#emc | Logs for 2005-06-02

[00:25:06] <lilo> [Global Notice] Hi all. Should we participate in Google's "Summer of Code"? I won't be very available for the next 20 hours, but please message me your comments. Thanks.
[00:25:47] <lilo> [Global Notice] (And we need help in Python and possibly in C, so be prepared to be personally involved if you say 'yes' ;)
[00:27:08] <lilo> [Global Notice] This is Google's link on their "Summer of Code" program.... we'd primarily need work done in Python at this point.... http://code.google.com/summerofcode.html
[00:31:07] <lilo> [Global Notice] Whoops, one final bit of information.... here's the project.... we'd have to figure out how to split it up into components, but it's intended to be modular anyway ;) http://partiallystapled.com/wiki/index.php/FreenodeRegistry
[01:09:12] <bpmw_> Hello all, does any one use BobCad/Cam V20?
[01:11:48] <Phydbleep> bpmw_: Not I. :\
[01:12:49] <bpmw_> Thanks, I also have a Emc question
[01:17:00] <bpmw_> I finally got my machine ready to run (try a program). I have a plate that needs a bunch of different size holes drilled in it. Here's the question: if you invoke a tool change command T? M6, does the program actually stop running for you to change the tool. And then just hit resume when you are done. What about setting z for new tool if not set up in tool library?
[01:25:13] <bpmw_> I would think it hard to set drill bits up in the library any ways, cause how could you chuck drill bits the same all the time as they are all different lengths.
[01:26:04] <asdfqwega> bpmw_: What's your machine, and are you using the latest emc?
[01:26:56] <bpmw_> I have a RF45 manual mill I converted and am running emc2
[01:28:41] <asdfqwega> Okay, if you've checked out emc2 from CVS, look in emc2/nc_files
[01:29:00] <asdfqwega> A good example to work from is skeleton.ngc
[01:29:24] <asdfqwega> If you need it to pause for tool change, you can make it do that
[01:29:56] <bpmw_> Ok, I will check it out. THANKS
[01:30:20] <Jymmm> Anyone need a 16gal 6.5HP shopvac with a 256MPH leaf blower?
[01:30:38] <asdfqwega> Also, there was just some work done with tool tables, so I don't know if they're working just yet
[01:35:07] <bpmw_> Good night Guys!
[01:35:18] <bpmw_> And thanks for the help
[01:37:06] <Phydbleep> Jymmm: What's that about a carpet remover?
[01:37:31] <Jymmm> Phydbleep $85
[01:37:50] <Jymmm> =)
[01:38:15] <Jymmm> too much for me, I'm goin for the next one lower
[01:38:21] <Jymmm> only $50
[02:10:15] <fenn> wow maybe google will pay me to fix EMC2 :)
[02:19:58] <Jymmm> joose on stoooooooooodent
[02:20:02] <Jymmm> no
[02:23:26] <Jymmm> fenn go for it, you have a chance
[02:24:07] <fenn> i graduated 2 years ago :(
[02:24:29] <fenn> they seem pretty irritable when you bring up the topic of not being a student
[02:25:06] <fenn> EMC is probably not "reputable" enough anyway.. we're no GNOME or Python
[02:25:25] <Phydbleep> * Phydbleep is a theology student for such purposes. :)
[02:26:22] <Jymmm> Jymmm is now known as Red70sShow
[02:26:22] <Red70sShow> Red70sShow is now known as Jymmm
[02:26:38] <fenn> Jymmm, how come you always do that?
[02:26:56] <Jymmm> you familure with the show?
[02:27:00] <fenn> no
[02:27:12] <Phydbleep> * Phydbleep temporarily cures Jymmm's cranio-rectal inversion..
[02:27:19] <Jymmm> The 70s Show, one of the characters 'Red'
[02:27:26] <Jymmm> Jymmm is now known as Red70sShow
[02:27:26] <Red70sShow> -----> fenn <----- DUMBASS
[02:27:26] <Red70sShow> Red70sShow is now known as Jymmm
[02:27:30] <Phydbleep> There.. You're no longer wearing your ass for a hat. :)
[02:27:41] <Jymmm> always calls someone a dumbass when appriot
[02:28:11] <Jymmm> and I'm in more than just #emc =)
[02:28:24] <Phydbleep> Sometimes he calls them a dumbass just for fun. :)
[02:29:05] <Jymmm> nah, someone was just being stupid in another channel is all.
[02:29:58] <A-L-P-H-A> aren't we proud to be Jymmm's aquaintance?
[02:30:09] <Phydbleep> Ther's A-L-P-H-A..
[02:30:31] <Phydbleep> A-L-P-H-A: I was just about to send a bin after oyu. :)
[02:30:57] <A-L-P-H-A> a bin? A Bin Ladin?
[02:31:11] <fenn> a wheelie bin!
[02:31:31] <fenn> /bin/wheelie
[02:32:39] <Phydbleep> A-L-P-H-A: Are you set up to mill steel?
[02:32:57] <A-L-P-H-A> yes.
[02:33:00] <A-L-P-H-A> mild steel.
[02:33:03] <A-L-P-H-A> what kinda of steel?
[02:33:14] <A-L-P-H-A> annealed tools steel.
[02:33:17] <A-L-P-H-A> nothing hardened.
[02:34:58] <Phydbleep> http://www.lathes.co.uk/globe/index.html
[02:35:10] <Phydbleep> have a look at the top pic.
[02:36:08] <A-L-P-H-A> what? you want the horizonatal mill?
[02:36:15] <Phydbleep> See that cross-slide block? I need something like that with more t-slots.
[02:36:54] <Phydbleep> Yeah, I'd love to have the whole unit, but I'd settle for the cross-slide. :)
[02:37:05] <A-L-P-H-A> oh... you need something with SUPER precision? Need the surface ground?
[02:37:34] <A-L-P-H-A> I know a guy that does wireEDM, but I requested something like that. but he was gonna charge me $600 bucks... tool steel, precision ground, the works.
[02:37:50] <A-L-P-H-A> you'd need a jib and a jib nut position too.
[02:38:01] <Phydbleep> ?
[02:38:10] <A-L-P-H-A> for the gantry...
[02:38:12] <Phydbleep> Cross-slide nut?
[02:38:15] <A-L-P-H-A> the cross slide.
[02:38:26] <A-L-P-H-A> cross slides have a jib, for taking up slack.
[02:38:46] <Phydbleep> OH.. the gib-strip.
[02:39:03] <A-L-P-H-A> I say tomatoe, you say tomatoe.
[02:39:10] <A-L-P-H-A> :)
[02:39:11] <Phydbleep> Yeah, the one the setscrews ride against.
[02:40:18] <A-L-P-H-A> cool unit.
[02:40:20] <A-L-P-H-A> I wouldn't mind.
[02:40:23] <A-L-P-H-A> but I got a mill.
[02:40:33] <A-L-P-H-A> and a milling attachment. :)
[02:40:39] <A-L-P-H-A> and a milling attachment for my lathe
[02:40:52] <fenn> Phydbleep, that globe thing is the dumbest milling attachment i've ever seen
[02:41:52] <fenn> http://www.lathes.co.uk/myford/page3.html <- now that's a real milling attachment
[02:41:58] <Phydbleep> Yeah, I like the palmgren style better, but that one has possibilities for metal spinning and a few other uses. :)
[02:42:45] <fenn> why do you need a milling attachment for metal spinning?
[02:42:52] <fenn> to increase swing?
[02:43:15] <A-L-P-H-A> I'm off.
[02:43:19] <A-L-P-H-A> ciao. I need sleep.
[02:43:35] <A-L-P-H-A> girls wore me out from just sitting around, doing nothing.
[02:43:43] <Phydbleep> We know you're off, Are you leaving too? <jk> :)
[02:45:08] <Phydbleep> fenn: Yeah, The Globe would be a nice way to increase the swing without re-blocking the head.
[02:46:23] <Phydbleep> fenn: What I actually want is the Myford plain style..
[02:48:04] <Phydbleep> But I need to get a t-slot table to replace the cross-slide so I have a place to mount it.
[02:48:16] <fenn> can you remove the compound slide and bolt it to an angle plate?
[02:49:33] <Phydbleep> I could bolt a 3" angle plate to t-slots and mill the mount for the cross-slide then. :)
[02:49:51] <Phydbleep> s/cross-slide/compound
[02:51:03] <fenn> sounds good
[02:51:36] <Phydbleep> fenn: Where are you at?
[02:51:51] <fenn> huh? in my basement/bedroom in bloomington IN
[02:51:57] <fenn> 10 feet from my lathe
[02:52:16] <Phydbleep> Ahh.. .bltnin = Bloonington. :)
[02:53:19] <fenn> wonder how i can change that to fenn.kicks-ass.nte
[02:54:30] <Phydbleep> register kicks-ass.net and set up reverse dns or spoof it.
[02:55:13] <fenn> kicks-ass.net is my dyndns domain
[02:55:24] <fenn> http://fenn.kicks-ass.net/
[02:56:36] <Phydbleep> So you just need to tweak your reverse dns..
[02:57:02] <Phydbleep> XP or linux?
[02:57:10] <fenn> what do you think>
[02:57:48] <Phydbleep> I think I want to go chase a cute redhead, but I'm stuck here yakking at you. :)
[02:57:59] <fenn> moron :)
[02:58:14] <fenn> linux.. fedora 3
[02:58:47] <Phydbleep> you want to tweak 'named'.. I THINK. :\
[02:59:21] <fenn> i dont have dns running.. nice try.
[02:59:54] <Phydbleep> Sorry, Not named.. 'identd'
[03:00:38] <fenn> okay, i don't really care anymore :/
[03:01:06] <fenn> i'm trying to figure out why my local pukiwiki doesn't like the password i gave it
[03:09:47] <Jymmm> fenn (caps lock =)
[03:10:14] <fenn> no, it's something weirder than that
[03:10:56] <fenn> i think it stored the md5 of the original password in the .ini.php file, and i erased that and wrote my new password in plaintext, since i didnt know what the original was
[03:11:19] <Jymmm> md5 is md5
[03:11:42] <fenn> gotta figure out how to make a new md5 from my new password
[03:11:50] <Jymmm> man md5sum
[03:12:08] <fenn> i did echo password | md5sum -
[03:12:15] <Jymmm> whats your new password?
[03:12:26] <fenn> and pasted the result into where it went before.. no luck
[03:12:36] <fenn> new password is "password"
[03:12:47] <Jymmm> 5f4dcc3b5aa765d61d8327deb882cf99
[03:12:58] <fenn> right
[03:13:21] <fenn> actually that's different
[03:13:30] <Jymmm> password:SHA256:b109f3bbbc244eb82441917ed06d618b9008dd09b3befd1b5e07394c706a8bb980b1d7785e5976ec049b46df5f1326af5a2ea6d103fd07c95385ffab0cacbc86
[03:13:55] <fenn> wasn't that long originally.. more like 32 characters
[03:14:08] <fenn> has a comment next to it that says md5 'pass'
[03:14:11] <SWP_Away> try crypt instead of md5sum
[03:14:13] <Jymmm> fenn read the first part again
[03:14:20] <Jymmm> SHA512
[03:14:30] <Jymmm> err SHA256
[03:15:02] <Jymmm> password:MD5:5f4dcc3b5aa765d61d8327deb882cf99 <-- that is the MD5 of 'password'
[03:15:46] <fenn> okay, it worked
[03:16:03] <fenn> i don't understand what's wrong with md5sum -
[03:16:09] <Jymmm> PEKBAC
[03:16:15] <Jymmm> PEBKAC
[03:16:44] <fenn> bless you
[03:17:01] <Jymmm> PEBKAC --> <fenn> i don't understand what's wrong with md5sum -
[03:17:12] <fenn> well, thanks anyway
[03:17:54] <Jymmm> btw you dont use md5sum in that way. check out man md5sum
[03:18:17] <fenn> weird no man page
[03:18:19] <Jymmm> if you pipe like that, you're introducing additional characters.
[03:18:25] <Jymmm> md5sum --help
[03:18:36] <fenn> oops n/m that was on my email server
[03:20:02] <fenn> so, how did you do it?
[03:20:52] <fenn> cksum md5 ... ?
[03:20:59] <Jymmm> I have a utility those does FAR more than just MD5
[03:21:03] <Jymmm> s/those/that/
[03:24:19] <Jymmm> I bet you that yours was password\n
[03:40:57] <Jymmm> Jymmm is now known as Red70sShow
[03:40:57] <Red70sShow> Red70sShow is now known as Jymmm
[04:36:03] <fenn> fenn is now known as fenn_afk
[04:41:13] <CIA-8> 03petev * 10emc2/configs/motenc_io.hal: Initial revision.
[04:41:49] <CIA-8> 03petev * 10emc2/configs/motenc_motion.hal: Initial revision.
[06:52:19] <Guest80FC68> Guest80FC68 is now known as lexx
[06:52:33] <lexx> lexx is now known as xlexx
[07:01:40] <alex_joni> morning guys
[07:08:56] <Phydbleep> Morning alex_joni :)
[07:09:28] <Phydbleep> * Phydbleep has been turning a nut to mount the new toolpost.
[07:11:47] <alex_joni> heh
[07:11:54] <alex_joni> you've been turning into a nut?
[07:13:05] <Phydbleep> Pretty close.. I've been aligning the lathe and making the nut/bolt to put a Groz type toolpost on it.
[07:14:32] <Phydbleep> http://homepage.ntlworld.com/v.ford/gt.htm
[07:14:49] <Phydbleep> That's the type of toolpost.
[07:28:02] <alex_joni> nice
[08:27:02] <alex_joni> meep?
[08:27:06] <alex_joni> http://dd0s.org/~wooolf/crazyfrog/
[08:27:16] <alex_joni> rrriing diing
[10:44:58] <alex_joni> quiet day ;)
[12:06:26] <les> morning
[12:06:53] <les> I am trying to help Till's student in stutgart get on this irc channel
[12:07:45] <les> what is the typical irc ap included on linux distros?
[12:07:49] <les> xchat?
[12:08:38] <alex_joni> sirc
[12:09:19] <les> googling alex...
[12:09:55] <alex_joni> hey les
[12:10:07] <alex_joni> ksirc that is
[12:10:11] <les> hi alex
[12:10:17] <les> ok
[12:10:28] <alex_joni> should be installed by default on most kde's
[12:10:36] <les> it would be already compiled on a typical linux distro?
[12:10:43] <les> ah ok
[12:11:28] <les> This guy is at U. of stutgart (sp)
[12:12:02] <les> I think Till assigned him the student project of improving the emc TP
[12:12:21] <les> So he ought to be on here for sure
[12:13:37] <alex_joni> nice
[12:13:44] <alex_joni> * alex_joni finished reading sonja's paper
[12:13:44] <les> I will email him about mitc on doze boxex and ksirc on linux
[12:13:54] <alex_joni> mirc is pretty ok
[12:13:57] <les> Ah a little light reading...
[12:14:11] <alex_joni> another option might be bitchx (another irc client on linux)
[12:14:14] <alex_joni> but non-x
[12:14:41] <alex_joni> and there's a firefox extension that does IRC
[12:15:00] <alex_joni> maybe the most easy to install if the other options fail
[12:15:08] <les> I will mention bitchx and firefox....the former works in terminal?
[12:15:24] <alex_joni> yup
[12:15:27] <les> k
[12:15:30] <alex_joni> iirc ;)
[12:15:36] <les> haha
[12:15:47] <alex_joni> as in if I remember corectly ;)
[12:16:24] <alex_joni> http://bitchx.org/screenshots/004.png
[12:17:28] <les> I am getting into the ultrasonic generator project a bit....have some research funded
[12:17:36] <alex_joni> nice
[12:18:04] <les> so waveguides, impedance matching, supersonic air flow stuff for me.
[12:18:30] <alex_joni> heh
[12:18:40] <alex_joni> what's this us generator for?
[12:18:44] <alex_joni> ndt?
[12:19:13] <alex_joni> as in non-destructive testing
[12:19:25] <les> I am getting $20,000 for a one month preliminary feasibility thing
[12:19:34] <les> pretty nice funding huh?
[12:19:40] <alex_joni> right
[12:19:51] <jacky^> hi
[12:19:54] <alex_joni> I don't make that much in a year
[12:19:56] <alex_joni> :)
[12:19:59] <les> It is for an electrostatic spray gun
[12:20:23] <les> generate 6 watts...100 kV from air
[12:20:28] <alex_joni> jacky^: hi
[12:20:35] <jacky^> hi alex_joni :)
[12:20:41] <les> without rotating generators and such
[12:20:48] <alex_joni> jacky^: got those limit switches working?
[12:20:51] <alex_joni> les: nice
[12:22:44] <jacky^> alex_joni: i tried using an IC DM7432 ... nothing to do :\
[12:22:55] <jacky^> i don't know what i'm missing..
[12:24:30] <jacky^> pin 3 of IC always show 3,8 volt, optointerrupter take no effect
[12:25:18] <alex_joni> jacky^: did you test the limit switches?
[12:25:18] <jacky^> also without input pin 3 is up to 3,8 volt..
[12:25:40] <jacky^> yes, the optointerruper work
[12:25:41] <alex_joni> jacky^: like in the schematic you pasted to the list?
[12:25:55] <jacky^> yes
[12:26:05] <alex_joni> what's the DM7432?
[12:26:14] <jacky^> switch from 0 to 5 V when an object cut the beam
[12:26:42] <alex_joni> I'll soon be gone home.. but maybe les can help out?
[12:27:05] <jacky^> Phydbleep: last day suggest to me to isolate the inputs using an IC 7432, but it won't work
[12:27:22] <alex_joni> les: still there?
[12:27:58] <jacky^> sending ouput of any switches to the input 1-2-5 of IC to get 1 only output
[12:28:27] <jacky^> maybe i'm missing something
[12:28:42] <alex_joni> got a link to a schematic?
[12:28:42] <alex_joni> for the 7432?
[12:29:24] <jacky^> yes, Phydbleep give me many lynks, i also have the datasheet
[12:29:46] <jacky^> pin 1-2 are input pin 3 is output
[12:30:21] <jacky^> i can't understand why pin 3 is always on 3,8 V, also in 'normal state'..
[12:30:56] <alex_joni> jacky^: send me the link for the datasheet
[12:30:59] <jacky^> maybe the problem is that my optoswitch are inverted
[12:31:13] <jacky^> they swith from 0 to 5 V
[12:31:36] <jacky^> ok, a moment..
[12:33:34] <jacky^> this is the circuit i build for any switch http://www.eleinmec.com/figures/007_04.gif
[12:34:12] <alex_joni> jacky^: yes.. that's the first stuff
[12:34:17] <alex_joni> now you need to connect 3 of those
[12:34:25] <jacky^> yes
[12:34:41] <jacky^> the ouptuts togheter
[12:34:54] <jacky^> to get only one
[12:35:47] <alex_joni> right
[12:35:57] <jacky^> i'm uplodaing a picure to the web..
[12:38:24] <jacky^> ok
[12:38:31] <jacky^> it is on my album now..
[12:39:07] <jacky^> http://www.roboitalia.com/modules.php?name=coppermine&file=displayimage&album=lastup&cat=&pos=0
[12:39:32] <jacky^> alex_joni: if you want to take a look go there..
[12:44:09] <jacky^> i tried to: connect +5 V to pin 14 VCC, ground to pin 7 GND ouput of first optoswitch to pin 1 of 7432, but on pin 3 of IC there are 3,8 volt, no change, on pin 1, instead, i can see the value swicth from 0 to about 5
[12:44:53] <jacky^> i'm go away.. for 15 min..
[12:45:01] <alex_joni> * alex_joni goes home
[12:45:05] <jacky^> jacky^ is now known as jacky^awy
[13:09:57] <jacky^awy> jacky^awy is now known as jacky^
[13:10:54] <jacky^> hi Phydbleep , are you ?
[14:57:22] <jacky^> Phydbleep: 7432 woork :-)))
[14:58:25] <jacky^> wow :D
[14:58:47] <jacky^> tnx a lot
[15:04:27] <fenn_afk> sigh.. the university retired their pine email servers.. now how am i supposed to hack into their system, huh? i should send a complaint letter.
[15:07:41] <jacky^> fenn_afk: your university will switch to ms server ? O_o
[15:08:45] <fenn_afk> they use webmail only. text-based email services were shut down on june 1st. I had a loophole that allowed me to login and breakout to a shell.
[15:09:02] <jacky^> mmmhhh :\
[15:09:26] <fenn_afk> my email server (sdf.lonestar.org - not univ) is down.. wanted to look at it from a different "angle"
[15:09:43] <SWP_Away> SWP_Away is now known as SWPadnos
[15:10:03] <fenn_afk> thought maybe it wasnt really down, but just blowing me off
[15:10:37] <jacky^> be patience..
[15:10:50] <jacky^> :-)
[15:11:26] <SWPadnos> jacky^: did you get the 7432 working? (it's hard to tell looking through the IRC log)
[15:11:34] <jacky^> SWPadnos: yeahh :P
[15:11:38] <SWPadnos> cool
[15:12:18] <jacky^> finally i get my limit switches working, tnx to Phydbleep
[15:13:01] <SWPadnos> great - did you find that you had to connect both inputs of the OR gate (or you always get a high output)?
[15:13:20] <jacky^> no no... solved
[15:13:39] <jacky^> it's right now
[15:13:59] <jacky^> i was tring only with one input at time
[15:14:03] <jacky^> :\
[15:14:34] <SWPadnos> right - and unconnected inputs are 1, so 1 .OR. (switch output) always results in a 1 output
[15:14:42] <jacky^> yes, :-)
[15:15:01] <SWPadnos> glad it was something simple :)
[15:15:08] <jacky^> hehe
[15:16:53] <ValarQ> emc2 is up and running! weeehooo!
[15:17:03] <ValarQ> * ValarQ dances around
[15:17:40] <SWPadnos> cool - maybe you can add information to the wiki on the Gentoo install process ;)
[15:18:39] <ValarQ> i didn't do anything very gentooish thought
[15:18:47] <SWPadnos> no matter
[15:19:00] <ValarQ> i just downgraded my kernel and compiled it
[15:19:20] <SWPadnos> the instructions will be more complex than 'emerge emc2', but what the heck :)
[15:19:22] <ValarQ> maybe if i create an ebuild which depends on 2.6.10
[15:20:25] <ValarQ> now i'm gonna try to implement my miller driver
[15:20:49] <ValarQ> * ValarQ searches for some friendly api
[15:21:06] <SWPadnos> friendly api - I think that's an oxymoron
[15:22:03] <ValarQ> friendly=good in this case
[15:22:11] <SWPadnos> :)
[15:22:28] <ValarQ> i don't search for a GUI called "driver construction kit"
[15:23:06] <ValarQ> there seems to be some support for parallel devices...
[15:23:21] <SWPadnos> yep - what hardware are you trying to drive?
[15:24:06] <ValarQ> some small stepengines with my own hardware
[15:24:13] <SWPadnos> ah
[15:24:37] <SWPadnos> you could just write a replacement for stepgen, which creates the signals you need
[15:24:42] <fenn_afk> does it take step and direction, or 4-bit patterns?
[15:24:45] <fenn_afk> fenn_afk is now known as fenn
[15:24:53] <SWPadnos> then connect the "pin" outputs to the parallel port driver
[15:25:11] <ValarQ> SWPadnos: oh, that would be great
[15:25:16] <ValarQ> * ValarQ searches for stepgen
[15:25:24] <SWPadnos> that's how HAL works
[15:25:44] <SWPadnos> if you connect something to the parallel port, you can use the parallel port driver to access that hardware
[15:26:05] <SWPadnos> you use a different module to generate the signals that get sent to the parport
[15:26:39] <SWPadnos> it's only if you need to do something other than input bits and output bits that you would write your own port driver
[15:26:52] <SWPadnos> (for an indexed byte output, for example)
[15:27:07] <fenn> indexed == addressed?
[15:27:18] <SWPadnos> yep - like you want to do :)
[15:27:30] <ValarQ> SWPadnos: i will have to do that later on
[15:27:37] <ValarQ> SWPadnos: now it is pretty straight
[15:27:51] <SWPadnos> ValarQ, meet fenn. fenn, meet ValarQ
[15:28:02] <ValarQ> * ValarQ waves at fenn
[15:28:04] <SWPadnos> (you both want the same parport driver :) )
[15:28:04] <fenn> * fenn tips his hat
[15:28:38] <ValarQ> SWPadnos: do fenn wan't my driver?
[15:28:41] <fenn> hmm, i find that highly unlikely swp
[15:28:59] <ValarQ> i haven't sold any I/O cards yet
[15:29:10] <ValarQ> fenn: you wanna buy a card? :)
[15:29:13] <SWPadnos> don't you both want to send bytes to the parallel port, with addresses output on the high byte of the port?
[15:29:29] <fenn> i am doing an array of 16 8-bit devices, addressed with 4 bits, multiplexed in time
[15:29:33] <ValarQ> SWPadnos: nope
[15:29:43] <SWPadnos> ah - well then, don't meet :)
[15:29:57] <SWPadnos> you've been unintroduced ;)
[15:29:58] <ValarQ> SWPadnos: i'm using a special mode with the 'select pin' to write the address
[15:30:05] <fenn> * fenn puts his hat back on.
[15:30:12] <SWPadnos> ah - so you get 256 addresses
[15:30:23] <ValarQ> SWPadnos: yeah, for a starter ;)
[15:30:44] <fenn> how does it remember which address it's on?
[15:30:51] <SWPadnos> you'll find that the outb instructions take a lot of time if you have a multi-stage output routine
[15:30:54] <SWPadnos> latch
[15:31:26] <fenn> swp: only twice as much time :)
[15:31:56] <ValarQ> SWPadnos: yeah, but i think it will work anyway
[15:32:02] <SWPadnos> not for ValarQ - he wants to output an address, then data, but there are also two outputs on the control port
[15:32:41] <SWPadnos> and, when you do one outb, the CPU continues chugging along (I think) - if you do another, then it waits for the first, and does the second
[15:32:58] <SWPadnos> it's in a wait state for around 1 microsecond
[15:33:01] <ValarQ> SWPadnos: but for now i just wan't it working straight
[15:33:17] <SWPadnos> 4 outb's make a 3 microsecond wait
[15:33:39] <ValarQ> SWPadnos: is there somewhere i can write the sequence i wan't sent to my parport?
[15:33:41] <SWPadnos> so a 3 GHz processor is sitting there doing nothing for 9000 cycles (!)
[15:34:04] <fenn> why does the cpu wait?
[15:34:20] <SWPadnos> you need to write a different driver that does the address / data output sequence - parport doesn't do that
[15:34:56] <ValarQ> hmm, i'll get to that i guess
[15:35:03] <SWPadnos> fenn: because IO instructions take 1 microsecond, regardless of the CPU speed (roughly - it varies from 800 ns to 1+ us)
[15:35:32] <fenn> ick i didn't know that
[15:35:37] <SWPadnos> yep - ick :)
[15:35:45] <fenn> that's what the latency charts on the RTAI website mean?
[15:35:50] <SWPadnos> nope
[15:36:13] <SWPadnos> that's the time it takes from when your RT code was supposed to start, until it actually does start executing
[15:36:44] <SWPadnos> when a timer interrupt occurs, there are a lot of things that hapen before the routine gets the CPU
[15:36:50] <SWPadnos> the RT routine
[15:37:31] <SWPadnos> 1) if interrupts are disabled, it waits until they are re-enabled (these are "critical sections" of code)
[15:38:27] <SWPadnos> 2) a context switch: save registers, update the task struct, load the registers from the RT code
[15:38:33] <SWPadnos> 3) jump to your code
[15:38:56] <SWPadnos> (1) can take forever (literally), and is pretty random - this causes the most jitter in response time
[15:39:13] <SWPadnos> (2) is fairly constant in 2.6, and probably better with Adeos
[15:39:42] <SWPadnos> the latency is a measure of how long 1 and 2 take (the jump is at CPU speed, so it's fast and constant :) )
[15:40:09] <SWPadnos> (got it?)
[15:40:41] <fenn> yeah
[15:41:01] <SWPadnos> heh - it seems so easy when you start at "main()" :)
[15:41:10] <fenn> what is defined usually as "critical code"
[15:41:31] <SWPadnos> there are parts of hardware drivers that have to execute uninterrupted - those are critical sections
[15:41:42] <SWPadnos> some other kernel bits are also uninterruptible
[15:41:59] <fenn> does emc's rt stuff have that flag?
[15:42:13] <SWPadnos> I don't think so, but I;m not sure
[15:42:40] <SWPadnos> it's an instruction (cli to clear interrup flag, sti to set interrupt flag)
[15:43:03] <fenn> heh shorter variable names make for faster execution :)
[15:43:11] <SWPadnos> I believe the kernel code has C functions as well - cli() and sti()
[15:43:14] <SWPadnos> yeah
[15:43:23] <fenn> you know something wrong when you have to abbreviate "set"
[15:43:26] <SWPadnos> (faster execution of hte assembler, at any rate :) )
[15:43:55] <SWPadnos> I think it's backwards as well - it's "Clear Interrupt Disable Flag"
[15:44:11] <SWPadnos> I haven't done x86 assembly in a long time though
[15:44:33] <Jymmm> SWPadnos : ColecoASM ?
[15:44:48] <SWPadnos> coleco - that was a Z80, right? :
[15:44:52] <SWPadnos> :)
[15:44:55] <Jymmm> I think so
[15:45:00] <SWPadnos> BTDT
[15:45:11] <Jymmm> PET ?
[15:45:13] <SWPadnos> How about the Exidy Sorcerer - remember that one
[15:45:19] <SWPadnos> PET was 6502
[15:45:24] <SWPadnos> BTDT
[15:45:26] <Jymmm> was it?!
[15:45:29] <SWPadnos> yep
[15:45:37] <Jymmm> damn
[15:45:39] <SWPadnos> I'm pretty sure it was - hmmm
[15:45:58] <Jymmm> dont really matter much other than to show how old we really are
[15:46:03] <SWPadnos> the Personal Electronic Transactor - cool, huh?
[15:46:32] <SWPadnos> yep - 1 MHz 6502
[15:47:11] <SWPadnos> same as Atari, Apple, and other Commodore (C64 and VIC20)
[15:47:14] <Jymmm> lol, even the burst srystal in TV's back then were higher than that!
[15:47:20] <SWPadnos> yep
[15:47:34] <SWPadnos> the TRS-80 Color Computer actually ran off a colorburst divided by 4 :)
[15:47:44] <SWPadnos> (6809 though - it was pretty fast)
[15:48:00] <Jymmm> Kapro
[15:48:19] <SWPadnos> Kaypro was a PC compatible (lluggable, as I recall)
[15:48:26] <Jymmm> CP/M
[15:48:37] <Jymmm> pre pc
[15:48:48] <SWPadnos> CP/M = Digital Research OS, used on a lot of Z80 machines
[15:48:57] <SWPadnos> MS copied this first :)
[15:49:07] <Jymmm> I still have CP/M around here - loved that thing
[15:50:20] <SWPadnos> wait for Steve or someone like him to talk about the "good old days" with the PDP-8 (or the old IBMs) :)
[15:50:49] <Jymmm> alright now.... dont make me grab a forklift and pull out a cray
[15:51:02] <SWPadnos> Crays are COOL!
[15:51:06] <Jymmm> or a IBM 360
[15:51:06] <SWPadnos> actaully, hot
[15:51:11] <Jymmm> VERY hot
[15:51:15] <SWPadnos> 360 - that's too new
[15:51:23] <SWPadnos> I'm talking about the 650
[15:51:28] <Jymmm> still need a forlift
[15:51:29] <SWPadnos> (circa 1967)
[15:51:39] <SWPadnos> need a Terex Titan to move that one
[15:51:45] <Jymmm> lol
[15:52:09] <Jymmm> * Jymmm is gonna to HF today (returns) and then buy a muffler for a vacum
[15:53:01] <A-L-P-H-A> I need a muffler for my vaccum.
[15:53:02] <steve_stallings> Saw, but never ran the 650, did run the IBM1620 mini and a 360-30. Actually designed channel attached devices and did hardware testing with 360.
[15:53:04] <A-L-P-H-A> it's sooo loud.
[15:53:04] <A-L-P-H-A> :/
[15:53:22] <SWPadnos> that's why they have special rooms for them :)
[15:53:30] <Jymmm> craftsman.com search for 'muffler' item# 16993 $11.99
[15:53:41] <SWPadnos> nevermind - eyeball crosstalk :)
[15:54:01] <steve_stallings> yea, 360s were actually kind of quiet
[15:54:26] <steve_stallings> it was the 1403 printers that would make you deaf
[15:54:34] <SWPadnos> my mother programmed on the Xerox Sigma 6 and the IBM 650 (before I was born)
[15:54:34] <Jymmm> Huh?
[15:55:02] <Jymmm> nm that didn't come out as expected
[15:55:34] <Jymmm> it's hard to say "Eh what ya say?" via irc
[15:55:48] <SWPadnos> speak up sonny, I can't HEAR YOU!!!
[15:57:27] <Jymmm> Alright Ladies and gents... I'm outta here for a while... Gots to buy me a router!
[15:57:38] <SWPadnos> hasta lavista, ba-bee
[15:57:44] <Jymmm> =)
[15:57:46] <SWPadnos> +' '
[17:57:14] <rayh> fenn: Got any problem with increasing the font size of case_studies ->glossary to match emc1 and emc2 font size?
[18:06:16] <fenn> actually right now i'm setting up my own wiki
[18:06:36] <fenn> it should be easier to keep track of everything.. not being able to rename a page is a pain in the ass
[18:07:47] <fenn> i'm sending a "press release" re: cnc-workshop to metalwebnews, mms online, cnczone.com, and HSM/village press.. can you think of anyone else?
[18:15:53] <rayh> Um. wiki .. wiki is going to be confusing.
[18:16:57] <rayh> Is this going to be a case of may the best wiki win?
[18:21:10] <cradek> making a second wiki is a terrible idea. There's lots of stuff on the first wiki already.
[18:21:46] <SWPadnos> there are already two - the German one came first :)
[18:22:03] <cradek> ... third
[18:22:31] <SWPadnos> :)
[18:22:44] <SWPadnos> as long as there's a reasonably easy way to get information from one to the other
[18:22:50] <SWPadnos> it shouldn't be a terrible thing
[18:23:17] <rayh> an i18n wiki. What you see is your default lang setting.
[18:23:58] <SWPadnos> An i18n wiki with l10n for all locales would be nice ;)
[18:24:17] <ValarQ> how did i use axis with emc2?
[18:24:29] <rayh> a bitch to edit...
[18:24:34] <ValarQ> the wiki says "It may be integrated into the EMC2 tree at some future time."
[18:25:07] <ValarQ> but i recall that someone here said that it was possible
[18:25:10] <SWPadnos> yeah -I'd hate to have to translate to 162 languages
[18:25:20] <SWPadnos> (or so)
[18:28:04] <cradek> ValarQ: it used to work. Now it doesn't because (I think) paul_c changed a bunch of stuff in emc2 to make it incompatible.
[18:28:19] <cradek> well, that surely was not his primary intent
[18:28:29] <SWPadnos> heh - I was going to say that :)
[18:28:29] <ValarQ> ok
[18:29:15] <cradek> but if you want to use it, I encourage you to get the cvs and fix it up and contribute patches. I will not accept patches that break interoperability with emc1, though.
[18:30:13] <ValarQ> nah, i can do without axis
[18:30:47] <ValarQ> it just looked so cute ;)
[18:30:50] <cradek> unfortunately, I can do without emc2, meaning it may not be fixed for a while.
[18:31:11] <cradek> it's not just cute. It has very valuable features that I'd be quite unhappy without.
[18:31:40] <SWPadnos> reminds me of a great quote on bash.org:
[18:31:47] <SWPadnos> <Sonium> someone speak python here?
[18:31:48] <SWPadnos> <lucky> HHHHHSSSSSHSSS
[18:31:50] <SWPadnos> <lucky> SSSSS
[18:31:51] <SWPadnos> <Sonium> the programming language
[18:32:00] <cradek> haha
[18:32:07] <cradek> I don't see the connection, but that's very funny
[18:32:11] <ValarQ> yeah, i think i saw that one :)
[18:32:21] <SWPadnos> I thought Axis was written in Python
[18:32:28] <ValarQ> it is
[18:32:33] <cradek> yeah
[18:32:39] <SWPadnos> connect()
[18:33:06] <ValarQ> woohoo, the syscall of the day :D
[18:38:20] <fenn> re: rayh Is this going to be a case of may the best wiki win?
[18:38:32] <fenn> no, i'm going to copy all the stuff from the first wiki over
[18:39:20] <fenn> then, hopefully someone with more permanent internet connection than me will host it
[18:39:33] <fenn> and we can change the dns for wiki.linuxcnc.org
[18:39:46] <cradek> but ... there's already a wiki, and these problems are all solved
[18:40:09] <fenn> but, it totally blows and doesn't work right
[18:40:46] <cradek> so far for me, it seems like both those assertions are false
[18:41:31] <cradek> I've added stuff and read stuff, both successfully
[18:41:35] <fenn> 1) cant upload files to the wiki, 2) can't rename pages
[18:41:51] <fenn> 3) no natural line breaks
[18:42:16] <fenn> 4) lavender background??
[18:42:29] <fenn> i guess that matches linuxcnc.org :P
[18:44:20] <cradek> what do you mean by natural line breaks?
[18:44:29] <fenn> btw it's not a big deal to set up a wiki.. you just copy all the files into your html directory and it just works
[18:45:23] <fenn> look at http://fenn.kicks-ass.net/wiki/index.php?test
[18:45:30] <fenn> and look at the source
[18:46:08] <fenn> okay?
[18:46:12] <cradek> ?
[18:46:12] <fenn> done that?
[18:46:25] <fenn> now hit reload
[18:46:53] <fenn> same source code, now the line breaks act the way they do on wiki.linuxcnc.org
[18:49:50] <fenn> oh yeah.. referer function is cool
[18:50:14] <fenn> but not really necessary
[18:50:25] <fenn> site map function is pretty cool too
[18:52:43] <fenn> pukiwiki supports .en and .ja right now.. probably wouldnt be hard to add .de support
[18:53:25] <fenn> dunno if it's i18n compliant
[19:38:06] <alex_joni> g'evening
[19:42:43] <A-L-P-H-A> hi
[19:43:35] <alex_joni> hey A-L-P-H-A
[19:43:56] <A-L-P-H-A> wuzzup?
[19:46:00] <alex_joni> nazzup
[19:47:27] <A-L-P-H-A> coo
[19:47:47] <alex_joni> how bout ya?
[19:48:28] <A-L-P-H-A> about to calculate how fast an emdmill should go running at 2050rpm, for a 3/16" endmill. I'm limited by the RPM, not the feedrate.
[19:48:37] <A-L-P-H-A> well, I guess I could go 2580rpm.
[19:48:50] <A-L-P-H-A> no where neat the 4074rpm I should be running at.
[19:48:53] <A-L-P-H-A> near
[19:49:06] <alex_joni> that's neat ;)
[19:53:05] <asdfqwega> logger_aj, bookmark
[19:53:06] <asdfqwega> See
[19:54:31] <anonimasu> hello
[19:55:56] <alex_joni> hey anders
[19:56:02] <alex_joni> finished your machine?
[19:56:38] <anonimasu> ValarQ: it's possible..
[19:56:46] <anonimasu> ValarQ: but you cant compile axis with emc2
[19:57:26] <anonimasu> alex_joni: yeah for now, I have a day tomorrow, to test the other operating mode
[19:57:27] <anonimasu> :
[19:58:00] <alex_joni> anonimasu: you might get axis to compile with emc2
[19:58:10] <alex_joni> but if you do.. send me the version aswell
[19:58:17] <alex_joni> as I decided it's too much trouble :P
[19:59:51] <anonimasu> alex_joni: but you can compile it with emc1 and run it with emc2 ;)
[20:01:17] <alex_joni> right
[20:01:32] <alex_joni> if you got a runnable emc1
[20:01:42] <alex_joni> on BDI4 that ain't gonna happen ;)
[20:02:25] <anonimasu> hm, I think I run bdi4
[20:02:26] <anonimasu> ;)
[20:02:53] <cradek> jeff set up nightly tarballs for axis
[20:03:01] <cradek> (instead of anoncvs)
[20:03:24] <alex_joni> cradek: thanks
[20:03:40] <alex_joni> you think I (you) could get jeff to modify axis to compile on emc2?
[20:04:02] <alex_joni> me|you me&you , either way
[20:04:11] <cradek> oh I'm sure I could do it, but I'm either too busy, or I don't care (not sure which)
[20:04:35] <alex_joni> heh
[20:04:46] <alex_joni> you're too busy to care
[20:04:50] <cradek> that must be it
[20:04:51] <alex_joni> I'm just too lazy ;)
[20:05:02] <cradek> and apparently nobody else wants it to work...
[20:05:27] <alex_joni> cradek: ain't like that
[20:05:30] <cradek> someday, maybe someone will care enough to bother.
[20:05:50] <alex_joni> well.. I started.. but got stuck on the way
[20:05:55] <alex_joni> and was too tired to keep going
[20:05:58] <alex_joni> maybe this weekend
[20:15:00] <ValarQ> hmm
[20:17:26] <alex_joni> ValarQ: you can run the axis that's included on the BDI
[20:20:26] <cradek> you also might be able to just replace axis.py/axis.nf/axis.tcl
[20:20:34] <cradek> (i.e. don't build emcmodule)
[20:21:29] <alex_joni> cradek: on the latest version.. what did you improve / fix ?
[20:21:46] <cradek> 1.0 has many bugfixes, I don't remember what they were
[20:21:58] <cradek> more recently I added nice support for G54 like was discussed in email
[20:22:09] <cradek> (relative coordinate system)
[20:22:48] <alex_joni> nice
[20:22:55] <alex_joni> how about tool offset?
[20:23:00] <alex_joni> G41 42?
[20:23:07] <alex_joni> G43?
[20:23:11] <cradek> it will work right, but there's no provision for manipulating the tool table
[20:23:40] <cradek> what version is on bdi? is it v1.0_b1?
[20:23:43] <alex_joni> right
[20:23:52] <alex_joni> think b2 or b3
[20:23:56] <alex_joni> from memory
[20:24:04] <alex_joni> if there is such a thing
[20:24:06] <alex_joni> :)
[20:24:13] <cradek> looks like there was b2
[20:25:18] <alex_joni> that might be it
[20:28:26] <cradek> installation made easier, keyboard shortcut fixes, display tool number and offset (?), g54, mouse scrollwheel support, some error handling bugfix, initial directory is programs/, read POSITION_OFFSET and POSITION_FEEDBACK from the ini, mdi entry fix, several bugs in the menus fixed
[20:28:31] <cradek> bored now
[20:28:45] <ValarQ> alex_joni: wasn't the BDI some sort of debian package?
[20:28:47] <alex_joni> oh my .. serious stuff
[20:28:48] <cradek> probably a few other things - tired of looking at the diff
[20:28:57] <alex_joni> the BDI is a debian based distro
[20:29:07] <alex_joni> cradek: it's enough ;)
[20:29:16] <ValarQ> alex_joni: ok, i want to run emc on gentoo
[20:29:21] <alex_joni> rayh: hello
[20:29:28] <alex_joni> ValarQ: be my guest ;)
[20:29:28] <rayh> Hi Alex.
[20:29:38] <alex_joni> how's it going?
[20:29:40] <ValarQ> alex_joni: i'm your guest ;)
[20:29:57] <alex_joni> rayh: are you aware of any other issues with tool comp. ?
[20:30:39] <rayh> Yes but I need to test. tool length gets added on each position change block for which there is NOT a z position specified.
[20:30:53] <ValarQ> alex_joni: actually i got emc2 running
[20:31:04] <alex_joni> ValarQ: nice...
[20:31:45] <ValarQ> now i'm trying to figure out how i should implement my driver
[20:31:47] <rayh> alex_joni: will test that now.
[20:32:20] <ValarQ> that HAL you keep talking about seems like a good thing
[20:32:47] <alex_joni> ValarQ: it's very good for its purpose
[20:33:39] <ValarQ> has it been missused?
[20:33:54] <alex_joni> nope
[20:33:59] <alex_joni> not yet ;)
[20:34:15] <ValarQ> then it's a good thing :)
[20:34:23] <alex_joni> it's not very usefull for tasks it hasn't been designed for
[20:34:25] <alex_joni> yes it is
[20:34:38] <SWPadnos> no it isn't (stop arguing with yourself :) )
[20:34:48] <ValarQ> hehe
[20:35:06] <alex_joni> lol
[20:35:18] <SWPadnos> has anyone here done much dynamic library loading with dlopen() and friends?
[20:35:19] <alex_joni> SWPadnos: you may lurk now :P
[20:35:29] <alex_joni> SWPadnos: not really
[20:35:32] <SWPadnos> heh
[20:35:50] <ValarQ> SWPadnos: not much
[20:36:09] <SWPadnos> OK - I was thinking of ways to get a more component-ized approach to the userspace parts of emc
[20:36:27] <SWPadnos> and less process-ized...
[20:37:14] <SWPadnos> (which would make it more HAL-ific)
[20:37:19] <fenn> * fenn coughs *object oriented* cough cough
[20:37:35] <SWPadnos> that's not the same thing, if you're talking about C vs. C++
[20:37:49] <fenn> what's the difference
[20:38:03] <alex_joni> * alex_joni hands fenn some medicine
[20:38:12] <SWPadnos> heh
[20:38:13] <ValarQ> SWPadnos: ordinary C can be "OO"
[20:38:20] <fenn> you can do non-OO stuff in c++ too
[20:38:23] <ValarQ> SWPadnos: but maybe it's a dull word
[20:38:25] <SWPadnos> a C program or a C++ program can be compiled and linked as a static binary
[20:38:51] <SWPadnos> or to use shared libraries (shared by any program on the system)
[20:39:06] <fenn> yes, but, you should still be able to access any class's public functions, no? whether static or dynamic
[20:39:26] <SWPadnos> when you compile a version of the task controller in EMC, you have to specify in the makefile which .o files get included in the resulting executable
[20:39:53] <SWPadnos> so you choose bridgeport.o and parport.o to get a version of bridgeporttask that uses the parallel port
[20:40:11] <SWPadnos> (I know that hardware uses HAL, but stay with me)
[20:40:27] <SWPadnos> actually - let me change to interpreters - that's a better example
[20:40:35] <fenn> i'm with you all the way, willie
[20:40:37] <alex_joni> * alex_joni stays
[20:40:49] <ValarQ> * ValarQ is lost
[20:40:59] <SWPadnos> if you want a version of bridgeportio that has a G-code interpreter, you link in the rs274ngc.o file
[20:41:46] <SWPadnos> if you want to use a STEP interpreter, you make a separate executable that has 99% of the same .o files as bridgeportGCode, but you change the one line in the makefle to switch to the STEP interpreter
[20:42:04] <SWPadnos> and you can't unload one interpreter and load the next on the fly
[20:42:14] <fenn> yep
[20:42:25] <alex_joni> SWPadnos: who would change the interp?
[20:42:27] <SWPadnos> dlopen() allows an ini setting to control which interpreter gets loaded
[20:42:33] <alex_joni> I mean from the users
[20:42:35] <fenn> maybe you should add this to future directions of EMC or as a feature request
[20:43:08] <alex_joni> it's not very different from having 2 executables with name specified in the ini
[20:43:15] <alex_joni> SWPadnos: don't get me wrong
[20:43:16] <SWPadnos> alex_joni: either the GUI can tell task to change interpreters (based on a fle type), or it can be specified in the .ini file
[20:43:21] <alex_joni> it's a nice thing
[20:43:24] <SWPadnos> assuming there are multiples to choose from :)
[20:43:31] <alex_joni> but not smthg that's very urgent for emc right now
[20:43:42] <SWPadnos> except that you have to terminate the task process and load a new one to change
[20:43:53] <alex_joni> when there will be too many to be handled by Makefile's .. I'll agree ;)
[20:44:03] <SWPadnos> not urgent - I'm looking ahead (I'm farsighted :) )
[20:44:19] <fenn> fenn is now known as fenn_afk
[20:44:24] <robin_sz> meep?
[20:44:29] <SWPadnos> mop
[20:45:15] <alex_joni> hey robin
[20:45:43] <SWPadnos> alex_joni: the other thing this can be used for is something like HAL components - load a userspace toolchange module, for instance (rather than having it in a separate process)
[20:46:04] <SWPadnos> it keeps the task controller as the single point of control
[20:46:57] <SWPadnos> though I'm sure it's a *huge* pain in the ass to get this kind of thing working robustly
[20:47:35] <alex_joni> * alex_joni doesn't want to think about that
[20:47:38] <alex_joni> :)
[20:47:56] <SWPadnos> heh - "SWPadnos - adding programming hemorrhoids for 17 years"
[20:48:17] <SWPadnos> no - wait - 27 years
[20:48:20] <SWPadnos> damn
[20:48:23] <alex_joni> SWPadnos: I think these days I'll start a todo list for emc
[20:48:35] <alex_joni> with priorities
[20:48:38] <alex_joni> on the wiki
[20:48:42] <alex_joni> what do you think?
[20:48:54] <SWPadnos> Isn't there a "future directions" wiki page right now?
[20:52:54] <fenn_afk> http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?Future
[20:53:13] <fenn_afk> fenn_afk is now known as fenn
[20:53:19] <SWPadnos> other stuff at http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?ProgrammerDecisions_Fest_2005
[20:53:24] <rayh> Wow Alex. The interpreter tool stuff works as promised.
[20:53:48] <alex_joni> rayh: how so?
[20:53:50] <alex_joni> :)
[20:54:04] <alex_joni> fenn, SWPadnos: I know about those
[20:54:49] <rayh> Calls the tool file and loads it.
[20:55:03] <rayh> Sets length properly with g43
[20:55:07] <robin_sz> thats "handy" ;)
[20:55:23] <SWPadnos> now if only G92 would work
[20:55:27] <rayh> Even shows the tool that you've prepped and loaded.
[20:55:29] <SWPadnos> * SWPadnos ducks and scurries off
[20:55:30] <robin_sz> didn't it always do that?
[20:55:51] <robin_sz> or has it been broken for a while?
[20:55:51] <SWPadnos> robin_sz: this is emc2, which had some trouble with it
[20:55:55] <robin_sz> ah yes
[20:56:05] <robin_sz> *now* I remember
[20:56:15] <rayh> mumblemumblemumble fs*&^%g
[20:56:41] <alex_joni> what's wrong?
[20:57:10] <robin_sz> swallowed a bee :)
[20:57:29] <alex_joni> heh
[20:57:33] <ValarQ> looks like perl code to me
[20:58:02] <robin_sz> shrug, since perl is almost sytactically identical to C ...
[20:58:13] <SWPadnos> har har
[20:58:22] <alex_joni> rayh: thought you were only up to tcl..
[20:58:29] <alex_joni> maybe it's perltcl
[20:58:31] <alex_joni> or perltk
[20:58:35] <ValarQ> robin_sz: yeah, but not perl _code_!
[20:58:42] <SWPadnos> (maybe tool length offsets aren't working right :o )
[20:58:51] <SWPadnos> :O
[20:58:53] <alex_joni> * alex_joni hides
[20:58:59] <SWPadnos> o_O
[20:59:03] <rayh> those little blue pills get me a ways above tickle
[20:59:05] <SWPadnos> there - that's the one I wanted
[20:59:21] <robin_sz> ValarQ: you mean perl bytecode?
[21:01:05] <ValarQ> robin_sz: you mean perl code can't look like rayh's statement?
[21:01:16] <ValarQ> thats my experience however
[21:01:22] <rayh> How much effort would it be to add dlopen to say the interp and the task.
[21:01:31] <ValarQ> * ValarQ mumbles something about weird perl code
[21:02:16] <rayh> and try out that library?
[21:05:10] <ValarQ> * ValarQ is of to bed
[21:05:15] <alex_joni> * alex_joni has been thinking about NML
[21:05:19] <robin_sz> ValarQ: Perl should look pretty much like C, or slightly easier to read ... if it looks liek "code" its bad perls
[21:05:22] <alex_joni> ValarQ: night
[21:06:01] <alex_joni> I think some day I'll start on a bit of rework of emc's libnml implementation
[21:06:13] <alex_joni> I think I'll take the class count down
[21:06:19] <alex_joni> have a few basic types
[21:06:26] <alex_joni> and do the others as sub-types
[21:06:38] <alex_joni> defined in the proper place (where they get used)
[21:14:30] <SWPadnos> rayh: not sure how much work it is
[21:14:37] <SWPadnos> there are (at least) three stages:
[21:14:57] <SWPadnos> 1) define the interface between modules (mostly done in header files)
[21:15:30] <SWPadnos> 2) make a version that can load a single library and use the functions in it correctly
[21:15:50] <SWPadnos> 3) extend that so that a library can be unloaded, and another one loaded, dynamically
[21:16:40] <rayh> fwiw http://www2.linuxjournal.com/article/3687 dlopen
[21:16:46] <fenn> alex_joni, i'm looking at replacing libnml with a real messaging system, developed by someone else (outside emc) particularly attractive are D-BUS, DCOP, and CORBA/TAO
[21:17:36] <alex_joni> fenn: tell me if you find a suitable one
[21:18:16] <SWPadnos> and make sure paul isn't around :)
[21:18:54] <fenn> DCOP looks sufficient, but it lacks in documentation and active development
[21:19:02] <fenn> but that might be a good thing, since it is more stable
[21:19:27] <fenn> DBUS is supposedly better at TCP/IP messaging, and has active development, good documentaation
[21:19:47] <fenn> CORBA of course, tries to be all things to all people, and is big and slow because of it
[21:20:10] <alex_joni> * alex_joni did some CORBA once
[21:20:18] <alex_joni> hacked up a project in 2-3 days
[21:20:32] <fenn> ICE (internet communications engine, not x-server ICE) was also big,bad, and ugly, but i don't like the vibe i get from it at all
[21:23:47] <fenn> too commercialized
[21:25:07] <fenn> here's a "discussion topic for ya"
[21:25:26] <fenn> how do you feel about making rs274 not monolithic?
[21:26:48] <fenn> right now the interpreter code is held to be some kind of holy artefact, and if we change the way it behaves it's some kind of sin against the standards organization
[21:27:21] <fenn> there's no way to control how the interpreter behaves, it's either you like it or you don't, end of story
[21:27:27] <rayh> rs -- recommended standard
[21:28:01] <alex_joni> fenn: it's one thing to want to fix smthg
[21:28:09] <alex_joni> and do some bugfixing
[21:28:25] <alex_joni> and another to submit code that might break what other people expect of it to do
[21:29:00] <fenn> i'm trying to get some ideas that will satisfy both camps
[21:29:07] <alex_joni> right
[21:29:09] <alex_joni> well..
[21:29:15] <alex_joni> regarding comm lib
[21:29:23] <alex_joni> if you're willing to do the coding..
[21:29:26] <alex_joni> go right ahead
[21:29:42] <alex_joni> if you think you're not up to the task.. make sure there are people who'll help
[21:29:59] <rayh> There would be nothing wrong with creating a new interpreter directory at sf
[21:30:03] <alex_joni> anyways... I need to get some sleep
[21:30:08] <rayh> and starting your interp in there.
[21:30:15] <alex_joni> rayh: a branch would be better
[21:30:25] <alex_joni> added dirs can't get removed
[21:30:40] <fenn> i like SWP's chained-together interpreter modules idea
[21:30:43] <rayh> I'm not certain that I agree. When we speak of making multiple interpreters.
[21:30:46] <alex_joni> and branches can help you touch other parts of emc too, without bugging anyone else
[21:30:55] <rayh> We don't want the code from one getting into the code from another.
[21:30:57] <alex_joni> rayh: in that case yes
[21:31:13] <alex_joni> thought you meant about changing stuff to the actual interp.
[21:31:53] <alex_joni> rayh: thx for letting Peter know I did some on classicladder & HAL
[21:31:59] <alex_joni> we exchanged a few emails
[21:32:07] <alex_joni> and I sent him what I have done
[21:32:07] <fenn> well, it would be ideal if our new interp could be configured to behave exactly like the old interp, based on some configuration file
[21:32:29] <alex_joni> fenn: why would you want that?
[21:32:41] <alex_joni> we don't need a new interp...
[21:32:47] <alex_joni> we might need another interp
[21:32:54] <alex_joni> like g-code & step
[21:33:05] <alex_joni> or maybe HPGL or what's that called
[21:33:08] <fenn> we were talking about how we need some way to implement tool change etc in a sensible manner
[21:33:20] <alex_joni> that's not interp. related
[21:33:30] <alex_joni> or shouldn't be
[21:33:32] <fenn> swp's idea was to have the interpreter modify the g-code into more and more low-level commands
[21:33:52] <alex_joni> you don't want the interp outputting io stuff with gcode
[21:33:59] <alex_joni> that'll turn out nasty
[21:34:04] <fenn> so an m06 would turn into g0 (tool change position),
[21:34:05] <rayh> IMO we have all the interp tool change commands we need.
[21:34:09] <fenn> no ? why not?
[21:34:19] <fenn> arg why not any io stuff with gcode?
[21:34:24] <alex_joni> I mean: set output12 for 1.2 msec
[21:34:29] <alex_joni> then wait for input 14
[21:34:31] <alex_joni> etc.
[21:34:35] <rayh> That is the way it is done on some european controls.
[21:34:39] <alex_joni> how do you do that in g-code?
[21:34:49] <fenn> you can't do that currently, that's my point
[21:35:00] <alex_joni> how does any control do that in gcode?
[21:35:12] <fenn> well, it's not gcode either, it's low-level messages
[21:35:15] <alex_joni> test input 253.1 for example
[21:35:16] <rayh> in hidden macros that get called from the interp when it hits a m6 or a txx word
[21:35:27] <alex_joni> ahhh.. now we're getting somewhere
[21:35:29] <alex_joni> so not g-code
[21:35:40] <alex_joni> then it's not interp
[21:35:54] <alex_joni> interp does only that.. it reads a file and interprets the signs it has read
[21:36:01] <alex_joni> and checks for syntax errors
[21:36:08] <alex_joni> and for permissible codes on the same line
[21:36:10] <rayh> We could do any of this now if we wanted to with M101+
[21:36:18] <alex_joni> for which commands have priority, etc
[21:36:32] <alex_joni> but what gets done with there commands, that's task
[21:36:36] <fenn> it just verifies the code?
[21:36:36] <alex_joni> rayh: right
[21:36:47] <alex_joni> fenn: that's what an interpreter does
[21:36:51] <alex_joni> it interprets
[21:36:56] <rayh> And there "which commands have priority" is the heart of the issue.
[21:36:59] <alex_joni> kinda like an compiler
[21:37:03] <fenn> um, i'm thinking of something more like a parser i guess
[21:37:13] <alex_joni> parser is ok too
[21:37:25] <alex_joni> but it doesn't do anything actually.. it only converts
[21:37:26] <rayh> Exactly or a translator
[21:37:35] <alex_joni> and does some ordering
[21:37:46] <alex_joni> and completes commands
[21:37:55] <alex_joni> say you have an F10 active
[21:38:12] <alex_joni> it makes sure that following movecommands use the F10
[21:38:13] <alex_joni> etc
[21:38:23] <alex_joni> anyways...
[21:38:26] <alex_joni> I'm pretty beat
[21:38:29] <alex_joni> going to bed now
[21:38:31] <alex_joni> night guys
[21:38:37] <fenn> g'night
[21:39:31] <fenn> i'm confused now
[21:40:04] <fenn> oh well
[21:40:13] <fenn> fenn is now known as fenn_afk
[21:48:39] <fenn_afk> fenn_afk is now known as fenn
[22:00:07] <fenn> it's code like this that pisses me off
[22:00:09] <fenn> void *hndl = dlopen("libnewshapes.so", RTLD_NOW);
[22:00:09] <fenn> if(hndl == NULL){
[22:00:09] <fenn> cerr << dlerror() << endl;
[22:00:09] <fenn> exit(-1);
[22:00:09] <fenn> }
[22:00:12] <fenn> void *mkr = dlsym(hndl, "maker");
[22:00:29] <fenn> in a tutorial of all places
[22:01:25] <fenn> it's like ancient hebrew.. just omit the vowels and only the initiated will know what you are saying
[22:03:22] <Phydbleep> * Phydbleep sacrifices a rubber chicken..
[22:08:40] <SWPadnos> a lot of that is because the names ned to be the same as they were 20 years ago, or old code won't compile
[22:08:46] <SWPadnos> need
[22:09:06] <SWPadnos> but the choice of HNDL for the handle is pretty stupid
[23:00:05] <robin_sz> SWPadnos: there a very good reason for it
[23:00:15] <SWPadnos> for hndl?
[23:00:49] <robin_sz> and all the other similar ones
[23:00:57] <SWPadnos> which is ... ?
[23:01:29] <robin_sz> /dev/vowels has only a certain capacity .. code with too many vowels in it overloads it. by cutting out the vowels, it runs quicker?
[23:01:48] <SWPadnos> Ah - they should just symlink to /dev/random instead :)
[23:01:59] <SWPadnos> (move the mouse, the program keeps going...)
[23:02:04] <robin_sz> heh
[23:02:28] <robin_sz> anyway, I agree, chopping out the vowels is a sign of madness
[23:02:35] <SWPadnos> so - I've determined that sharp endmills cut better than dull ones
[23:02:49] <robin_sz> no shit :)
[23:02:53] <SWPadnos> nope
[23:03:13] <robin_sz> did you use the traditional method?
[23:03:17] <SWPadnos> How much smoke should you get when machining aluminum? :)
[23:03:29] <SWPadnos> (with a cobalt cutter)
[23:03:31] <robin_sz> yep, you used the traditional method ;)
[23:03:43] <SWPadnos> and tap-magic + spray coolant
[23:04:14] <SWPadnos> (hand mist coolant, actually - spray bottles work wonders)
[23:05:18] <robin_sz> well, we were looking at the spray mist unit on the mill today ...
[23:05:29] <SWPadnos> it's a bummer - I bought a couple of 1.5" 6-flute cobalt rougher/finisher mills off ebay - and now that I've dulled them, the seller has no more :(
[23:05:31] <robin_sz> trying to decide it it woudl work or not
[23:05:50] <robin_sz> oh, shame
[23:06:03] <robin_sz> I bet you can buy them from your regular supplier
[23:06:04] <SWPadnos> yeah - they were like $15 from this seller
[23:06:18] <SWPadnos> Niagara brand too
[23:06:22] <robin_sz> for like $80?
[23:06:59] <SWPadnos> $15 each, I bought two
[23:07:05] <SWPadnos> they had plenty in stock
[23:07:09] <SWPadnos> back then
[23:07:10] <robin_sz> regular price is $80?
[23:07:16] <robin_sz> or more?
[23:07:42] <SWPadnos> checking now
[23:08:14] <SWPadnos> from $76.23 to $117.95 :(
[23:08:26] <robin_sz> haha :) .. pricey.
[23:09:34] <SWPadnos> these were pretty nice too - they leave a nice bevel at the bottom of the pocket
[23:09:45] <SWPadnos> (45 deg angle or so)
[23:10:15] <SWPadnos> I guess I'll have to switch to the 1" or 3/4" ones now (and crank a bit more)
[23:11:06] <SWPadnos> crap - I forgot
[23:11:25] <SWPadnos> the 1" ones are double-ended, and don't fit in my tool holders
[23:11:36] <robin_sz> 1.25" is fairly big ...
[23:11:43] <robin_sz> bt50 holder?
[23:11:55] <SWPadnos> R8
[23:11:58] <SWPadnos> Bridgeport
[23:12:06] <robin_sz> wo there ...
[23:12:15] <robin_sz> thats a prtty big cutter for R8
[23:12:35] <SWPadnos> the 1.5" cutters have a 3/4 shank
[23:12:54] <SWPadnos> the 1" is a 1" shank, but too long since it's double-ended
[23:12:57] <robin_sz> must give a 3hp bridgport a hard time
[23:13:19] <SWPadnos> 2HP, and I was taking fairly shallow cuts
[23:13:37] <SWPadnos> I did pop the overload after a whil - I should probably revisit my depth and feed calculations
[23:13:41] <SWPadnos> while
[23:13:56] <robin_sz> in that size, you could use replaceable tip tooling
[23:14:12] <SWPadnos> yeah - but I only spent $30 on these ... :)
[23:31:31] <SWPadnos> Hi Paul - how are you?
[23:31:59] <paul_c> hrmmm... so, so.
[23:32:14] <SWPadnos> bummer - I noticed you've been absent from IRC for a while
[23:32:44] <paul_c> been messing around with emcmot.c and fusion.
[23:33:06] <SWPadnos> ah - that sounds like enough for a headache
[23:33:58] <SWPadnos> I talked to my mother about the spline smoothing problem - she said it sounds like an enjoyable problem to solve :)
[23:34:30] <paul_c> She's game for it then ?
[23:34:35] <SWPadnos> yep
[23:34:49] <SWPadnos> I'll probably see her this weekend, and we can go over it a bit more
[23:35:15] <paul_c> also been exploring some of the NURBS libs....
[23:35:28] <SWPadnos> she remembered a course she took called something like "interpolation algorithms"
[23:35:48] <SWPadnos> she ordered a couple of books so she can read up on the subject :)
[23:35:54] <paul_c> The most usefull one is libnurbs++, except for one minor detail.
[23:36:11] <SWPadnos> minor, huh?
[23:36:17] <SWPadnos> (skeptical about that)
[23:36:22] <paul_c> fsking STL
[23:36:34] <SWPadnos> ah
[23:36:56] <SWPadnos> is there a short version of what you don't like about STL?
[23:37:11] <paul_c> It sucks.
[23:37:16] <SWPadnos> that's short
[23:37:34] <SWPadnos> is there a short, yet slightly technical, version of why you don't like STL?
[23:37:46] <paul_c> it slows compile times down...
[23:38:21] <paul_c> it's a freakon' nightmare trying to figure out how the damned code works...
[23:38:32] <SWPadnos> heh - I'll give you that
[23:38:47] <SWPadnos> I wish I could find my C++ book.
[23:38:53] <robin_sz> * robin_sz notes it ships as a pre-built library on debian sarge
[23:39:10] <paul_c> My C++ book has a short chapter on STL
[23:39:14] <SWPadnos> The template overloading stuff is very odd
[23:39:25] <paul_c> The first few lines say....
[23:39:40] <paul_c> Do not use STL for new projects.
[23:39:54] <SWPadnos> ?? how old is your book?
[23:41:08] <paul_c> couple of months.
[23:41:12] <SWPadnos> hmmm
[23:43:00] <SWPadnos> is there a reason they say not to use it?
[23:45:42] <paul_c> STL, like XML, and so many other "buzz" methods was just a passing fad
[23:46:01] <SWPadnos> ah
[23:46:15] <Ircaddict> mmm, xml
[23:46:18] <Ircaddict> * Ircaddict checks his RSS feeds now
[23:46:37] <Jymmm> * Jymmm agrees with paul_c
[23:46:53] <SWPadnos> it did seem to have some value, though there were as many "hack-ish" things done in the STL as anywhere else
[23:47:06] <SWPadnos> it saved people from re-doing the hacks
[23:47:28] <paul_c> C++ still allows you to reuse the "hacks"
[23:47:41] <Jymmm> or perl, or python
[23:48:23] <paul_c> Pearls are for hanging round a girl's necks.
[23:48:43] <Jymmm> PERL2EXE =)
[23:48:43] <SWPadnos> do the files in the numerical/ and nurbs/ directories use STL?
[23:48:46] <robin_sz> ummm, i dont think thats quite what is meant by a "pealr necklace" these days ...
[23:49:04] <Jymmm> robin_sz I was thinking that, but you said it =)
[23:49:06] <SWPadnos> a cursory look at webcvs tells me that those are the only parts we'd need
[23:49:19] <A-L-P-H-A> boo
[23:50:11] <robin_sz> STL, while non-abvious at first, works pretty well, although, admittedly, it can be a bugger to debug
[23:50:12] <paul_c> STL trough and through.
[23:50:28] <SWPadnos> ah
[23:50:41] <paul_c> All that "we" would need is the curve fitting routines.
[23:51:03] <robin_sz> but, its pretty efficient, and more efficient than many home-brewed alternatives
[23:52:36] <paul_c> but I fear it is a non-starter for other reasons.
[23:52:59] <SWPadnos> it looks unmaintained
[23:53:22] <robin_sz> it must be failry well maintaniend to have made it into Debian Sarge ...
[23:53:51] <robin_sz> at least at the 'package' level
[23:54:27] <paul_c> and it doesn't do more than three axis.
[23:54:31] <SWPadnos> most of the core directories haven't been modified in 2 or 3 years, according to sourceforge CVS
[23:54:48] <robin_sz> it could just be stable?
[23:55:22] <SWPadnos> could be, but the last change was adding an HTML reference, 2 years ago
[23:55:28] <les> hi guys
[23:55:36] <Jymmm> lo les
[23:55:41] <SWPadnos> prior to that was "3.0.11 will break your project", a year before :)
[23:55:44] <SWPadnos> Hi Les
[23:56:42] <Jymmm> les I picked up a different router today =)
[23:56:43] <paul_c> ditto.
[23:56:57] <les> Kinda expecting Till's student to show up here any time
[23:57:09] <Jymmm> Profeddor Les
[23:57:13] <les> He is assigned to work on emc's TP
[23:57:14] <Jymmm> Professor Les
[23:57:18] <les> haha
[23:57:22] <les> which router?
[23:57:49] <robin_sz> les: wow, what did he do wrong to deserve *THAT* ? ;))
[23:58:12] <Jymmm> les http://craftsman.com search for 926620
[23:58:14] <les> heheh...remember Till offered a while back.....
[23:58:46] <robin_sz> yeah, well, Till must be quite bright .. bright enough to know not to sendyou a female student ;)
[23:58:58] <Jymmm> rotf
[23:59:08] <les> ha
[23:59:17] <Jymmm> les : I got it for $178 + tax
[23:59:25] <les> I had female interns working for me in the past....
[23:59:33] <les> never touched them...I swear
[23:59:42] <les> not easy either
[23:59:47] <robin_sz> heh
[23:59:51] <les> they always get a crush.
[23:59:57] <Jymmm> les so, when will the videos be repleased?