Treatment of integer constants in subtraction

Technical Discussion
1

Hello,

I am new to Go and I am learning my way through its features. I have a question about the use of integer constants in certain operations. I tried some code in Go Playground which had the following lines

var i int8 = 0
var j int8 = i - 128

The statement defining variable j gives the error “constant 128 overflows int8” because Go deals with 128 as the constant as not -128. On the other hand, the following statements work fine presumably because Go is looking at the constant -128 and not 128.

var i int8 = 0
var j int8 = -128 + i

var i int8 = 0
var j int8 = i + (-128)

When subtracting an integer constant (as in i - 128) doesn’t it make more sense to take the negative of the constant for limit checking? Conversely, I see that var j int8 = i - -128 does not register an error, which it would if one were to take the negative of the constant -128 for limit checking.

2

Not to me.

The behaviour of a computer programming language should be defined by a language specification. For Go that is The Go Programming Language Specification.

Your first example applies a rule for binary operators.

package main

func main() {
	var i int8 = 0
	// constant 128 overflows int8
	var j int8 = i - 128
	// constant 128 overflows int8
	var k int8 = i - int8(128)

	_, _, _ = i, j, k
}

Your second and third examples apply a rule for unary operators before applying a rule for binary operators. Unary operators have the highest precedence.

package main

func main() {
	var i int8 = 0
	var j int8 = -128 + i
	var k = int8(-128) + i

	_, _, _ = i, j, k
}

package main

func main() {
	var i int8 = 0
	var j int8 = i + (-128)
	var k int8 = i + int8(-128)

	_, _, _ = i, j, k
}
3

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.