#linuxcnc-devel | Logs for 2015-05-23

[10:30:24] <mozmck> should I get more input on the haltcl doc update or just push it? http://pasteboard.co/ELuHHYR.png
[11:19:52] <seb_kuzminsky> mozmck: i dont know what it used to look like, do you have a patch?
[11:21:35] <mozmck> I can make one. The difference is the text jepler and I added starting right under "and as $::AXIS_0(MAX_VELOCITY) in .tcl files for haltcl" and going to the end of the image
[11:23:30] <mozmck> Jepler explained the problem better, and I contributed the concrete example. I think my part is accurate, but not 100% sure about the "actual command that is run"
[11:25:27] <mozmck> If I print the command with puts it gives me the loadrt line that I put in this doc, but the linuxcnc error in the terminal shows it as in this doc (my first version before jepler's help): http://pasteboard.co/EFjWybv.png
[11:26:48] <mozmck> I think the "puts" version is really what is being executed, but I don't know why linuxcnc would print it out differently.
[16:30:58] <andypugh> Which is the most up-to-date JA branch?
[16:32:54] <andypugh> seb_kuzm1nsky: Do you remebe http://git.linuxcnc.org/gitweb?p=linuxcnc.git;a=commit;h=eb57c6c842adf0e7ba90db2507ae20daec257425 ?
[16:33:30] <andypugh> Well, perhaps there should have been a check to not print it every servo thread…
[16:33:42] <andypugh> http://pastebin.com/Hf3Cmprp
[16:34:11] <andypugh> However I think it might be a symptom of a 5i23 / PCI problem in this case
[16:43:14] <skunkworks> I think I was getting that error when the communication to the 7i80 wasn't quite right
[17:04:12] <seb_kuzm1nsky> that bug was fixed in b38797a7d4b, in September 2014, in 2.6 and newer
[17:19:26] <Tom_itx> was there an issue with early wheezy iso with usb?
[17:20:14] <seb_kuzm1nsky> Tom_itx: yes, on baytrail machines
[17:20:22] <seb_kuzm1nsky> it's fixed in the current image
[17:20:26] <andypugh> seb_kuzm1nsky: I should have said, I found that it had been fixed when I found where it was.
[17:20:28] <Tom_itx> ok i'll have to get it
[17:21:26] <andypugh> My Mill is running JA6 (at the time I was expecting 2.7 to be a JA version, wrote the converter script, and conveted my config
[17:23:15] <Tom_itx> i suppose there's no reason to keep the early iso cd
[17:23:24] <Tom_itx> dvd*
[17:25:24] <mozmck> andypugh: I'm pretty sure the latest JA is cradek/joints_axes7
[17:25:29] <Tom_itx> do you know the date it was fixed? i have a 2.6 iso image download dated 1/11/15
[17:25:39] <Tom_itx> or was it after that?
[17:26:16] <mozmck> seb_kuzm1nsky: any objections to my pushing the haltcl documentation update?
[17:32:28] <seb_kuzm1nsky> Tom_itx: it was early november 2014
[17:32:33] <cradek> Tom_itx: use zsync to easily get the latest iso
[17:33:00] <Tom_itx> nearly got it already
[17:33:22] <seb_kuzm1nsky> mozmck: did you post a patch and i missed it?
[17:33:39] <mozmck> no
[17:33:59] <seb_kuzminsky> "git diff | nopaste -"
[17:34:18] <cradek> andypugh: I haven't seen that warning for years, but long ago, I used to see it frequently
[17:35:02] <seb_kuzminsky> that warning happened when there were two edges in opposite directions in between two subsequent hm2_read calls
[17:35:10] <seb_kuzminsky> then the edge count would not change, but the timestamp would
[17:35:12] <seb_kuzminsky> iirc
[17:35:40] <seb_kuzminsky> time to go crawl up in the attic and look for leaks...
[17:36:13] <mozmck> seb_kuzminsky: http://pastie.org/10204220
[17:36:15] <Tom_itx> seb, where are you located?
[17:36:24] <Tom_itx> been raining here all day..
[17:37:41] <nojepler> $ git log -1 --oneline b38797a7d4bc81b0c0f9ea5e1df71e79cf770409
[17:37:41] <nojepler> b38797a hm2: fix long standing encoder velocity estimation error (thanks pcw and jepler)
[17:37:44] <nojepler> $ git describe --contains b38797a
[17:37:45] <nojepler> v2.6.4~41
[17:37:52] <nojepler> so v2.6.4 is the first one with this bugfix
[17:38:12] <andypugh> I think that in my case it was due to communications problems
[17:38:42] <andypugh> Because now I can’t start hm2_pci “input/output error”
[17:39:58] <nojepler> mostly what I recall is that I wrote a special program on an avr and drove the encoder pins of hm2 with it
[17:40:06] <nojepler> and then it reproduced consistently
[17:40:27] <cradek> you're a wizard
[17:41:02] <cradek> welp I bought two more parallel port cards for some reason
[17:41:13] <cradek> but I have screws to make before I get distracted trying them
[17:42:03] <cradek> one's a netmos (tinkerer said his works with epp_wide=0) and the other is that new? chinese chipset I don't think we have data on yet
[17:42:12] <Tom_itx> cradek did you find data on them
[17:42:18] <cradek> nope
[17:42:18] <nojepler> so the fix looks like "oh well let's not print that error message then"
[17:42:42] <andypugh> Well, it does fill the disk with kernel logs pretty quickly
[17:43:37] <andypugh> And the fix is also “don’t divide by zero if you have been given a zero to divide by”
[17:51:46] <Tom_itx> ok that didn't burn right
[17:53:34] <Tom_itx> get a grub rescue prompt
[17:58:19] <seb_kuzminsky> mozmck: that's a good improvement, it's welcome in the oldest applicable branch (2.6?)
[17:59:47] <seb_kuzminsky> Tom_itx: i'm in boulder, colorado
[17:59:49] <mozmck> seb_kuzminsky: ok. thanks to jepler for most of it. I'll try and push to 2.6 soon, have to run shortly.
[17:59:59] <seb_kuzminsky> they advertised 300 sunny days per year and a nice arid climate
[18:00:01] <seb_kuzminsky> i want my money back
[18:00:26] <mozmck> haha! We've been having rain almost every day here for a while in TX
[18:01:37] <seb_kuzminsky> in a few years we'll all be really good at retrofitting french drains to old houses
[18:03:47] <cradek> ok now I'm sidetracked fixing a touchy bug
[18:03:51] <cradek> good grieeefff
[18:04:16] <cradek> stupid free software. you can't just use it and feel powerless to fix the problems.
[18:06:28] <Tom_itx> cradek how do you run zsync?
[18:06:41] <Tom_itx> i guess i'm gonna have to burn this on the linux box
[18:07:47] <cradek> http://linuxcnc.org/docs/devel/html/getting-started/index.html#_getting_linuxcnc
[18:08:58] <Tom_itx> i want 2.7 or 2.6 image?
[18:09:05] <Tom_itx> 2.7 is fine i suppose
[18:10:36] <cradek> ii linuxcnc 1:2.6.1
[18:10:42] <cradek> heh I should probably feel bad about this
[18:11:16] <Tom_itx> where does it put the file? the default folder you're in the terminal?
[18:12:56] <seb_kuzminsky> Tom_itx: yeah, zsync downloads to the current directory
[18:14:20] <Tom_itx> any special way to burn it since it's a hybrid?
[18:14:43] <Tom_itx> nm.. reading
[18:14:52] <seb_kuzminsky> just dd it to a usb stick
[18:15:38] <cradek> Tom_itx: http://linuxcnc.org/docs/devel/html/getting-started/index.html#_getting_linuxcnc
[18:17:00] <Tom_itx> yep reading it
[18:58:39] <KGB-linuxcnc> 03Chris Radek 052.6 9850544 06linuxcnc 10src/emc/usr_intf/touchy/mdi.py Fix touchy's Set Tool/Origin defaults on lathes * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=9850544
[19:01:50] <cradek> buildbot, don't fail me now - no way am I building that on the lathe's pc
[19:02:07] <andypugh> cradek: Where does Touchy keep the defaults for “set tool”? I use a tool setting gauge, to need G10 L1 not G10 L10. (or the other way round, I just remember that I need to change it)
[19:02:43] <cradek> andypugh: on the prefs tab you can pick the tool setting method
[19:03:19] <andypugh> I should have thought of checking that.
[19:03:53] <andypugh> Users like me drive developers mad, don’t we?
[19:04:04] <cradek> haha
[19:04:33] <andypugh> Oh, one thing that I don’t know how typical it is…
[19:04:45] <cradek> actually we like you the best
[19:05:17] <andypugh> In cradek/ja7 if I open a config with no [EMC] VERSION the startup script tries to open the conversion script, and can’t find it.
[19:05:39] <cradek> weird, I thought that worked
[19:05:41] <cradek> did I bust it?
[19:05:56] <andypugh> I am not sure , it could just be me.
[19:06:42] <andypugh> I configured with prefix /usr and then make-installed. Perhpas the script gets left behind?
[19:07:43] <cradek> ooh I wonder if it gets installed/packaged right
[19:07:46] <cradek> surely we tried that...?
[19:12:48] <andypugh> Should I go out and start the PC and look or it?
[19:13:06] <andypugh> (it’s 0050 and I would rather not start looking at it now)
[19:14:16] <cradek> andypugh: sorry I can't right now, I'm already down one rabbit hole
[19:14:26] <cradek> and hungry
[19:56:17] <Tom_itx> somebody remember where to set the autologin in wheezy?
[20:02:19] <andypugh> I haven’t managed to figure it out yet.
[20:02:32] <andypugh> I tried something that didn’t work.
[20:03:33] <andypugh> If it is using lightDM then it’s here: https://wiki.debian.org/LightDM
[20:03:54] <Tom_itx> well i did it once
[20:06:52] <Tom_itx> this is using xfce
[20:09:40] <andypugh> I just realised why it didn’t work for me, I edited the examples section, not the real section.
[20:36:42] <cradek> https://www.google.com/search?q=linuxcnc+wheezy+autologin
[20:37:21] <cradek> not that it's linuxcnc-specific; that's just what I typed first and it worked
[20:39:06] <Tom_itx> i tried one that didn't
[20:39:48] <cradek> there's certainly some luck involved
[20:40:11] <Tom_itx> well i did it once and we're here again
[20:41:00] <Tom_itx> it's on the test ssd drive anyway, not too concerned
[22:00:35] <cradek> if the spindle encoder isn't plugged in, the thread doesn't cut very fast
[22:27:27] <Tom_itx> any of you per chance using a 5i24 / 6i24?
[22:27:45] <Tom_itx> making a bitfile for one and it seems alot bigger than the others
[22:28:01] <Tom_itx> maybe it's due to the 256 pins instead of 144
[22:28:24] <Tom_itx> would be kinda nice to know the file size of another one