#emc | Logs for 2007-02-15

[00:02:45] <tomp> before i mess with this any more. i think axis is a great tool, but i think i can't expect it to show the path actually followed vs programmed when i zoom way in. (where .001 radius fills a 1024x768 screen). tho i thought the 'torture' test used axis to show the same idea, i may be a magnitude too far.
[00:07:57] <Twingy> k, who's proficient in gtk
[00:07:58] <jepler> in the preview plot, axis shows lines with pixel accuracy, and arcs with somewhat less accuracy (because it translates an arc into a number of short line segments)
[00:08:26] <jepler> for the backplot, it samples the position at intervals, and the sample rate is well under the 1kHz that is the typical "servo rate" of emc
[00:08:33] <erDiZz> Twingy, what's up?
[00:08:44] <Twingy> erDiZz, gtk widget properties, g_object_set () ?
[00:09:04] <erDiZz> I saw the implementation
[00:09:09] <Twingy> (<unknown>:30818): GLib-GObject-WARNING **: IA__g_object_set_valist: object class `GtkTreeView' has no property named `expander-size'
[00:09:18] <Twingy> g_object_set (G_OBJECT(gui.gcode_block_treeview), "expander-size", 24, NULL);
[00:09:34] <Twingy> am I calling wrong function to set properties?
[00:10:01] <erDiZz> let me see..
[00:11:25] <tomp> jepler: yeah, i'm not using it correctly, if i expect arcs to be pixel accurate ( not correct if i get excited when i seeing white dots in the plot over the path ;) )
[00:11:28] <Twingy> I get the feeling I'm calling the wrong function to set widget properties listed at the end of each widget's documentation on the gtk+ reference manual
[00:12:42] <erDiZz> Twingy, yes, the wrong function
[00:12:47] <erDiZz> it's a "style property"
[00:12:55] <Twingy> what's the function for that?
[00:12:58] <erDiZz> installed with gtk_widget_class_install_style_property
[00:13:30] <erDiZz> Twingy, I don't know anything about that, probably you could find something in the reference for GtkWidget class
[00:13:50] <erDiZz> my guess is that it should be set with gtkrc...
[00:15:58] <owhite> hey people, I'm trying to link a pin motion.spindle-forward which is found in configs/univstep/univstep_io.hal to my pyvcp file which can be seen at http://pastebin.ca/356281
[00:16:16] <owhite> if anyone could assist that'd be great.
[00:17:11] <erDiZz> Twingy, in the docs, "Properties" are set with g_object_set, I did it before, but in your case it's a "Style property", so...
[00:17:23] <Twingy> yes...
[00:17:33] <tomp> owhite: describe the net you'd like to have
[00:17:39] <Twingy> let me know if you figure it out :)
[00:18:59] <Twingy> erDiZz, more importantly, try to figure out how to set the HideSelection property on a treeview to false
[00:19:14] <owhite> tomp: I think I want a situation where pyvcp.chbutton-sample1 is driving motion.spindle-forward, and lighting up pyvcp.led-sample and parport.0.pin-02-out
[00:19:54] <jepler> owhite: motion.spindle-forward is strictly an output, which motion sets in response to spindle m-words
[00:20:09] <jepler> or the GUI "spindle forward" button I suppose
[00:20:32] <owhite> oh. let me think about that for a sec.
[00:21:21] <erDiZz> Twingy, I've grepped mync sources and got this:
[00:21:22] <erDiZz> gtk_tree_selection_set_mode (
[00:21:24] <erDiZz> gtk_tree_view_get_selection (GTK_TREE_VIEW (parts_tree)),
[00:21:24] <erDiZz> (GtkSelectionMode) GTK_SELECTION_BROWSE);
[00:21:39] <Twingy> erDiZz, http://forums.microsoft.com/MSDN/ShowPost.aspx?PostID=569095&SiteID=1
[00:21:46] <Twingy> read that real quick
[00:22:06] <owhite> jepler: okay so that means since g-codes drive motion-spindle-forward, I cant expect to use the pyvcp to change the status of that pin?
[00:22:18] <jepler> owhite: no, you can't
[00:22:28] <owhite> okay, that makes sense.
[00:22:39] <erDiZz> Twingy, thats exactly what I do in myconf-ui
[00:23:02] <Twingy> erDiZz, ?
[00:23:18] <jepler> owhite: two things come to mind: you could OR motion.spindle-on with pyvcp.chbutton so that either one could turn on the spindle
[00:23:46] <erDiZz> Twingy, something like this? http://mync.sourceforge.net/f2.png
[00:24:00] <erDiZz> here's the code... http://mync.svn.sourceforge.net/svnroot/mync/trunk/myconf-ui/myconf-ui/control.cpp
[00:24:04] <jepler> owhite: or you could also learn about 'halui', which would allow you to do the same thing the GUI does when the 'spindle forward' or 'spindle stop' buttons are pressed.
[00:24:12] <Twingy> erDiZz, take this to private chat
[00:24:20] <erDiZz> ok
[00:25:00] <owhite> jepler: the second sounds more flexible. could I use that strategy with pyvcp?
[00:26:45] <jepler> owhite: yes. you can see an example of this using the old vcp, in the sample configuration 'halui_halvcp'
[00:26:58] <jepler> I think this includes a spindle forward and spindle stop button, as well as spindle indicator
[00:27:33] <Twingy> got that erDiZz ?
[00:27:40] <erDiZz> Twingy, got nothing
[00:28:23] <owhite> jepler: that looks pretty good. I 'll take a shot at it.
[00:28:26] <Twingy> how about now
[00:28:32] <jepler> freenode has restrictions on "unidentified" users for sending private messages
[00:28:34] <erDiZz> now, yes
[00:28:59] <jepler> owhite: you should be able to see the HAL pins to link in that example, but of course the vcp pin names will be different from the old vcp to pyvcp
[00:29:23] <owhite> jepler: oh wait. I know why I didnt go that route. I was told that halvcp is not going to be supported over time.
[00:29:37] <jepler> owhite: yes but halui is
[00:29:48] <jepler> owhite: halui + pyvcp
[00:29:59] <jepler> someone should re-do that example in pyvcp, but not me
[00:30:01] <owhite> okay. my mistake.
[00:30:26] <owhite> just trying to be a good citizen/user. :-)
[00:32:38] <owhite> jepler: should I put "loadusr -W halvcp halui.vcp" in my pyvcp file (see http://pastebin.ca/356281)
[00:33:05] <owhite> I guess the problem is that I dont get the purpose of the loadusr command.
[00:33:08] <jepler> owhite: no, you should read the halui.vcp example and learn from it
[00:33:28] <owhite> okay. will take a look.
[00:34:03] <jepler> when you want to apply this to your pyvcp configuration, you need to enable halui
[00:34:13] <jepler> you add a line to the [HAL] section that says: HALUI = halui
[00:34:59] <jepler> you can learn about the halui pin names by looking at halvcp.hal or by reading the documentation
[00:35:12] <jepler> http://linuxcnc.org/docs/2.1/html/gui/halui/index.html
[00:35:48] <owhite> okay. thanks much.
[00:37:04] <tomp> owhite: i'll try to follow it too, now reading the halui.vcp where the 2 btns get pins an d 1 led indicates spindle on & off
[00:37:48] <owhite> tomp: okay.
[00:42:34] <tomp> owhite: so far there's 3 files, halui. vcp, .hal, & .ini , we have no need to look into .ini, leave it alone ( we wont really change any of them, the ini is of no interest to this investigation )
[00:42:40] <owhite> is there documentation on the net command?
[00:43:17] <jepler> owhite: yes, it's documented in 'man halcmd'
[00:43:41] <jepler> net signame pinname ...
[00:43:41] <jepler> Create signname to match the type of pinname if it does not yet
[00:43:41] <jepler> exist. Then, link signame to each pinname in turn. Arrows may
[00:43:41] <jepler> be used as in linkps.
[00:44:13] <owhite> tomp: which config directory are you in? I think I'll have to change the .ini file to addd HALUI = halui in my application. I'm starting with univstep.ini
[00:44:59] <tomp> owhite: ~/emc2-head/configs/halui_halvcp/
[00:45:06] <owhite> jepler is was wondering if you can add more than two pinnames to a net. (e.g., net signame pin1 pin2 pin3)
[00:46:12] <tomp> owhite: we can try that interactively, and then we'll know
[00:46:33] <jepler> owhite: yes, if the same would work as separate 'link' commands
[00:46:49] <owhite> got it.
[00:46:50] <jepler> "..." is supposed to signify that you can put as many pinnames as you like
[00:47:06] <owhite> oh crap it was in the docs. sorry.
[00:47:40] <tomp> owhite: halui.vcp creates widgets and halpins, of interest is halpin = button.spindle-start, halpin = button.spindle-stop, halpin = led.spindle-is-on those are the gui generated pins we have to work with for your investigation
[00:48:02] <owhite> got it.
[00:49:14] <CIA-19> 03jepler 07v2_1_branch * 10emc2/docs/html/gcode.html: Mike Sheldrake reports that G93 and G94 were transposed
[00:49:59] <tomp> owhite: halui.hal creates pins common to all emc configs, so we get button.spindle-start , button.spindle-stop, and led.spindle-is-on,,, then it links the gui pins to the emc hal pins with...
[00:50:37] <tomp> owhiet: cntd linkpp button.spindle-start halui.spindle.start, linkpp button.spindle-stop halui.spindle.stop linkpp led.spindle-is-on halui.spindle.is-on
[00:50:48] <tomp> owhite: sorry
[00:51:17] <CIA-19> 03jepler 07TRUNK * 10emc2/docs/html/gcode.html: merge rev Mike Sheldrake reports that G93 and G94 were transposed
[00:51:23] <owhite> tomp: I'm taking a shot at altering the file I posted at http://pastebin.ca/356281
[00:51:47] <tomp> owhite: lookin now
[00:52:33] <tomp> owhite: same as before
[00:52:58] <tomp> ok, you try or post the changes
[00:54:12] <owhite> seems like its working. got to fire up my machine.... :-)
[00:54:58] <tomp> why do you say it loks like it's working? did a physical button light up the gui led?
[00:55:36] <owhite> oh yah. the gui LED lit, and it flipped the correct solenoid.
[00:55:46] <tomp> woh! cool
[00:55:49] <owhite> its great. as usual.
[00:56:17] <tomp> make t-shirts for all the devs with that slogon
[00:56:48] <jepler> "emc2: flipping the correct solenoid since 2006"
[00:56:49] <jepler> ?
[00:57:13] <tomp> emc2: it's great , as usual
[00:57:34] <owhite> the thing that (this bush-league, low-level user) has done is to a bunch of wall boxes with relays, that drive various appliances. I made a larger number of them, some are driven by an extra parallel port...
[00:58:31] <owhite> ...that was stuff people helped me do a week or so ago. now I'm tying my pyvcp to existing equipment that was originally just driven by the univstep_io to USC board.
[00:59:07] <owhite> er, "has done is _make_ a bunch of wall boxes"
[00:59:21] <tomp> yeh yeh yeh, we'll get the hal-cappuchino machine running later... does the spindle spin? ;)
[00:59:43] <owhite> I my laser? spindle was tied to turning the oxygen on, actually.
[00:59:57] <owhite> er, _IN_ my laser...
[01:00:03] <tomp> oh, yeh, i use M03 for spark on :)
[01:00:09] <owhite> yah.
[01:01:28] <owhite> now I can control ventilation, chiller, laser power supply, power to the motor power supply, power to the laser controller, and lighting, all with software.
[01:01:51] <owhite> its nice, less getting up to perform a run. *shrug*
[01:02:33] <tomp> nice, someday get some fotos for the wiki
[01:03:19] <owhite> well that leads me to a question. I think the hardware part of what I have made is of interest (too small minded blokes like me) but I assumed it wouldnt have a place for the wiki because its mostly about the software.
[01:04:02] <owhite> I have a blog, and I'd put it up there, but would it be any good for the wiki?
[01:04:24] <jepler> I think that would be fine
[01:04:34] <tomp> the soft-ware connected to the hard-ware...
[01:05:12] <jepler> there's a section "DIY hardware known to work with emc" on here: http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?EMC2_Supported_Hardware
[01:05:31] <jepler> I'd write the blog entries and link from the wiki to the blog
[01:05:38] <owhite> got it.
[01:05:39] <tomp> owhite: did it turn on the O2?
[01:05:52] <owhite> tomp: yep. first time everytime.
[01:06:20] <tomp> way cool, yes, please doc this, i wanna read
[01:06:20] <jepler> e.g., "schematic for parport to 8 relays for 110VAC devices" or whatever it is
[01:07:04] <owhite> okay. well I'll keep it mind. I have been hoping I could help with some sort of documentation because you guys help me so much.
[01:07:23] <owhite> you could put me to work in other areas too. I dunno.
[01:07:48] <tomp> helping others to know how to use the tools that the devs make is good
[01:09:17] <jepler> owhite: writing blog entries on pyvcp or whatever you're working with on emc is good -- it helps "get the word out" about the software
[01:09:37] <owhite> are you crazy? get more low-end users like me?
[01:09:57] <jepler> :-P
[01:10:05] <owhite> I doubt you could support more than one or two more guys at my level. :-)
[01:10:57] <tomp> ok, bye 4 now
[01:11:01] <jepler> by that time you'll have matured into a power user and you'll be helping the newcomers
[01:12:02] <owhite> be better off out sourcing to india.
[01:14:58] <jepler> if you're paying, why not
[01:15:02] <jepler> we could have a whole IRC "call" center
[01:15:25] <jepler> "I am ted, and I will be happy to help you if you will put the entire error into pastebin.ca"
[01:15:41] <owhite> piped to /dev/null...
[01:17:03] <cradek> jepler: it's impossible to read that without hearing the indian accent in my head
[01:17:58] <fenn> i think y'all need a brazilian call center
[01:18:24] <owhite> with web cams
[01:18:25] <jepler> fenn: friends ...
[01:18:35] <fenn> friend tis is joao, how can help you?
[01:18:43] <tomp> lol
[01:18:51] <cradek> oh come on guys
[01:18:57] <tomp> heard that across the room
[01:19:41] <jepler> that's interesting. my cat just managed to draw a rectangular text selection in the terminal, by flopping over on the touchpad
[01:19:46] <jepler> I don't even know how to do that!
[01:19:54] <cradek> haha
[01:20:16] <jepler> he won't tell me what he did, either
[01:20:44] <cradek> try selecting with your rear
[01:20:55] <jepler> seems to be ctrl+alt+lmb
[01:21:05] <jepler> I am impressed that he was able to do that
[01:21:12] <cradek> xterm just beeps
[01:21:20] <jepler> this is rxvt-unicode with perl enabled
[01:22:18] <tomp> touchpad? some can be touched twice at same time, defining a rect ( laptops syn-something driver for linux supports it )
[01:22:49] <jepler> I think his "flop" pressed my palm against ctrl+alt, and his furry parts did a drag on the touchpad
[01:22:53] <jepler> not entirely sure
[01:23:04] <jepler> but the important thing is that it was a learning experience for me
[01:23:05] <tomp> hard to dupli-cat :)
[01:23:22] <jepler> indeed
[01:23:26] <fenn> no luck here either
[01:23:50] <tomp> bye no more jokes
[01:42:05] <robin_sz> ick, valentines day
[01:42:31] <robin_sz> remeber .... to get that girl you'll need:
[01:42:38] <robin_sz> 1) roses
[01:42:45] <robin_sz> 2) chocolates
[01:42:53] <robin_sz> 3) champagne
[01:43:00] <robin_sz> 4) rohypnol
[01:45:44] <jepler> is this a multiple choice test?
[01:45:58] <owhite> sorry if this is repeating a quesion people ignored, seemed like the server bounced...the question is my laser enable is tied to M09. is it possible to pulse that for a short duration, like 100ms?
[01:47:19] <jepler> if you want just one pulse each time the enable is turned on, look at the 'oneshot' component. If you want something more like PWM, look at the 'pwmgen' component.
[01:48:20] <owhite> can you control the length of oneshot?
[01:48:30] <jepler> did you read 'man oneshot' yet?
[01:48:52] <owhite> gak. you guys document more than I thought.
[01:49:36] <fenn> manpages are nice.. i wish they existed for more than 50% of the linux-compatible software out there
[01:49:57] <jepler> fenn: I agree with you, that's why I made a big effort towards manpages in emc2.1
[01:50:20] <owhite> fenn: my mistake is that I use your outlook and make the wrong assumption about this crew.
[01:50:42] <jepler> there are manpages for most of the HAL realtime components and most of the HAL and RTAPI API calls in emc2.1
[01:50:44] <fenn> well, info and html-based help is just really painful to use
[01:54:44] <jepler> info isn't bad, it's much better than manpages when the information would run to more than a few sheets of paper printed
[01:55:30] <jepler> maybe if I knew a few tricks for searching out man headings in 'less'...
[01:55:49] <fenn> yes but i'd rather have an executive summary than "GNU is a bunch of bigots; go read the info pages you twirp"
[01:56:42] <owhite> jepler: you reckon there would be any gotchas to using oneshot in pyvcp, like could I tie a oneshot pin to a slider?
[01:57:21] <fenn> dont you want to be able to control the laser from gcode though?
[01:57:51] <jepler> owhite: a slider for the pulse duration? That is not possible, because 'width' is a parameter, not a pin
[01:58:04] <owhite> ah.
[01:58:28] <jepler> perhaps it should have been a pin
[01:58:44] <fenn> everything should be a pin
[01:58:45] <owhite> fenn: not in this application. basically I need to send short pulses when I am "sighting in" the laser. right now I use electronics to control the pulse duration and a footswitch.
[01:59:07] <fenn> and you should be able to set the state of a pin
[01:59:15] <jepler> owhite: are you working from an emc you built from source?
[01:59:19] <owhite> yes.
[01:59:28] <owhite> er, jepler: yes. :-)
[01:59:41] <jepler> in that case, you can easily change 'width' to be a pin, so that you can hook it up to a slider
[02:00:13] <jepler> in src/hal/components/oneshot.comp, find the line that says "param rw float width..." and change the first two words to "pin in"
[02:00:35] <owhite> is that in the man page? :-)
[02:00:38] <jepler> then re-make to get a new oneshot component
[02:00:47] <jepler> the manaual page will even change to reflect the modification you made
[02:01:08] <owhite> after I wtfm.
[02:02:40] <fenn> heh
[02:02:56] <jepler> w=write? no, the manual page is automatically generated from the .comp file when you 'make'.
[02:04:52] <owhite> you have GOT TO BE KIDDING.
[02:04:56] <owhite> you guys are crazy.
[02:05:18] <owhite> documentation doesnt make it self. that's like machines fixing themselves.
[02:05:21] <jepler> uh ... is that a compliment?
[02:05:34] <owhite> yes. its a compliment.
[02:05:40] <jepler> actually there are a lot of systems for putting the documentation and the source code together
[02:05:55] <jepler> by now there's an <every language name>doc
[02:06:05] <CIA-19> 03jepler 07v2_1_branch * 10emc2/src/hal/components/conv.comp.in: note GPL license
[02:06:11] <owhite> pod/perl
[02:06:28] <CIA-19> 03jepler 07TRUNK * 10emc2/src/hal/components/conv.comp.in: merge rev note GPL license
[02:06:50] <jepler> ok so you weren't completely serious -- thank goodness
[02:08:29] <owhite> hey, I suppose today is a red-letter day. its the first time I "wrote" emc2 documentation. :-)
[02:08:46] <jepler> I guess so
[02:09:02] <jepler> is it the first time you've written real-time software?
[02:09:35] <owhite> well I've compiled emc2 before.
[02:10:00] <owhite> and I made a rt-linux kernal before you guys moved to ubuntu.
[02:10:20] <jepler> but this time you're writing it
[02:10:22] <jepler> not just compiling
[02:10:41] <owhite> yes.
[02:10:50] <jepler> never mind that I told you exactly which line to change and how :-P
[02:10:52] <owhite> do I get to wear a developers t-shirt?
[02:11:06] <jepler> uh if you want
[02:12:24] <jepler> the real test will come when you ask for something that can't be done easily, except by writing a new .comp
[02:13:37] <owhite> that would be a real test alright. so I dont know if the change to man oneshot took. mine says this...
[02:13:58] <owhite> oneshot.N.width float in
[02:13:58] <owhite> Pulse width in seconds
[02:13:57] <jepler> did 'width' move up to the 'pins' section?
[02:14:08] <owhite> yes.
[02:14:11] <jepler> then it did
[02:14:17] <jepler> before it was under 'parameters'
[02:14:22] <owhite> got it.
[02:14:33] <jepler> 'pins' you can link, 'parameters' you can't -- and that's what you changed in the .comp file
[02:15:19] <jepler> so now you can hook a pyvcp slider (floating point value) to oneshot.0.width, and a pyvcp momentary button to oneshot.N.in
[02:15:33] <jepler> and oneshot.0.out to your laser enable
[02:15:33] <owhite> got it. as an excersize the student shall attempt to flash an LED using pyvcp.
[02:16:02] <jepler> I'll be back, time to tend the laundry..
[03:04:46] <jepler> anyone know what wire gauge is used in hard drive power connectors?
[03:06:08] <jepler> maybe 18 AWG?
[03:06:24] <cradek> that seems right
[03:06:39] <cradek> or maybe a little smaller
[03:06:59] <jepler> though this document says 10AWG is best: http://www.codeproject.com/Purgatory/humorarticleone.asp
[03:07:22] <jepler> (oh darn, the URL kinda gives away that it's not serious)
[03:07:32] <cradek> kinda
[03:07:48] <cradek> "Improving the speed of you PC"
[03:08:43] <skunkworks> funny. are you guys grossed out easy?
[03:09:30] <cradek> well I don't think so, but I'm hesitant to answer
[03:09:34] <skunkworks> http://www.ebaumsworld.com/2007/01/nasty-bot-fly.html
[03:09:34] <jepler> so in any case it's probably a bad idea to do 22A 5V over a single drive power connector
[03:09:48] <cradek> skunkworks: pass!
[03:10:38] <skunkworks> 14 guage is good for 15a
[03:10:48] <skunkworks> rated that is
[03:10:57] <skunkworks> 12gauge is good for 20
[03:11:07] <jepler> what does your table say for 18?
[03:11:26] <jepler> my table says 16A, but for 14 gauge it has 32 so clearly our tables are different
[03:13:14] <skunkworks> I am going by house wiring standards
[03:13:50] <crepincdotcom> cradek: i just milled a board i exported from eagle using the ulp
[03:13:55] <crepincdotcom> it was, shall we say, "phail"
[03:13:58] <cradek> yay
[03:14:10] <cradek> I don't know that word
[03:14:16] <crepincdotcom> lol
[03:14:17] <crepincdotcom> fail
[03:14:22] <cradek> didn't go well?
[03:14:28] <crepincdotcom> the code seemed to think that the milling bit was much smaller
[03:14:46] <crepincdotcom> even though, in your example you used a 0.0625 and i used a 0.0394
[03:14:51] <cradek> you give it the width when you output it
[03:15:01] <crepincdotcom> there are no traces, the bit on either side took them away
[03:15:07] <cradek> you should measure exactly what separation you get
[03:15:12] <crepincdotcom> wait, not in drill.cfg? i thought the first was the milling bit
[03:15:23] <cradek> no, that's only about holes
[03:15:27] <crepincdotcom> damn it
[03:15:29] <cradek> the first is an end mill made to cut big holes
[03:15:31] <crepincdotcom> * crepincdotcom == noob
[03:16:19] <crepincdotcom> well i guess ill redo the code and work on that
[03:16:18] <crepincdotcom> thanks
[03:16:33] <cradek> what separation are you getting?
[03:16:45] <crepincdotcom> separation on what. there are _no_ traces
[03:16:57] <crepincdotcom> the mill going by both sides removed them
[03:16:59] <cradek> if you cut one pass how wide is the cut?
[03:17:22] <crepincdotcom> 0.0394 is the bit diameter, if thats what you mean...
[03:17:30] <cradek> oh you're using an end mill?
[03:17:50] <crepincdotcom> i am this time, but since I don't have anything smaller i'm going to use a 60 degree V next time
[03:17:55] <cradek> that's a big mill
[03:18:01] <crepincdotcom> so i found out
[03:18:02] <cradek> ok you're cutting big stuff :-)
[03:18:13] <crepincdotcom> not really...
[03:18:27] <cradek> yeah the default is probably for .008 or .010 cuts
[03:18:34] <crepincdotcom> ahhh ok
[03:18:44] <crepincdotcom> im cutting this, with a few slight editations
[03:18:45] <crepincdotcom> http://users.wpi.edu/~jackc/stuffs/Chopper%20Locker%201.png
[03:19:14] <cradek> that's going to be impossible
[03:19:27] <crepincdotcom> so cut of a V bit is x=depth/(arctan(60)) right?
[03:19:31] <crepincdotcom> cradek: why is that?
[03:19:43] <skunkworks> you can get a 556 that is 2 555 in one package
[03:19:59] <crepincdotcom> yeah, but i have a bunch of 555s sitting around
[03:20:03] <cradek> the pins are .100 apart (or less for some parts) and you are trying to have two cuts and a trace between pads
[03:20:03] <crepincdotcom> ;-)
[03:20:17] <crepincdotcom> * crepincdotcom smacks face
[03:20:20] <jepler> cradek: you mean impossible with a .0394 mill
[03:20:24] <cradek> straightforward math will tell you your mill is way too big
[03:20:37] <crepincdotcom> or straightforward logic, that i seem to have misplaced
[03:20:43] <crepincdotcom> well, this was a learning experience
[03:21:09] <cradek> you could make some boards with that, but nothing with traces between pads etc
[03:21:29] <crepincdotcom> well ill just load up my V bit and get going with that
[03:22:26] <cradek> there you go
[03:22:37] <crepincdotcom> thanks a lot for the help
[03:22:42] <crepincdotcom> talk to you guys soon
[03:23:49] <cradek> no problem
[03:23:57] <cradek> good luck :-)
[03:24:14] <cradek> try to cut only a few thousandths deep with the V
[03:25:24] <crepincdotcom> yeah, the main issue with this was that my PCB stock is bent a bit... in both axis
[03:25:43] <crepincdotcom> so i have to make the cuts deep enough to make sure it took the surface off at the (lower) edges
[03:26:01] <cradek> that's why you have to use carpet tape
[03:26:02] <crepincdotcom> heavens forbid i actually *buy* stock
[03:26:05] <crepincdotcom> carpet tape?
[03:26:05] <cradek> they're never flat
[03:26:13] <crepincdotcom> oh
[03:26:20] <crepincdotcom> wells thats good to know
[03:26:32] <crepincdotcom> so the tape holds it from the bottom then
[03:26:51] <cradek> yes
[03:27:11] <crepincdotcom> I suppose Home Depot would have said tape?
[03:27:12] <cradek> and use an extra piece of pcb, or something else (flat), so you can drill your pcb
[03:27:20] <cradek> yeah or any hardware store
[03:27:33] <cradek> the kind with the filaments in it comes off easier
[03:27:37] <crepincdotcom> under it, that is, so that the bit can travel all the way through i imagine
[03:27:40] <crepincdotcom> ok thanks
[03:28:01] <cradek> yeah
[03:28:16] <jepler> in this picture of cradek's mill you can see the perfboard that is quasi-permanently taped to the actual mill table: http://timeguy.com/cradek-files/cnc/milling-setup.jpg
[03:28:43] <crepincdotcom> aha
[03:28:54] <cradek> yeah that's taped down with "more permanent" tape (the kind without the filaments)
[03:29:03] <cradek> I use the other tape to attach the pcb on top of it
[03:29:08] <crepincdotcom> ok
[03:29:22] <cradek> you can see how it's milled away and drilled (I cut out the pcb that way too)
[03:29:23] <jepler> the smart money says that this is the PCB shown on the monitor in axis: http://timeguy.com/cradek-files/cnc/pcb/stepdriver2.jpg (bottom center of circuit board is onscreen)
[03:30:19] <crepincdotcom> so the white is areas the mill has been?
[03:30:29] <crepincdotcom> thats quite a lot of area for a 0.008" bit
[03:31:00] <cradek> which white?
[03:31:03] <jepler> and here's a simulation of the isolation traces being cut with a V tool: http://axis.unpy.net/files/01169521961/gdepth7-pcb-multitool.png
[03:31:15] <crepincdotcom> cradek: in the picture jepler just linkes
[03:31:16] <crepincdotcom> *d
[03:31:37] <cradek> yes the white lines on the screen are the tool paths for the small V bit
[03:31:42] <jepler> crepincdotcom: yes, the vast majority of the distance milled is with the V tool
[03:31:50] <crepincdotcom> wow
[03:32:03] <crepincdotcom> i guess i'll be sitting here for a couple hours while this mills :-p
[03:32:06] <cradek> that's why it's carbide...
[03:32:12] <crepincdotcom> ahh
[03:32:25] <cradek> yeah a complex board can take a while
[03:32:27] <crepincdotcom> well the girlfreind is poking me to stop milling things and pay attention to her
[03:32:34] <jepler> you should do that, at least one day a year
[03:32:36] <crepincdotcom> i guess i should attend to that and finish this tomorrow
[03:32:39] <crepincdotcom> yeah
[03:32:44] <cradek> goodnight
[03:32:51] <crepincdotcom> night night :-p
[03:35:20] <Twingy> k, 1 more bug to fix and next release
[04:01:01] <owhite> mr jepler: got your ears on?
[04:01:29] <cradek> I doubt he'll be back tonight
[04:02:40] <owhite> ah. perhaps I someone else will know - when I use 'addf oneshot.0 base-thread' in my hal file, it reports: "HAL: ERROR: function 'oneshot.O' needs FP"
[04:03:03] <cradek> the base thread doesn't have floating point. you'll have to stick it in the servo thread instead
[04:03:13] <owhite> okay.
[04:03:30] <cradek> the base thread is only used for things that have to run very fast, like step generation
[04:03:56] <cradek> it's best to keep stuff out of there unless it's really necessary, since it runs so often
[04:04:46] <owhite> thanks. could I get a thumbnail picture of what exactly a base thread and a servo thread is?
[04:05:18] <cradek> the servo thread is what updates the requested position, usually at 1kHz
[04:05:52] <owhite> how thread types are there?
[04:05:54] <cradek> base thread runs much faster (often as fast as possible) to generate step pulses, or read raw quadrature signals, or anything else that must happen very fast
[04:06:06] <owhite> sorry, how _many_ thread types are there?
[04:06:38] <cradek> emc2 usually uses only those two threads, but you can have as many as you want
[04:06:48] <owhite> so its sort of like a queue, and you can tie things to those queues.
[04:07:07] <owhite> like hardware in a computer gets tied to queues.
[04:07:14] <cradek> when you make a thread, it has a period (run me every so often)
[04:07:26] <cradek> when you addf, you arrange for a certain function to be called every time the thread runs
[04:07:33] <owhite> oh okay. that makes sense.
[04:07:47] <cradek> if that function requires FP, you need a thread with FP
[04:08:46] <owhite> earlier tonight I was told about oneshot, which looked really good for producing a pulse to go my laser. so I was working on a slider in pyvcp that sends a float to oneshot.
[04:09:11] <owhite> jepler gave me a suggestion to convert pulse to a pin by altering the component code.
[04:09:25] <cradek> ok
[04:09:25] <owhite> thanks to your suggestion, it looks like its working :-)
[04:09:36] <cradek> great
[04:09:59] <cradek> in the servo thread, you'll get multiples of 1ms on your output
[04:10:12] <cradek> if you need more resolution, you'll have to make a faster thread
[04:10:22] <owhite> oneshot gets passed a variable for length, which is seconds.
[04:10:32] <owhite> do you think I could pass it 0.001 of a second?
[04:10:43] <cradek> right but it only has the opportunity to change its output when its function runs on that thread
[04:10:52] <cradek> yes you should be able to get 1ms
[04:10:59] <owhite> so cool.
[04:11:35] <owhite> before this to send a short signal I was using electonics and a foot switch. now it will be pyvcp and software. its the simple pleasures in life.
[04:12:19] <cradek> neat
[04:12:40] <cradek> the flexibility you get with software is nice
[04:12:58] <owhite> sure is. wire wrap is dead :-)
[04:16:13] <owhite> thanks. gonna run.
[05:53:12] <ejholmgren_> alo?
[06:22:15] <paragon36_> paragon36_ is now known as paragon36
[07:38:08] <tray> I'm wondering if anyone can help config a PCI parallel port card I've added to my computer??
[07:52:37] <tray> anyone??
[07:59:09] <anonimasu> tray: they are all sleeping
[07:59:37] <tray> fair enough, one must sleep.
[08:00:02] <tray> Are you running a machine from your parallel port?
[08:00:09] <anonimasu> yes
[08:00:10] <anonimasu> but my internal one
[08:00:22] <tray> Did you have any troubles getting started?
[08:02:51] <anonimasu> no
[08:03:22] <anonimasu> though im using a hardware step/dir generator
[08:03:38] <anonimasu> but, I've run emc with steppers directly from the port before
[08:04:02] <anonimasu> if you can get your port working it's simply modifying your hal file to fit or your pins to fit the hal config..
[08:04:15] <anonimasu> then setting up your ini :)
[08:06:58] <ejholmgren_> http://imagebin.org/7305
[08:07:16] <ejholmgren_> think that will work for a cheap-ass end bearing on my leadscrew?
[08:07:47] <anonimasu> if your I'd use dual nuts..
[08:07:58] <anonimasu> so you can lock them aginst eachother
[08:08:01] <ejholmgren_> acme nuts and lock washers sandwiching a pair of bearings
[08:09:30] <tray> Thanks, anonimasu. That's confirmed the basic approach.
[08:09:41] <tray> Are you familar with using the halscope?
[08:09:59] <anonimasu> not very
[08:10:47] <anonimasu> ejholmgren_: that'll work :)
[08:11:40] <ejholmgren_> split/spring lock washers ... or whatever they're called
[08:12:06] <anonimasu> hm, im not sure how thoose look
[08:12:11] <ejholmgren_> http://images.google.com/images?q=tbn:1tOlsCfPSxIcvM:http://www.mfgsupply.com/img/rotary/17-141.jpg
[08:12:25] <anonimasu> oh
[08:12:34] <anonimasu> I think it would be better with dual nuts..
[08:12:49] <anonimasu> as you wouldnt have to tighten aginst the bearings to get it to lock..
[08:13:19] <anonimasu> or http://www.nordlock.de/web/jpg/advantag.jpg'
[08:13:23] <anonimasu> http://www.nordlock.de/web/jpg/advantag.jpg
[08:13:50] <ejholmgren_> want to tighten it a little with that approach though I think ... acme OD is the bearing ID
[08:14:07] <anonimasu> yeah, but you want to tighten it to get preload then lock it..
[08:14:09] <anonimasu> bbl..
[08:17:00] <ejholmgren_> interesting, haven't seen those before
[09:00:07] <anonimasu> ejholmgren_: thoose lock without much force..
[09:00:29] <anonimasu> we have this test thing at work and if you tighten them it takes 3 times the force to get them to loosen
[09:00:48] <anonimasu> you know toy with nut :) and a washer..
[13:22:54] <CIA-19> 03jepler 07v2_1_branch * 10emc2/docs/src/hal/drivers.lyx: improve discussion of parport x-mode. explicitly mention 'lspci -v' for finding port addresses
[13:24:46] <CIA-19> 03jepler 07TRUNK * 10emc2/docs/src/hal/drivers.lyx: merge rev improve discussion of parport x-mode. explicitly mention 'lspci -v' for finding port addresses
[13:28:56] <Twingy> erDiZz, switched from notebook to vbox and problem went away
[13:29:20] <erDiZz> Twingy, so you didn't file a bug?
[13:29:49] <Twingy> nope
[13:30:20] <Twingy> if you think about it, I shouldn't be using a notebook in my usage
[13:30:40] <Twingy> it's the wrong container
[13:30:58] <Twingy> once upon a time it made sense to use one though
[13:31:16] <erDiZz> yeah, that was the first thing I tried to figure out - "why notebook" :)
[13:31:49] <Twingy> 6 months ago when the design was different it was suitable
[13:31:58] <Twingy> as things changed it made less and less sense to be there
[13:32:11] <Twingy> * Twingy goes to work
[13:32:22] <erDiZz> well, no problems - no bugs. gtknotebook.c can be fun for nobody
[13:36:52] <CIA-19> 03jepler 07TRUNK * 10emc2/docs/src/hal/drivers.lyx:
[13:36:51] <CIA-19> simplify the discussion of loading drivers, since they're all loaded with
[13:36:51] <CIA-19> 'halcmd loadrt' and unloaded with 'halcmd unload'.
[14:31:57] <SKUNKWORKS> SKUNKWORKS is now known as skunkworks
[15:04:51] <cradek> * cradek hates having to fish for a complete bug report
[15:06:32] <jepler> cradek: not half as much as people hate giving complete bug reports
[15:06:44] <cradek> why??
[15:07:13] <jepler> I dunno, but clearly that must be the case
[15:07:51] <erDiZz> bureaucracy the open source way
[15:13:29] <skunkworks> * skunkworks loves making bug reports.
[15:13:37] <skunkworks> evil laugh
[15:17:21] <cradek> The flaw was in a "malware protection engine" that helps several Microsoft security products -- including "Windows Defender" for Vista -- guard against online threats. The problem could let an outsider "take complete control" of a victim's computer, according to Microsoft's security advisory.
[15:17:43] <cradek> thanks, first patch tuesday
[15:20:40] <skunkworks> It doesn't suprise me
[15:30:00] <Dallur> cradek: They should have phrased it like: The problem code has let outsiders take complete control of victim's computers
[15:47:07] <CIA-19> 03jepler 07TRUNK * 10emc2/docs/src/Submakefile:
[15:47:07] <CIA-19> improve configure: get rid of stuff that corresponds to old systems we no longer support.
[15:47:07] <CIA-19> make it scream if python support is not explicitly disabled, but something required isn't found
[15:47:07] <CIA-19> make --enable-build-documentation=pdf and --enable-build-documentation-html build only one type of documentation
[15:47:07] <CIA-19> make it scream if documentation is explicitly requested, but something required isn't found
[15:47:09] <CIA-19> try to make 'make install' work when python support is disabled (untested)
[15:47:09] <CIA-19> 03jepler 07TRUNK * 10emc2/src/ (configure configure.in Makefile Makefile.inc.in):
[15:47:11] <CIA-19> improve configure: get rid of stuff that corresponds to old systems we no longer support.
[15:47:13] <CIA-19> make it scream if python support is not explicitly disabled, but something required isn't found
[15:47:15] <CIA-19> make --enable-build-documentation=pdf and --enable-build-documentation-html build only one type of documentation
[15:47:17] <CIA-19> make it scream if documentation is explicitly requested, but something required isn't found
[15:47:19] <CIA-19> try to make 'make install' work when python support is disabled (untested)
[15:47:21] <CIA-19> 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/axis/Submakefile:
[15:47:23] <CIA-19> improve configure: get rid of stuff that corresponds to old systems we no longer support.
[15:47:25] <CIA-19> make it scream if python support is not explicitly disabled, but something required isn't found
[15:47:31] <CIA-19> make --enable-build-documentation=pdf and --enable-build-documentation-html build only one type of documentation
[15:47:33] <CIA-19> make it scream if documentation is explicitly requested, but something required isn't found
[15:47:35] <CIA-19> try to make 'make install' work when python support is disabled (untested)
[17:13:33] <CIA-19> 03jepler 07TRUNK * 10emc2/src/ (configure configure.in): fix simulator to build again
[17:49:11] <awallin> I'm trying to host an SVG file on my blog, it's not going great and support says "You will need to make a mime type in order for the svg file to work. " anyone know what I should do?
[17:51:12] <awallin> this is under wordpress
[17:58:58] <a-l-p-h-a> awallin, .htaccess file add a mimetype.
[17:59:02] <a-l-p-h-a> apache setting.
[17:59:12] <a-l-p-h-a> awallin, you do realize that it'll need support from the browser as well.
[17:59:31] <a-l-p-h-a> msie doesn't support it, and ff needs a plugin I think. opera works though I think
[17:59:53] <awallin> on ff it works fine from the hdd
[18:00:06] <awallin> so what would I add to the .htaccess file ?
[18:01:04] <a-l-p-h-a> let me check
[18:02:06] <a-l-p-h-a> http://wiki.svg.org/MIME_Type
[18:03:56] <awallin> does the .htaccess need to be in the same dir as the svg ? I only have one .htaccess in the root foled
[18:04:00] <awallin> folder
[18:07:42] <awallin> still doesn't work...
[18:08:01] <a-l-p-h-a> awallin, depends on how apache is setup.
[18:08:13] <a-l-p-h-a> sometimes it can use the parent .htaccess... sometimes you need it in the same folder.
[18:17:56] <a-l-p-h-a> awallin, working?
[18:18:29] <pier> BitchX-1.1-final+ by panasync - Linux
[18:18:47] <awallin> nope, I put the two lines suggested on the webpage in both the root htaccess and a new htaccess in the svg file dir - but no success.
[18:19:28] <awallin> now support email some further suggestions for changing settings in the admin panel of the site...
[18:23:57] <awallin> dman svg
[18:24:03] <awallin> damn
[18:26:12] <a-l-p-h-a> why svg?
[18:26:18] <a-l-p-h-a> just toying around?
[18:26:36] <lerneaen_hydra> 'lo all
[18:26:38] <a-l-p-h-a> I'm looking into a capchat to prevent spam on my site...
[18:26:43] <a-l-p-h-a> hey lerneaen_hydra.
[18:28:27] <awallin> going home. bbl
[18:43:51] <lerneaen_hydra> random question, what's a bullnose mill shaped like?
[18:51:17] <SWPadnos> lerflat on the bottom but with rounded corners, I think
[18:51:28] <SWPadnos> err - lerneaen_hydra, flat ...
[18:52:59] <a-l-p-h-a> it's not flat
[18:53:04] <a-l-p-h-a> http://www.millstar.com/techarticles/article_004.htm
[18:53:45] <SWPadnos> oh, so the exact opposite of what I thought. good. :)
[18:53:54] <a-l-p-h-a> SWPadnos,. :) good man.
[18:54:18] <a-l-p-h-a> SWPadnos,. :) next time, lets both say "playing with live electrical wires in your mouth is uber fun!" :)
[18:54:34] <SWPadnos> it is - how do you think I got so smart?
[18:55:30] <a-l-p-h-a> thought it was just good genes.
[18:55:38] <SWPadnos> oh - right
[18:55:45] <a-l-p-h-a> heh... reading bash.
[18:56:11] <a-l-p-h-a> never donate sperm... cause when some 18yr old finds you later in life, you'll have to explain that you did it for beer money.
[18:57:08] <anonimasu> heh
[19:04:37] <fenn> SWPadnos: you were right you just didnt know it
[19:04:45] <SWPadnos> I rarely do
[20:29:00] <a-l-p-h-a> quiet day in here
[20:33:33] <SWPadnos_> SWPadnos_ is now known as SWPadnos
[20:34:06] <SWPadnos> well. I guess it's time to replace the battery in the UPS that my router and DSL modem are plugged into
[20:41:51] <skunkworks> I found out today that my ups will not keep my computer and 2 monitors running after the breaker tripped.
[20:43:37] <jepler> I don't plug monitors into the UPSes
[20:49:07] <a-l-p-h-a> jepler, you prob should do one. :)
[20:49:25] <a-l-p-h-a> unless you have super human ability to move a mouse around blindly. :)
[20:49:47] <jepler> nah, I mostly just cross my fingers and hope the power comes on again before the battery fails
[20:49:56] <a-l-p-h-a> jepler,. hahaha
[20:50:24] <a-l-p-h-a> some reason, my ups just tripped... a few weeks back... it was just dead. I couldn't get it back on, till I turned off the UPS, and reset it.
[20:50:34] <a-l-p-h-a> never had to do that before... but did then
[22:30:43] <robin_sz> meep?
[22:31:41] <robin_sz> heres how i see it
[22:31:52] <robin_sz> the electricity supply is very stable
[22:32:18] <robin_sz> in order to offer an improvement, the ups needs to be an order of magnitude more reliable
[22:32:32] <robin_sz> any ups less reliable, is counter productive
[22:33:54] <fenn> its easy enough to test if they are going to work or not (pull the plug)
[22:34:12] <robin_sz> nope
[22:34:32] <robin_sz> thats a measure if 'if it will work at time point x'
[22:34:53] <robin_sz> im moreinterest in failures to work
[22:35:06] <robin_sz> and work includes just passing through
[22:35:32] <fenn> the ups can fail 50% of the time and still be better than nothing
[22:35:59] <fenn> if it never works at all, its worse than nothing
[22:36:15] <robin_sz> ok,
[22:36:18] <fenn> somewhere between 0 and 50% is a magic number where you break even
[22:36:39] <robin_sz> so .. if I offered you a UPS that worked 75% of the time .. you'd buy it?
[22:36:48] <fenn> sure, how much?
[22:36:54] <robin_sz> $10
[22:36:57] <robin_sz> 5kva
[22:36:57] <fenn> ok
[22:37:01] <robin_sz> great!
[22:37:03] <fenn> i bet shipping's a bitch tho :(
[22:37:24] <robin_sz> its down for about 15 minutes in every hour,
[22:37:39] <fenn> it actually shuts off power?
[22:37:44] <robin_sz> yeah
[22:37:54] <fenn> how is that even possible?
[22:38:20] <robin_sz> because all, or nearly all modern UPS's work that way
[22:38:40] <robin_sz> they dont just switch from mains to battery
[22:38:49] <robin_sz> they can top up the mains if it dips a little
[22:38:56] <robin_sz> or reduce it if its over voltage
[22:39:14] <robin_sz> they are known as "online UPS's"
[22:39:27] <robin_sz> the APC smartUPS range is the most popular
[22:40:30] <robin_sz> so, still think an uptime of 50% is OK?
[22:41:07] <robin_sz> persoanlly, I;d want less than 1 "event" in say, 2 years
[22:41:58] <robin_sz> id estimate the number of critical power events in our offices, to be roughly 6 a year
[22:42:15] <robin_sz> id want a UPS to be a factor of 10 better than that
[22:42:20] <robin_sz> at least
[22:42:21] <fenn> of course the (mains + ups) total combined reliability has to be better than mains alone
[22:42:28] <robin_sz> exsactly
[22:42:45] <robin_sz> which, is what I originally said:)
[22:42:59] <fenn> the ups needs to be an order of magnitude more reliable
[22:43:27] <fenn> is what you said
[22:43:28] <robin_sz> ... shall I paste?
[22:43:36] <robin_sz> <robin_sz> meep?
[22:43:36] <robin_sz> <robin_sz> heres how i see it
[22:43:36] <robin_sz> <robin_sz> the electricity supply is very stable
[22:43:36] <robin_sz> <robin_sz> in order to offer an improvement, the ups needs to be an order of magnitude more reliable
[22:43:43] <fenn> yes i can read
[22:43:46] <robin_sz> so we agree, :)
[22:44:03] <robin_sz> * robin_sz shakes fenn warmly byu the hand and buys him a drink
[22:44:11] <fenn> i didnt know that a ups could bring down the mains along with it
[22:44:15] <robin_sz> yeah
[22:44:38] <fenn> so, how's your robot doing?
[22:44:59] <robin_sz> fine .... got the base plate drilled for it this week
[22:45:20] <robin_sz> going to fabricate that
[22:45:57] <SWPadnos> any UPS that isn't expected to be 100% reliable (allowing for the occasional unexpected failure) is worse than no UPS at all, IMO
[22:46:10] <robin_sz> well, 100% is a dangerous figure
[22:46:23] <robin_sz> and unacheivable ....
[22:46:32] <SWPadnos> the reason being that you operate as though the UPS will save you, so some things may not be done that would otherwise be done if there were no UPPS at all
[22:46:46] <SWPadnos> sure, that's why I said 100% expected, less unexpected failures ...
[22:46:59] <robin_sz> the way most datacentres identify rleiability is in "nines"
[22:47:05] <SWPadnos> sure
[22:47:10] <robin_sz> 3 nines, 4 nines ..
[22:47:11] <fenn> nah nobody would expect the ups to be 100% reliabile, that's why you write log files to disk, etc
[22:47:35] <robin_sz> meaning 99.9999% uptime for a 4 nines data centre
[22:47:37] <fenn> six sigma
[22:47:51] <robin_sz> * robin_sz puts on his blackbelt
[22:47:52] <fenn> * fenn karate chops robin_sz in the gut
[22:47:55] <SWPadnos> right, but with a UPS, you expect to get a signal that you need to save state and close all programs since the system will be shut down soon. without one, you might design a more reliable system in the first place
[22:48:23] <fenn> too quick for ya
[22:48:23] <robin_sz> sure, and most UPS's do that
[22:48:45] <SWPadnos> unless they don't work :)
[22:48:57] <SWPadnos> which is why an unreliable UPS is worse than no UPS at all
[22:49:44] <robin_sz> fenn, nah, white crane 3rd dan. you died already but yor body hasn;t noticed ;)
[22:50:28] <robin_sz> that was a complete lie about 3rd dan etc ;)
[22:50:54] <fenn> its ok, i didnt really karate chop you either
[22:51:28] <robin_sz> I got apcupsd running on the main server, the two other servers listen in and it lets them shut down 5 minutes before it does
[22:51:35] <robin_sz> * robin_sz bows to fenn
[22:52:31] <robin_sz> although ... I did buy my 6yr old a karate suit for christmas and have been teaching him the basic kicks etc
[22:53:15] <robin_sz> yoko geri, mashwari geri etc hes quite good with his kick, but his punches really suck
[22:54:17] <robin_sz> * robin_sz orders his new laptop
[22:59:04] <Goslowjimbo> How do I tell which version of head I have? I'm told to need to update to have waitusr available, but I tried an update, and when I executed it, it said I had the latest already installed.
[22:59:08] <lerman> Are any Python programmers around? I have a question.
[22:59:46] <lerman> jepler: cradek: anyone?
[23:00:24] <SWPadnos> Goslowjimbo, you tried using synaptic to do the update?
[23:00:51] <fenn> lerman: dataja?
[23:01:14] <Goslowjimbo> Yes, the first time, then I tried emc-install.sh
[23:01:19] <lerman> Is there a maximum length to a Python 'string'?
[23:01:41] <Goslowjimbo> syanptic told me I had the latest version.
[23:02:02] <SWPadnos> Goslowjimbo, ok. the 2.1 packages are marked as being different from 2.05, since there are some required changes to user configurations
[23:02:22] <SWPadnos> I think you need to uninstall emc2-2.0.5 (though I'm not positive of this)
[23:02:40] <SWPadnos> and you need to enable a different repository for emc 2.1, I believe
[23:02:57] <SWPadnos> actually, before doing anything like that, do this instead:
[23:03:04] <SWPadnos> run synaptic, and search for emc
[23:03:17] <SWPadnos> see if there is a package called emc2.1 or similar
[23:05:04] <Goslowjimbo> All I see is EMC2 with a 1:2.05 version
[23:05:40] <SWPadnos> ok. you'll have to edit the sources list
[23:06:13] <SWPadnos> under Settings->Repositories
[23:06:44] <SWPadnos> click on the emc repository (the binary one first, though you'll probably want to do the same change on the source repo line)
[23:07:20] <jepler> lerman: yes, until you get a 64 bit computer and python2.5, the limit of strings and other containers is 2^31-1 items.
[23:07:39] <SWPadnos> it probably shows dapper as the distribution, and emc2 as the components line
[23:07:47] <Goslowjimbo> Yew
[23:07:51] <Goslowjimbo> Yes
[23:07:54] <SWPadnos> you'll need to change the components line to emc2.1
[23:08:10] <lerman> That isn't my problem. I'm editing an item in a ListCtrl (wxPython) and it seems to beep at me when I have a few hundred characters.
[23:08:18] <lerman> in a line.
[23:08:48] <SWPadnos> jepler, cradek, or alex_joni could tell you if anything else needs to be done (I suspect that emc2.1 conflicts with emc2.0, so synaptic should warn you that you can'r install both)
[23:09:24] <SWPadnos> lerman, it wouldn't be 255+-1 chars, would it?
[23:09:58] <lerman> Actually, I'm using an EditableListBox. I don't know. I think it's fewer chars than that, though. I'd guess around 200.
[23:10:36] <SWPadnos> I'dd bet that it's a wxPython limitation, not a language limitation
[23:10:54] <lerman> For now, I'll leave it be. I can go back and fix it later. (This is for my latest version of GWiz -- my GCode Wizard.)
[23:11:18] <SWPadnos> cool. how's it coming along?
[23:11:18] <lerman> I think I've decided on wxPython. I didn't much care of TKinter.
[23:12:00] <lerman> Well, making some progress. But learning wxWidgets and Python while trying to get something done is *interesting* to say the least.
[23:12:14] <lerman> At least I already have an Object Oriented background.
[23:12:23] <SWPadnos> heh
[23:12:30] <SWPadnos> just not the same objects ;)
[23:12:33] <lerman> Without that, I'd be dead in the water.
[23:12:56] <lerman> Subclassing is still subclassing, though.
[23:13:15] <lerman> (Objective-C is my o-o language from the past.)
[23:13:33] <SWPadnos> NeXTStep rulez!
[23:13:59] <lerman> I was one of the developers of ICpak201 which was the basis for Next's gui stuff.
[23:14:05] <SWPadnos> cool
[23:14:17] <Goslowjimbo> OK, I see 2.1.0 on the synaptic.
[23:14:40] <SWPadnos> I have NeXTstep 3.x (whichever one was released for the PC) - the interface builder on that still hasn't been surpassed, IMO
[23:15:01] <lerman> I never used it much.
[23:15:09] <erDiZz> my experience with wxwidgets (Gtk+ backend) was highly negative, problems here and where... I switched to plain Gtk+ then
[23:15:11] <SWPadnos> Goslowjimbo, try marking it for installation, and see if Synaptic tells you that you'll have to uninstall EMC2-2.0.5
[23:15:32] <lerman> I'm actually enjoying the wxPython stuff. So far, no bugs detected.
[23:15:47] <lerman> And they have a lot of demo code to steal.
[23:16:18] <SWPadnos> that's helpful
[23:16:41] <lerman> I'm developing on the PC (my notebook), but testing periodically on Linux.
[23:16:41] <Goslowjimbo> No, it said it would leave EMC 2.0.5 and add 2.1.0.
[23:16:56] <SWPadnos> hmmm. that sounds like a bad thing to me
[23:17:12] <lerman> See you guys later.... I have a Kempo class to go to.
[23:17:37] <SWPadnos> I'm not sure how you need to proceed, but it's likely that you need to tell it to uninstall emc2.0.5 and friends (like emc2-dev 2.05
[23:17:41] <SWPadnos> see you
[23:18:01] <erDiZz> lerman, I was particularly unsatisfied with look&feel on Linux, didn't like the dialogs, custom widgets were rendered with some fixed-color elements, ignoring the theme engine... black letters on black background...
[23:18:10] <Goslowjimbo> This was when I selected emc2-dev 2.1.0
[23:19:06] <lerman> I'm not into themes. I might even be anti theme. I'll build something that I like the color of and then use it.
[23:19:47] <lerman> Generally I find black and white with some gray to suit me well.
[23:19:59] <lerman> Old eye syndrome, I suspect.
[23:20:19] <erDiZz> I think that if claimed as native, it should honor the themes, since they are there as _native_ ...
[23:21:04] <lerman> I can't argue with that. Since I'm not selling wxPython; just using it, it isn't a problem for me.
[23:21:41] <jepler> SWPadnos: in the case of emc 2.0 vs emc 2.1, the debian package name is the same ("emc2") so installing the version 2.1.0 package will remove the 2.0.5 package, just like installing 2.0.5 would remove an earlier 2.0.x package
[23:21:51] <erDiZz> yep. In C it feels a lot like MFC. Memories of MFC programming finally switched me off :)
[23:22:15] <SWPadnos> jepler, hmm. shouldn't it tell Goslowjimbo that 2.0.5 will be removed then (or upgraded or something)?
[23:27:11] <jepler> SWPadnos: btfoom -- I don't use those GUI interfaces
[23:27:25] <SWPadnos> ok :)
[23:27:53] <SWPadnos> wouldn't apt-get install ... also tell you that emc2-2.1.0 conflicts with 2.0.5, so 2.0.5 will be removed?
[23:28:00] <jepler> hm -- it may be a pacage bug. emc2-dev 1:2.1.0 does not depend on emc2 = 1:2.1.0
[23:28:13] <jepler> SWPadnos: they don't conflict -- they're different versions of the same package
[23:28:24] <jepler> you can only have one version of a package installed at a time
[23:28:46] <SWPadnos> ok, so apt-get should "upgrade" to the later version ...
[23:29:42] <SWPadnos> Goslowjimbo, I assume you reloaded the package lists after changing the sources entries?
[23:29:50] <SWPadnos> (I think you get a prompt, just making sure)
[23:30:08] <Goslowjimbo> Yes, I did.
[23:31:03] <SWPadnos> well, if you hit "apply" in synaptic, a window qill show up telling you what will be added, what will be removed, and what will be upgraded
[23:31:19] <SWPadnos> does emc2-2.1.0 show up as an upgrade to emc2 2.0.5 now?
[23:32:49] <robin_sz> right .. laptop ordered :)
[23:33:01] <robin_sz> * robin_sz looks outside the door
[23:33:05] <Goslowjimbo> I hit return twice, so I lost that screen. I have both EMC2 and EMC2-dev on the same drive. Different directories, but the same drive. Is this messing me up?
[23:33:07] <robin_sz> nope, not here yet
[23:33:37] <SWPadnos> Goslowjimbo, no - emc2-dev just depends on the packages needed to build emc
[23:33:43] <robin_sz> Goslowjimbo, shouldnt do
[23:33:54] <SWPadnos> so having both makes sense if you may want to recompile emc2
[23:34:36] <Goslowjimbo> I need to recompile it or remove it?
[23:34:48] <SWPadnos> you don't need to do any compilation
[23:35:15] <Goslowjimbo> OK, update it or remove it.
[23:35:30] <SWPadnos> if you look at packages under status / upgradable, does emc2 show up there now?
[23:35:46] <Goslowjimbo> Yes, as 2.0.5
[23:35:50] <jepler> what is the actual goal here? If it is upgrading from emc 2.0 to emc 2.1, did something go wrong with the steps shown on the wiki?
[23:36:28] <SWPadnos> jepler, I think it's possible that too many wiki steps were followed - 2.0.5 didn't show up as upgradable
[23:36:56] <robin_sz> oh, goals? well if you are going to get picky and expect people to know what the actual goals are ... ;))))
[23:37:35] <SWPadnos> Goslowjimbo, is this the wiki page you used? http://wiki.linuxcnc.org/cgi-bin/emcinfo.pl?UpdatingTo2.1
[23:37:38] <Goslowjimbo> Funny. I just want to run a DRO display.
[23:37:47] <SWPadnos> I don't think so, since you mentioned emc-install.sh
[23:38:32] <robin_sz> * robin_sz is now confused
[23:39:58] <Goslowjimbo> I haven't seen that page of the wiki before now. I followed the instructions on one of the main pages.
[23:40:31] <SWPadnos> ok. basically, you should ne able to unmark all, then mark all upgrades and go
[23:40:36] <SWPadnos> s/ne/be/
[23:40:53] <SWPadnos> unmark all only because I don't know what you've marked already :)
[23:41:10] <robin_sz> before you go too far ... is it worht just stopping and checking that the version of emc you actually have cant do what you actually want?
[23:41:35] <SWPadnos> robin_sz, no, that's already been done. he's looking for a specific feature of 2.1
[23:41:41] <robin_sz> right
[23:41:45] <robin_sz> just sanity checking
[23:41:54] <SWPadnos> no prob
[23:42:02] <SWPadnos> but, are you qualified for that?
[23:42:06] <jepler> Goslowjimbo: reading back, I think you are confused between the process of getting emc through synaptic / the debian package system, and compiling a version of emc yourself.
[23:42:07] <robin_sz> sanity?
[23:42:16] <robin_sz> doubtful
[23:42:28] <jepler> earlier you said "How do I tell which version of head I have", which indicates you're using a self-compiled version of emc
[23:42:43] <robin_sz> * robin_sz nods
[23:42:54] <jepler> if that is the case, then doing any number of actions in the debian package manager will not update the self-compiled version
[23:43:29] <robin_sz> unless of course you built emc, packaged it and then dpkg -i 'd it ...
[23:43:49] <SWPadnos> robin_sz, shhhh!
[23:45:10] <robin_sz> SWPadnos, you have to admit, if people "make install" an emc on top of a dpkg installed package, it gets a bit messy ... packageing your own is probably a snae option
[23:45:21] <robin_sz> sane
[23:45:25] <fenn> yep checkinstall is nice for that
[23:45:44] <fenn> dont want my machine cluttered up by stupid binary installers
[23:46:08] <fenn> emc -v should tell you something
[23:46:09] <SWPadnos> it could be, but it's a bit more complex than even compiling from CVS
[23:46:57] <robin_sz> there is alot to be said for only installing packages on a dpkg managed system, yes it is a bit more complex for "many builds" style developemnt, but for users, even users of "head CVS" it might make a lot of sense
[23:47:34] <robin_sz> probably worht considering for user installs as opposed to development installs ??
[23:48:47] <robin_sz> or, would there be any mileage in a nightly package build from CVS head? failry easy to automate??
[23:49:13] <SWPadnos> probably could be automated, but I don't think the userbase is the kind that would need that
[23:49:26] <ejholmgren_> base base base base
[23:49:33] <SWPadnos> since it's pretty easy to compile/run-in-place, the whole package thing is a red herring
[23:49:52] <SWPadnos> (at least, the self-built package thing)
[23:50:12] <robin_sz> I was just thinking of the problems that can begin when packages and make-installs get to fighting with each other
[23:50:43] <SWPadnos> sure, but a reinstall of the package would (should) overwrite the make install files
[23:51:02] <robin_sz> let me give you a non-emc example ...
[23:51:12] <fenn> all your userbase..
[23:51:15] <robin_sz> I have a asterisk install from packages
[23:51:29] <robin_sz> and then I built one by hand and make-installed it
[23:51:43] <fenn> you fool!
[23:51:47] <robin_sz> now, I cant apt-get remove it .. incase it screws my hand-built install
[23:51:58] <ejholmgren_> I posted this fairly late last evening .. but more people are on now http://imagebin.org/7305
[23:52:06] <jepler> (this is one reason that emc2 has --enable-run-in-place -- no file is created outside the emc2 top level directory)
[23:52:19] <ejholmgren_> any problems with sandwiching two bearings with acme nuts on the leadscrew?
[23:52:28] <robin_sz> and it keeps wanting to upgrade various sub-packages like zaptel ... which is also frightening
[23:52:56] <fenn> robin_sz: cant you just remove the package and re-install the build?
[23:53:12] <robin_sz> jepler, sure, I understand ... but .. I'm willing to bet that people do occasionally do a full make install and get screwed
[23:53:40] <fenn> ejholmgren_: acme nuts dont lock very well (by design)
[23:53:42] <robin_sz> fenn, I could .. but ... well, its working with 60 users on that phone system now ... its a bit "worrying"
[23:53:51] <jepler> robin_sz: yes, so don't do that
[23:54:09] <SWPadnos> wasn't that an example of why you shouldn't make your own packages?
[23:54:19] <SWPadnos> or am I missing something?
[23:54:21] <robin_sz> nope
[23:54:32] <fenn> ejholmgren_: you can drill a setscrew hole perpendicular to the leadscrew, put a piece of plastic in so it doesn't mar the thread
[23:54:34] <ejholmgren_> fenn: double them up? lockwashers?
[23:54:48] <robin_sz> If I had made my own package, i would have had to remove the debian supplied one before installing mine
[23:54:49] <SWPadnos> ah - you built asterisk and make-installed it, you didn't build a new package by hand and install it
[23:54:55] <fenn> ejholmgren_: in the acme nut
[23:54:57] <robin_sz> exactly
[23:55:07] <robin_sz> SWPadnos, exactly .. stupid I know ...
[23:55:09] <alex_joni> good evening gents
[23:55:16] <SWPadnos> hi Alex. back so soon?
[23:55:21] <robin_sz> gents?
[23:55:27] <robin_sz> * robin_sz straightens his dress
[23:55:32] <SWPadnos> it's OK. I understood him :)
[23:55:36] <alex_joni> yeah.. got home a bit earlier
[23:55:37] <robin_sz> * robin_sz straightens HER dress
[23:55:50] <ejholmgren_> just trying to figure out how to secure my acme rod without machining
[23:55:52] <alex_joni> robin_sz: want my kilt? :P
[23:56:00] <fenn> ejholmgren_: also i find that setscrews are easier to use than trying to adjust locknuts
[23:56:09] <fenn> just preload, lock
[23:56:21] <robin_sz> alex_joni, and the traditional scottish underwear?
[23:56:22] <fenn> not, preload, tighten, find out it's too tight, order replacement bearings, etc..
[23:56:25] <alex_joni> SWPadnos: took a 1.5 day auto ride in one day
[23:56:29] <alex_joni> robin_sz: that too
[23:56:49] <SWPadnos> robin_sz, you should be able to back up your configs, apt-get uninstall, then redo the make-install
[23:56:52] <ejholmgren_> had a neighbour on the corner of my block growing up who would play the bagpipes on weekends ...
[23:57:09] <SWPadnos> and then of course restore the configs and data, and hope it all works :)
[23:57:37] <crepincdotcom> cradek jepler: im about to do a board with a 1mm roughing endmill and a 60 degree V for the final, at 0.017 depth (which, by my calculations, gives 10-mil depth)
[23:57:40] <SWPadnos> there is some apt-foo that can force the system to think that the package is not installed, but I don't know it offhand
[23:57:42] <crepincdotcom> does this sound ok?
[23:57:46] <robin_sz> SWPadnos, yeah, I think I should have built it as apackage, then at least I gould have removed it, tried another, put it back nicely etc when upgradign or adding new features etc .. packages are nice like that
[23:58:30] <SWPadnos> package machinations wouldn't help with configs anyway, for the most part
[23:58:42] <robin_sz> configs no
[23:58:51] <robin_sz> as they should be detected as hand-modified and left alone