#emc-devel | Logs for 2010-05-18

[01:17:36] <CIA-2> EMC: 03jthornton 07v2.4_branch * rfcedf562b881 10/docs/src/ (Submakefile docs.xml index.tmpl): add pncconf to html docs
[01:17:37] <CIA-2> EMC: 03jthornton 07v2.4_branch * ra8fe342f04cb 10/docs/src/gcode/coordinates.lyx: try and clear up vars a bit
[02:56:16] <CIA-2> EMC: 03jepler 07v2.4_branch * re002a40a94b7 10/debian/changelog: correct errors in changelog
[02:56:17] <CIA-2> EMC: 03jepler 07v2.4_branch * r66a54f5c50f6 10/configs/classicladder/cl-estop/cl-estop.ini: use default EMC_NMLFILE
[02:56:18] <CIA-2> EMC: 03jepler 07v2.4_branch * r68fc9f72caa9 10/debian/update-dch-from-git: don't prepend to deb changelog if it's up-to-date
[02:56:22] <CIA-2> EMC: 03jepler 07v2.4_branch * r35a106191515 10/src/emc/rs274ngc/interp_convert.cc: fix touch-off when loaded program uses %
[06:10:06] <ries_> ries_ is now known as ries
[06:19:10] <CIA-2> EMC: 03cmorley 07master * re02fa5dc302c 10/src/hal/components/mux16.comp: fix mux16 component - once again
[14:00:35] <jepler> yes, ubuntu 10.04's lyx 1.6.5 writes a different data format than 8.04 (345 vs 276). However, the docs (including html docs) seem to build on 10.04 when the files are in format 345. IMO this means there are no technical barriers against editing lyx docs with 10.04 when we decide to abandon hardy.
[14:02:10] <cradek> yay
[14:02:12] <JT-Work> cool
[14:19:46] <jepler> there's a new lyx->html converter available in 10.04. Here's a sample of its output: http://media.unpythonic.net/emergent-files/sandbox/Getting_Started.html.LyXconv/Getting_Started.html
[14:20:45] <jepler> there are some unfortunate things about it but the software ('elyxer') might be worth improving, rather than relying on the framework I wrote
[14:20:56] <jepler> (what the HECK is with that section numbering, for instance..)
[14:21:18] <jepler> <h3 class="Subsection">
[14:21:18] <jepler> <a class="toc" name="toc-Subsection-8.1.1">8.1.1</a> <a class="toc" name="toc-Subsection-3.1.1">3.1.1</a> Sharing Files
[14:24:02] <jepler> http://wiki.lyx.org/Tools/ELyXer
[14:30:34] <JT-Work> chapter numbers are messed up too
[14:34:39] <alex_joni> and image tags
[14:35:08] <alex_joni> SWPadnos: I think you used some 12V PSUs?
[14:35:12] <alex_joni> or was it jmkasunich ?
[14:35:17] <SWPadnos> I have
[14:35:20] <SWPadnos> and 24V
[14:36:26] <alex_joni> how about the D945GCLF2 application you did a while ago?
[14:36:33] <alex_joni> I think you used those.. right?
[14:36:36] <SWPadnos> yep
[14:36:42] <SWPadnos> though they're discontinued now
[14:36:46] <SWPadnos> (bastards)
[14:36:50] <alex_joni> did you by any chance measure consumption?
[14:36:58] <SWPadnos> 26W under load
[14:37:09] <SWPadnos> including the hard drive and I think an ethernet switch
[14:37:09] <alex_joni> cool
[14:37:23] <SWPadnos> yes
[14:37:40] <SWPadnos> there's still a fan on these boards, I think the D510-based replacement has only a giant heatsink
[14:37:58] <SWPadnos> and I believe it's a little faster and slightly lower power as well
[14:38:06] <alex_joni> cool
[14:38:12] <SWPadnos> only has 3 USB ports, so it's useless for me
[14:38:46] <alex_joni> http://www.newegg.com/Product/Product.aspx?Item=N82E16813130281&cm_re=d510-_-13-130-281-_-Product
[14:39:15] <SWPadnos> yep, that's one
[14:40:10] <SWPadnos> here's the one from Intel: http://www.newegg.com/Product/Product.aspx?Item=N82E16813121399
[14:40:35] <alex_joni> it says 4 USB
[14:40:42] <SWPadnos> so I see. I'm looking now :)
[14:41:01] <SWPadnos> yep, there they are
[14:41:58] <skunkworks_> I thought it had 4 external and only 3 internal.. you needed 4 internal
[14:42:54] <SWPadnos> the MSI has 4 ports on two headers
[14:43:01] <alex_joni> wonder if it has any parports
[14:43:13] <alex_joni> there's a PCI, so an add-on is possible
[14:43:14] <SWPadnos> plus the 4 ports on the back (though they only list 2)
[14:43:28] <SWPadnos> oh no, they do say there are 4 on the back
[14:44:01] <SWPadnos> yeah, you'd have to get a PCI Card
[14:44:15] <SWPadnos> actually there is an LPT header on the MSI board
[14:44:25] <SWPadnos> just behind the serial port
[14:45:27] <SWPadnos> it looks like the Intel board also has a parport header
[14:45:41] <alex_joni> there is a something, but not sure it's a parport header
[14:45:55] <SWPadnos> it says "parallel
[14:45:56] <SWPadnos> "
[14:46:09] <SWPadnos> so there's a good chance :)
[14:46:42] <SWPadnos> hmm. the Intel board also has a mini-PCIe "slot"
[14:47:03] <jepler> and it's 26 with a key in the last pin, which is also a good sign
[14:47:20] <jepler> and it's right next to the winbond chip
[14:47:36] <SWPadnos> and it says "PARALLEL" next to it :)
[14:47:56] <SWPadnos> though they could be describing the pin arrangement
[14:51:19] <jepler> One parallel port header with Extended Capabilities Port (ECP) and Enhanced
[14:51:19] <jepler> Parallel Port (EPP) support
[14:56:06] <jepler> SWPadnos: I'd say it's more of a perpendicular connector, wouldn't you?
[14:56:22] <SWPadnos> woth parallel rows, so it's a hybrid
[14:57:49] <SWPadnos> with
[15:08:25] <CIA-2> EMC: 03cradek 07v2.4_branch * rfe03d1c0432d 10/src/emc/rs274ngc/gcodemodule.cc: remove debugging print
[15:36:07] <Dave911> That D510 MSI board must be very new ... I haven't seen that one before....
[15:37:14] <Dave911> Good to see more D510 type ITX boards coming out...
[15:41:53] <Dave911> With the D510 Intel board, the video flakes out with the standard intel drive for Linux that comes with 9.10. It causes a i810 module not found error on boot up. I found that using the vesa driver works fine. I put a note on the Wiki page that I setup to reflect that. Others are having similar problems with Ubuntu 10.4 also. The Ubuntu intel IGP driver apparently needs some work.
[15:44:21] <mozmck_work> Seems like intel video linux drivers are generally more flaky than others. There have been lots of problems with i915 as well.
[15:47:01] <jepler> Ingrid's laptop with intel video (915?) is flaky under 10.04 but my laptop works well (GM965).
[15:47:46] <jepler> micges has seen crashes with intel video on 9.10 as well, though that specific one had a fix on the debian bugtracker
[15:55:37] <cradek> I don't believe I could have screwed up that commit
[15:57:29] <cradek> oh, maybe it's mux16
[15:57:41] <Dave911> I didn't realize the Jetway ITX D510 boards don't require a conventional power supply with 12/5/3.3 volts etc.... only 12 volts!
[15:58:31] <Dave911> You would think that the Ubuntu guys would have fixed the intel driver issue since the intel IGP seems to be very popular
[16:21:48] <dgarr> Dave911: jepler posted a patch http://emergent.unpy.net/files/sandbox/0001-axis-avoid-leaving-lots-of-zombie-processes.patch
[16:40:37] <micges> anyone compiled emc-sim in 10.04 ?
[16:41:05] <jepler> yes. no surprises.
[16:41:15] <jepler> well, I haven't built a package
[16:41:36] <dgarr> yes -- on both 32 bit and 64 bit machines
[16:42:13] <micges> what dependencies can have hal tools? they doesn't get compiled here
[16:42:21] <micges> halscope, halmeter
[16:43:14] <jepler> looking at debian/control.in, probably a
[16:43:15] <jepler> a
[16:43:17] <jepler> argh
[16:43:20] <jepler> libgtk2.0-dev
[16:43:29] <jepler> but when in doubt run configure and look at the output
[16:44:25] <micges> thanks
[16:44:40] <Dave911> dgarr: Thanks ... didn't catch that before ... I'll do that. Thanks.
[16:45:54] <jepler> dgarr: have you been running with that patch?
[16:46:10] <jepler> if you are and no trouble turned up, I'm going to put it on v2.4_branch
[16:46:12] <jepler> thanks for bringing it up again
[16:46:15] <dgarr> yes
[16:47:14] <dgarr> also, you suggested a reminder re mem leaks for another patch
[16:47:37] <jepler> which patch is that?
[16:48:07] <dgarr> http://www.panix.com/~dgarrett/stuff/0001-tooledit-conditionally-add-button-to-write-file-AND-.patch
[16:50:33] <jepler> + if [info exists ::env(EMC2_HOME)] {
[16:50:38] <jepler> does this line mean "if started inside emc"?
[16:53:36] <dgarr> yes or executed the emc-environment script or otherwise set EMC2_HOME
[16:54:17] <dgarr> so not necessarily started inside emc, if EMC2_HOME is set, one can start tooledit from a shell and have a chance of finding emcsvr running
[16:54:39] <jepler> for a run-installed system and tooledit started from a shell, it won't look for emcsvr
[16:56:18] <dgarr> true -- because it wouldn't know how to find EMC2_TCL_DIR to find emc.so
[17:00:07] <jepler> so what we need is a better way to locate tcl-related files (for instance, by doing some kind of sed to get tooledit into bin/, not just a copy)
[17:00:16] <jepler> bbl, lunchtime here
[17:15:57] <dgarr> emc.so is found by EMC2_TCL_DIR for rip, found by hardcoded path for install, so EMC2_HOME requirement can be eliminated, revised:
[17:15:59] <dgarr> http://www.panix.com/~dgarrett/stuff/0001-tooledit-conditionally-add-button-to-write-file-AND-.patch
[17:16:23] <dgarr> but you may object to hardcoding of /usr/share/emc/tcl/emc.so
[17:19:51] <dgarr> starting from an emc gui like axis or tkemc is surely the normal way for most users; starting independently from a shell requires caution as one can specify a different tooltable file than specified in the emd ini file
[17:34:28] <SWPadnos> editing another file isn't an error in itself, since you might want to edit tools for a different job than is running
[17:34:39] <SWPadnos> it could be confusing though, if that's not your intent :)
[17:36:10] <dgarr> "requires caution"-- particularly if you hit the button to load the tool table (new button in the patch)
[17:44:43] <alex_joni> Dave911: where did you see that? (12V)
[17:55:33] <CIA-2> EMC: 03jepler 07v2.4_branch * rfadc3ca0691a 10/src/emc/usr_intf/axis/scripts/axis.py: axis: avoid leaving lots of zombie processes
[17:59:40] <jepler> it looks like to me there's an ATX-looking power connector at the bottom center of that motherboard.
[18:03:32] <SWPadnos> which one?
[18:03:57] <jepler> SWPadnos: I think the one we're talking about is http://www.newegg.com/Product/Product.aspx?Item=N82E16813130281&cm_re=d510-_-13-130-281-_-Product
[18:04:30] <SWPadnos> a few of the Jetway ones say "power onboard" in the one-liner description
[18:05:08] <SWPadnos> that isn't one of them :)
[18:05:29] <jepler> ah
[18:05:31] <SWPadnos> http://www.newegg.com/Product/Product.aspx?Item=N82E16813153171
[18:06:14] <jepler> dgarr: we should find a tcl expert to tell us how to handle this. it would be nice if run-installed would have "package require emc" always work, and rip would too as long as you'd sourced emc-environment.
[18:06:28] <jepler> but to my shame I still don't understand tcl 'package' after all these years
[18:10:39] <jepler> intel just can't resist high TDP, can they. The first Atom CPUs in 2008 had TDP <=2.4W; the latest ones in January 2010 have TDP up to 13W (Atom D510)
[18:10:44] <jepler> I thought the whole point of Atom was low TDP
[18:11:18] <jepler> I guess D510 includes a GPU on the package so it's not quite a fair comparison.
[18:11:24] <SWPadnos> and dual cores
[18:11:40] <SWPadnos> and much higher speed
[18:12:24] <SWPadnos> the funny thing is that if you use an athlon and underclock it, you get similar power usage
[18:12:47] <Dave911> alex_joni: http://www.newegg.com/Product/Product.aspx?Item=N82E16813153175 Jetway has a couple of them, they must be relatively new. I checked out their website and they don't say what power regulation is required for the 12 volt input which I find strange so I sent them an email asking them.
[18:12:48] <jepler> underclock+undervolt has long been a recognized technique for reducing power consumption
[18:12:48] <SWPadnos> especially when you look at overall system power, since there is no northbridge in an AMD system
[18:12:54] <SWPadnos> yep
[18:13:13] <SWPadnos> which is what Intel is doing with the ULV chips - they run on 1.2V or so
[18:15:28] <alex_joni> I think less
[18:16:24] <SWPadnos> maybe down to 0.9V
[18:16:45] <Dave911> How far do you have to underclock an AMD chip to get into Atom territory? Which AMD chip can you do that to?
[18:17:00] <jepler> lowest voltage I can find is the Atom Z500/Z515 at 0.712V
[18:17:12] <jepler> 0.65W TDP @ 800MHz
[18:17:18] <SWPadnos> I think the article I read used an Athlon something at 1 GHz
[18:17:37] <SWPadnos> CPU power consumption was 7W or so
[18:17:50] <Dave911> That's slow
[18:17:54] <SWPadnos> overall system consumption was a little lower than an Atom 270 board
[18:18:06] <Dave911> wow
[18:18:14] <SWPadnos> sure, but it's a real 64-bit CPU, with real cache etc.
[18:19:23] <Dave911> So you can still get more hp out of the AMD even though the clock is that slow ... is that a 1.8 ghz chip running at 1 ghz or similar?
[18:19:41] <SWPadnos> something like that, I don't remember the details
[18:22:09] <jepler> http://www.silentpcreview.com/article164-page1.html Deep UnderVolt/Clock: 4.7W CPU PC " I have my Athlon XP1700+ at 1.1 V and 300 Mhz"
[18:22:22] <SWPadnos> heh
[18:22:26] <jepler> not sure if this is the one you had in mind, but 300MHz is going to be pretty slowwwwww
[18:22:26] <SWPadnos> that sounds a bit too slow ;)
[18:22:31] <SWPadnos> no, it's not
[18:28:08] <Dave911> I just recompiled the master as run in place and I have one file that hangs up EMC2 when starting ... "emc_module_helper" If I look at the permissions from the gui, it isn't execute or not. It is something different. If I set it to execute then emc2 will run. After I change it to execute and I look at it with ls -l, I see
[18:28:10] <Dave911> rwsr-x--x?? What is read- write - S?? There are few other files with the same permissions, but not many. What did I screw up??
[18:28:37] <cradek> emc_module_helper should be setuid root and that's what the s means.
[18:29:18] <Dave911> ok so it isn't set for execute though and that hangs emc2.
[18:30:03] <Dave911> Shouldn't things be correct after I do a sudo make setuid or is that for something else?
[18:30:30] <cradek> wait wait wait
[18:30:37] <cradek> tell me what problem you are having
[18:31:25] <Dave911> Ok... I have rip directory, did a git clone etc compiling the master with the patches from dgarr and jepler..
[18:32:06] <Dave911> I follow the usual routine to compile, get to the end and emc2 won't run, I get a message that I don't have permission to run that file -
[18:32:20] <cradek> what is the exact message??
[18:32:31] <Dave911> emc_module_helper
[18:32:54] <Dave911> I can probably recreate this is just a minute .. and post the results
[18:33:41] <Dave911> I " fixed" it by setting the file to execute via the Ubuntu gui logged in as root..
[18:42:26] <Dave_911> http://pastebin.com/vWfVK7Fm Here it is ... the emc2 errors window content
[18:43:09] <jepler> doesn't work. - Unknown paste ID, it may have expired or been deleted! THE SITE IS IN READ-ONLY MODE AS WE ARE CURRENTLY MOVING HOSTER, SOME PASTES MIGHT BE MISSING FOR A FEW HOURS!
[18:44:02] <Dave_911> Geez.. I can see it.. ;-) I'll try .ca
[18:44:34] <andypugh> Works for me...
[18:45:15] <cradek> pasties?
[18:45:28] <cradek> hoster?
[18:45:54] <Dave_911> I can't get into pastebin.ca ???
[18:46:10] <jepler> pastebin.ca has been broken for me for weeks
[18:47:04] <mozmck_work> try http://pastebin.org/
[18:48:17] <Dave_911> http://pastebin.org/249392 That worked :-)
[18:49:52] <jepler> you need to retrace your steps and figure out why emc_module_helper got bad permissions. rm /home/user/emc2-dev/bin/emc_module_helper; make; sudo make setuid; ls -l /home/user/emc2-dev/bin/emc_module_helper
[18:50:24] <jepler> after that step it should have permissions similar to the installed file
[18:50:24] <jepler> -rwsr-xr-x 1 root root 5788 2010-05-09 12:31 /usr/bin/emc_module_helper
[18:50:45] <Dave_911> So changing to the permission to execute on that file via the Ubuntu gui changed the last letter to x
[18:51:17] <Dave_911> So I should delete the file, recompile etc? and then look at the file?
[18:51:26] <jepler> why were you changing the permissions manually in the first place? the same message, or a different message?
[18:51:53] <Dave_911> I changed the permissions after the run failed after a compile
[18:51:59] <jepler> you should never have to manually change the permissions of emc_module_helper; sudo make setuid is the way to set the special permissions required by emc_module_helper
[18:52:21] <Dave_911> I agree ...
[18:52:26] <jepler> did you "sudo make setuid"?
[18:52:32] <Dave_911> Yes
[18:52:56] <Dave_911> This just popped up when I did a compile last night after I updated my git clone
[18:53:21] <Dave_911> I can delete that file and recompile .. it will just take a few minutes
[18:54:05] <Dave_911> Actually I just did a recompile and it reset the permissions from what I had manually changed to "fix" the problem
[18:56:18] <Dave_911> I did a sudo make clean last night thinking something was stuck but that didn't affect anything ..
[19:00:54] <jepler> after "make clean" you should be informed to "sudo make setuid" again.
[19:01:11] <jepler> usually the only think you need to sudo is "sudo make setuid"; "sudo make clean" shouldn't be necessary
[19:02:59] <Dave_911> OK.
[19:03:01] <Dave_911> I just did what you suggested and I have the file back after I deleted it..
[19:03:03] <Dave_911> -rwsr-x--- emc_module_helper
[19:03:31] <cradek> the whole line please?
[19:04:17] <Dave_911> -rwsr-x--- 1 root user 18902 201--05-18 15:00 emc_module_helper
[19:04:36] <cradek> are you logged in as user?
[19:04:42] <jepler> is emc working now or still busted?
[19:04:45] <Dave_911> yes
[19:04:45] <cradek> and what the hell is with your date?
[19:05:05] <cradek> oh I bet you typed it
[19:05:07] <Dave_911> hey .. now it works
[19:05:24] <Dave_911> I typed it ..
[19:05:52] <jepler> if it works now then my diagnosis is: you probably did something (like make clean; make) that removed the setuid bit, and didn't notice it at the time
[19:06:04] <Dave_911> I left out a "0"
[19:06:30] <Dave_911> ok ... as my fault .. sorry
[19:08:51] <Dave_911> wait a minute ... so I should do a ......make clean ..... sudo make setuid .... make I thought the setuid came after the make?
[19:09:13] <jepler> the correct order is: make; sudo make setuid
[19:10:44] <Dave_911> h mmmmm... possible I screwed that up.... but if so I did it wrong several times ... ok...
[19:11:42] <Dave_911> deleting the file seems to have straightened it out .. but ... who knows .. I'll try once more to replicate it and then forget about it .. :-) Thanks
[19:17:33] <alex_joni> SWPadnos: http://gigapan.org/gigapans/48492/
[19:20:08] <SWPadnos> that's - um - a bit much :)
[19:23:19] <alex_joni> ha, there's a ghost there
[19:23:49] <alex_joni> in the shortcut list on the bottom (second from the left is "man with a red towel")
[19:24:05] <alex_joni> actuall 'tawl'
[19:24:39] <alex_joni> if you zoom out a bit after that, you'll see a ghost :D
[20:04:31] <alex_joni> is this still valid? http://sourceforge.net/tracker/?func=detail&aid=1886652&group_id=6744&atid=356744
[20:06:09] <alex_joni> I'm pretty sure this one is fixed http://sourceforge.net/tracker/?func=detail&aid=1849496&group_id=6744&atid=356744
[20:06:19] <alex_joni> jepler: can you confirm? ^^
[20:07:00] <alex_joni> http://sourceforge.net/tracker/?func=detail&aid=1846328&group_id=6744&atid=356744 <- also fixed iirc
[20:16:39] <andypugh> I am surprised not to see a bunch of demands from Steve B on that list.
[20:20:43] <alex_joni> I was gonna point him to that list, and that his demand is missing, but that would just mean pouring gas on the fire
[20:23:07] <andypugh> And you don't mean Halon do you?
[20:24:43] <alex_joni> bofh-style
[20:32:57] <cradek> alex_joni: if you keep giving him hoops to jump through you'll renew his mistaken belief that his jumping through a hoop obliges someone to do what he says. that's a silly distraction at best, and like you say, fuel for the fire at worst
[20:34:44] <alex_joni> right, that's why I won't do it
[20:35:04] <alex_joni> kinda sorry I did it last time, I was under the delusion I might enjoy working on this
[20:35:08] <alex_joni> or have the time to
[20:37:08] <SWPadnos> it's a hard (set of) problem(s)
[20:40:07] <jepler> alex_joni: I'm looking at those now
[20:40:23] <alex_joni> SWPadnos: it could have been a nice challenge
[20:40:33] <alex_joni> but now I surely feel no joy even considering it
[20:41:09] <SWPadnos> yeah, it's less exciting when you're more or less getting yelled at, called lazy, and berated for lack of talent
[20:41:33] <cradek> well, what if I *am* talentless and lazy?
[20:41:35] <jepler> alex_joni: 1846328 stepconf now has a limited ability to configure a second parport -- you can select the base address but you can't create connections in the GUI
[20:41:50] <SWPadnos> cradek, then you should join in the conversation ;)
[20:42:01] <jepler> alex_joni: 1849496 is implemented and I'm not aware of any outstanding bugs in it (there were during the 2.3 series)
[20:42:04] <alex_joni> and get yelled at
[20:42:38] <SWPadnos> of course, the yelling is mandatory
[20:42:51] <SWPadnos> otherwise why would any lazy talentless people do anything
[20:44:07] <jepler> alex_joni: 1886652 first item is not implemented (and I think I would reject it) -- you can always do it in custom.hal if you really want to harm the reliability of the machine.. second seem to have been.
[20:48:20] <jepler> cradek: jt actually did nurbs docs which I excised when the underlying code seemed broken
[20:48:40] <jepler> cradek: if you think it's fixed in 2.4 and have some kind of committment to keeping it working, we can restore the docs for 2.4.1.
[20:50:27] <cradek> I don't know of any problem other than (like I said on the list) how it isn't very good at rejecting invalid stuff
[20:51:30] <cradek> I can't say I'm committed to maintaining it, but the feature interests me a bit and I kind of know how it works after fixing that quadrant-transition glitch
[20:52:23] <jepler> obviously you can't predict whether you *can* fix any serious bug that might be discovered. but you can say whether you predict future-you will at least try.
[20:53:13] <jepler> it's the latter I'm trying to gauge
[20:54:27] <SWPadnos> woohoo - I may be able to stay an extra day at the CNC workshop
[20:54:31] <cradek> I think I'm as likely to work on any problem that pops up in the interpreter as any other. [is that a dodge?]
[20:54:38] <SWPadnos> it looks like I don't need to be in LA until Friday
[20:55:53] <cradek> SWPadnos: neat
[20:56:11] <SWPadnos> yes indeed
[20:56:13] <jepler> the docs that jt had for G5.2/5.3 can be seen in this version of the gcode docs: http://emergent.unpy.net/files/sandbox/gcode_main.html#sec:G5.2-G5.3-NURBs
[20:56:21] <jepler> looks like G5.1 is not tested/documented
[20:57:56] <cradek> it wfm
[21:10:47] <dgarr> jepler: I think i've learned how to make package for emc.so if it is located in a subdir of a directory specified in auto_path.
[21:11:02] <dgarr> jepler: Do you know why emc.so is currently located in PREFIX/share/emc/tcl?
[21:11:22] <jepler> dgarr: because I am lazy and that's the first arrangement that got it to work
[21:11:38] <dgarr> A better location for emc.so might be /usr/lib/tcltk/emc/ or /usr/share/tcltk/emc
[21:11:48] <jepler> dgarr: before emc.so all the tcl files got put in share/ in installed systems
[21:12:22] <jepler> when I added emc.so I put it next to the .tcl files in RIP, then wrote code to load it from the same directory as the .tcl files, then kept that arrangement (even though it was inappropriate) in the package/installed version
[21:12:50] <dgarr> would you consider a change that relocates emc.so?
[21:12:54] <jepler> yes!
[21:13:41] <jepler> the stuff in emc.tcl should probably be in the package too .. does it move to lib/ as well?
[21:14:03] <jepler> (a smaller sin by far than having arch-dependent stuff in share/!)
[21:15:25] <dgarr> not sure about emc.tcl, probably hal.so too
[21:15:54] <jepler> all that stuff should come from a nice easy "package", not load or source with ugly path manipulation!
[21:32:55] <andypugh> I can see how Steve's attitude could be demotivating, but I don't think he is the only one who wants pause/jog/resume. It would be a shame if it was sidelined _only_ because of a (struggles for polite term) "personality clash".
[21:34:47] <SWPadnos> there are a number of technical issues that are unrelated to any personality clash
[21:35:39] <andypugh> Granted. I was talking hypothetically.
[21:36:57] <SWPadnos> sure. it's pretty rare (I think) that personality problems materially get in the way of actual coding
[21:42:32] <andypugh> Talking of actual coding: I had a brief look at the COMEDI drivers today. Is anyone looking to integrate those into HAL? I saw the idea on a list on the Wiki and was curious enough to investigate.
[21:43:12] <SWPadnos> we noticed the mass commit of COMEDI drivers to RTAI, and would definitely accept patches that let us use all that hardware :)
[21:44:24] <jepler> I couldn't tell if COMEDI is rich enough to let you write one driver for "any COMEDI-compatible hardware", or if you have to write a little HAL shim for each distinct COMEDI driver
[21:44:44] <SWPadnos> it's hard to know at the in-kernel level
[21:45:19] <SWPadnos> certainly userspace code can open devices as it sees fit, but I don't know how much of that part of the API is available in-kernel;
[21:45:51] <jepler> I also don't know if COMEDI needs us to recompile our kernel or rtai -- I suspect it does. If so, now is a good time to help mozmck make the right decisions in that regard.
[21:46:16] <SWPadnos> we'd also have to consider that a person might not want to use all of the attached devices for EMC2/HAL, so we'd probably need some way of selecting which devices to use
[21:46:39] <SWPadnos> weren't the COMEDI drivers committed to RTAI some few weeks ago?
[21:46:54] <jepler> the thing that happened recently was that a bunch of them were put in 'staging' in the kernel git
[21:46:58] <jepler> but that won't be in the ubuntu kernel
[21:47:02] <SWPadnos> oh, right
[21:52:42] <andypugh> I might look at a little harder then. (I have other, higher, priorities, like something to actually use the 3PPWM).
[21:53:09] <andypugh> But I will keep an eye out for a cheap NI board on eBay.
[21:54:09] <jepler> andypugh: oh, that reminds me -- I added TPEN6_6 pinouts for the 3-connector mesa boards, so if you have someone wanting to test with one of those cards I can put those bitfiles online somewhere
[21:54:52] <andypugh> Good. PCW offered to send me a 7i39 to test with too.
[21:55:31] <jepler> he seems very willing to share his toys when it helps emc start supporting them
[21:55:41] <andypugh> How odd :-)
[22:00:58] <andypugh> Is it easy to add analogue voltage measurement to the FPGA configs, or are they not in the existing "building blocks"? It seems that almost nobody will want 6 servo axes, but 5 axes + current feedback would be very useful.
[22:01:51] <SWPadnos> someone with a hexapod might want 6 servo axes
[22:02:17] <andypugh> Can't they just use a pentapod?
[22:02:21] <SWPadnos> analog is a different ball game. you need hardware for that (A/D converters)
[22:02:27] <SWPadnos> they could, but that would be lame
[22:02:31] <jepler> andypugh: the FPGAs have no ADCs. There are some ADC daugherboards but I think they're not supported at the moment.
[22:02:36] <SWPadnos> (rimshot)
[22:02:55] <SWPadnos> I have one that is, but it's super expensive :)
[22:02:57] <andypugh> OK. I wasn't clear about how configurable FPGAs were.
[22:03:12] <SWPadnos> it's not the configuration, it's that they're digital, not analog
[22:03:16] <jepler> they can do most anything that is digital logic, but not analog
[22:03:19] <SWPadnos> unless you get a Lattice Fusion IIRC
[22:03:53] <SWPadnos> I think those have 32-channel A/Ds in them
[22:03:55] <jepler> you can get primitive DAC with resistor ladders, PWM, or other tricks that are digital-based, or hook the right analog stuff to the digital I/O pins
[22:04:08] <SWPadnos> but they're still low voltage, so you need signal conditioning and protection on them
[22:04:30] <andypugh> SPI...
[22:04:37] <SWPadnos> yep, that's what I used
[22:04:56] <SWPadnos> 6-channel 16-bit ADC and dual 4-channel 16-bit DAC
[22:05:19] <SWPadnos> at ~175k updates/second
[22:05:27] <andypugh> Seb said Mesa SPI was on his list, but timescale is unknown.
[22:05:34] <jepler> 7i65 is the mesa daughtercard that has servo interface + ADCs. It requires SPI which is available in hostmot2 but not in the emc hostmot2 driver..
[22:05:50] <SWPadnos> yes, that's another "simple for 60% but really hard hard for the other 40%" problems
[22:06:01] <SWPadnos> err, really really hard
[22:06:02] <jepler> hopefully at cnc workshop we can chain seb to a chair and make him write it. (I recently learned that it may be my perogative to do so, *evil laugh*)
[22:06:11] <SWPadnos> hmmm
[22:06:15] <SWPadnos> no more kool-aid
[22:08:00] <dgarr> jepler: when is master be updated from the 2.4v_branch?
[22:08:47] <jepler> dgarr: not on any particular schedule.
[22:09:33] <dgarr> if i work on package for tcl, i presume it should be on master
[22:09:54] <jepler> yes, I probably won't want to put that on 2.4.
[22:10:33] <dgarr> i would like changes from v2.4_branch in master but my git foo is too weak for that
[22:11:31] <jepler> (huh -- openoffice still doesn't know how to show its splash screen in the middle of a monitor on multimonitor systems)
[22:12:04] <jepler> (and TOTAL FAIL importing this xls file!)
[22:16:58] <jepler> dgarr: here you go.
[22:17:08] <CIA-2> EMC: 03jepler 07master * r64b064649a3b 10/docs/src/lathe/lathe-user.lyx: Add a bit more on tool changes
[22:17:09] <CIA-2> EMC: 03jepler 07master * r7f5c8030222a 10/docs/src/config/ini_config.lyx: add info on debug and task
[22:17:11] <jepler> .. any minute now ..
[22:17:11] <CIA-2> EMC: 03jepler 07master * r768f4f801bbf 10/debian/changelog: note new changes
[22:17:12] <CIA-2> EMC: 03jepler 07master * rf9cd27781bf4 10/tcl/tkemc.tcl: tkemc: fix tool editor
[22:17:13] <CIA-2> EMC: 03jepler 07master * r10815acca9ce 10/ (VERSION debian/changelog): bump version number for release
[22:17:14] <CIA-2> EMC: 03jepler 07master * r29393fe2a3aa 10/tcl/tkemc.tcl: tkemc: make the tool items more like axis
[22:17:14] <CIA-2> EMC: 03jepler 07master * ra661c173144d 10/src/hal/drivers/parport_common.h: parport: Don't believe linux parport feature bits
[22:17:15] <CIA-2> EMC: 03jepler 07master * rda78bc3a1a83 10/src/emc/usr_intf/tooledit.tcl: tooledit: fix angle format
[22:17:16] <CIA-2> EMC: 03jepler 07master * re8bf553388da 10/src/modsilent: silence several bogus kernel build warnings
[22:17:18] <CIA-2> EMC: 03jepler 07master * r6aab8a531f67 10/docs/src/gcode/main.lyx: update G2/3 G41/42 description
[22:17:24] <CIA-2> EMC: 03jepler 07master * r50e968d58c2b 10/ (15 files in 12 dirs): Merge branch 'v2.4_branch' of ssh://jthornton@git.linuxcnc.org/git/emc2 into v2.4_branch
[22:17:25] <CIA-2> EMC: 03jepler 07master * r6bd6136142d4 10/docs/src/hal/basic_hal.lyx: fix some typos pointed out by ries
[22:17:26] <CIA-2> EMC: 03jepler 07master * rce5e4f94908d 10/docs/src/gcode/main.lyx: Opps I removed this by mistake
[22:17:29] <CIA-2> EMC: 03jepler 07master * r275edfacb384 10/docs/src/config/ini_config.lyx: add info on homing with index
[22:17:29] <CIA-2> EMC: 03jepler 07master * r0a39101f0a0b 10/docs/src/drivers/hostmot2.lyx: add svst2 4 7i47 to 5i20 chart
[22:17:30] <CIA-2> EMC: 03jepler 07master * r67a462bd47da 10/docs/src/config/ini_config.lyx: expand a bit on machine units
[22:17:31] <CIA-2> EMC: 03jepler 07master * ra8fe342f04cb 10/docs/src/gcode/coordinates.lyx: try and clear up vars a bit
[22:17:35] <CIA-2> EMC: 03jepler 07master * re002a40a94b7 10/debian/changelog: correct errors in changelog
[22:17:36] <CIA-2> EMC: 03jepler 07master * r66a54f5c50f6 10/configs/classicladder/cl-estop/cl-estop.ini: use default EMC_NMLFILE
[22:17:38] <CIA-2> EMC: 03jepler 07master * r68fc9f72caa9 10/debian/update-dch-from-git: don't prepend to deb changelog if it's up-to-date
[22:17:38] <CIA-2> EMC: 03jepler 07master * r35a106191515 10/src/emc/rs274ngc/interp_convert.cc: fix touch-off when loaded program uses %
[22:17:39] <CIA-2> EMC: 03jepler 07master * rc14ae1145382 10/src/emc/usr_intf/pncconf/pncconf.py: Add a splash screen / progress bar
[22:17:40] <CIA-2> EMC: 03jepler 07master * rfe03d1c0432d 10/src/emc/rs274ngc/gcodemodule.cc: remove debugging print
[22:17:46] <CIA-2> EMC: 03jepler 07master * r76a676aa2d47 10/ (35 files in 16 dirs): Merge remote branch 'origin/v2.4_branch'
[22:17:47] <CIA-2> EMC: 03jepler 07master * rfadc3ca0691a 10/src/emc/usr_intf/axis/scripts/axis.py: axis: avoid leaving lots of zombie processes
[22:17:48] <CIA-2> EMC: 03jepler 07master * rf667e9b60e93 10/docs/src/ (16 files in 3 dirs): add point n click to getting started guide
[22:17:58] <CIA-2> EMC: 03jepler 07master * rfcedf562b881 10/docs/src/ (Submakefile docs.xml index.tmpl): add pncconf to html docs
[22:18:18] <jepler> you know, the way our scripts treat merges is not ideal
[22:18:55] <SWPadnos> blast blast blast why would you say that blast blast blast
[22:19:07] <dgarr> jepler: thanks!