Can anyone tell me what this is doing

Question

hey this is part of some parser code from an earlier java forum post
http://forum.java.sun.com/thread.jspa?threadID=506162&messageID=2399163#239 9163
i modified the code slightly to work with my program which involves its use in an applet and to solve logical equations. However i'm still uncertain as to what certain sections of the code are doing
i know that this code evaluates the variables within the given equation, begins a for loop that does not exit until i = codeSize which is the length of the given String. However i'm not so certain as to how it achieves this. I'd like to know
private int eval(int variableX, int variableY, int variableZ)
                try {
                        int top = 0;
                        for (int i = 0; i < codeSize; i++)
                                if (code[i] >= 0)
                                        stack[top++] = constants[code];
else if (code[i] >= POWER)
int y = stack[--top];
int x = stack[--top];
int ans = (int)Double.NaN;
switch (code[i])
case PLUS: ans = x | y; break;
case MINUS: ans = x & y; break;
case TIMES: ans = x & y; break;
case DIVIDE: ans = x / y; break;
// case POWER: ans = Math.pow(x,y); break;
if (Double.isNaN(ans))
return ans;
stack[top++] = ans;
else if (code[i] == VARIABLEX)
stack[top++] = variableX;
else if (code[i] == VARIABLEY)
stack[top++] = variableY;
else if(code[i] == VARIABLEZ)
stack[top++] = variableZ;
/* else {
double x = stack[--top];
double ans = Double.NaN;
int d;
d = (int) ans;
switch (code[i])
case SIN: ans = Math.sin(x); break;
case COS: ans = Math.cos(x); break;
case TAN: ans = Math.tan(x); break;
case COT: ans = Math.cos(x)/Math.sin(x); break;
case SEC: ans = 1.0/Math.cos(x); break;
case CSC: ans = 1.0/Math.sin(x); break;
case ARCSIN: if (Math.abs(x) <= 1.0) ans = Math.asin(x); break;
case ARCCOS: if (Math.abs(x) <= 1.0) ans = Math.acos(x); break;
case ARCTAN: ans = Math.atan(x); break;
case EXP: ans = Math.exp(x); break;
case LN: if (x > 0.0) ans = Math.log(x); break;
case LOG2: if (x > 0.0) ans = Math.log(x)/Math.log(2); break;
case LOG10: if (x > 0.0) ans = Math.log(x)/Math.log(10); break;
case ABS: ans = Math.abs(x); break;
case SQRT: if (x >= 0.0) ans = Math.sqrt(x); break;
case UNARYMINUS: ans = -x; break;
if (Double.isNaN(ans))
return d;
stack[top++] = d;
catch (Exception e)
double b = Double.NaN;
int i = (int)b;
return i;
} double d2;
if (Double.isInfinite(stack[0]))
d2 = Double.NaN;
int i2 = (int) d2;
return i2;
}else
return stack[0];
any help much appreciated
cheers
podger

Answer 1

Well this is a very basic interpreter. What this piece of code does is:
1) take an opcode from the code array, and retrieves the description for it from the constants array. This description is stored in the stack array.
stack[top++] = constants[code];
2) check what opcode it is
switch (code)
3) "execute" the opcode with simple java math operators.
case PLUS:    ans = x | y;  break;
case MINUS:   ans = x & y;  break;
case TIMES:   ans = x & y;  break;
case DIVIDE:  ans = x / y;  break;4) store the result of the opcode in the stack array, after the description of the opcode retrieved in step 1)
stack[top++] = ans;There is also some sort of error check in there: initially the result of the opcode is "NaN", (not a number). If the opcode passed to this code is not known, the result is not added to the stack according to this piece of code:
if (Double.isNaN(ans))
return ans;