unable to find a register to spill in class 'POINT

Forum 2005-2010 (read only) Software Syntax & Programs
1

Hey Guys,
I'm working on a gcode interpreter for the arduino and also making a library so it can be a controller for a 3axis cnc machine. I have some code and am getting an odd error that I've never seen. Here's the error message I'm getting.

CNC3AXIS.cpp: In member function 'void CNC3AXIS::moveToPosUnits(float, float, float)':
CNC3AXIS.cpp:276: error: unable to find a register to spill in class 'POINTER_REGS'
CNC3AXIS.cpp:276: error: this is the insn:
(insn 29 27 30 0 CNC3AXIS.cpp:268 (set (mem/s:SF (post_inc:HI (reg/f:HI 10 r10 [orig:41 this ] [41])) [4 <variable>.y_move_to_pos_units+0 S4 A8])
        (reg/v:SF 43 [ ypos ])) 15 {*movsf} (insn_list:REG_DEP_TRUE 5 (insn_list:REG_DEP_TRUE 28 (nil)))
    (expr_list:REG_INC (reg/f:HI 10 r10 [orig:41 this ] [41])
        (nil)))
CNC3AXIS.cpp:276: confused by earlier errors, bailing out

Now I've found a line that if I comment i won't get that error anymore but it's not making sense to me why its happening.

void CNC3AXIS::moveToPosUnits( float xpos, float ypos, float zpos )
{
    x_move_to_pos_units = xpos;
    x_move_to_pos_steps = (long)(steps_per_unit * xpos);
    y_move_to_pos_units = ypos;
    y_move_to_pos_steps = (long)(steps_per_unit * ypos);
    z_move_to_pos_units = zpos;
    z_move_to_pos_steps = (long)(steps_per_unit * zpos);

    moveUnits( (x_move_to_pos_units - x_cur_pos_units),
               (y_move_to_pos_units - y_cur_pos_units),
               (z_move_to_pos_units - z_cur_pos_units) );
}

It's this line that causes the error to show up. If I comment it there's no problem.

x_move_to_pos_steps = (long)(steps_per_unit * xpos);

Thanks for any help!

Blair

2

I wonder if you have a control character in your file, or there is a problem in code you have not included. The code below compiled ok for me ( I had to guess at some of the types)

float x_move_to_pos_units;
float y_move_to_pos_units;
float z_move_to_pos_units;

long x_move_to_pos_steps;
long y_move_to_pos_steps;
long z_move_to_pos_steps;

float x_cur_pos_units;
float y_cur_pos_units;
float z_cur_pos_units;

long steps_per_unit = 1;

void  moveUnits( float xpos, float ypos, float zpos ){
  
}
               
void moveToPosUnits( float xpos, float ypos, float zpos )
{
    x_move_to_pos_units = xpos;
    x_move_to_pos_steps = (long)(steps_per_unit * xpos);
    y_move_to_pos_units = ypos;
    y_move_to_pos_steps = (long)(steps_per_unit * ypos);
    z_move_to_pos_units = zpos;
    z_move_to_pos_steps = (long)(steps_per_unit * zpos);

    moveUnits( (x_move_to_pos_units - x_cur_pos_units),
               (y_move_to_pos_units - y_cur_pos_units),
               (z_move_to_pos_units - z_cur_pos_units) );
}
3

Its really quite odd because I have another function that is almost identical same variables and everything. The only difference is instead of multiplication its division and it has no problems.

4

You must have been right about the special character thing. I just rearranged the code and I'm not have any error messages now. Is there a way to see those special characters?

Heres the code now and it works fine.

void CNC3AXIS::moveToPosUnits( float xpos, float ypos, float zpos )
{
    x_move_to_pos_units = xpos;
    y_move_to_pos_units = ypos;
    z_move_to_pos_units = zpos;
    x_move_to_pos_steps = (long)(steps_per_unit * xpos);
    y_move_to_pos_steps = (long)(steps_per_unit * ypos);
    z_move_to_pos_steps = (long)(steps_per_unit * zpos);

    moveUnits( (x_move_to_pos_units - x_cur_pos_units),
               (y_move_to_pos_units - y_cur_pos_units),
                   (z_move_to_pos_units - z_cur_pos_units) );
}
5

Good to hear you got it going.

Sometimes a weird error message like “unable to find a register to spill in class 'POINTER_REGS'
Is a good indication that the compiler has gotten particularly confused :wink:

I have resorted to a hex editor once or twice when I have had a similar problem that I couldn't track down.

Speaking of weird error messages, my favorite error message of all time was:
“The world is coming to an end on device number 3” when I was programming a Datapoint 2200 back in the 70's. That was the only time I ever started laughing when discovering a bug in my code.