#emc-devel | Logs for 2009-06-28

[01:00:23] <cradek_> what have I been missing?
[01:00:33] <cradek_> cradek_ is now known as cradek
[01:00:52] <skunkworks> heh
[01:01:13] <skunkworks> Lots of cradek has joined #emc-devel..
[01:01:28] <skunkworks> ;)
[01:02:20] <cradek> we're a sturdy lot
[01:02:31] <skunkworks> then some - cradek has quit #emc-devel. <- about 50%
[01:03:07] <cradek> if that's all I've missed, it doesn't seem like much.
[01:03:15] <skunkworks> nope
[01:04:13] <skunkworks> hw id the bus go?
[01:04:23] <cradek> lots of work, but all done
[01:04:32] <skunkworks> *how did
[01:04:36] <skunkworks> Nice!
[01:04:58] <cradek> I have a few photos - I want to take some more of the various tools I made for the job, then I will write a little blog post
[01:05:34] <skunkworks> did you make the 3.5 inch socket?
[01:05:56] <cradek> yes
[01:06:35] <skunkworks> what ended up being your internet problem?
[01:07:16] <cradek> it's not fixed yet. now my router is rebooting itself a few times an hour, and sometimes stopping working altogether. it is just a bad one I hope. I will get a new one monday.
[01:07:28] <skunkworks> ah
[01:07:32] <cradek> there have been so many different routers and problems that I've lost track.
[01:07:40] <skunkworks> wow
[01:08:01] <cradek> we'll get it soon.
[01:08:17] <skunkworks> struck by lightning?
[01:08:20] <cradek> I've spent pretty much two solid days on it
[01:08:48] <cradek> nope, the original problem is a firmware bug in the router that's tickled when you have one machine with more than one IP
[01:09:00] <cradek> it just goes nuts
[01:09:19] <cradek> I worked around that by adding another network card and futzing with static arp tables
[01:09:43] <cradek> now the resetting periodically is a separate problem I think.
[01:10:31] <skunkworks> yecky
[01:17:23] <cradek> skunkworks: special tool made for seal insertion: http://timeguy.com/cradek-files/emc/IMAG0032.jpg
[01:17:51] <cradek> new seal in place: http://timeguy.com/cradek-files/emc/IMAG0033.jpg
[01:19:21] <skunkworks> very nice!
[01:20:11] <cradek> counterweight tool made for getting the spline lined up in the differential without taking the differential cover off: http://timeguy.com/cradek-files/emc/IMAG0038.jpg
[01:20:45] <skunkworks> Now that is thinking
[01:21:08] <cradek> I was a hero for thinking of that :-)
[01:21:19] <skunkworks> :)
[02:22:45] <jmkasunich> I didn't realise this was a rear axel
[02:22:49] <jmkasunich> axle evel
[02:22:52] <jmkasunich> even
[02:24:44] <jmkasunich> big heavy parts
[02:29:32] <skunkworks> jmkasunich: how have you been?
[02:48:41] <jmkasunich> busy
[02:49:00] <jmkasunich> went to a nephew's graduation party in PA today
[02:54:34] <skunkworks> very far? only a state away ;)
[02:55:58] <jmkasunich> 2.5 hours each way, not bad
[02:56:09] <jmkasunich> but it still puts a huge dent in the day
[02:57:24] <skunkworks> yes
[02:57:56] <jmkasunich> got a long day tomorrow too... .bedtime
[03:02:15] <skunkworks> night
[07:19:58] <CIA-3> EMC: 03micges 07master * rf4f0ed834856 10/src/emc/ (4 files in 3 dirs): Move task delayLeft value to correct place
[09:21:59] <CIA-3> EMC: 03micges 07master * r0c71683e6490 10/src/emc/ (4 files in 3 dirs): Change all boolean fields of TRAJ_STAT to have bool type
[09:25:52] <micges> odd buildbot reports again git error
[09:46:31] <awallin> this looks odd: return *(new myObj()); is there a better way?
[13:09:04] <jepler> awallin: you probably want return myObj()
[13:56:47] <awallin> jepler: I'm copy/pasting from an operator overload tutorial. they suggest += -= *= = ans similar operators have return type of &MyClass then return myObj() doesn't work (I think)
[13:56:54] <awallin> and
[13:57:52] <awallin> I think they do that to allow things like this: obj.method1().method2()
[13:59:31] <SWPadnos> err
[14:00:11] <awallin> * awallin has about 2 days of c++ behind him :)
[14:00:21] <SWPadnos> me too, but it was years ago ;)
[14:01:35] <jepler> awallin: for an operator like =, +=, return *this
[14:01:59] <jepler> same for method1, method2 if you want to be able to chain them in that way
[14:02:15] <jepler> c++ is a baffling language for the first 5 years or so, then it levels off
[14:02:45] <jepler> (or else they come out with the next standard rev, and it's bafflement all over again...)
[14:06:23] <awallin> I'm trying to writhe the DropCutter CAM-stuff in c++ now. It's the third time around (first in Matlab, then in Python) so I should know it by now :) Hopefully this is the final version and they will be able to use it from HeeksCAD/CNC. Then I want to do the octree simulated cutting. (lot's of plans for the vacation :) )
[16:22:23] <Goslowjimbo> Is it proper to add the same function at 2 different points in a thread?
[16:22:48] <CIA-3> EMC: 03bigjohnt 07v2_3_branch * r69df55b539e8 10/docs/src/common/User_Concepts.lyx: Add more info on Homing and Touch Off
[16:23:31] <Goslowjimbo> I want to add a hm2_7i43.0.write to my servo thread, and 1 already exists for another purpose.
[16:26:26] <SWPadnos> Goslowjimbo, I believe it's impossible to do that
[16:27:23] <SWPadnos> if you try, you should get an error stating that the function is already part of a thread
[16:30:01] <Goslowjimbo> So if I use the write function for the servo system, there's no way to use it for raw reads and raw writes on the same thread cycle?
[16:34:37] <SWPadnos> I don't know that I understand the question
[16:34:58] <SWPadnos> if the write function writes both servo settings and SPI data, then both will be done in one call
[16:35:22] <SWPadnos> you can't do two SPI things in the same servo cycle though, unless there are two separate HAL functions
[16:38:24] <Goslowjimbo> I may be the confused one. When I read in the hostmot2 man, I see that a write_address tell where the .write function sends the write_data.
[16:39:08] <SWPadnos> that's a register, not a HAL function (right?)
[16:40:11] <Goslowjimbo> The .write is a function according to the man page. The address and data are registers.
[16:40:26] <SWPadnos> ah, ok
[16:40:50] <SWPadnos> you can write exactly once per thread execution, and the write function can be in only one thread
[16:41:12] <SWPadnos> if the .write function also writes servo data, then it makes sense to put it in the servo thread
[16:41:40] <Goslowjimbo> No argument there.
[16:43:25] <Goslowjimbo> I may be reading too much into it. In the description of the .write function, it says it updates the rates, cycles, and GPIO details.
[16:44:46] <Goslowjimbo> But in the raw mode description, it say when a write fuction is called, It uses the address and data specified by write_address and write_data.
[16:46:04] <Goslowjimbo> Maybe that's just in context of raw mode, and the write function does much more ? (I hope).
[16:46:54] <SWPadnos> I'd have to look harder than I have time for at the moment :)
[16:47:19] <Goslowjimbo> I understand.
[16:47:50] <SWPadnos> it looks like you can do whatever you want by making a component that has output pins for the address to write and the data to be written
[16:47:59] <SWPadnos> plus an I/O pin for the strobe
[16:48:17] <Goslowjimbo> That's what I'm hoping.
[16:49:34] <SWPadnos> readback may be a little trickier, since you'd have to detect when your data is valid somehow
[16:49:46] <SWPadnos> and the function order has to be pretty particular
[16:50:22] <SWPadnos> (whatever sets the read address, then the hm2 read function, then whatever actually uses the data and sets the write address, then the hm2 write function)
[16:52:14] <SWPadnos> gotta run. good luck with it
[16:53:06] <Goslowjimbo> Actually, pcw explained to me that readback is easy. the SPI interface is so quick, the read data is always valid.
[16:53:14] <Goslowjimbo> Thanks
[17:22:44] <noel> noel is now known as Guest29584
[17:22:48] <Guest29584> Guest29584 is now known as Roguish
[17:37:10] <Roguish> SWPadnos: is git down or up? seems unresponsive.
[17:41:32] <micges> there are some problems with net where git is hosted but here (Poland) it works for me
[17:41:51] <micges> pull works
[17:47:50] <Roguish> i get an error: fatal: The remote end hung up unexpectedly
[17:49:03] <micges> where are you?
[17:50:01] <Roguish> SF bay area
[17:51:37] <micges> seems some net routing problems
[17:52:48] <Roguish> ghost of MJ
[17:54:22] <micges> heh
[18:02:57] <CIA-3> EMC: 03cmorley 07master * rab153ba58023 10/src/hal/utils/meter.c: Fix halmeter command line options
[18:12:04] <alex_joni> Roguish: wait a bit, then retry
[18:12:08] <alex_joni> it usually works then
[18:17:13] <micges> alex_joni: buildbot can't connect with git
[18:20:01] <Roguish> alex_joni: still no luck. just did a 'pull' with same error.
[18:30:53] <cradek> I"m working on the git problems. like SWPadnos said on the list, the ISP is having trouble with a router. Please be patient.
[18:31:18] <cradek> I think it's up and down, so if you keep trying, it will work sometimes.
[18:32:11] <cradek> (actually, for me it seems like it is working right now)
[20:49:15] <Roguish> anybody: little help with hostmod2....
[20:49:44] <SWPadnos> what's the trouble?
[20:49:51] <Roguish> constant error at start of emc: hm2-servo.hal:36: parameter or pin 'hm2_5i20.0.pwmgen.pwm_frequency' not found
[20:50:08] <Roguish> good, the pros from dover...
[20:50:26] <Roguish> stock emc install works with sample.
[20:52:01] <Roguish> stock emc (released 2.3.1) does not work with my config.
[20:52:02] <SWPadnos> how many pwmgens did you enable on the loadrt line?
[20:52:07] <Roguish> used to work.
[20:52:47] <Roguish> 4 encoders, 4 pwms
[20:52:51] <SWPadnos> ok
[20:52:52] <Roguish> no stepgens
[20:53:19] <SWPadnos> did you copy that error or re-type it?
[20:53:23] <Roguish> copy
[20:53:28] <SWPadnos> ok
[20:53:40] <Roguish> typnng is not thet goood
[20:53:45] <SWPadnos> heeeeh
[20:53:59] <SWPadnos> can you pastebin the hal and ini files
[20:54:10] <Roguish> is this something with the symlink to the bit files ???
[20:54:39] <SWPadnos> probably not, unless there are no PWMs in the bitfile it loads
[20:54:44] <SWPadnos> (assuming it loads a bitfile at all)
[20:55:08] <SWPadnos> I'd expect some other error in that case though
[20:55:17] <SWPadnos> (don't know what, but I'd expect something different)
[20:56:01] <Roguish> http://pastebin.com/dc01042f
[20:56:44] <Roguish> http://pastebin.com/d10b6577b
[20:57:47] <Roguish> i really don't follow that symlink thing
[20:58:00] <Roguish> like and alias?
[20:58:14] <SWPadnos> what symlink are you referring to?
[20:59:43] <Roguish> to the bit file. referred to in the beginnign of one of the sample files for the hostmod2 config
[21:00:11] <Roguish> maybe it's left over text....
[21:00:24] <SWPadnos> are you calling this a symlink: #
[21:00:26] <SWPadnos> loadrt [HOSTMOT2](DRIVER) config=[HOSTMOT2](CONFIG)
[21:00:46] <Roguish> no, it's referred to in a comment .
[21:00:54] <SWPadnos> ah
[21:00:55] <SWPadnos> ok
[21:01:00] <Roguish> at the begging of a file.
[21:01:09] <SWPadnos> yep, I see that comment now
[21:01:11] <Roguish> confuses the heck out of us idiots.
[21:01:41] <SWPadnos> if you're running installed it shouldn't matter
[21:01:49] <SWPadnos> if you're compiling for yourself then it does
[21:02:17] <Roguish> well, my config is not runnning from either.
[21:02:33] <SWPadnos> the config is immaterial, unless you have custom bitfiles
[21:03:08] <Roguish> not me.
[21:03:23] <SWPadnos> well OK then :)
[21:03:41] <SWPadnos> I'm pretty sure the installed version creates the proper symlink for you
[21:03:53] <SWPadnos> or just puts the files where they need to go
[21:04:40] <Roguish> yeah, i checked . the bit files are in the /lib/firmware directory
[21:05:07] <Roguish> which leads back to what the heck is going on or not with my config?
[21:06:31] <SWPadnos> well, the param name sure looks right to me
[21:06:44] <Roguish> i thought maybe the board wasn't found, but it is 'cause it works on the standard installed emc2
[21:08:09] <Roguish> is a variable getting walked on somewhere?
[21:08:14] <SWPadnos> could you pastebin the error?
[21:08:19] <Roguish> ok
[21:08:36] <SWPadnos> I mean the terminal text from when you type emc until you have a prompt back
[21:09:13] <Roguish> http://pastebin.com/d62b73387
[21:09:50] <SWPadnos> how about dmesg?
[21:09:59] <Roguish> debug = 27
[21:10:09] <SWPadnos> you shouldn't have to run "scripts/emc". if you do, there's something wrong
[21:10:18] <Roguish> the whole thing?
[21:10:34] <SWPadnos> well, probably just the last 100 lines or so
[21:10:52] <SWPadnos> or if you don't need to keep it, you can dmesg -C (or -c - I don't recall) to clear it, then run emc agaion
[21:10:54] <SWPadnos> again
[21:11:59] <Roguish> [ 4716.896465] hm2/hm2_5i20.0: firmware hm2/5i20/SVST8_4.bit not found
[21:12:00] <Roguish> [ 4716.896474] hm2/hm2_5i20.0: install the package containing the firmware, or link your RIP sandbox into /lib/firmware manually
[21:12:02] <Roguish> [ 4716.896480] hm2_5i20.0: board fails HM2 registration
[21:12:15] <Roguish> that darn link thing.
[21:12:16] <SWPadnos> BIT
[21:12:18] <SWPadnos> not bit
[21:12:39] <Roguish> ????
[21:12:43] <SWPadnos> I didn't catch that in the ini file
[21:13:12] <SWPadnos> change the bitfile name on line 14 of the ini file so it ends with "BIT", not "bit"
[21:13:24] <SWPadnos> Linux names are case-sensitive ...
[21:13:56] <Roguish> oh crap.
[21:14:21] <Roguish> what ever happened to good old dos.....
[21:14:29] <SWPadnos> death, luckily :)
[21:14:48] <Roguish> freeDOS is alive....
[21:15:11] <SWPadnos> yeah. it's still useful for flashing BIOS chips and stuff :)
[21:15:44] <Roguish> runnning again. thanks.
[21:15:47] <SWPadnos> sure
[21:15:59] <Roguish> hey, whos worked on the integrators manual
[21:16:11] <SWPadnos> mostly bigjohnt lately
[21:16:12] <Roguish> it is REALLY GOOD !!!!!!!!
[21:16:36] <Roguish> heck of a clean up and improvement. OUTSTANDING.
[21:16:39] <SWPadnos> great - let him know :)
[21:17:17] <Roguish> git stilll seems to be unresponsive.
[21:18:06] <SWPadnos> yep
[21:18:31] <SWPadnos> I think a router is getting replaced tomorrow - we'll have to see what happens then
[21:19:27] <Roguish> ok. no problem.
[21:19:58] <Roguish> at least i can get back to some moderately productive screen layout. thanks swpadnos.
[21:20:06] <SWPadnos> sure. enjoy