#linuxcnc-devel Logs

Aug 14 2019

02:55 PM mozmck: I'm not making much headway on the move-after-abort bug (#579)
03:00 PM mozmck: I'm pretty sure I don't fully understand the sequence of events in task/interp. Here is the relevant debug output to stdout: https://pastebin.com/3uy0WYR3
03:03 PM JT-Shop: wow if you don't understand I sure won't but I'll look anyway
03:07 PM mozmck: I'm wondering if it might have to do with the naive cam detector. There seems to be a final linear move that is not getting cleared out on an ABORT, and it is getting added to the interplist.
03:08 PM mozmck: It is before the first command in the on_abort subroutine.
03:11 PM mozmck: Have to run for now - bbl
03:34 PM andypugh: mozmck: If you are there, have you tried chatting to rellenberg?
03:34 PM mozmck: Hi andypugh, no I thought about it though.
03:35 PM mozmck: I should email him and see if maybe he can give me some pointers. Right now I'm trying to figure out how to disable the naive cam detector for testing.
03:39 PM JT-Shop: doesn't G61, G61.1 turn off the naive cam detector?
03:39 PM JT-Shop: G64 is just blending and the naive
03:39 PM JT-Shop: cam detector is not enabled.
03:40 PM JT-Shop: G64 P- Q- - (Blend With Tolerance Mode) This enables the naive cam detector and enables blending with a tolerance.
03:43 PM mozmck: Yes, just found and tried G61
03:44 PM mozmck: It does affect where the code jumps to, but does not fix the problem.
03:46 PM mozmck: So I'm back to the readahead. It is executing the last line that readahead read. But it does it before it executes the code in the on_abort subroutine.
05:32 PM KimK: mozmck: Hang in there, it sounds like you might be closing in on something significant.
05:32 PM mozmck: thanks!
05:33 PM mozmck: Seems like I'm close. I emailed Rob and he gave me a couple of pointers just now so maybe soon here...
05:34 PM jthornton: cool, good that he can help