float

MudOS LPC now provides support for the floating point type. Declare variables like this:

float pi;

In general the same operations are supported for floats as are for integers. Float literals are written with a decimal point, a trailing dot, or exponent notation (with or without a fraction), and digits may be grouped with _ separators:

pi = 3.14159265;
x = 1.;          // 1.0
big = 2.5e6;     // 2500000.0
tiny = 2.5e-3;   // 0.0025
k = 1_000.25;

Note that casts do not convert values — use to_float() / to_int() to convert between floats and integers.

A float that the compiler knows is a float is truncated, fractional part discarded, when it lands in an int slot: an assignment, an initializer, an op= such as x *= 0.85 (the 0.85 becomes 0 before the multiply), or a return from an int function. That is deliberate, but it is easy to write by accident, so the compiler warns:

int b = 2.75;          // warning: Float value truncated to int ...
int x = 1000;
x *= 0.85;             // warning -- x becomes 0, not 850
int ok = to_int(2.75); // explicit: no warning
int i = 2.0;           // nothing is lost: no warning

The warning is on by default and #pragma no_warnings silences it (see diagnostics). It only fires when the value is statically a float; a mixed value, a call_other result or a mapping/array element has no static type to check.

The LPC float type is a C double: about fifteen (15) significant decimal digits.