Florian Finkernagel
Florian Finkernagel
I can also offer a pretty stripped down gcode (and umc) file - got it down to 300 odd lines, with the second to last (functional) line triggering the error....
Bisection testing shows me that I can print with release 0.5.0, but I can't print with release 0.6.0. Gotta get the compilation set up to pinpoint the exact commit...
ok, so 63e5b3cf13dfce4526748cccc99721dc73979f33 fails, and 9ac3112eed854e6d06af2ad223cddc4f566c1643 works. I fail to see how the reduced acceleration settings create the issue though?!?
Possibly. I can now confirm, patching the current master with acceleration = 2000, 2000 leads to a working umc. Please find attached the invader_minimal gcode from above in both ok...
> Might be that there's a rounding error that generates an acceleration of 0 or something like that. sa_x, sa_y and sa_a (line 175, hoststepper.c) are sometimes 0 with both...
I instrumented the umc writer and it's helper. At acceleration = 1500, this is a diff of the result of Invader.gcode with the line triggering the bug and without: ```...
Disregard my earlier analysis with regard to 'it didn't happen with versions
Any ideas what I could try next? The issue appears to be rotation invariant. Neither 45 nor 15 degree rotation in Slic3r seem to make a difference.
@link0007 you can always try a test print without any filament to see if it finishes :(.
Well, this would involve listening to the new-window event (easy), and then having the layout decide where the container should go (doable), and then examining the current layout tree and...