#linuxcnc-devel | Logs for 2016-07-26

[08:10:35] <KGB-linuxcnc> 03John Thornton 052.7 3402849 06linuxcnc 10docs/src/getting-started/getting-linuxcnc.txt Docs: remove a cut and paste error * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3402849
[09:42:58] <tinkerer> hello experts!
[09:45:30] <tinkerer> Question about adding changes to a closed PR:
[09:52:39] <tinkerer> I've found a forgotten change on my raspian image and I would like to hand in later this change. How can I do this?
[09:55:23] <jepler> tinkerer: open a new PR
[09:56:07] <tinkerer> OK, and may I use the same branch?
[09:58:10] <jepler> I'm not sure, but here's how you can check: When creating the PR, look at how many commits it says the pull request includes
[09:58:32] <jepler> if it shows that the pull request will include only the new commit then it's OK
[09:58:53] <jepler> if not, you should create a new branch starting at the linuxcnc master branch, cherry-pick your commit, push to github, and use that branch as the basis of the PR
[09:59:24] <tinkerer> ok thanks!
[11:22:09] <mozmck> I presume I can use PROGRAM_EXTENSION = .ngc and run standard ngc files through a filter program?
[11:23:00] <jepler> mozmck: dunno!
[11:23:14] <mozmck> guess I'll just have to try it!
[11:23:58] <mozmck> I want to make a filter program to automatically update some things in gcode files.
[11:24:43] <mozmck> One thing for instance, we use M0 right now, and I'd like to make it change that to M1 without users having to re-make all of their files.
[12:08:42] <tinkerer> jepler: very strange, the commit 7875ed8 is in the last pull request and in the new one. but just this commit. ????
[12:17:20] <jepler> you may not have compared to the correct branch? I see one commit when I compare your hm2_raspi branch to our master branch. https://github.com/LinuxCNC/linuxcnc/compare/master...tinkercnc:hm2_raspi
[12:17:29] <jepler> but when I first arrived at that page, it compared your branch to our hm2_raspi branch instead
[12:17:52] <KGB-linuxcnc> 05hm2_raspi c5bcd08 06linuxcnc 04. branch deleted * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=c5bcd08
[12:18:12] <jepler> .. a branch I have now deleted, since seb_kuzminsky created it just to test the branch with our CI system.
[12:21:51] <tinkerer> ???? =)
[12:30:00] <tinkerer> and when will github recognize the changes?
[12:30:20] <tinkerer> ah, now :)
[12:30:22] <tinkerer> thanks
[12:32:07] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15tinkercnc opened pull request #124: Raspian-8.x Distro added... (06master...06hm2_raspi) 02https://github.com/LinuxCNC/linuxcnc/pull/124
[12:35:24] <KGB-linuxcnc> 03Jeff Epler 05master 3100f6e 06linuxcnc Merge branch 'hm2_raspi' of https://github.com/tinkercnc/linuxcnc-mirror-old * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=3100f6e
[12:35:24] <jepler> thank you
[12:35:34] <jepler> tinkerer: thank you
[12:36:00] <tinkerer> *bow
[12:39:25] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15cradek closed pull request #124: Raspian-8.x Distro added... (06master...06hm2_raspi) 02https://github.com/LinuxCNC/linuxcnc/pull/124
[13:15:01] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master 178059c 06linuxcnc Merge branch 'master' of ssh://norbert@git.linuxcnc.org/git/linuxcnc.git into gmoccapy_JA_based_on_master * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=178059c
[13:15:01] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master 03e7b27 06linuxcnc 10configs/sim/gmoccapy/gmoccapy-JA/table-rotary-tilting/sim-xyzac-trt.pref 10src/emc/usr_intf/gmoccapy/gmoccapy.py gmoccapy_JA_master - trying to find halmdihistory bug * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=03e7b27
[13:15:01] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master d56b31a 06linuxcnc 10lib/python/gladevcp/hal_mdihistory.py 10lib/python/hal_glib.py gmoccapy_JA_master - solved the MDI History bug * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=d56b31a
[13:22:42] <KGB-linuxcnc> 03Norbert Schechner 05gmoccapy_JA_based_on_master de8dae3 06linuxcnc 10lib/python/gladevcp/hal_mdihistory.py GladeVCP - hal_mdi_history - deleted unneded function * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=de8dae3
[14:18:51] <tinkerer> That was a close call! Unicorn! ยท GitHub...
[21:02:26] <jepler> nobody seems to have raised any concerns about uspace-plus. I think I'll merge it soon, even though I don't think anybody's using it.
[21:07:08] <jepler> the main possibility of problems for good old fashioned preempt-rt is something regarding the new way messages are exfiltrated from realtime
[21:10:18] <seb_kuzminsky> i haven't looked at it
[21:10:28] <seb_kuzminsky> been spending my time on issue 117 :-/
[21:11:00] <seb_kuzminsky> and regretting my Task changes in 2.6 and 2.7
[21:13:47] <jepler> :-/ indeed
[21:14:27] <jepler> is there a way I could be helpful wrt 117?
[21:19:27] <skunkworks_> http://dilbert.com/strip/1995-11-13
[21:31:51] <seb_kuzminsky> jepler: you could talk me into reverting those changes in 2.6 & 2.7
[21:33:27] <jepler> skunkworks_: I suspected the punchline was gonna be "minivan"
[21:44:44] <KGB-linuxcnc> 03Jeff Epler 05master fdc7c25 06linuxcnc 10docs/src/config/python-interface.txt docs: don't give numeric values for enumerated constants * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=fdc7c25
[21:44:45] <KGB-linuxcnc> 03Jeff Epler 05master de26234 06linuxcnc 10docs/src/config/python-interface.txt docs: update linuxcnc module docs for JA * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=de26234
[21:44:46] <jepler> seb_kuzminsky: that is sure a possibility.
[21:45:49] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15jepler closed issue #121: documentation of linuxcnc.stat.attributes does not reflect the JA changes 02https://github.com/LinuxCNC/linuxcnc/issues/121
[21:46:11] <jepler> sigh I went testing image2gcode on my jessie amd64 system and with or without my branch fixing it for new numpy, it crashes in a way I don't get
[21:46:19] <jepler> maybe pertaining to transferring the image from PIL to Tk, not sure
[21:48:17] <jepler> I think the Python Imaging Library that it uses is deprecated; the replacement is called "pillow" :-/ and they promise that you just change your import and then your software works again forever
[22:05:45] <seb_kuzminsky> this un-breaks #117: http://paste.ubuntu.com/21093288/
[22:06:17] <seb_kuzminsky> there's something i still don't understand
[22:06:51] <seb_kuzminsky> my runtests-test that tries to reproduce the axis situation jepler's and JT-Shop's config demonstrates, runs without any problems
[22:08:41] <seb_kuzminsky> the per-axis error checking code in taskintf.cc's emcMotionUpdate() is super squint-worthy
[22:08:50] <seb_kuzminsky> bbl
[22:21:13] <tinkerer> hmmm? "Missing <max linear speed> specifier"
[22:23:12] <tinkerer> and: Error in startup script: can't read "env(LINUXCNCVERSION)": no such element in array
[22:23:39] <tinkerer> are this JA issues?
[22:24:04] <seb_kuzminsky> tinkerer: not one i've seen
[22:31:17] <tinkerer> what is the "max linear speed" specifier?
[22:31:50] <tinkerer> MAX_VELOCITY?
[22:31:55] <seb_kuzminsky> probably
[22:32:43] <seb_kuzminsky> [DISPLAY]MAX_LINEAR_VELOCITY maybe?
[22:33:00] <seb_kuzminsky> i'd say, go look at whatever script printed that error, and see what it's really talking about
[22:46:26] <tinkerer> ok: http://linuxcnc.org/docs/devel/html/getting-started/updating-linuxcnc.html#_traj_velocities_accelerations_names
[22:48:36] <seb_kuzminsky> dammit why does motion have both joint->on_pos_limit and (joint->flag & EMCMOT_JOINT_MAX_HARDLIMIT)????!!?
[22:50:18] <seb_kuzminsky> i feel like a LibreSSL-style refactor is in order here
[22:50:46] <mozmck> Is one a hal pin?
[22:51:10] <mozmck> hmm, no
[22:51:20] <archivist> is one the soft limit the other the switch
[22:51:46] <seb_kuzminsky> no, these are the hard limits
[22:58:36] <mozmck> where do you see joint->flag & EMCMOT_JOINT_MAX_HARDLIMIT?
[23:01:58] <seb_kuzminsky> mozmck: it's in a macro in mot_priv.h
[23:03:03] <mozmck> You mean EMCMOT_JOINT_MAX_HARD_LIMIT_BIT ?
[23:03:25] <seb_kuzminsky> yeah, sorry
[23:04:44] <mozmck> Well, looks like joint->on_pos_limit is set from that macro in process_inputs(), and only used in homing
[23:05:13] <mozmck> but I wouldn't know why they didn't just use GET_JOINT_PHL_FLAG()
[23:05:33] <seb_kuzminsky> or just use the struct field
[23:05:42] <seb_kuzminsky> pick one or the other, folks!
[23:05:57] <seb_kuzminsky> i haven't git blamed it yet so i dont know what's up
[23:06:01] <mozmck> struct field?
[23:06:05] <seb_kuzminsky> maybe it made sense at the time
[23:06:18] <seb_kuzminsky> the on_pos_limit field of the joint status struct
[23:07:30] <mozmck> my guess is that the flag bit-field came first because it has all the macros which are used all over the place.
[23:07:51] <seb_kuzminsky> good point
[23:07:55] <seb_kuzminsky> i bet you're right
[23:09:21] <mozmck> maybe someone added the struct fields because it would be less cpu cycles when doing homing ;-)
[23:16:54] <mozmck> seb_kuzminsky: what is wrong with my python regex to search for o<touchoff> strings? re.compile(r'\b[Oo]<touchoff>\b')
[23:18:57] <seb_kuzminsky> my first guess is that < and > have special meaning
[23:19:59] <seb_kuzminsky> hmm, not that i can see
[23:21:48] <mozmck> Yeah, that's what I thought. I just found this site: http://pythex.org/
[23:22:04] <mozmck> and playing with it I found if I leave of the last \b it works
[23:22:37] <seb_kuzminsky> \b doesn't match end-of-line?
[23:22:43] <mozmck> But I wonder if that will find sub-strings?
[23:22:54] <seb_kuzminsky> no, that's not it
[23:23:18] <mozmck> I don't know, but the line will usually be: o<touchoff> call
[23:23:21] <seb_kuzminsky> maybe the ">" doesn't count as a word, so the trailing \b doesn't match
[23:23:22] <mozmck> and maybe a comment after
[23:24:27] <mozmck> It looks like removing the \b will find substrings which I don't want.
[23:25:12] <mozmck> so you're saying the ">" may be considered punctuation or something?
[23:26:10] <seb_kuzminsky> that is what i said, but now i think i was wrong about it
[23:26:21] <seb_kuzminsky> sorry i'm not more help :-/
[23:26:55] <mozmck> heh, no problem. I think I can work with what I have now.
[23:41:03] <KGB-linuxcnc> 03Dewey Garrett 05master 5f6b37c 06linuxcnc 10tcl/bin/emccalib.tcl emccalib.tcl parsing [HAL]HALFILE,POSTGUI_HALFILE * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=5f6b37c