#emc-devel | Logs for 2005-11-08

[00:57:42] <jblunzen> hidiho
[01:07:40] <jblunzen> jblunzen is now known as cncuser
[01:08:16] <cncuser> "Welcome to the Enhanced Machine Control development place. | Regular Developers' meetings 24/7 !"
[01:08:22] <cncuser> seems over ;)
[01:08:27] <cncuser> good night folks
[20:58:53] <chinamill> Hello
[20:59:05] <chinamill> No crash today...
[20:59:53] <chinamill> The red line in axis do not follow the psykadilic white (4 axis plot)
[21:00:14] <cradek> I looked at your picture
[21:00:20] <cradek> it doesn't tell me anything really
[21:00:30] <cradek> what I need is a simple (!) program that reproduces the problem
[21:00:39] <cradek> I would think the right ten lines or so should do it
[21:00:43] <chinamill> I arranged a new screenshot, but it is on the mill
[21:01:06] <alex_joni> chinamill: try to look XY, XZ or other views and say where it first appears
[21:01:08] <chinamill> I'll deliver it latar
[21:01:16] <chinamill> Same thing
[21:01:19] <cradek> I'm not actually sure what axis *should* do
[21:01:24] <cradek> what it *does* is a whole other store
[21:01:25] <cradek> story
[21:02:13] <chinamill> Axis is much more good looking than mini, but I miss some features
[21:02:26] <cradek> like what?
[21:03:10] <chinamill> The offset view (or what I shall call it) is good with mini
[21:03:25] <cradek> I don't know what that is
[21:03:43] <alex_joni> clicking on an axis lets you offset it
[21:03:53] <alex_joni> G10 iirc? although I never used that really
[21:04:09] <cradek> well, I like the axis way better
[21:04:21] <chinamill> I'll try and fix some screenshots and then we can talk about it, Axis is a great job thou!
[21:04:51] <cradek> I use offsets all the time with axis
[21:05:14] <chinamill> It's more powerful with mini
[21:05:20] <chinamill> easyer
[21:05:43] <cradek> I dislike "hidden" functionality - guess where to click to offset the axis? Not a button, you click on a label, nothing on the screen to document it
[21:05:54] <chinamill> There is one more this I miss in Axis, it the ability to set the speed when jogging manual style
[21:06:03] <chinamill> :)
[21:06:10] <cradek> you can do that with feed override
[21:06:35] <cradek> I see no reason to have two sliders
[21:06:48] <chinamill> I guess... I have to lear how to tweak the ini file to get a better Axis setup
[21:07:36] <chinamill> Now my ini values does not make sense
[21:07:49] <chinamill> When I run Axis...
[21:08:22] <chinamill> The default speed when jogging is slower than slow (running emc metric)
[21:08:50] <cradek> Honor [TRAJ] DEFAULT_VELOCITY for jogs. I think it's silly to have a
[21:08:50] <cradek> separate slider for jogs ("axis speed" in tkemc), so I suggest setting
[21:08:50] <cradek> DEFAULT_VELOCITY to approximately MAX_VELOCITY and using feed override
[21:08:50] <cradek> to slow down when necessary.
[21:09:04] <cradek> (from the axis cvs)
[21:09:13] <cradek> obviously, not the best place for documentation
[21:10:03] <chinamill> And to move a axis 0.1 is not so much in mm:s would be graet with a option to get more metric style values for stepping values
[21:10:17] <chinamill> You get what I'm thinking?
[21:10:32] <cradek> oh, the steps are too small? I suppose you're right
[21:10:39] <chinamill> yep
[21:11:13] <alex_joni> use MDI :)
[21:11:20] <cradek> what should they be?
[21:11:21] <chinamill> I wonder, are all You people using inch g-code?
[21:11:26] <cradek> I do
[21:11:34] <chinamill> ok... MDI?
[21:12:31] <chinamill> Maybe 0.01 ; 0.1 ; 1 ; 5 ; 10
[21:13:04] <chinamill> One inch is about 25mm
[21:13:28] <cradek> I'm in the US, but I do know how big a mm is :-)
[21:13:44] <chinamill> Are You sure ;)
[21:13:50] <cradek> yes, quite
[21:14:12] <chinamill> Only joking...
[21:14:23] <chinamill> I cant spell
[21:14:25] <cradek> how about 1.5875, 3.175, 6.35, 12.7, 25.4 mm
[21:14:34] <cradek> * cradek snickers
[21:14:59] <alex_joni> rofl.. how about those values for inches?
[21:15:10] <alex_joni> so no-one is happy :D
[21:15:16] <cradek> haha
[21:15:48] <cradek> actually, once in a while, fractional inch jogs would be nice, since our tooling is fractional
[21:15:56] <cradek> I never thought about it really.
[21:16:05] <alex_joni> yeah.. but no cheating about 1/8 inches and such
[21:16:14] <alex_joni> you need 0.27531 "
[21:16:14] <cradek> we measure in thousandths, but the tooling is fractional
[21:16:15] <alex_joni> :D
[21:16:21] <cradek> it's asinine actually
[21:17:20] <cradek> metric tooling is very expensive here, or else I'd probably use it
[21:17:39] <cradek> my manual mini mill and lathe are metric
[21:17:46] <cradek> my manual full size lathe is inch
[21:18:05] <cradek> I can use either system comfortably
[21:18:06] <jepler> back to the size of incremental jogs, maybe we should read it from the INI file
[21:18:22] <chinamill> Would be nice!
[21:18:28] <cradek> jepler: you're probably right
[21:19:51] <chinamill> Some support for copy and paste would also be nice
[21:19:59] <chinamill> Anyone miss that?
[21:20:14] <cradek> copy and paste what?
[21:20:59] <chinamill> Like cody a g line from history to paste it in the mdi line, edit alittle and run it
[21:21:11] <cradek> you can already do that
[21:21:29] <jepler> does it work with middle-button paste, like old X apps?
[21:21:32] <chinamill> I guess I have to update my axis
[21:21:38] <cradek> testing...
[21:22:10] <chinamill> Does copy apply to the g-code interpreter window
[21:22:24] <cradek> yeah you can copy and paste previous MDI commands
[21:23:00] <cradek> no, you can't copy from the program display
[21:23:31] <chinamill> How would I upgrade Axis to run on a 4.30 BDI system, the brain dead way?
[21:23:57] <chinamill> I miss that, copy from the program display
[21:23:58] <cradek> get a cvs snapshot from axis.unpy.net
[21:24:15] <cradek> clicking in the program display highlights that line in the preview window
[21:24:20] <chinamill> Not brain dead enogh
[21:25:09] <chinamill> I cant get a deb packge somewhere?
[21:25:50] <cradek> no.
[21:26:10] <chinamill> Another thing I miss i Axis is, abort - back, forward, restart
[21:26:40] <chinamill> You get what I think?
[21:26:55] <alex_joni> A)bort , R)etry, I)gnore
[21:26:55] <alex_joni> =))
[21:31:14] <chinamill> ok, If I try hard not to be so much brain dead, I can use cvs to get latest code. Do I need to compile anything to upgrade? (Sorry for the stupid questions)
[21:31:44] <chinamill> (upgrading axis that is)
[21:37:06] <cradek> no, you don't have to compile
[21:37:19] <cradek> I don't understand what you mean by abort - back, forward, restart
[21:39:35] <cradek> to update, get the latest cvs snapshot, unpack it, read README, look for the section about BDI
[21:39:54] <chinamill> It is functionallity from mini
[21:40:03] <jepler> what should this .ini line for incremental jog distances look like? I think that "[DISPLAY] JOG_INCREMENTS = .1 .01 .001" isn't good enough, because the value that should be shown to the user and the value in machine units might be different.
[21:40:41] <jepler> JOG_INCREMENT0 = .1 JOG_LABEL0 = 1/10 inch ?
[21:41:05] <jepler> JOG_INCREMENT1 etc
[21:41:12] <cradek> I can imagine wanting increments of .001", 1/16", .01", 1mm, ...
[21:41:32] <jepler> cradek: yes, I can see wanting a mix of metric, decimal inch and binary fractional inch as well
[21:41:36] <cradek> but I have no idea how to specify them
[21:41:37] <chinamill> nope I think
[21:41:38] <jepler> for those of us stuck in inchland, anyway
[21:42:08] <cradek> chinamill: I have never used mini. Can you describe what you mean?
[21:42:21] <chinamill> sure.
[21:42:46] <jepler> set activedecr [button $programrestart.b1 -text "Back" -command {incr restartlin
[21:42:49] <jepler> e -1} ]
[21:42:49] <jepler> oops, excuse the bad wrapping
[21:43:13] <jepler> it looks like Back / Forward are buttons you can press to change the line emc will run from when you start milling.
[21:43:21] <cradek> ugh
[21:43:29] <alex_joni> cradek: think it's related about running a program in Auto, then stopping it and beeing able to go back/forward one line
[21:43:38] <cradek> in axis, you click on the line you want to start from
[21:43:50] <jepler> to set the desired line to run from in Axis, click the line in the preview plot or code area, then choose Program > Set Next Line from themenu.
[21:44:07] <alex_joni> chris: can you double-click a line and make it run in MDI ?
[21:44:27] <cradek> alex_joni: no, but you can double-click to copy, middle click to paste, click run
[21:44:29] <alex_joni> that would be too cool ;)
[21:44:31] <chinamill> ok, I'm not yet a power user...
[21:44:48] <alex_joni> cradek: how about on a MAC mouse? :D
[21:44:56] <cradek> troll!
[21:45:22] <alex_joni> did you hear they have 2-button mouses now?
[21:45:26] <alex_joni> with scroll even :D
[21:45:36] <cradek> OT OT OT
[21:45:52] <alex_joni> OT?
[21:45:58] <cradek> "off topic"
[21:46:11] <alex_joni> it's GUI related :D
[21:46:20] <jepler> AXIS doesn't support mac
[21:46:27] <jepler> does mini?
[21:46:43] <alex_joni> jepler: you sure it wouldn't run on a mac?
[21:47:17] <jepler> alex_joni: no, but if it does that's a bug I'll fix immediately j/k
[21:47:19] <alex_joni> * alex_joni will shut up now, and leave porting rcslib to mac os-x to someone else
[21:47:55] <alex_joni> jepler: (OT) I even started to compile/install AXIS on doze ;) but gave up after a while..
[21:48:13] <alex_joni> but rest assured .. it's still in the back of my head :))
[21:54:40] <cradek> I put explicit install instructions for bdi4.30 on axis.unpy.net and updated the README in cvs
[22:32:49] <jepler> I just checked in a few more bugfixes to axis
[22:33:01] <cradek> yay
[22:33:30] <alex_joni> ditto
[22:36:11] <jepler> fix an infrequent error where axis would die at startup because linear_units was inexplicably 0.
[22:36:14] <jepler> fix the worst of the "my mouse wiggled while I was clicking the code display" problems by stomping on the B1-Motion and B1-Leave Text widget bindings
[22:38:36] <alex_joni> axis gets installed into a path-bin dir, right?
[22:39:14] <jepler> it gets installed into $EMCROOT/plat/$PLAT/bin/ or so
[22:39:20] <alex_joni> so probably "if [-x $EMC2_BIN_DIR/$EMCDISPLAY ] ;" won't find it
[22:39:27] <alex_joni> jepler: scrap emc :D
[22:39:30] <alex_joni> I mean emc2
[22:39:58] <alex_joni> on the BDI I've seen it in /usr/bin ?
[22:40:28] <jepler> huh. I don't think it's installed there in ages
[22:40:42] <jepler> a long time ago it was installed there and the user had to put a symlink from plat/$PLAT/bin
[22:40:51] <alex_joni> I see, so not anymore?
[22:40:57] <alex_joni> how about .deb packages?
[22:41:45] <jepler> I don't have the extra file paul uses to build the deb packages. Maybe he does something in them that we don't do by default.
[22:42:55] <cradek> alex_joni: neither of us use debian, and the SIMPLEINSTALL update is trivial (see the web page)
[22:43:11] <alex_joni> yeah.. not talking about that right now
[22:43:11] <alex_joni> :D
[22:43:21] <alex_joni> me is touching the emc.run script for emc2
[22:43:45] <alex_joni> and I was wondering if I should test for AXIS (or any other display) in the standard user path, or only in the bindir in emc2
[22:44:36] <jepler> what recovery do you perform when DISPLAY doesn't actually exist?
[22:44:46] <alex_joni> recovery?
[22:44:49] <alex_joni> it barfs :D
[22:45:09] <jepler> if I were writing the script, I'd just prepend the platform binary directory to the existing PATH and blindly try to run DISPLAY.
[22:45:28] <jepler> PATH=$EMC2_BIN_DIR:$PATH; export PATH; $DISPLAY -ini ... &
[22:45:43] <jepler> I suppose it's probably not called $DISPLAY, but you get the idea.
[22:46:24] <alex_joni> yup.. I got the idea
[22:47:22] <jepler> In setup.py I have to force python to install the scripts in a non-default location. However, I don't know if it's possible to later *stop* doing this, because under the scenario I suggested users would continue to get an old version of axis after they tried to upgrade to a version which installs /usr/bin/axis or /usr/local/bin/axis...
[22:47:57] <alex_joni> or /usr/share/bin ;)
[22:48:09] <cradek> /usr/local/libexec
[22:48:31] <cradek> sorry, my bsd is showing
[22:48:37] <jepler> it is true, since it doesn't make sense to run it directly, it shouldn't be in a .../bin directory.
[22:48:43] <alex_joni> yep.. you need an installer that removes old versions
[22:48:46] <cradek> right
[22:48:59] <cradek> having these sub-binaries in the path is stupid.
[22:49:08] <alex_joni> * alex_joni faces that problem on emc2 too
[22:49:09] <jepler> cradek: libexec isn't just a bsdism. linux has started to use it. on fc4:
[22:49:10] <jepler> $ ls /usr/libexec/ | wc -l
[22:49:10] <jepler> 80
[22:49:20] <cradek> huh
[22:49:22] <jepler> (compared to 2787 in /usr/bin)
[22:49:44] <cradek> I guess I have some (mostly gnome stuff) there in my redhat9 also
[22:50:29] <alex_joni> cradek: remember that hack you gave me to ignore CVS?
[22:50:41] <alex_joni> couldn't get it to work from a Makefile :/
[22:50:50] <jepler> OK, so I take back everything I said about wanting to install axis in a regular system binary directory. $EMC2_BIN_DIR/axis is probably the location where it should remain.
[22:50:52] <cradek> it sucked anyway
[22:51:21] <cradek> depending unnecessarily on bash is a bad idea
[22:51:28] <alex_joni> I did another hack (also bad)..
[22:51:38] <alex_joni> I used cp -f foo/*
[22:51:49] <cradek> a programmer friend of mine used to say "when in doubt, do it right"
[22:51:51] <alex_joni> which only copies files, but fails on directories
[22:52:02] <alex_joni> so I did: -cp -f foo/*
[22:52:21] <alex_joni> to let make continue, but it's not pretty :(
[22:52:33] <cradek> tell me again why you don't have a list of the binaries you want to install?
[22:52:45] <alex_joni> it's not just binaries
[22:52:53] <alex_joni> also config files, documentation, etc
[22:53:28] <alex_joni> and most of those change very often (at one moment someone will forget to add them to the list of installable)
[22:54:56] <jepler> I don't know what cradek's trick is, but here's a way to list files-and-directories that match a glob (default "*") but aren't named CVS:
[22:54:56] <cradek> I still think you should be using install(1), not cp
[22:54:59] <jepler> globnocvs () { find . -maxdepth 1 -not -name CVS -name "${1-*}" }
[22:55:31] <jepler> fixing the assumption of 'in the current directory' is left as an exercies
[22:55:50] <alex_joni> jepler: you're no fun :D
[22:55:58] <alex_joni> j/k
[22:56:11] <cradek> hey look, it's long after 4
[22:56:18] <jepler> yays
[22:56:26] <cradek> g'bye alex
[22:56:31] <alex_joni> bye chris
[22:56:40] <alex_joni> jepler: would that work to get included in a Makefile?
[22:58:05] <cradek> in a Makefile, you can use $(wildcard) and $(filter-out)
[22:58:23] <jepler> cradek: assuming GNU make
[22:58:35] <cradek> I'm sure that's already assumed
[22:59:28] <cradek> then maybe $(foreach i,$FILES,install ... $i) or whatever the format is
[22:59:55] <cradek> anyway, gotta go
[23:00:09] <alex_joni> ok.. thanks again