#emc-devel | Logs for 2006-10-03

[01:31:35] <jmkasunich> jepler: misleading text on compile farm fixed (and RCSLIB/EMC1 removed)
[01:31:50] <jepler> jmkasunich: thanks
[01:32:08] <jmkasunich> emc1 is dead, long live emc2 ;-)
[01:32:27] <jepler> may it rest in peace
[01:34:15] <skunkworks> :)
[01:37:16] <cradek> ray is using cvs head on the mazak; he may be the first to try it on a machine
[01:37:52] <cradek> I mean since the sim stuff
[01:39:00] <jepler> hum, valgrind gets a lot of errors running rtapi_app
[01:39:11] <jmkasunich> valgrind?
[01:39:49] <jepler> "Valgrind is an award-winning suite of tools for debugging and profiling Linux programs. With the tools that come with Valgrind, you can automatically detect many memory management and threading bugs, avoiding hours of frustrating bug-hunting, making your programs more stable. You can also perform detailed profiling, to speed up and reduce memory use of your programs." http://valgrind.org/
[01:39:52] <cradek> yay, we can use real development tools!
[01:40:06] <cradek> this kicks ass, jepler
[01:40:14] <jepler> it says things like "you used this memory after freeing it", "you used memory before/after this allocation", "you used this value before initializing it" at runtime
[01:40:39] <jmkasunich> is it referring to hal shared memory perhaps?
[01:40:46] <jmkasunich> it probably has no clue about that
[01:40:48] <jepler> hm, that's possible
[01:42:34] <jepler> so far these messages don't give me an address, except one that says it's inside a malloc'd block
[01:43:06] <jmkasunich> ls
[01:43:55] <jepler> . .. a.out core
[02:34:32] <skunkworks> Thanks for the education :)
[02:36:45] <jmkasunich> no prob
[02:37:59] <jmkasunich> I over simplified things. I completely ignored switching loss, which makes things worse, but also ignored duty cycle, which makes them better - hopefully they cancel out and those calcs get you close - then you run it while monitoring temperature
[02:39:49] <skunkworks> sounds like a plan.
[02:52:17] <skunkworks> bed time - night. Thanks again.
[02:54:19] <jmkasunich> you're welcome
[11:56:13] <alex_joni> morning jeff
[11:57:19] <jepler> hi alex
[11:57:30] <alex_joni> so early up?
[11:57:36] <jepler> it's not that early
[11:57:41] <alex_joni> 7am ?
[11:57:45] <jepler> yeah
[11:57:56] <alex_joni> like I said.. so "early" up? ;)
[11:58:37] <alex_joni> I was still in bed at 8:30 this morning :D
[12:00:52] <jepler> that must be because you're lazy.
[12:01:06] <alex_joni> :P
[12:01:22] <alex_joni> not really ;) I just shifted my hours
[15:20:58] <skunkworks> jmkasunich: Just to let you know - there is a temp-current graph on the datasheet that shows if your running 44 amps - the case temp need to be 25c :)
[15:42:57] <alex_joni> you can probably do that with nitrogen :)
[15:50:47] <skunkworks> I was wondering if cooled oil would work :)
[15:53:30] <skunkworks> alex_joni: the thing I keep forgetting is that it will not be at the higer amps for long most likely - only during shuttle accelleration most likely
[15:53:41] <skunkworks> most likely?
[15:53:49] <skunkworks> :)
[15:55:36] <cradek> it's easy to have full accel for a longer time, imagine cutting smallish circles
[15:55:48] <cradek> like helical hole milling for instance
[15:57:11] <cradek> I guess that's constant magnitude accel, but it would still be spread out among all your transistors
[15:58:55] <skunkworks> right - but that would not be at 200 ipm most likely
[15:59:52] <jepler> you could tie some temperature sensors on the FETs to the external feed override, to slow down before you destroy anything
[16:00:01] <jepler> </wacky idea 33>
[16:00:54] <skunkworks> I like wacky ideas - that is pretty cool
[16:01:20] <cradek> except if it's bogging down on actual cuts, slowing might make worse
[16:01:44] <cradek> maybe you would just want estop on an overheat
[16:05:32] <skunkworks> right right
[16:29:45] <cradek> jepler: should 64 bit stuff be done on a branch? I think we might be getting close to starting a 2.1 branch/feature freeze
[16:32:21] <alex_joni> cradek: I think so
[16:32:31] <cradek> hi alex
[16:32:35] <alex_joni> hi chris
[16:32:40] <alex_joni> just finished dinner :)
[16:33:28] <cradek> about to go to lunch here
[16:34:05] <alex_joni> nice .. what'cha having?
[16:34:14] <cradek> grease sandwich and pie
[16:34:29] <alex_joni> grease sandwich ?
[16:34:38] <cradek> well that's not what it's actually called
[16:34:52] <cradek> and I probably won't have it, but jepler will
[16:35:20] <cradek> it's a veggie burger patty with cheese and grilled onions, with buttered rye bread that's then fried on the outside
[16:35:22] <alex_joni> what kind of grease does it have on it?
[16:35:34] <cradek> it's very good, but too greasy
[16:35:37] <alex_joni> oh.. sounds "interesting: :)
[16:35:51] <cradek> I think they use each of the kinds of grease for this sandwich
[16:36:30] <cradek> I'll order breakfast, egg and potatoes, coffee, maybe pie
[16:36:49] <cradek> this place has easily the best pie in the city
[16:36:58] <alex_joni> I don't recall when I ate pie the last time :)
[16:37:01] <cradek> the owner is an older lady and she makes all the pies
[16:38:40] <jepler> I'll try not to do anything invasive on the trunk
[16:39:53] <jepler> I would say something about "obviously right", but that's a funny phrase
[16:40:28] <alex_joni> jepler: ?
[16:40:54] <jepler> alex_joni: at work some changes I recently made that were "obviously right" actually caused a problem that consumed a lot of developer hours
[16:41:14] <alex_joni> ouch ..
[16:41:57] <cradek> yep, 3 days to recover - maybe a new record?
[16:43:48] <alex_joni> one question: shouldn't axis switch the slider for jog speed to mm when switching from inch to mm ? (either when G20->G21, or by chosing machine->Display MM) ?
[16:44:19] <jepler> alex_joni: yes, probably
[16:44:23] <cradek> I think definitely not G20/G21, but maybe with machine->display
[16:44:29] <jepler> I mean, the code is not there, but it would be best if it did
[16:46:41] <cradek> another interesting problem is dealing with rotary and linear units
[16:46:56] <cradek> if you run stepper-xyza you will see that it works poorly
[16:47:10] <cradek> talking about the jog slider still
[16:47:21] <cradek> seems like there needs to be two of them
[16:47:37] <cradek> machine->inches/mm would just switch the linear one of course
[16:49:45] <cradek> alex_joni: we'd love it if you would fix it :-)
[16:51:43] <alex_joni> wish I knew enough python not to bork it completely :)
[16:52:29] <jepler> I guess you'll have to learn
[16:53:14] <cradek> you'll figure it out, it's just a programming lanaguage
[16:53:23] <cradek> for the few strange things it has nice online docs/tutorial
[16:53:33] <cradek> but mostly it's pretty easy to see how it works
[16:54:11] <SWPadnos> I often find that leanring a language is easy, but finding out what you don't have to write (ie, what's available in libs) is time consuming
[16:54:31] <SWPadnos> I often find that typing correctly is harder than thinking
[16:57:09] <cradek> in a new language I tend to write each of the wrong things that all the new users write, because they seem perfectly fine
[16:57:16] <cradek> so maybe I should learn to read the faq first
[16:58:01] <cradek> a python example --
[16:58:04] <cradek> for line in ...:
[16:58:11] <cradek> if line.fine("string"):
[16:58:15] <cradek> something()
[16:58:26] <jepler> .find
[16:58:28] <cradek> err I mean line.find
[16:59:38] <cradek> I don't remember how .find works, but I remember it's not like this, and I better look it up
[17:03:29] <jepler> >>> "abc".find("x")
[17:03:29] <jepler> -1
[17:03:29] <jepler> >>> "abc".find("a")
[17:03:29] <jepler> 0
[17:04:30] <alex_joni> cia has something against me
[17:04:45] <alex_joni> jepler: do you also have lots of commits not noted ?
[17:06:08] <alex_joni> hmm.. this time it was only slow :)
[17:19:01] <alex_joni> jepler: what happened with hallelujah ?
[18:23:33] <jepler> alex_joni: I got distracted by sim-rtapi
[18:24:02] <alex_joni> jepler: ok.. was cleaning up a bit, and wondered about emc2/src/hal/utils/halgui
[18:24:11] <jepler> leave it alone for now...
[18:24:17] <alex_joni> sounds like the place to put hallelujah :)
[18:24:32] <alex_joni> I'm still tempted to remove that.. it's not used anyways
[18:24:43] <alex_joni> but right now I'm working on something else
[18:46:17] <alex_joni> maybe jmkasunich will turn up today
[18:46:31] <alex_joni> jmkasunich: bet you know a thing or two about alter & comp
[20:31:26] <alex_joni> jmkasunich: I'll read the logs tomorrow morning.. if there's something about alter & comp I can do.. let me know
[23:09:59] <jmkasunich> * jmkasunich is here
[23:38:21] <jmkasunich_> nick jmkasunich
[23:38:32] <jmkasunich_> jmkasunich_ is now known as jmkasunich