BSD License" http://code.google.com/p/heekscnc/ "[libarea is] Derived from the kbool library written by Klaas Holwerda" "libarea code license: GNU General Public License v3" http://code.google.com/p/libarea/ "
[17:50:30] <jepler> "The [kbool library] code is written by Klaas Holwerda, it is free to use for non commercial open source projects licensed as GPL." http://boolean.klaasholwerda.nl/bool.html#copyright
[17:50:47] <jepler> so it's a BSD-licensed program that depends on a GPL3 library that is "non commercial GPL"
[17:50:52] <jepler> * jepler 's head explodes
[17:51:39] <alex_joni> heh
[17:55:59] <jepler> (to be fair to the kbool project, it includes a copy of GPLv3 and each source file I looked at refers the reader to that file for the full license text. it's his wording concerning the availability of the code also under some proprietary license that confuses things)
[18:06:43] <fenn> i chalk it up to language barrier
[18:07:17] <fenn> what's wrong with BSD code depending (for some functions) on a GPL library?
[19:27:39] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/cooltool/ (17 files): initial version
[19:29:32] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/cooltool/README: add README. will need to add more info later
[19:30:36] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/src/Makefile: install emc.nml and core_stepper.hal to cooltool sample configs dir
[19:31:38] <CIA-2> EMC: 03alex_joni 07TRUNK * 10emc2/configs/cooltool/.cvsignore: silence
[19:47:02] <jepler> alex_joni: cool
[19:50:31] <jepler> fenn: if I send you a.out compiled from bsd.c and gpl.c, I can't take advantage of the main feature of the BSD license compared to the GPL: to withhold the source bsd.c from you
[19:50:50] <jepler> fenn: because to obey the license of gpl.c, I have to follow one of the various "complete corresponding source code" provisions of the GPL
[20:17:02] <alex_joni> jepler: what is?
[20:17:15] <jepler> alex_joni: having a cooltool config
[20:17:51] <alex_joni> ah ;)
[20:18:31] <alex_joni> cd cooltool && ls *.ini | wc -l
[20:18:32] <alex_joni> -> 9
[20:30:00] <invite|MkS> if one wants to develop a C++ frontend for emc2
[20:30:17] <invite|MkS> (lie tkemc)
[20:30:48] <invite|MkS> it will take how many sleepless nights to do a minimalistic GUI
[20:31:08] <invite|MkS> interacting with hal & sending nml messages
[20:31:11] <alex_joni> that depends on your C++ skills
[20:31:21] <invite|MkS> just a rought approximation
[20:31:35] <invite|MkS> for an intermediate c++ devel
[20:31:52] <alex_joni> maybe 2-3 weeks for a rough approximation ;)
[20:32:05] <alex_joni> under a week for an more advanced devel
[20:32:25] <cradek> it is silly to ask someone you don't know how long it will take you to do something
[20:32:26] <invite|MkS> and for a total noob ;)
[20:32:27] <invite|MkS> ?
[20:32:36] <cradek> you should study the problem and then you will know
[20:32:41] <alex_joni> maybe 2-3 lifetimes :P
[20:32:45] <invite|MkS> :p
[20:32:49] <jepler> look at xemc.cc and shcom.hh
[20:32:54] <alex_joni> obviously the answers above are pretty useless
[20:33:20] <cradek> I suppose it's fun to see the range of answers
[20:33:45] <jepler> you take all the parts out of xemc that aren't the GUI you want, then add in all the parts that are
[20:34:10] <jepler> or else you start with the APIs defined in shcom.cc plus some amount of setup code that you would glean from xemc.cc or a user of shcom and add to it everything that makes your GUI
[20:35:29] <invite|MkS> hmm
[20:35:41] <invite|MkS> I hav totaly forgot about xemc
[20:35:51] <invite|MkS> have only studied tkemc
[20:36:08] <invite|MkS> code and though that I would have to mix tk with c++
[20:36:30] <alex_joni> with tkemc you need to look at the tk part and emcsh for the connection to emc2
[20:36:32] <invite|MkS> (somehow like axis with tk)
[20:36:52] <jepler> axis only uses tk because I chose it. there's nobody with a gun forcing you to use tcl/tk
[20:37:51] <SWPLinux> there's some old sample code written with Qt
[20:38:12] <SWPLinux> I'm not sure where it is
[20:38:26] <invite|MkS> It would be very handy to see it
[20:39:02] <invite|MkS> especially that I will 90% use Qt
[20:39:34] <invite|MkS> does anybody else knows?
[20:40:23] <SWPLinux> I think it's on the linuxcnc.org site somewhere, but I'm not sure where
[20:41:36] <SWPLinux> it's far from complete. I think we managed to get it to show some numbers in a DRO component, via NML
[20:41:50] <SWPLinux> I don't know how much more was ever done
[20:43:16] <alex_joni> I think there is a newer Qt-lib that allows connecting to emc2
[20:43:29] <alex_joni> but I'm not sure it has been published yet
[20:44:41] <SWPLinux> eztrol?
[20:44:55] <alex_joni> yeah, part of that
[20:45:23] <SWPLinux> It's a little more appealing now that Qt45 is also LGPL
[20:45:27] <SWPLinux> 4.5
[20:45:37] <SWPLinux> stupid laptop keyboards
[20:45:51] <alex_joni> [EZTROL]
[20:45:51] <alex_joni> QTEMC_PLUGIN = /usr/share/eztrol/lib/libqtemc-2.3.cvs-head.so
[20:45:51] <alex_joni> QTEMCHAL_PLUGIN = /usr/share/eztrol/lib/libqtemchal-2.3.cvs-head.so
[20:45:52] <alex_joni> EZTROLGUI= /usr/share/eztrol/eztrolgui/libguimill1.so
[20:47:47] <invite|MkS> and who's developping eztrol?
[20:48:14] <invite|MkS> (don't want to reinvinte the weel)
[20:48:44] <invite|MkS> (sorry I'm french don't know if it's the approriate
[20:49:08] <invite|MkS> word reinvinte maybe rediviseis better)
[20:49:14] <jepler> "reinvent the wheel"
[20:49:55] <SWPLinux> as long as it's a wheel of brie or cheddar, I'm happy
[20:50:14] <invite|MkS> (I have a problem with the "e" even redivise is wrong :p)
[20:50:29] <SWPLinux> blame it on the keyboard
[20:50:33] <SWPLinux> I usully do :)
[20:50:35] <SWPLinux> +a
[20:50:52] <SWPLinux> stupid laptop keyboards
[20:50:54] <invite|MkS> stupid laptop keyboards
[20:50:59] <invite|MkS> :p
[20:51:02] <SWPLinux> heh
[20:51:28] <fenn> * fenn wonders why the topic keeps coming back to cnc machined cheese
[20:51:37] <jepler> this C++ program uses shcom.cc to send an nml message (estop) and look at nml status (print position: http://emergent.unpy.net/files/sandbox/sp.tar.gz
[20:51:41] <SWPLinux> if you want an automatic cheese cutter, that's your deal
[20:51:59] <jepler> compile it, then while emc is running, invoke it: "/path/to/sp axis.ini"
[20:52:15] <jepler> there is very little boilerplate required, as you can see
[20:52:38] <jepler> you can quickly get to the business of writing Qt code
[20:53:14] <invite|MkS> thanks jepler
[20:53:36] <invite|MkS> but I'm currently still working on my Qt app for converting
[20:53:51] <invite|MkS> DXF to g-code specifically for plasma cutters
[20:54:01] <invite|MkS> will leave the Qt Gui for later
[20:54:10] <SWPLinux> have you looked at dxf-to-gcode?
[20:54:18] <SWPLinux> (possibly without the dashes)
[20:54:26] <invite|MkS> dxf2g
[20:55:00] <invite|MkS> yes but it python
[20:55:11] <invite|MkS> it's python
[20:55:21] <SWPLinux> oh. well in that case ...
[20:55:47] <invite|MkS> I mean for windows users it will be less complicated to get a Qt
[20:56:12] <invite|MkS> app to work under Ms win than a python one
[20:56:18] <fenn> for windows users why not wxwidgets?
[20:56:19] <SWPLinux> could be
[20:56:31] <SWPLinux> because Qt is now LGPL also, duh! :)
[20:56:42] <fenn> but, er.. huh.
[20:56:59] <invite|MkS> and in addition I'm doing this as part of my
[20:57:21] <invite|MkS> end of studies project
[20:57:34] <fenn> an even better reason to keep the code secret
[20:57:42] <invite|MkS> (dunno if my literal french to english tranlation is good)
[20:57:42] <fenn> those professors can be so nosy
[20:58:25] <invite|MkS> :p
[20:58:54] <invite|MkS> Will close source everything than and switch to Visual studio :p
[20:59:36] <invite|MkS> no seriously I consier that my little work will help the community
[20:59:48] <invite|MkS> to whom I'm greatfull
[21:00:20] <invite|MkS> and of course satisfy by the way those nosy profs
[21:02:09] <alex_joni> there is also http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?GCncCam
[21:04:13] <invite|MkS> have took a look at it
[21:04:21] <alex_joni> invite|MkS: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Cam
[21:04:29] <invite|MkS> but I'll use ribbonSoft libDxf
[21:04:46] <invite|MkS> it had proven itself for a while now
[21:04:57] <invite|MkS> as Qcad uses it flawlesly
[21:11:24] <alex_joni> good night all :)
[21:11:50] <SWPLinux> see ya Alex
[21:12:10] <alex_joni> invite|MkS: bon courage!
[21:16:05] <invite|MkS> <alex_joni>: Merci!