. operator ?

[sorex]

Hello,

Can anyone tell me what the . (dot) operator was doing in an old TP/BP program or even in freepascal?

like

b:=21167.0;

[Phil]

Hello,

Can anyone tell me what the . (dot) operator was doing in an old TP/BP program or even in freepascal?

like

b:=21167.0;

That's a decimal point, not an operator. Perhaps you're thinking of the string concatenation dot operator in PHP.

Thanks.

-Phil

[sorex]

no it's not, when you remove it you'll get totally different results.

when you run this in TP

var N:Integer;
begin
  writeln ('with .');
  for N := 0 to 10 do
  writeln ( Round(Sqr(Sin(N*21167.0))*63));
  writeln;
  writeln ('without .');
  for N := 0 to 10 do
  writeln ( Round(Sqr(Sin(N*21167))*63) );
end.

you'll get

with .
0
47
47
0
48
46
0
49
45
0
49

without .
0
47
60
29
5
19
63
12
2
38
55

[felipemdc]

That's probably a problem involving precision and the internal implementation of floating point in TP. I'd recommend you ask at a TP forum, if they still exist.

Anyway, . do is a decimal separator in your code.

It is also indication of access of a function, or variable inside a class, unit or record, but that's in a completely different context.

[Vincent Snijders]

Code: [Select]

21167.0 is a floating point constant,

Code: [Select]

21167 is an integer constant.

[Bart]

And 2*21167 will overflow because it is large then MaxInt (32768 in TP) so you calculate the sinus of a very different angle then with 10*21167.0 (which will not overflow).
Try turning on overflow checking (not sure if TP has this, cannot check it for yu right now) and see what it outputs then (it should cause a runtime error, I guess).

Bart