#emc-devel | Logs for 2007-12-27

[01:27:38] <rayh> I need to tar up a run in place and move to another computer.
[01:27:52] <rayh> When I do that I loose the setuid root stuff.
[01:28:00] <rayh> What do I have to do to get it back.
[01:28:06] <rayh> I don't have make on that box.
[01:31:59] <rayh> I got it. Just have to learn to read make files.
[01:33:53] <fenn> you can use tar -p to preserve permissions
[01:34:07] <fenn> might need to be root to do that
[01:35:59] <rayh> Ah I didn't know that possibility.
[01:36:06] <rayh> Thanks fenn
[01:36:18] <rayh> It's running.
[02:00:05] <rayh> I'm having trouble running the latency test on a run in place system.
[02:00:33] <rayh> I moved timedelta to where the script was expecting to find it but it complains about pyvcp.
[02:00:42] <rayh> any hints?
[02:03:21] <fenn> i have PYTHONPATH=/usr/local/emc2/lib/python
[02:03:48] <fenn> without that it doesnt know where to find pyvcp_widgets.py
[02:04:09] <fenn> should be part of emc-environment though
[02:04:19] <fenn> what's the error?
[02:05:47] <rayh> pyvcp exited without becoming ready.
[02:06:14] <rayh> There is a hint before that something about the xml file.
[02:13:35] <rayh> The whole message is;
[02:13:36] <rayh> cnc:~/emc2-new/emc2-trunk/rtlib$ latency-test
[02:13:38] <rayh> Usage: pyvcp -c hal_component_name myfile.xml
[02:13:39] <rayh> HAL:6: ERROR: pyvcp exited without becoming ready
[03:04:51] <jepler> rayh: you should always ". scripts/emc-environment" before running anything else in an RIP system.
[03:05:16] <rayh> ah okay.
[03:05:39] <rayh> is that . (space) scripts...
[03:05:48] <jepler> dot space scripts slash emc dash environment
[03:05:59] <rayh> okay
[03:06:21] <rayh> I matched the directory structure between source and target machines but that didn't do it.
[03:07:34] <cradek> hmm, mounting the switches and cams was easy compared to the wiring
[03:08:02] <jmkasunich> cradek: limit switches for max5?
[03:08:06] <cradek> yes
[03:08:23] <jepler> rayh: I am confused by what you said about having to move the timedelta module to a different location -- it should get put in the same location as all the other hal components by 'make'
[03:09:28] <rayh> Yea. I think the environment was looking for it someplace other than the rip.
[03:09:46] <rayh> latency test does seem to work now. Thanks a bunch.
[03:10:13] <jepler> ok always remember to use ". scripts/emc-environment" before running from RIP (in fact I modified the configure procedure in TRUNK to remind people, because often it is overlooked)
[03:12:26] <rayh> Okay.
[03:14:02] <rayh> Ah I see it there in the notes after configure. Thanks again Jeff.
[03:14:08] <cradek> jmkasunich: http://timeguy.com/cradek-files/emc/limitswitches.jpg
[03:15:08] <jmkasunich> pretty
[03:15:25] <jmkasunich> the shallow angle on the cam will hurt repeatability a bit tho
[03:16:29] <cradek> yeah, it's a compromise. the switches are not really meant for that
[03:16:53] <cradek> I got all limits/homes with those two switches, it's a nice design I think
[03:17:00] <jmkasunich> does the machine have any hard stops?
[03:17:14] <jmkasunich> five axes with two switches? what am I missing?
[03:17:18] <cradek> sort of
[03:17:21] <cradek> sorry, only XY
[03:17:28] <cradek> the others I'll do later, or not at all
[03:17:35] <jmkasunich> oh, duh, limits and homes
[03:17:49] <jmkasunich> I'm not gonna bother with limits
[03:18:15] <cradek> yeah I wouldn't have, except I saw a simple way to get them all without too much extra work
[03:22:37] <jmkasunich> on quill and table I have hard stops. so I can place the switches right next to them
[03:22:54] <jmkasunich> the table can press the switch, but the hard stop prevents it from smashing the switch if it goes to far
[03:23:10] <jmkasunich> that means I can come straight at the switch, no cam
[03:23:23] <cradek> yeah that's nice, the one switch on my lathe is that way
[03:23:31] <cradek> I created a hard stop for that
[10:23:43] <fenn> bc is required for latency test? (but not emc)
[13:01:03] <jepler> fenn: hm is bc not part of the base system? I can change it to something else that is already required (like python) if bc isn't..
[13:01:20] <jepler> Package: bc
[13:01:21] <jepler> Priority: standard
[13:04:55] <jepler> so I'd say just add bc to the depends so it's sure to get installed .. it won't represent an additional package to install for very many people
[15:42:27] <CIA-20> EMC: 03jepler 07TRUNK * 10emc2/src/hal/components/ (Submakefile mkconv.sh): mkconv.sh has bashisms in it (apparently) so invoke with bash instead of sh
[15:49:08] <jmkasunich> question: if "make install" leaves a mess and there is no "make uninstall", why don't we remove "make install"?
[15:49:17] <jmkasunich> tell people to either use packages or use RIP
[15:49:29] <cradek> make install is a necessary part of creating packages
[15:49:50] <cradek> (part of the process is that you install into a false root directory)
[15:49:55] <jmkasunich> oh
[15:50:08] <cradek> and I don't think it leaves a mess - it probably works fine
[15:50:15] <jmkasunich> ISTR that people wanting to un-install aren't the only ones who've had problems
[15:50:27] <jmkasunich> even those who just do make install and expect it to work are sometimes dissapointed
[15:50:38] <jmkasunich> (and I sure don't know how to help them)
[15:51:23] <cradek> people often crap up their nice package-managed system with various source packages' make install
[15:51:35] <cradek> it's not just an emc problem
[15:51:53] <cradek> we have run-in-place which is much more than most programs have
[15:51:55] <jmkasunich> I've probably crapped up my system that way
[15:52:07] <jmkasunich> I agree its not just our problem
[15:52:10] <cradek> yeah and it's always hard to recover from
[15:52:31] <jmkasunich> it may be a little more severe for us tho, because we attract former doze users who are only running linux because of emc
[15:52:49] <cradek> no, those people don't build emc
[15:52:57] <jmkasunich> you'd think that
[15:53:13] <jepler> this fellow is a die-hard believer in an OS distro without a packaging system .. I don't think he's a new refugee from the windows world
[15:53:24] <jmkasunich> true in this case
[15:58:59] <jmkasunich> so cradek, did you get your switches wired up?
[15:59:06] <cradek> nope
[15:59:17] <jmkasunich> at work today?
[15:59:24] <cradek> I found a proper wire/plug though
[15:59:25] <cradek> yes
[16:02:20] <jmkasunich> well, I'm now the spindle baron - I bought the rest of the surplus HSK spindles
[16:02:28] <cradek> ha
[16:02:56] <cradek> party favors?
[16:03:37] <jmkasunich> I have visions of making spindle assemblies that take Tormach tooling, and then selling them at the CNC workshop or to similar people online/ebay
[16:04:17] <cradek> fun
[16:04:26] <jmkasunich> maybe - time will tell
[16:13:32] <jmkasunich> at $120 total for 14 spindles, I'm not exactly in over my head
[16:13:50] <jmkasunich> I've wasted that much money at least once or twice before
[16:13:58] <cradek> haha, yeah at least.
[16:20:27] <jmkasunich> I suppose I should turn down the sarcasm
[16:24:09] <jmkasunich> I think he's kvetching
[16:24:33] <cradek> when I answer a question (correctly), he says that's not the answer he wanted. I guess that means the question was wrong...?
[16:24:44] <jmkasunich> probably
[16:24:49] <cradek> so I'll back off and see if he asks for something more specific
[16:24:57] <cradek> I don't know what he's even trying to do
[16:39:14] <alex_joni> hi guys
[17:03:58] <cradek> hi
[17:04:49] <alex_joni> what's up?
[17:05:10] <alex_joni> * alex_joni is happy .. no more holding the cell in my hand