#emc-devel | Logs for 2007-12-07

[00:22:05] <tissf> good night all
[00:37:58] <skunkworks> http://www.cnczone.com/forums/showthread.php?p=376339#post376339
[04:13:09] <cradek> I'm trying to rekey some locks and I'm ashamed I'm going to buy the pins instead of making them
[04:15:29] <cradek> argh the gnome calculator has 'advanced' and 'scientific' modes but no RPN
[04:16:01] <SWPadnos> heh
[04:16:11] <SWPadnos> no "truly advanced" mode I guess
[04:19:16] <cradek> what good is that, I ask you
[04:19:52] <cradek> * cradek has a disorder that makes non-RPN calculators impossible to use
[04:19:54] <SWPadnos> good (push) is what() you I ask()
[04:20:13] <cradek> well not impossible, but it's like driving someone else's car ... on the moon
[04:20:47] <SWPadnos> heh
[05:20:19] <christel> [Global Notice] Hi all! One of our sponsors experienced a rather large ddos attack just now, which in turn caused disturbance on the network. As a result that leaf server is not accepting client connections atm. Apologies for the inconvenience and have a good day.
[14:03:56] <skunkworks_> logger_dev: bookmark
[14:03:56] <skunkworks_> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2007-12-07.txt
[14:05:35] <SWPadnos> nothing to see here, move along
[14:09:09] <skunkworks_> heh
[14:09:11] <skunkworks_> morning
[14:09:23] <skunkworks_> we had a house blow up in lacross yesterday.
[14:09:25] <skunkworks_> lacrosse
[14:09:32] <SWPadnos> that's odd
[14:09:35] <SWPadnos> gas leak?
[14:09:42] <skunkworks_> I don't think they know yet
[14:10:08] <alex_joni> that it leaked?
[14:10:23] <alex_joni> I would suspect the blowing up should be a hint
[14:10:26] <skunkworks_> http://www.wkbt.com/global/story.asp?s=7459511
[14:10:39] <skunkworks_> the roof landed in the front yard,
[14:11:02] <skunkworks_> the dad and son descided to go swimming at the Y
[14:12:54] <SWPadnos> here's "Steve's Unfounded Analysis":
[14:13:08] <SWPadnos> they left the oven or stove on, and a phone call sparked it into oblivion
[14:13:44] <jepler> it could have been terarists
[14:13:54] <SWPadnos> at least it wasn't nukular
[14:14:06] <jepler> how can you be sure?
[14:14:18] <SWPadnos> there's no mushroom cloud in the photos
[14:19:50] <skunkworks_> you guyzzz
[14:20:59] <alex_joni> oh, and skunkworks_ is still typing :)
[14:22:03] <skunkworks_> my hair is falling out though...
[14:22:20] <SWPadnos> excellent. muahahahahaha
[14:22:33] <cradek> * cradek scans the article for the owners thanking god that they went swimming (but not thanking him for blowing up the house)
[14:22:49] <SWPadnos> heh
[14:22:55] <SWPadnos> the lord provides
[14:23:00] <SWPadnos> something or other
[14:23:11] <cradek> it's just bad reporting if they don't cover that
[14:23:22] <SWPadnos> it looks like they missed it though
[14:23:42] <jepler> this must be a joke: http://www.zetona.org/markwahl/20070615_01.shtml
[14:24:18] <SWPadnos> hmmm. I don't recognize a lot of those characters, let alone know how to type them
[14:24:38] <cradek> haha
[14:24:55] <cradek> This helps prevent -automated- registrations
[14:25:00] <cradek> we need strikethrough in irc
[14:25:27] <SWPadnos> that's what the magic marker is for (the one next to the monitor)
[14:26:06] <skunkworks_> I have a hell of a time with those.. There is one message board that requires you to type in captcha. The are caps and small letter + different colors. I am color blind and some blend right into the background. takes me about 5 tries to do a search.
[14:28:25] <SWPadnos> ouch
[14:28:40] <SWPadnos> are you fully color-blind or something like red-green?
[14:31:15] <skunkworks_> I am not sure anymore.. I just had my eyes checked and they gave me a a color blind test (colored dots you have to make out the numbers) I only saw one.
[14:31:31] <SWPadnos> hmmm. that's probably "bad" :)
[14:31:34] <skunkworks_> out of about 20 pages.
[14:31:44] <skunkworks_> I didn't think I was that bad..
[14:31:59] <SWPadnos> don't they make those so you see one number if you're color blind anda different number if you're not?
[14:32:13] <SWPadnos> (for the particular color pair they're testing with that card)
[14:33:29] <skunkworks_> I don't know - We do ink matching here - and they gave me a color blindness test they have here. I know one of them was 'hey I can see that number' and the lady said - that means your color blind ;)
[14:33:37] <SWPadnos> heh
[14:36:07] <skunkworks_> http://www.toledo-bend.com/colorblind/Ishihara.html
[14:36:12] <skunkworks_> I can see the 25
[14:36:23] <skunkworks_> I cannot see the 56
[14:37:11] <skunkworks_> If I squinch - I sort of make out the 2 down below though - if I didn't know it was a 2 I don't know if I could have figured it out.
[14:37:24] <SWPadnos> hmm
[14:37:40] <SWPadnos> that's also dependent on your monitor calibration, but that's the kind of test I was thinking of
[14:37:47] <skunkworks_> yes
[14:37:58] <skunkworks_> my screen may not be ther greatest.
[14:40:07] <SWPadnos> the best way to test for color blindness would be to make sure someone gets you an Eizo ColorEdge LCD monitor, then take the test again :)
[14:41:31] <SWPadnos> preferably the 24" one
[14:43:39] <skunkworks_> I think so..
[14:43:40] <skunkworks_> :)
[15:01:23] <Guest916> Guest916 is now known as skunkworks__
[15:01:53] <skunkworks__> I really should do something about this router..
[15:05:28] <alex_joni> woodrouter?
[15:07:00] <alex_joni> bbl
[15:17:08] <skunkworks__> heh
[16:10:45] <steve_stallings> steve_stallings is now known as steves_logging
[16:46:08] <alex_joni> jepler: around?
[16:46:23] <jepler> alex_joni: yeah
[16:47:12] <alex_joni> jepler: more users report successfully running stepconf
[16:47:21] <jepler> good
[16:47:25] <alex_joni> one small "annoyance" mentioned by one of them though
[16:47:32] <alex_joni> he uses HALUI = halui in the ini file
[16:47:43] <alex_joni> and needs to reenter that every time after he runs stepconf
[16:47:48] <alex_joni> (which is probably expected)
[16:48:04] <jepler> he should 'loadusr halui' in his custom.hal instead
[16:48:18] <jepler> unless there's some reason that it has to be in the inifile that I'm unaware of
[16:48:18] <alex_joni> ok, that might work too
[16:48:26] <alex_joni> jepler: nope, no real reason
[16:48:33] <alex_joni> err..
[16:48:37] <jepler> it's just historical that it was put in the inifile
[16:48:39] <alex_joni> it needs a NML file
[16:48:51] <alex_joni> so you need to loadusr halui -ini ..ini
[16:49:05] <jepler> OK that's still no big deal especially since you can just use an inifile reference to get the right file
[16:49:16] <jepler> and anyone using halui has to have a custom.hal file anyhow
[16:49:26] <alex_joni> I agree
[16:49:42] <alex_joni> (and as a temp/perm fix, this should work good enough)
[16:49:57] <jepler> I am not going to add a "halui" check box to the GUI
[16:50:06] <alex_joni> otoh, maybe there are things inside an ini file which should be treated as persistent
[16:50:14] <alex_joni> no checkbox to the GUI
[16:50:29] <alex_joni> I was thinking more like grub/menu.lst does it
[16:50:55] <alex_joni> parts between ## foo and ## baz, simply get remembered to the next ini file
[16:53:01] <jepler> I don't like that scheme and I don't plan to implement anything like it for 2.3.
[16:53:04] <jepler> er, for 2.2
[16:53:15] <alex_joni> * alex_joni wasn't talking about 2.2
[16:53:41] <alex_joni> for 2.2 and even 2.3 the loadusr halui should be enough
[16:55:35] <jepler> first we need to understand better what modifications users might want to make to the .ini file (are there any besides halui?); only then can we figure out a good way to make it work with stepconf
[16:55:36] <alex_joni> I was only thinking that it might be used for other "advanced" stuff.. like HOMING params
[16:58:28] <jepler> which homing params specifically?
[16:59:51] <alex_joni> hmm.. homing sequences?
[17:00:29] <alex_joni> or maybe tool-change stuff.. (when we might have tool-rack changers)
[17:00:57] <jepler> stepconf is not meant to support people with machines complex enough to have tool changers.
[17:01:28] <alex_joni> jepler: I agree there are lots of things which stepconf shouldn't even try to support
[17:02:19] <jepler> presently, stepconf writes an arbitrary homing sequence: Z, X, Y, A
[17:02:53] <jepler> so in that case it's not good enough to not overwrite a line in the inifile when it is regenerated; stepconf would have to know that the line was present (added by a user) and not overwrite it
[17:03:32] <alex_joni> does it always write that out?
[17:03:39] <jepler> yes
[17:07:10] <alex_joni> gotta run for a while.. will think about it
[17:07:35] <alex_joni> but I'm thinking about marking unchangeable lines with a special comment (at the end of the line?, on the previous line, etc)
[17:07:46] <jepler> that doesn't work
[17:07:46] <jepler> X = foo # bar
[17:07:48] <jepler> the # bar is not a comment
[17:08:03] <alex_joni> #bar
[17:08:05] <alex_joni> X = foo
[17:08:13] <alex_joni> (although I admit that's ugly)
[17:08:19] <jepler> yes I think it's a terrible scheme
[17:08:21] <jepler> easy for users to get wrong
[17:08:24] <jepler> hard for stepconf to parse
[17:08:34] <alex_joni> right
[17:09:06] <alex_joni> mind if I brainstorm other ideas?
[17:09:28] <jepler> maybe you'd better wait for another day when I am more receptive to new ideas
[17:09:30] <jepler> I don't seem to be today
[17:09:30] <cradek> I think halui, tool changers, estop, etc are entirely outside stepconf's domain
[17:09:53] <alex_joni> cradek: right..
[17:10:03] <alex_joni> and the idea was to allow the user to write those by hand
[17:10:12] <alex_joni> and still be able to run stepconf to easily change the rest
[17:10:44] <jepler> all those things can be done in custom.hal
[17:14:31] <alex_joni> bbl :)
[17:15:13] <jepler> here's one thing I thought about: store the generated .ini file in the .xml; do a 3-way merge between that file, the new generated .ini file and the user-modified .ini file
[17:15:28] <jepler> but then you have to include a GUI for resolving merge(1)-style conflicts *blech*
[17:15:42] <alex_joni> yeah.. *blech* alright
[17:15:43] <cradek> yargh
[17:16:01] <alex_joni> how about having a ini.templ with lines that will always overwrite the lines generated by stepconf ?
[17:16:23] <alex_joni> if they don't exist in the output from stepconf, they only get appended
[18:31:10] <SWPadnos> how about a mini editor on the last page of stepconf, to let you add your own ini lines (the text of which is saved)
[18:48:22] <alex_joni> SWPadnos: that still doesn't take care of the initial problem (running stepconf on existing configs (which the user already tweaked))
[18:48:41] <SWPadnos> true
[18:49:01] <SWPadnos> though it is possible to pre-populate the text editor with everything that isn't going to be overwritten by stepconf
[18:49:14] <SWPadnos> (confusing, but possible)
[18:50:05] <SWPadnos> I just thought about writing an ini file editor, but I would hope it would obsolete by the time 2.3 or 2.4 comes out
[18:50:26] <SWPadnos> ie, that we would have moved to a unified config scheme of some sort
[18:51:20] <alex_joni> * alex_joni starts to think it was a stupid idea in the first place
[18:52:06] <SWPadnos> does stepconf recreate the ini file each time, or does it write the keys it wants to an existing file?
[18:53:17] <alex_joni> it recreates it afaik
[18:54:35] <SWPadnos> ok. modifying it to inifile.write() each var/value pair instead of recreating the file (when the file already exists) would probably solve a lot of customization issues
[19:06:43] <skunkworks__> alex_joni: wasn't there someone with a robot - maybe in germany that somehow was able to 'teach' it?
[19:08:27] <SWPadnos> teach-in mode is pretty common on robots, I think
[19:08:47] <skunkworks__> (using emc)
[19:09:00] <SWPadnos> oh, well that's different ;)
[19:09:00] <skunkworks__> ;)
[19:09:26] <SWPadnos> I'd think you could do it with some sort of probing maneuver - use a button to capture the joint positions
[19:11:17] <SWPadnos> hmmm. maybe there should be a G38.<whetver wr'te up to> that would disable the motor drives and ferror detection
[19:11:22] <SWPadnos> we're
[19:12:13] <alex_joni> skunkworks__: yup
[19:12:20] <alex_joni> there's a teach-in script in emc2
[19:12:34] <alex_joni> basicly you start emc2, switch to teleop, jog it to the place you like
[19:12:49] <alex_joni> hit a button on the teach-in script (it outputs the current position)
[19:13:08] <alex_joni> then you can change the produced data to make it into a program
[19:19:30] <cradek> tool pocket "forward" substitution would be very easy to put in the interp
[19:20:13] <cradek> you ask for Tn. It requests CHANGE_TOOL #[xxx+n] where xxx is some number that represents the start of the tool array
[19:22:16] <SWPadnos> I'm not sure I understood Stuart's original request
[19:22:47] <SWPadnos> it seems he wants to be able to swap tool numbers around for a given machine, without operator intervention (since that's prone to error)
[19:22:54] <cradek> yeah, sometimes he touches a lot of things in one message and doesn't quite clearly explain them
[19:23:32] <SWPadnos> it seems to me he's going to need some master table that knows what the tools are and can swap numbers around as necessary to match the tool shapes
[19:23:41] <cradek> I think he wants gcode T1, for instance, to refer to pockets other than 1, depending on the machine
[19:23:58] <cradek> yes I agree
[19:24:11] <cradek> but when I back up I see the problem is bigger
[19:24:11] <SWPadnos> oh sure, that's supposed to more or less almost be in there - I think that's what FMS was meant for
[19:24:14] <SWPadnos> yep
[19:24:51] <SWPadnos> somewhere, there has to be intelligence to know that tool 1 in this program means a 3" long 1/4" diameter cutter, and that really is tool number 7 on the other machine
[19:24:54] <cradek> what I think one really wants is the gcode to ask for "3/4 nominal roughing mill, it's OK if the diameter is a bit smaller, with flute length at least 1.5 inch"
[19:24:59] <alex_joni> wonder how they do it on a DNC with N x CNC machines, with one central tool-store
[19:25:15] <SWPadnos> there's only one tool table in that case
[19:25:18] <SWPadnos> I think
[19:25:27] <SWPadnos> what's a DNC? :)
[19:26:21] <alex_joni> http://en.wikipedia.org/wiki/Direct_Numerical_Control
[19:26:35] <SWPadnos> ah, ok
[19:26:42] <SWPadnos> I figured, but didn't know the acronym
[19:27:17] <alex_joni> it also means "Distributed Numerical Control"
[19:29:07] <cradek> haha "Many companies were still punching programs on paper tape well into the 1980s"
[19:29:20] <cradek> uh, you mean well into the 2000s
[19:29:31] <SWPadnos> eek
[19:29:41] <cradek> I talked to a guy at fest who is still doing this
[19:29:46] <SWPadnos> eek
[19:29:57] <cradek> he really wanted to do something else :-)
[19:29:59] <alex_joni> yuck
[19:30:50] <alex_joni> throw step in the mix with DNC
[19:30:59] <alex_joni> and you'll be wanting to go back to punching cards :D
[19:39:31] <alex_joni> ROFLMAO..
[19:39:44] <alex_joni> we should change BASE_PERIOD and SERVO_PERIOD to work in paramanu
[19:39:56] <alex_joni> 1 paramanu - 3.29 usec
[19:40:26] <jepler> Unknown unit 'parmanu'
[19:40:34] <alex_joni> http://home.earthlink.net/~mrob/pub/math/numbers-15.html#l_p1_1000e100
[19:41:04] <alex_joni> jepler: it's from Hindu mythology.. doubt units would know about it
[19:41:19] <alex_joni> seems crazy to have that many units..
[19:42:02] <alex_joni> there is also 'thruti' (a bit more down on that page)
[19:43:07] <jepler> you should file a bug report against 'units'
[19:44:26] <SWPadnos> try telling your boss that you're only taking a 1 yama lunch break
[19:45:14] <SWPadnos> because you only got a muhoorta yesterday
[19:45:21] <alex_joni> lol
[19:46:06] <alex_joni> I'll only be gone 5 ghati's
[19:46:15] <SWPadnos> that's what they all say
[19:46:28] <SWPadnos> I'm hoping to get a computer that has only one paramanu of latency
[19:48:23] <alex_joni> SWPadnos: have some lunch, I'll buy: http://www.banknotes.com/YU137.JPG
[19:48:38] <SWPadnos> heh
[19:48:46] <SWPadnos> you only need two to be a trillionaire!
[19:51:02] <alex_joni> "on January 1, 1994 when the currency was devalued at a rate of one billion old dinars to one "new dinar". Three and a half weeks later, on January 24, 1994, the "new dinar" was yet again devalued at a rate of (if I get this right) ten trillion to one"
[19:51:41] <SWPadnos> ouch
[19:52:26] <cradek> wow
[19:52:45] <cradek> seems something is going wrong there
[19:52:46] <SWPadnos> if only we could devalue our national debt that way
[19:52:55] <alex_joni> cradek: how about a war?
[19:53:07] <alex_joni> this was in yugoslavia/serbia
[21:28:09] <SWPadnos___> SWPadnos___ is now known as SWPadnos
[23:13:45] <tissf> hi
[23:13:58] <alex_joni> hi
[23:15:46] <alex_joni> bonne nuit tissf
[23:16:05] <tissf> tu vas te coucher ?
[23:16:36] <alex_joni> oui.. tres fatiguee jour
[23:17:01] <tissf> ok bonne nuit, à demain :)
[23:17:19] <alex_joni> á demain
[23:19:17] <tissf> jepler: No problem to add checklinks.py in v2_2_branch?
[23:23:27] <jepler> tissf: I had not planned to add it .. probably the same problems will show up in 2.2 as TRUNK
[23:25:06] <tissf> jepler: the compiler claims it
[23:26:12] <tissf> After I copied the files of a new translation
[23:27:08] <tissf> when I add checklinks.py, all is well
[23:27:49] <tissf> I want your agreement for cvs add it