#emc-devel | Logs for 2009-03-23

[02:23:33] <jepler> alex_joni: sorry to step on your toes by building all the packages today. if you need a project that will doubtless require days of your time, you should work on a 2.3.x live cd with emc2-dev and emc2 build-dep .. start by removing ooffice, I imagine.
[08:56:05] <alex_joni> * alex_joni points micges at the topic
[09:03:21] <micges> alex_joni: I'll update translation this evening
[09:23:06] <alex_joni> cool, thanks
[09:23:26] <alex_joni> although I was more interested in getting some feedback from you after running beta2 ;)
[09:25:55] <micges_emc> 2xTRUNK is working now here with no problems
[09:26:55] <micges_emc> when there are package I'll install them on them
[09:28:03] <micges_emc> alex_joni: about joints/axes problem:
[09:28:20] <micges_emc> I have project with 18 individual joints
[09:29:14] <micges_emc> I have no success with modifying emc to allow this
[09:30:24] <micges_emc> I've traced some strange hang up in task so far
[09:30:58] <micges_emc> when you have some time I'll be very appreciated with your help
[10:15:15] <alex_joni> micges_emc: I don't think it can be done with current TRUNK
[10:15:34] <alex_joni> it could be done once the joints_axes branch is done/merged
[10:28:34] <micges_emc> oh
[10:33:46] <micges_emc> not good :|
[13:00:14] <jepler> cvs is unavailable until I resolve a problem with my isp
[13:00:27] <alex_joni> good to know ;)
[13:00:31] <SWPadnos> bummer
[13:00:33] <jepler> they upgraded my connection but failed to keep me on my static ip
[13:00:37] <SWPadnos> not that I'm doing anything right now
[13:00:47] <jepler> their help line just opened so I'm about to call them .. more news as I have it
[13:00:48] <alex_joni> I'm sure it's something that'll be fixed soon
[13:01:01] <SWPadnos> oh hmmm. I've had that problem too (minus the upgraded connection)
[13:01:09] <alex_joni> jepler: if the new ip is static, then we can just point cvs.linuxcnc.org there
[13:01:19] <SWPadnos> if you have a new IP, let me know. I can change the what he said
[13:01:40] <alex_joni> although it doesn't sound like a static IP :)
[13:01:50] <alex_joni> jepler@dsl-dynamic-206-222-195-89.inebraska.com
[13:01:58] <SWPadnos> indeed
[13:14:03] <jepler_> OK, I think cvs service should be restored
[13:17:08] <SWPadnos> yep. that was quick
[13:18:12] <jepler_> I have to admit I expected much worse
[13:20:40] <SWPadnos> yeah. my ISP changed my fixed IP once, and when I asked them about it (months later, I'll admit), they didn't understand the question
[13:24:37] <SWPadnos> I wonder if it is worse :)
[13:27:54] <cradek> his ISP is fairly small and local
[13:28:11] <cradek> a big advantage at times like this
[13:28:26] <SWPadnos> mine is (was) too
[13:28:37] <SWPadnos> I have a 4-digit customer number :)
[13:29:33] <skunkworks_> I am just glad I don't have satellite anymore
[13:29:55] <alex_joni> jepler_: wfm
[13:34:24] <jepler_> problem is my upload speed seems to be rather low
[13:34:30] <jepler_> oh well, I can put out that fire later
[15:19:50] <skunkworks_> how does the montech cards dak work? is a scale of 1 give you 0-1 = 0-10v?
[15:28:30] <SWPadnos> montech?
[15:33:19] <skunkworks_> motenc heh
[15:33:36] <SWPadnos> heh
[15:37:10] <skunkworks_> http://www.cnczone.com/forums/showthread.php?p=586970#post586970
[15:38:52] <SWPadnos> looks like gain=1.0 is supposed to be volts
[15:40:56] <SWPadnos> the correct calculation is likely gain=10/4000 (or whatever)
[15:41:06] <skunkworks_> so my first option is probably correct
[15:41:10] <SWPadnos> I think so
[15:41:23] <skunkworks_> Thank you
[15:41:45] <SWPadnos> hmmm. I note that there's no max output parameter
[15:41:57] <SWPadnos> or min. I wonder if that should be considered a bug
[16:16:16] <skunkworks_> http://www.cnczone.com/forums/showthread.php?p=586989#post586989
[16:16:19] <skunkworks_> cool :)
[17:09:22] <skunkworks_> iocontrol.0.user-enable-out could be used instead of the xenable,yenable,zenable (comes on at the same time?)
[17:09:55] <skunkworks_> (axis.0.amp-enable-out)
[17:29:08] <skunkworks_> that wouldn't work
[18:14:21] <alex_joni> skunkworks_: since Xenable, Yenable and Zenable _always_ come on at the same time, he can simply just use one of them
[18:14:39] <alex_joni> currently emc2 only know how to turn them on/off simultaneously
[18:26:28] <BJT-Work> wow you can get endmills that are 0.001" in diameter
[18:26:53] <BJT-Work> LOC 0.003
[18:27:32] <jepler> I recently used a .010" endmill, .020" LOC. I can't imagine using one 1/10 as big
[18:28:09] <BJT-Work> if you dropped it it would break I think
[18:28:29] <jepler> quite possibly
[18:28:34] <BJT-Work> I've drilled some 0.015" diameter holes and that was not fun
[18:28:49] <jepler> this was trace isolation milling
[18:29:10] <BJT-Work> www.harveytoo.com
[18:29:17] <BJT-Work> www.harveytool.com
[18:29:30] <jepler> it gave a more uniform cut than vee bits, since my table isn't quite parallel to Z
[18:29:46] <BJT-Work> I just got a catalog in the mail
[18:30:21] <BJT-Work> heh you can get diamond coated down to 0.010"
[18:31:08] <jepler> this is the site I bought from: http://precisebits.com/products/carbidebits/precisebit-stub.asp
[18:31:29] <jepler> Material - solid submicrograin carbide
[18:53:52] <SWPadnos> does that mean nanograin?
[18:58:36] <BJT-Work> I'm sure it is at least femtograin
[19:41:54] <skunkworks_> alex_joni: thanks - passed it along
[21:24:18] <alex_joni> skunkworks_: guru my ass
[21:24:38] <SWPadnos> ewww
[21:24:50] <alex_joni> SWPadnos: not like that :P
[21:24:53] <SWPadnos> heh
[21:25:17] <skunkworks_> heh
[21:28:27] <alex_joni> jepler: no sweat about the package building, it's something I can gladly help with (but if you want to do it yourself, that's fine too), regarding the 2.3 LiveCD I'll probably get around to do that some day, but it's harder to start :)
[21:37:38] <micges> while translating stepconf: "Additional signal conditioning or isolation such as optocouplers and RC filters\n"
[21:37:52] <micges> what RC filters mean ?
[21:38:43] <skunkworks_> http://en.wikipedia.org/wiki/RC_circuit
[21:38:48] <skunkworks_> does that help?
[21:40:21] <micges> I see thanks
[21:44:02] <jepler> more cvs downtime may be coming .. I have to make some modifications to the firewall rules on a machine between cvs and the internet
[21:44:04] <jepler> thank you for your patience
[21:45:51] <alex_joni> are you home?
[21:50:50] <SWPadnos> hmmm
[21:51:00] <SWPadnos> I have emc almost running on Jaunty
[21:51:12] <SWPadnos> RTAI 3.7, kernel 2.6.28-8
[21:51:25] <SWPadnos> the problem I had before was the memlock limit
[21:52:06] <SWPadnos> now motmod won't load, due to some missing math symbols (cos, atan2, acos)
[21:52:44] <SWPadnos> I did notice some compiler messages go by with missing symbols, but assumed those were normal
[21:53:06] <SWPadnos> is there some magic that has to be done to get math functions on x86_64?
[21:53:32] <cradek> maybe it's an rtai build option? I thought those came from rtai.
[21:54:33] <SWPadnos> hmmm. I thought I had selected math, btu I guess I hadn't
[21:54:35] <SWPadnos> duh
[21:55:41] <SWPadnos> do you know offhand if C99 support is needed (or the opposite?)
[21:55:54] <cradek> I don't think it matters - it will compile either way
[21:56:07] <jepler> yes they will often be shown during the build as unknown symbols
[21:56:07] <jepler> since kbuild doesn't always pick up symbols offered by rtai modules
[21:56:07] <jepler> unless you do a special sexy dance
[21:56:07] <jepler> s/sexy//
[21:56:08] <jepler> is there an rtai_math.ko? is it loaded?
[21:56:40] <SWPadnos> I'm sure there isn't, RTAI_MATH was unset
[21:56:55] <SWPadnos> any clue on RTAI_MATH_C99?
[21:59:55] <jepler> c99 added additional math functions such as fpclassify
[22:00:00] <SWPadnos> I wonder what C99 would have to do with it (I'm building without that now)
[22:00:00] <jepler> I don't think we need it, but it can't hurt
[22:00:01] <SWPadnos> ah, ok
[22:00:17] <SWPadnos> I wasn't sure if there might have been some name mangling changes or something
[22:01:00] <SWPadnos> hmmm. do you suppose sincos is part of C99?
[22:01:11] <jepler> CONFORMING TO
[22:01:11] <jepler> This function is a GNU extension.
[22:01:29] <alex_joni> SWPadnos: for some odd reason rtai_math is disabled by default :)
[22:01:43] <SWPadnos> there are a few places where I get the "sincos" undefined message
[22:01:55] <alex_joni> while compiling or running?
[22:02:02] <alex_joni> compiling is usually fine
[22:02:07] <SWPadnos> (siggen, rotatekins, maxkins, biquad, 5axiskins)
[22:02:11] <SWPadnos> compiling
[22:04:00] <SWPadnos> hmmm. still doesn't load. I guess I'll have to think now
[22:04:53] <alex_joni> pastebin?
[22:04:58] <SWPadnos> one sec
[22:05:05] <SWPadnos> what would you like to see?
[22:05:12] <alex_joni> the error :)
[22:05:16] <SWPadnos> :P
[22:06:00] <SWPLinux> http://pastebin.ca/1369638
[22:06:10] <SWPLinux> well, it's poorly titled, but that's what I see in dmesg
[22:06:42] <alex_joni> did you rerun emc2's configure?
[22:06:48] <SWPLinux> hmmm. did I
[22:06:52] <alex_joni> it needs to run again, so it can pick rtai_math
[22:06:58] <alex_joni> check rtapi.conf to see if you did
[22:06:59] <SWPLinux> probably not, though I did have the sense to rebuild :)
[22:07:07] <alex_joni> rebuild won't help here
[22:07:14] <alex_joni> without a rerun of configure
[22:07:52] <SWPLinux> well, we'll know in a minute or so
[22:08:21] <SWPLinux> well, now the sincos messages stop the build :)
[22:08:37] <SWPLinux> maybe I'll try that with C99 now
[22:08:43] <alex_joni> odd :)
[22:09:47] <SWPLinux> maybe in RTAI 3.7 you need the C99 option to get it
[22:10:19] <SWPLinux> stupid parallel make sensitive builds
[22:10:25] <SWPLinux> hmmm. or not
[22:25:39] <SWPLinux> well, it's a little odd
[22:25:55] <SWPLinux> for some reason, LONG_MAX and LONG_MIN aren't defined
[22:26:05] <SWPLinux> even though limits.h is being included
[22:26:21] <SWPLinux> In file included from /download/rtai-3.7-test2/base/math/rndint.c:47:
[22:26:23] <SWPLinux> /usr/include/limits.h:125:26: error: limits.h: No such file or directory
[22:26:43] <SWPLinux> limits.h includes features.h, which also exists
[22:29:38] <alex_joni> that's not ok
[22:29:44] <SWPadnos> I agree
[22:29:53] <alex_joni> the /usr/include/limits.h isn't the file to include
[22:30:14] <alex_joni> rtai should include /usr/src/linux/*
[22:30:25] <alex_joni> at least for kernel stuff
[22:30:27] <SWPLinux> yeah, it seemed to me that a more kernel-y include should be there
[22:30:29] <SWPLinux> yep
[22:30:43] <alex_joni> got a ssh open?
[22:31:05] <SWPLinux> hmmm. probably not through the firewall/router
[22:31:24] <alex_joni> right, probably too complicated to be worth it :)
[22:31:34] <SWPLinux> heh, probably :)
[22:32:01] <SWPLinux> this is all in some stupid rounding code too, which I very likely don't care aout
[22:32:22] <alex_joni> I'm pretty sure I didn't use C99 for hardy
[22:32:35] <SWPLinux> was the option there?
[22:32:44] <alex_joni> I think so
[22:32:45] <SWPLinux> it's marked as new here, but you never knwo
[22:32:47] <SWPLinux> know
[22:33:53] <SWPLinux> well, I saw the math function modules build before the error came up
[22:34:20] <SWPLinux> make
[22:34:23] <SWPLinux> oops
[22:35:27] <SWPLinux> well, getting rid of the c99 option ts it build
[22:35:29] <SWPLinux> lets
[22:36:07] <alex_joni> if emc2 runs, then you're set :)
[22:36:09] <alex_joni> debs next
[22:36:18] <SWPLinux> I agree
[22:36:23] <SWPLinux> :)
[22:36:35] <alex_joni> what I usually do, and it works great:
[22:36:50] <alex_joni> grab the latest rtai package source from linuxcnc.org/hardy
[22:36:59] <alex_joni> and copy the debian/ folder to the new rtai folder
[22:37:13] <alex_joni> then 2-3 lines to adjust (kernel name & such)
[22:37:21] <alex_joni> then debuild to build it, and you're done
[22:37:30] <SWPLinux> I think it's a little more complicated here, since this is Jaunty
[22:37:38] <SWPLinux> the deps aren't all right
[22:37:53] <alex_joni> for rtai?
[22:38:02] <SWPLinux> oh, those might be correct
[22:38:09] <SWPLinux> I was thinking EMC
[22:38:10] <alex_joni> there shouldn't be much of a dep besides the kernel header
[22:38:15] <alex_joni> emc comes later :)
[22:38:19] <SWPLinux> heh
[22:38:20] <alex_joni> kernel & rtai are the hard part
[22:38:29] <alex_joni> emc2 runs in place greatly :P
[22:38:40] <SWPLinux> not yet ;)
[22:38:45] <alex_joni> but it will
[22:38:49] <SWPLinux> I'm sure it will
[22:38:56] <SWPLinux> this PC is interesting for latency
[22:39:13] <SWPLinux> it's really great, until nvidia accelerated 3D kicks in
[22:39:22] <SWPLinux> glxgears or desktop effects or whetever
[22:39:36] <SWPLinux> but as soon as you stop using the 3D hardware, the latencies get excellent again
[22:39:57] <alex_joni> < 10 usec ?
[22:40:07] <SWPLinux> yeah, around 4 or so
[22:40:10] <alex_joni> cool
[22:40:10] <SWPLinux> often less
[22:40:22] <alex_joni> I get 10-15usec on this laptop
[22:40:31] <alex_joni> everything is great, no overruns & all
[22:40:43] <alex_joni> unless I change the screen brightness (with Fn-F4/F5)
[22:40:52] <alex_joni> then it overruns to a couple hundred msec
[22:41:08] <SWPLinux> heh
[22:41:13] <SWPLinux> SMI, anyone?
[22:41:32] <alex_joni> don't think it's SMI..
[22:41:45] <alex_joni> maybe it is.. but I let it run for 1-2h without hickups
[22:41:50] <SWPLinux> http://pastebin.ca/1369684
[22:42:12] <SWPLinux> no real information on the console or in ~/emc_*.txt
[22:42:29] <alex_joni> try a sim-only config
[22:42:33] <SWPLinux> oh, maybe there is
[22:42:41] <alex_joni> sounds (gut feeling) that it tried to load a parport
[22:42:48] <alex_joni> and had an issue with the IO port
[22:42:54] <alex_joni> maybe parport_pc is installed
[22:42:56] <SWPLinux> http://pastebin.ca/1369687
[22:43:01] <SWPLinux> could be
[22:43:30] <SWPLinux> yeah, halrun works
[22:43:46] <alex_joni> what config did you run?
[22:43:56] <alex_joni> n/m
[22:44:12] <SWPLinux> and it doesn't have aparport, so ...
[22:44:23] <SWPLinux> maybe I should stick in the PCIe parport I have and try that :)
[22:44:27] <alex_joni> that one looks like a tcl error
[22:44:52] <jepler> if (blockPtr->magicNum1 != MAGIC || blockPtr->magicNum2 != MAGIC) {
[22:44:53] <jepler> Tcl_Panic("alloc: invalid block: %p: %x %x",
[22:44:53] <alex_joni> http://osdir.com/ml/lang.tcl.core/2005-06/msg00002.html
[22:44:53] <jepler> blockPtr, blockPtr->magicNum1, blockPtr->magicNum2);
[22:44:58] <jepler> I agree, it looks like it's coming from inside tcl
[22:45:08] <jepler> could indicate bad memory management in a program extending tcl with c, though
[22:45:11] <jepler> and emc sure does that
[22:45:35] <alex_joni> stepper_inch.ini should run AXIS though
[22:45:45] <SWPLinux> all the stepper/* do
[22:45:47] <alex_joni> jepler: tcl extending with c is mainly emcsh .. right?
[22:46:01] <alex_joni> e.g. mini & tkemc
[22:46:12] <alex_joni> SWPLinux: just for fun.. try sim/tkemc.ini
[22:46:13] <jepler> SWPLinux: hm, you're repeating yourself
[22:46:14] <jepler> http://www.linuxcnc.org/irc/irc.freenode.net:6667/emcdevel/2009-02-19.txt
[22:46:19] <jepler> 17:26
[22:46:35] <SWPLinux> alex_joni: I just did, and it works
[22:47:48] <jepler> alex_joni: axis has some tcl extensions
[22:48:01] <jepler> I see that last month SWPLinux pastebinned a gdb traceback but now it's gone
[22:48:11] <jepler> bbl
[22:48:22] <alex_joni> <cradek> then the program that dumped core was /usr/bin/python or so
[22:51:09] <SWPLinux> I wonder what I did to (a) get that and/or (b) fix it :)
[22:51:50] <alex_joni> I don't
[22:51:51] <alex_joni> :P
[22:52:14] <SWPLinux> ah, I did nothing - I waited for some Jaunty fixes :)
[22:53:28] <alex_joni> apt-get update && upgrade ? :D
[22:53:58] <SWPLinux> oh, it's up to date
[22:54:15] <SWPLinux> at least it was this morning. I'm seeing if much has changed since then ;)
[22:54:45] <SWPLinux> heh. new kernel :)
[22:55:47] <CIA-2> EMC: 03seb 07TRUNK * 10infrastructure/farm-scripts/emc2_mail.py: dont complain so much when everything's working well
[23:02:23] <SWPLinux> ok, it's in the same place, _toglmodule.cc:21
[23:02:29] <SWPLinux> now I remember :)
[23:03:14] <SWPLinux> http://pastebin.ca/1369714
[23:04:09] <SWPLinux> I remember having weird issues with it as well. I added a little code, like a printf, to the function, and the error location changed significantly - not just 1 line later
[23:04:25] <SWPLinux> bbiab
[23:13:43] <alex_joni> good night all
[23:15:06] <micges> good night
[23:29:07] <micges> emc-devel translation is for trunk
[23:29:18] <micges> forgot to metion this
[23:30:42] <micges> good night all
[23:58:08] <CIA-2> EMC: 03jepler 07TRUNK * 10emc2/src/emc/usr_intf/stepconf/stepconf.py: i18n for some strings written to ini and hal files