Problem with variable_bed_surfaces

[appleimperio]

I recently add a klicky probe now every time I enable the variable_bed_surfaces. I'm getting this error: Error evaluating 'gcode_macro set_surface_active:gcode': TypeError: a float is required

[Variables]
bed_surfaces = {'active': u'smooth', 'available': {u'smooth': {'offset': '0'}, u'texture': {'offset': 0.0}}}

[appleimperio]

I found that one of the klicky probe macros was resetting the gcode offset to 0 which i think is the command you use to set the bed surface offset. The strange thing it is saving the gcode state before set it to 0 but when it restored it i get the error it completely erase the values on the variable file. For now I just comment that line and its working (hoping it don't crash when its doing the docking) if you have a better solution let me know thanks.

[gcode_macro _entry_point]
gcode:
    {% set function  = 'pre_' ~ params.FUNCTION %}
    {% set move_accel = printer["gcode_macro _User_Variables"].move_accel|default(1000) %}
    # mandatory to save the new safe position
    M400
    SAVE_GCODE_STATE NAME={function}
    # removes the Z offset for better bed based docking
    #SET_GCODE_OFFSET Z=0
    # all the macros initially assume absolute positioning
    G90
    # set a safe(sane) Acceleration
    SET_VELOCITY_LIMIT ACCEL={move_accel}

[jschuh]

Thanks for the report. It looks like I need to make my SET_GCODE_OFFSET wrapper ensure it's converting to the correct data type prior to saving.

That stated, I don't think that klicky macro should be blindly the clearing the gcode offset. I get what they're trying to do, but the right approach would be to perform the docking operation accounting for the gcode offset. Or, I suppose save, the offset and restore it when done (although that could still clear it if something aborts the operation at the wrong time).

Is there a github for the klicky probe? I might see about writing a PR.

[jschuh]

Should be fixed so that it won't save the incorrect type in the future (and thus won't have this issue again).

However, you may have to edit your save variables file and change {'offset': '0'} to {'offset': 0.0}.

[appleimperio]

Thank You! here is the github for the klicky probe https://github.com/jlas1/Klicky-Probe

[jschuh]

I thought about it a bit more and I think klicky is doing the right thing here. So, I added code to track when SET_GCODE_OFFSET is used in an ephemeral context (i.e. when SAVE_GCODE_STATE has been called without a matching call to RESTORE_GCODE_STATE). In that case the macro simply doesn't persist the offset change into the saved variables.

You should be able to uncomment that line now, and everything should work as expected.