Re: [Emc-developers] reports on rt-preempt / latency-test: max jitter going down - FIXED

2014-01-27 Thread EBo
On Jan 26 2014 8:22 PM, Jeff Epler wrote: On Sun, Jan 26, 2014 at 09:07:56PM +0100, Michael Haberler wrote: I have a fix which works for me: https://github.com/mhaberler/linuxcnc/commit/10924b5509a2d2e38533a9f9e876d139a2cf5d4b (branch unified-build-candidate-3) Good catch. That it may

Re: [Emc-developers] ngcgui cutter path compensation

2014-01-27 Thread Dewey Garrett
A simplified example (using no subroutines) shows the type of failure that bruno reported. g42.1 is used so no specific tool is needed to show the fault. The failing test is in interp_convert.cc: CHKS((fabs(cos(A_ang)) TOLERANCE_EQUAL), NCE_TOOL_RADIUS_NOT_LESS_THAN_ARC_RADIUS_WITH_COMP);

Re: [Emc-developers] reports on rt-preempt / latency-test: max jitter going down - FIXED

2014-01-27 Thread Jeff Epler
On Mon, Jan 27, 2014 at 06:04:06AM -0700, EBo wrote: Hmmm... it might be a good idea to write a filter and search for large constants and review them not only for simplifying, but also for appropriate size. It's worth noting that gcc 4.7 compiles both of these to the same code when -O or

Re: [Emc-developers] ngcgui cutter path compensation

2014-01-27 Thread andy pugh
On 27 January 2014 15:42, Dewey Garrett dgarr...@panix.com wrote: CHKS((fabs(cos(A_ang)) TOLERANCE_EQUAL), NCE_TOOL_RADIUS_NOT_LESS_THAN_ARC_RADIUS_WITH_COMP); with TOLERANCE_EQUAL == .0001 in interp_internal.hh My initial thought is that it isn't necessary to use a tolerance for a less

Re: [Emc-developers] ubc threads.0 test failure

2014-01-27 Thread Sebastian Kuzminsky
Thanks for pushing that fix. But Michael (and everyone), please stop adding new features to ubc. I'm thinking specifically of Jeff's lineardeltakins that were included in the threads.0 bugfix push. Let's all focus on getting that branch mergable, without adding more stuff to it. On

Re: [Emc-developers] ngcgui cutter path compensation

2014-01-27 Thread EBo
On Jan 27 2014 9:18 AM, andy pugh wrote: On 27 January 2014 15:42, Dewey Garrett dgarr...@panix.com wrote: CHKS((fabs(cos(A_ang)) TOLERANCE_EQUAL), NCE_TOOL_RADIUS_NOT_LESS_THAN_ARC_RADIUS_WITH_COMP); with TOLERANCE_EQUAL == .0001 in interp_internal.hh My initial thought is that it isn't

Re: [Emc-developers] Emc-developers Digest, Vol 93, Issue 47

2014-01-27 Thread bruno
Glad you can reproduce the error ! Obviously it is important to fix linuxcnc, but do you think the gcode generated by qpocket is also a little too close to an edge case (programatically and also geometrically :-) ) I mean in the example I gave, if you disable path compensation, the entry

[Emc-developers] web site critique

2014-01-27 Thread Chris Morley
To those with linuxcnc web site control: http://linuxcnc.org/index.php/english/forum/32-documents/27447-source-code-access he surely has some good points. Chris M --

[Emc-developers] house tomp

2014-01-27 Thread TJoseph Powderly
Dennis Thanks for stopping by, sorry I was grumpy, no good excuse. Please dont worry about the house maybe drive by and see if its still there :) I'm in Ky now, Mon night got one up and running but I'm not happy its too sensitive to flush pressure other one is running too its problem is, IF it