#linuxcnc-devel | Logs for 2014-10-05

Back
[15:56:12] <skunkworks> well - that part was painless.. http://electronicsam.com/images/KandT/testing/Screenshot%20from%202014-10-05%2015:32:57.png
[15:56:28] <skunkworks> (I look rough I know... I am fighting shingles..)
[15:59:17] <memleak> how did you do that...
[15:59:19] <skunkworks> the laptop camera is surprisingly fast.. Hardly no lag..
[15:59:53] <skunkworks> Followed your dads directions on the wiki
[16:00:06] <memleak> i thought he only had mplayer working..
[16:00:25] <memleak> i guess im out of date as to whats going on with camview
[16:00:31] <memleak> do the plugins work for you?
[16:00:33] <skunkworks> http://wiki.linuxcnc.org/cgi-bin/wiki.pl?Adding_Camview
[16:00:55] <skunkworks> I added the repository and did a apt-get install camview-emc
[16:01:38] <memleak> camview-emc doesnt replace the entire linuxcnc package on the system?
[16:01:55] <memleak> i see you're using 2.6.3 so it must not
[16:02:31] <skunkworks> camunits-plugins-emc doesn't seem to be supported on precise though..
[16:02:31] <CaptHindsight> camview-emc is just camview with some mods to work with Linuxcnc
[16:03:17] <memleak> and its compatible with newer linuxcnc releases apparently.
[16:05:32] <skunkworks> the camunits-plugins-emc has the cool stuff like shape recognition?
[16:10:02] <CaptHindsight> http://psha.org.ru/b/camview-emc.html "List of plugins" near the bottom of the page
[16:11:28] <CaptHindsight> camview hasn't been touched in a few years, most have moved on to using OpenCV http://opencv.org/
[16:12:05] <CaptHindsight> camview/camunits https://code.google.com/p/camunits/
[16:12:48] <CaptHindsight> lest update was jan 2010
[16:14:27] <CaptHindsight> http://delta.firepick.org/ is using OpenCV but with Java, so unfortunately we can't share much
[16:15:10] <CaptHindsight> unless you want to have a Java gui for Linuxcnc
[16:26:29] <CaptHindsight> skunkworks: circle, ellipse, features and markers http://psha.org.ru/cgit/psha/cu-plugins/tag/?id=debian/0.0.37
[16:31:36] <skunkworks> CaptHindsight: did you get camunits-plugins-emc working on precise?
[16:33:05] <CaptHindsight> skunkworks: no, I gave up working with Precise after everyone jumped to Wheezy
[16:33:43] <skunkworks> oh - I would prbably like that better
[16:34:24] <cmorley> jepler: Peter says probe_parport causes error with uspace - is this a bug?
[16:34:24] <skunkworks> sucess on wheezy?
[16:35:36] <CaptHindsight> skunkworks: just getting mplayer and camview video in the an Axis tab so far, no other features yet
[16:36:17] <CaptHindsight> it probably makes more sense at this point to use OpenCV with Linuxcnc
[16:51:53] <skunkworks> I am sure... I wonder if that is something I could play with... I think that is probably above my pay grade..
[17:33:00] <tjtr33> in sim/axis/vismach/5axis joint3 isnt 'wired' to anything,
[17:33:01] <tjtr33> is it supposed to be the 2nd servo on the Y of gantry?
[17:54:50] <CaptHindsight> skunkworks: https://github.com/firepick1/FireSight it's probably better to have something like this run in Axis and hook into Linuxcnc for auto positioning
[17:57:14] <CaptHindsight> except for JSON http://en.wikipedia.org/wiki/JSON
[17:57:57] <jepler> cmorley: probe_parport is not built for uspace
[17:58:53] <jepler> cmorley: I don't know if any systems need it since we cooperate with linux parport drivers. in the days where we had people not load the kernel parport drivers, it was necessary for "ISAPNP" ports
[17:59:03] <jepler> cmorley: I don't think this describes a situation that many users deal with anymore
[18:00:00] <jepler> so in the context of pncconf/stepconf we probably should stop generating the load line for it
[18:00:21] <jepler> alternately, it would be possible to offer a dummy probe_parport for uspace I suppose
[18:01:38] <pcw_home_> maybe at least not generate it for non LPT configs
[18:01:39] <cmorley> i guess it depends if the probe is useful or not anymore... I can certainly remove it from step and pnc...
[18:01:54] <cmorley> did the 7i43 need it?
[18:02:12] <pcw_home_> probably not
[18:02:51] <cmorley> well i can remove it from master and see if anyone complains....
[18:03:14] <jepler> probe_parport is for anything that talks to the parport, when the standard kernel parport_pc is not loaded, and the port is of an ISAPNP type
[18:03:38] <jepler> but these days we ship distributions with the kernel parport_pc driver enabled
[18:04:00] <jepler> though there are still some signposts for users to follow to deactivate it if they believe it causes a problem
[18:04:22] <jepler> given that we don't hear about this much from users, I doubt it comes up. Maybe 1/4 of them would find the signposts without asking..
[18:04:36] <pcw_home_> But in any case it should only be there if a LPT port is used
[18:04:48] <jepler> pcw_home_: yes
[18:04:54] <jepler> but that would include 7i43
[18:05:02] <cmorley> at present stepconf and pnc _always_ add the line so if it's needed nobody would complain
[18:07:40] <pcw_home_> I suspect the number of 7I43 users that don't use the motherboard port is vanishing small
[18:07:54] <cmorley> forgetting to commit with -s is very very annoying :(
[18:08:34] <pcw_home_> (and the Motherboard port doesn't require probe-parport, at least no I've seen do)
[18:08:55] <pcw_home_> s/no/none/
[18:08:55] <jepler> pcw_home_: I think it's most frequently the onboard ones that were ISAPNP
[18:09:21] <pcw_home_> Must be pretty old to not have the BIOS set it up
[18:09:23] <jepler> unless you had an actual ISA add-on parport card, which you should sell as a vintage antique if you do
[18:09:32] <cmorley> well I'm for removing it and see if there are complaints.
[18:09:46] <jepler> cmorley: yay (unless there are complaints in which case boo)
[18:10:30] <jepler> if there are problems that arise out of that, the next solution is probably to restore writing it with any parport config, and put a dummy module for uspace
[18:10:49] <cmorley> ok i will remove it from master and we can go from there. thanks
[18:11:28] <jepler> thank yu
[18:12:14] <cmorley> is there any easy way to add signoff to commits? The only way I found was to un commit and try again
[18:13:15] <jepler> cmorley: just one commit? git commit --amend -CHEAD -s
[18:13:22] <cmorley> nope 7
[18:14:17] <jepler> untested: git rebase --interactive -x "git commit --amend -CHEAD -s" origin/master
[18:14:26] <cmorley> or is there away to make git always add the sign off automatically ?
[18:14:34] <cmorley> ok i'll try that
[18:14:35] <jepler> cmorley: as far as I can tell, there is not :-/
[18:14:44] <cmorley> annoying
[18:15:43] <micges-dev> cmorley: create ~/.gitconfig file
[18:16:04] <micges-dev> add those:
[18:16:28] <micges-dev> [user]
[18:16:28] <micges-dev> name = Michael Geszkiewicz
[18:16:28] <micges-dev> email = micges@wp.pl
[18:16:28] <micges-dev> [alias]
[18:16:28] <micges-dev> ci = commit -s
[18:16:58] <micges-dev> then 'git ci' will be enough
[18:17:11] <cmorley> i'll try that thanks
[18:34:31] <cmorley> ahh one can convert the commits to a patch then re-apply with git am --signoff
[18:49:00] <KGB-linuxcnc> 03Chris Morley 05master fcb9ff7 06linuxcnc 10src/emc/usr_intf/pncconf/s_motor.glade pncconf -fix GTK warning from spindle page * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=fcb9ff7
[18:49:00] <KGB-linuxcnc> 03Chris Morley 05master 1edd0f1 06linuxcnc 10src/emc/usr_intf/pncconf/build_HAL.py 10src/emc/usr_intf/pncconf/pncconf.py pncconf -don't always add probe_parport * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=1edd0f1
[18:49:01] <KGB-linuxcnc> 03Chris Morley 05master 770f59d 06linuxcnc 10src/emc/usr_intf/stepconf/build_HAL.py stepconf -rename the spindle command signal * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=770f59d
[18:49:03] <KGB-linuxcnc> 03Chris Morley 05master a4e3473 06linuxcnc 10src/emc/usr_intf/stepconf/build_HAL.py stepconf -fix the at speed comparison * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=a4e3473
[18:49:06] <KGB-linuxcnc> 03Chris Morley 05master 53903db 06linuxcnc 10src/emc/usr_intf/stepconf/build_HAL.py stepconf -change spindle-velocity signal name * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=53903db
[18:49:10] <KGB-linuxcnc> 03Chris Morley 05master f10b933 06linuxcnc 10src/emc/usr_intf/stepconf/build_HAL.py stepconf -don't use abs component if we don't have to * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=f10b933
[18:49:14] <KGB-linuxcnc> 03Chris Morley 05master 71d04c4 06linuxcnc 10src/emc/usr_intf/pncconf/private_data.py pncconf -change a dialog text from stepconf to pncconf * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=71d04c4
[18:49:18] <KGB-linuxcnc> 03Chris Morley 05master b51fe94 06linuxcnc 10src/emc/usr_intf/stepconf/build_HAL.py stepconf -remove probe_parport HAL command * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=b51fe94
[18:49:22] <KGB-linuxcnc> 03Chris Morley 05master 09ba2f1 06linuxcnc 10src/emc/usr_intf/pncconf/build_HAL.py pncconf -remove probe_parport command * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=09ba2f1
[18:49:26] <KGB-linuxcnc> 03Chris Morley 05master 498eef5 06linuxcnc 10src/emc/usr_intf/pncconf/pncconf.py pncconf -use the linux parport designation instead of address * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=498eef5
[21:31:27] <skunkworks> wow SUCCESS : firesight installed
[21:31:32] <skunkworks> that was painless.
[21:31:36] <skunkworks> how what?
[21:31:43] <skunkworks> *now what?
[21:31:45] <skunkworks> :)
[21:35:22] <tjtr33> https://github.com/firepick1/FireSight/wiki/op-HoleRecognizer looked interesting for registration
[21:37:41] <CaptHindsight> skunkworks: http://www.youtube.com/watch?v=Hzi3dY2WU0k I mentioned Firesight as an example of imaging features that could be an asset to linuxcnc for edge finding and pick-n-place
[21:39:11] <CaptHindsight> https://github.com/firepick1/FireSight
[21:41:37] <CaptHindsight> https://github.com/firepick1/FireSight/wiki
[21:43:52] <tjtr33> weird that (http://www.youtube.com/watch?v=Hzi3dY2WU0k) freezes ffox20.0 at 12 secs 3x inarow (or hangs flash i spose )
[21:45:12] <CaptHindsight> plays fine here
[21:46:00] <skunkworks> https://github.com/firepick1/FireSight
[21:46:02] <CaptHindsight> https://github.com/firepick1/FireSight/wiki/op-warpAffine locate features and rotate. But how to offset the machine to fit the parts?
[21:47:05] <CaptHindsight> you can have the camera tell you how to reposition the part to square it to the machine
[21:49:50] <skunkworks> wow - the example works.. http://electronicsam.com/images/KandT/testing/pipeline0.jpg
[21:50:33] <tjtr33> i asked about rotatekins and the advise is its not a good solution. but a troyke under an xy table could work.
[21:50:41] <skunkworks> starting with https://github.com/firepick1/FireSight/blob/master/img/cam.jpg
[21:51:28] <tjtr33> ? is the test on a jpg? not a live img?
[21:52:03] <tjtr33> impressive in any case
[21:52:40] <skunkworks> just using their jpeg
[21:53:26] <skunkworks> gives me a center for the 6 holes (that where within the tolerance..)
[21:53:50] <tjtr33> hmm the troyke solution looses work area as it translates to the new center ( so best to be sorta close , not sloppy throw it on the table )
[21:54:04] <CaptHindsight> you can use a different jpg
[21:55:39] <tjtr33> http://www.mini-lathe.com/Mini_lathe/Accessories/123_blocks.jpg
[22:18:24] <skunkworks> I would be more interested on finding the location of shapes. Say 2 circles on a printed sheet for registration.
[22:18:51] <skunkworks> If I have the 2 locations - then I rotate/move the program to fit the sheet.
[22:19:11] <skunkworks> if that makes sense
[22:19:28] <skunkworks> we have machines that do this (but are very expensive...)
[22:20:00] <skunkworks> if I could come up with a 'low rent' way of doing it...
[22:20:46] <skunkworks> that is the hard part I think - finding the 'targets' once I have a way to know the target locations in relation to the camera/tool - I am golden
[22:26:32] <tjtr33> http://www.practicalmachinist.com/vb/attachments/f38/18847d1263136238-rotary-table-sizing-advanced-20cross-20slide-20rotary-20table-20pixels.jpg
[22:26:47] <tjtr33> you turn the rotary till the holes line upo to an axis
[22:27:03] <tjtr33> ]then you move the xy till one hole is centered
[22:32:36] <tjtr33> ooh a local resource i didnt know http://www.chicagopartsnetwork.com/By-County-/DuPage-County-/Household-goods-parts-/Troyke-dmt-15-cross-slide-rotary-table-bridgeport-nice.ASPX
[23:31:14] <tjtr33> gnite