#linuxcnc-devel | Logs for 2016-12-09

[09:49:00] <skunkworks> Is G72/72.1 working with andys last push? I can'
[09:49:06] <skunkworks> can't seem to make it work
[14:56:00] <skunkworks_> zlog
[14:56:22] <skunkworks_> andypugh, is g72/72.1 working in your last push?
[14:56:59] <andypugh> skunkworks_: Probably
[14:57:11] <andypugh> I only really checked G71 / G71.1
[14:57:39] <skunkworks_> I was just taking one of my programs and switching it to G72 and I get an error - let me post the error
[14:58:02] <andypugh> On reflection, gouge detection will be wrong in G72 G72.1
[15:00:33] <skunkworks_> andypugh, http://pastebin.com/vy1GxLvh
[15:01:22] <skunkworks_> http://pastebin.com/6LuSUxcg
[15:17:17] <andypugh> You’re right. I have broken it.
[15:17:51] <skunkworks_> that is good - normally I am the one doing the breaking...
[15:50:50] <KGB-linuxcnc> 03andypugh 05andypugh/g71type2remap 106f730 06linuxcnc 10configs/sim/axis/g71/python/remap.py G71 / G72 Fix a problem with max_depth searching which was preventing G72 from working * 14http://git.linuxcnc.org/?p=linuxcnc.git;a=commitdiff;h=106f730
[15:52:15] <andypugh> skunkworks_: That makes the G72 test file work, but your file has a profile that starts at the wrong end.
[15:52:52] <Tom_L> conventional vs climb?
[15:53:06] <andypugh> (If you start at Z-2.4 and rapid to 0.2 as your first move in the block, then you get a left-hand G72 with it)
[15:53:12] <Tom_L> oh wait.. these are lathe routines..
[16:02:33] <skunkworks_> andypugh, oh.. I might have been confused. I thought G72 was clearing paths perpendicular to the part
[16:07:30] <andypugh> It feeds in X and increments along Z
[16:08:28] <andypugh> But the first move defines the line that every cut starts from. And then things go wrong if you can’t get from the end of that line to the start of your profile.
[16:12:59] <skunkworks_> andypugh, awesome.. Have to play with it to understand - but it is working
[16:13:48] <skunkworks_> I am not understanding why it didn't initally clear right to left? (with the original program set to g72)
[16:14:01] <andypugh> skunkworks_: G0z-2.4x1
[16:14:02] <andypugh> (Bottom to top, right to left)
[16:14:03] <andypugh> G72 P100 D.02 F900 J0 L0 I0 K0 R0.01
[16:14:05] <andypugh> O0100 SUB
[16:14:06] <andypugh> G0 Z0.2
[16:14:07] <andypugh> G0 X0 Z0.2
[16:14:08] <andypugh> G1 x0 Z0
[16:14:47] <andypugh> skunkworks_: Because it didn’t know where to stop in the Z increments.
[16:15:18] <andypugh> It’s a standard. Not a very good one.
[16:18:17] <andypugh> But, I need to put in a halpful error message.
[16:18:50] <linuxcnc-build> build #2799 of 1400.rip-wheezy-i386 is complete: Failure [4failed compile runtests] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/1400.rip-wheezy-i386/builds/2799 blamelist: andypugh <andy@bodgesoc.org>
[16:20:35] <skunkworks_> it is really amazing. this plus native cam - you could do lath ops pretty quick. (it has a poly line feature also)
[16:20:41] <skunkworks_> lathe
[16:22:23] <bpuk> step 1: G71/G72. Step 2: G70. Step 3: G73 - that covers most major lathe stuff
[16:23:03] <bpuk> G74/G75 would also be handy
[16:23:28] <skunkworks_> andypugh, it does go into the part...
[16:24:25] <andypugh> skunkworks_: It’s a faeature. :-)
[16:24:34] <andypugh> Picture?
[16:25:19] <skunkworks_> hold on
[16:26:11] <skunkworks_> http://electronicsam.com/images/KandT/testing/Screenshot-34.png
[16:26:34] <skunkworks_> this is clearing from left to right
[16:32:28] <linuxcnc-build> build #4655 of 0000.checkin is complete: Failure [4failed] Build details are at http://buildbot.linuxcnc.org/buildbot/builders/0000.checkin/builds/4655 blamelist: andypugh <andy@bodgesoc.org>
[16:32:37] <andypugh> What’s your R word?
[16:41:08] <andypugh> Ignore that, it isn’t the retract, is it?
[16:56:13] <skunkworks_> G72 P100 D.02 F900 J0 L0 I0 K0 R0.01
[16:57:48] <skunkworks_> andypugh, it just looks like the retract is going the wrong direction for some of the profile
[16:57:50] <skunkworks_> http://electronicsam.com/images/KandT/testing/Screenshot-35.png
[16:58:40] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15andypugh commented on issue #217: I don't really understand remap and/or stdglue.py but I have a feeling that INTERP_EXECUTE_FINISH is intended to cause a "synch" after a tool change, ie make sure that the readeahed queue is emptied and recalculated with new tool data. ... 02https://github.com/LinuxCNC/linuxcnc/issues/217#issuecomment-266140953
[16:59:11] <andypugh> I don’t think that Type11 G71 even uses the R-word.
[16:59:45] <andypugh> But you are right, my exit-finding code is broken.
[17:03:11] <andypugh> I had not considered that somebody would use G71 on a profile where the line segments are short compared to the cut depth.
[18:51:10] <andypugh> Frustrating. I have a Python problem with the G71 now. Something is not re-initialising properly. I have spoofed a C-strict as a class in the Python, with no initialiser. The effect is that the code works once, then re-loading the same file fails. This would notbe an issue in C or C++, but I need to fix it to carry on with the real task.
[20:51:35] <andypugh> Help! does anyone speak parseltongue?
[20:55:16] <andypugh> This class (modelling a C struct) turns out to be annoyingly persistent between invocations. How do I fix that? I suspect it is actually because it is global in some way. Definition here:
[20:55:17] <andypugh> https://github.com/LinuxCNC/linuxcnc/blob/andypugh/g71type2remap/configs/sim/axis/g71/python/remap.py#L18
[20:56:56] <andypugh> Because this code is intended to transfer to C++ I have been unpythonic. it seems that scope and “static” are different in Python.
[22:36:37] <jepler> andypugh: assign all class attributes in __init__. Putting them in the class like that does something that is well-defined in Python but is surprising to many programmers anyway
[23:05:55] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15dhweizel commented on issue #217: If I undertand correctly the error is related to the re execution of the epilog.... 02https://github.com/LinuxCNC/linuxcnc/issues/217#issuecomment-266178554
[23:39:50] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15dhweizel commented on issue #217: If I undertand correctly the error is related to the re execution of the epilog.... 02https://github.com/LinuxCNC/linuxcnc/issues/217#issuecomment-266178554
[23:55:56] -linuxcnc-github:#linuxcnc-devel- [13linuxcnc] 15zultron commented on issue #217: @dhweizel Thanks for reporting this. I'm going to have a little trouble finding time to look at this before January, but I'll try. I'm very sorry you must have had to look at this extremely hairy part of the codebase! 02https://github.com/LinuxCNC/linuxcnc/issues/217#issuecomment-266180442