#emc-devel | Logs for 2007-11-09

[00:08:38] <fenn_> fenn_ is now known as fenn
[02:36:24] <cradek> http://timeguy.com/cradek-files/emc/coin.png
[02:36:41] <cradek> this is a penny. I need to make a new probe tip - this one's not sharp enough for any detail but the edge
[03:52:19] <jmkasunich__> jmkasunich__ is now known as jmkasunich
[15:17:54] <skunkworks_> logger_dev: bookmark
[15:17:54] <skunkworks_> Just this once .. here's the log: http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2007-11-09.txt
[15:23:20] <jepler> cradek: does loading (previewing) a gcode file with (PROBEOPEN) nuke the file?
[15:24:02] <jepler> I guess it doesn't, because (PROBEOPEN) is handled post-canon
[15:24:15] <jepler> while (debug) and (message) or whatever they're called are handled pre-canon
[15:24:31] <SWPadnos> jepler, did you see my conversation last night with Wowbagger?
[15:24:51] <SWPadnos> about AXIS hanging at NC file load
[15:25:29] <jepler> SWPadnos: yes but I didn't read it very closely
[15:25:50] <cradek> what's in the file?
[15:26:14] <SWPadnos> ok. I don't think there's a problem with AXIS, I'm more suspicious of the guy's 233MHz laptop with neomagic display drivers and only 160M RAM
[15:26:15] <SWPadnos> any file
[15:26:22] <SWPadnos> I had him reload the axis splash file
[15:26:36] <cradek> maybe his opengl is bad
[15:26:54] <SWPadnos> it loads fine at startup, but if you load a file after that it gives you an endless hourglass
[15:26:53] <cradek> (display drivers?)
[15:27:07] <cradek> huh that sounds strange
[15:27:11] <SWPadnos> yep
[15:27:36] <SWPadnos> and he was ssh'ed into it, using the GL on an Intel 915 setup last night
[15:28:37] <SWPadnos> I couldn't figure out why it would work at startup and not afterwards
[15:31:09] <jepler> if it can take the whole X server with it (can't switch to text console), it's an X server bug
[15:31:17] <SWPadnos> nope
[15:31:36] <SWPadnos> though this was remote, so who knows what might have been in there
[15:32:04] <SWPadnos> but it would load the emc2/AXIS splash code fine
[15:32:04] <jepler> oh for some reason I thought it broke the whole system
[15:32:13] <SWPadnos> pressing the reload button would hang it
[15:32:16] <SWPadnos> nope, just axis
[15:32:34] <SWPadnos> I had to ask him to clarify what "hang" meant though ;)
[15:34:02] <jepler> you might get a traceback from axis if you kill -INT it
[16:11:00] <jepler> cradek: what's it take to run the sim probe?
[16:11:28] <jepler> I take it there's some configuration stuff I don't have
[16:14:30] <cradek> http://timeguy.com/cradek-files/emc/probe.tar.gz
[16:14:33] <cradek> http://timeguy.com/cradek-files/emc/hemisphere.comp
[16:15:21] <skunkworks_> so - your digitizing of the sphere was virtual? coolk
[16:15:21] <skunkworks_> cool
[16:17:19] <fenn> soon we'll have the secret service in here investigating cradek's counterfeiting racket
[16:17:50] <skunkworks_> :)
[16:18:06] <skunkworks_> he did try to scan a coin..
[16:40:40] <jepler> tail probe-results.txt
[16:40:41] <jepler> G1 X0.120000 Y-0.960000 Z0.253132
[16:40:41] <jepler> G1 X0.080000 Y-0.960000 Z0.268646
[16:40:41] <jepler> G1 X0.040000 Y-0.960000 Z0.277190
[16:42:49] <cradek> neat!
[16:43:52] <jepler> now I am debating what to do about opening the log for append vs clearing it if it has contents (fopen "w" vs "a")
[16:45:06] <jepler> probeopen clears the file's existing contents unconditionally
[16:45:32] <cradek> I don't see any problem with that
[16:54:58] <cradek> that's great jeff, thanks
[16:55:03] <cradek> * cradek kicks CIA-18
[17:17:37] <SWPadnos> I guess we have two partially-working CIAs now
[20:03:47] <Guest718> Guest718 is now known as skunkworks___
[20:31:12] <skunkworks___> skunkworks___ is now known as skunkworks_
[20:43:45] <SWPadnos> hmmm. at the moment, are tool offsets part of the interpreter or canon/traj?
[20:44:23] <SWPadnos> or, does canon operate in machine coordinates, and offsets / coordinate systems / other neat things are left as an exercise for the interpreter?
[20:44:44] <SWPadnos> (or being another question phrasing, not an either/or option for canon)
[20:45:52] <eugenio>
[20:46:14] <skunkworks_>
[20:46:33] <SWPadnos> be quiet, will you?
[20:46:55] <skunkworks_> sound of silence..
[20:47:12] <SWPadnos> "Hello darkness, my old friend"
[20:48:09] <skunkworks_> hmm - I actually think I have that cd in the car..
[20:48:23] <SWPadnos> mine is in the CD rack about 6 feet to my left
[20:48:27] <SWPadnos> and on the PC :)
[20:52:55] <skunkworks_> I think I only have 3 cd's right now in the car. Simon & Garfunkel, Feist, and tom petty ;) wife took the cd case..
[20:53:12] <SWPadnos> heh
[20:59:19] <SWPadnos> hmmm. there seems to be a dichotomy in HAL drivers and pin naming/use
[21:00:34] <SWPadnos> the fact that you (the user) need to explicitly tell HAL which I/O bit to use, for example, seems to conflict with the concept of drivers that auto-detect waht hardware is available and export different HAL pins depending on what's found
[21:01:15] <SWPadnos> the former case conceptually is "the integrator should define exactly where the I/Os connect"
[21:01:35] <SWPadnos> whereas the latter case is "the integrator doesn't have to know what hardware is connected"
[21:03:28] <SWPadnos> this is reasonably easy to reconcile in the case of the parport, for example. just specify the order the ports should be scanned and their I/O modes, and you should get the same HAL pin <-> physical hardware pin connections every time
[21:03:51] <SWPadnos> (unless you have PCI parports and the PCI subsystem decides to change the scan order or something)
[21:05:24] <SWPadnos> in the case of more complex hardware, such as PPMC, Mesa, Motenc or STG, it's not so easy
[21:39:44] <jepler> maybe the "improved" latency test should be removed .. it sounds like a lot of people run emc 2.1 with timings that trigger this error, but never noticed a problem..
[21:50:26] <SWPadnos> yeah, it's hard to tell whether it's better to warn people, or if they'r ehappier not knowing
[23:13:05] <SWPadnos> hmmm. Do we really support RTLinux still?
[23:13:23] <cradek> it may still work
[23:13:27] <SWPadnos> I see the rtl_??api.c files, but they're considerably older than the RTAI ones
[23:13:33] <SWPadnos> ok, that was my conclusion
[23:13:34] <cradek> I have not heard this word "support" before - what does it mean?
[23:13:45] <SWPadnos> err - attempt to make it work with ...
[23:13:53] <cradek> :-)
[23:13:57] <cradek> afaik nobody is using it.
[23:14:13] <SWPadnos> I'm writing the abstract for the potential ESC talk, so I'm actually looking at it ATM :)
[23:14:14] <cradek> I think I was the last hold-out
[23:14:17] <SWPadnos> heh
[23:14:42] <SWPadnos> so it's safe to say that although it started out somewhat cross-platform, it's Linux/RTAI only at the moment (sim excluded)
[23:14:50] <cradek> if you really want, it might be possible to talk me into booting that old disk. I probably still have it. I maybe can find it.
[23:15:02] <SWPadnos> oh no, I don't really care about RTLinux these days
[23:15:05] <cradek> I'm pretty sure I kept it for a future day such as that
[23:15:10] <cradek> ok
[23:15:31] <SWPadnos> in fact, I haven't cared much since they started being a money-grubbing company
[23:18:55] <SWPadnos> cradek, would you mind looking at the abstract (and commenting on it)?
[23:19:49] <cradek> sure
[23:20:03] <cradek> can you send it? I'm about to run off for a bit
[23:20:35] <cradek> bbl
[23:20:39] <SWPadnos> oh - I'm in a time crunch as well. I'll pastebin and see later if I've made any glaring errors :)
[23:20:41] <SWPadnos> see you
[23:21:26] <alex_joni> SWPadnos: I can look it over too, not sure I'll be much help though :)
[23:21:30] <SWPadnos> heh
[23:21:37] <SWPadnos> ok. if only DNS were working again ...
[23:22:03] <alex_joni> stick other servers in there
[23:22:26] <SWPadnos> yeah, no kidding
[23:49:45] <alex_joni> good night all