OMCompiler doesn't remove mathematical signs when flattening model + libraries

[anonymous]

Using OMCompiler compiled from master, I'm trying to generate a flat modelica model through combination of the two following files:

-Ball_Platform_im.mo:

model Ball_Platform_im
  Ball_Platform 	Ball_Platform1;
  OutPutPort 	OutPutPort1;
  OutPutPort 	OutPutPort2;
equation
  Ball_Platform1.y2 = OutPutPort1.vi;
  Ball_Platform1.y1 = OutPutPort2.vi;
end Ball_Platform_im;

-Ball_Platform.mo:

class Ball_Platform
 parameter Real g=9.8; 
 parameter Real m1=0.50; //platformKg 
 parameter Real m2=0.30;//Kg
 parameter Real k=2; //Kg/sec
 Real y1(start=11),v1(start=0);//Platform
 Real y2(start=15),v2(start=1);//ball
 Real y0;

 discrete Real v1p,v2p;
equation
 y0=10; //Nominal position
//-----------------------------------
 der(y1)=v1;
 m1*der(v1)=if noEvent(v1<0.001 and v1>-0.001) then 0
 else -m1*g-k*(y1-y0)-0.2*v1;
 der(y2)=v2;
 der(v2)=if noEvent(v2<0.001 and v2>-0.001) then 0
 else -g;

//-----------------------------------
 when y2<y1 then 
 v1p=(m1*v1+2*m2*v2-m2*v1)/(m1+m2);
 v2p=(m2*v2+2*m1*v1-m1*v2)/(m1+m2);
 reinit(v1,v1p*0.98);
 reinit(v2,v2p*0.98);
 end when;
//-----------------------------------
end Ball_Platform;

using the following shell command:

$ omc /tmp/SCI_TMP_1358_LuFlNI/Ball_Platform_im.mo /tmp/SCI_TMP_1358_LuFlNI/modelica/Ball_Platform.mo --modelicaOutput --useLocalDirection --reduceTerms

which runs fine, but produces a line with side to side mathematical signs (in line under comment: "+ -0.2") that prevents further compilation.

class Ball_Platform_im
  parameter Real Ball_Platform1__g = 9.800000000000001;
  parameter Real Ball_Platform1__m1 = 0.5;
  parameter Real Ball_Platform1__m2 = 0.3;
  parameter Real Ball_Platform1__k = 2.0;
  Real Ball_Platform1__y1(start = 11.0);
  Real Ball_Platform1__v1(start = 0.0);
  Real Ball_Platform1__y2(start = 15.0);
  Real Ball_Platform1__v2(start = 1.0);
  Real Ball_Platform1__y0;
  discrete Real Ball_Platform1__v1p;
  discrete Real Ball_Platform1__v2p;
  output Real OutPutPort1__vo;
  Real OutPutPort1__vi;
  output Real OutPutPort2__vo;
  Real OutPutPort2__vi;
equation
  Ball_Platform1__y0 = 10.0;
  der(Ball_Platform1__y1) = Ball_Platform1__v1;
  // Minus sign following plus sign and not isolated by parentheses
  Ball_Platform1__m1 * der(Ball_Platform1__v1) = if noEvent(Ball_Platform1__v1 < 0.001) and noEvent(Ball_Platform1__v1 > -0.001) then 0.0 else Ball_Platform1__k * (Ball_Platform1__y0 - Ball_Platform1__y1) + -0.2 * Ball_Platform1__v1 - Ball_Platform1__m1 * Ball_Platform1__g;
  der(Ball_Platform1__y2) = Ball_Platform1__v2;
  der(Ball_Platform1__v2) = if noEvent(Ball_Platform1__v2 < 0.001) and noEvent(Ball_Platform1__v2 > -0.001) then 0.0 else -Ball_Platform1__g;
  when Ball_Platform1__y2 < Ball_Platform1__y1 then
    Ball_Platform1__v1p = (Ball_Platform1__m1 * Ball_Platform1__v1 + Ball_Platform1__m2 * (2.0 * Ball_Platform1__v2 - Ball_Platform1__v1)) / (Ball_Platform1__m1 + Ball_Platform1__m2);
    Ball_Platform1__v2p = (Ball_Platform1__m2 * Ball_Platform1__v2 + Ball_Platform1__m1 * (2.0 * Ball_Platform1__v1 - Ball_Platform1__v2)) / (Ball_Platform1__m1 + Ball_Platform1__m2);
    reinit(Ball_Platform1__v1, 0.98 * Ball_Platform1__v1p);
    reinit(Ball_Platform1__v2, 0.98 * Ball_Platform1__v2p);
  end when;
  OutPutPort1__vi = OutPutPort1__vo;
  OutPutPort2__vi = OutPutPort2__vo;
  Ball_Platform1__y2 = OutPutPort1__vi;
  Ball_Platform1__y1 = OutPutPort2__vi;
end Ball_Platform_im;

It's a simple thing, but as the flattening should be part of an automated process for generation of FMI2 libraries, modifying it each time by hand wouldn't be feasible.

Thanks in advance.

[Martin Sjölund]

Fixed in ​https://github.com/OpenModelica/OMCompiler/pull/2096