Consider adding a compiler flag -vet-conversion to warn about implicit conversions that may result in loss of data, including converting integers to floats #5495
Description
MIRSA discourages implicit conversions outright between integral and floating-point types.
GCC has the -Wconversion flag that can be used to warn about implicit conversions that may result in loss of data, including converting integers to floats.
// compile time values and types with implicit conversions
main :: proc() {
i :: 1000000160
f: f32 : i
fmt.printfln("% 20.1f", f) // 1000000128.0
// i :: 1000000160
}// with explicit typing constant and dynamic cast
main :: proc() {
f: f64
i: int : 9223372036854775807
f = cast(f64)i
fmt.printfln("% 20.1f", f) // 9223372036854775808.0
// i : int : 9223372036854775807
}// same as above but with a variable vs. constant for both
package main
import "core:fmt"
main :: proc() {
f: f64
i: int = 9223372036854775807
f = cast(f64)i
fmt.printfln("% 20.1f", f) // 9223372036854775808.0
// i := 9223372036854775807
}