1f86e7cd04
These characters get classified as "East Asian Mixed" by Go's `text/width` package, and thus assumed that their printed version occupies a width of 2 characters, whereas they each only occupy one. I'm not sure why they are classified as East Asian, but I did not have the energy to dive into Go's Unicode tables, so here is a workaround based on an exclusion list.
80 lines
1.3 KiB
Go
80 lines
1.3 KiB
Go
package text
|
||
|
||
import (
|
||
"golang.org/x/text/width"
|
||
)
|
||
|
||
// DisplayWidth calculates what the rendered width of a string may be
|
||
func DisplayWidth(s string) int {
|
||
w := 0
|
||
for _, r := range s {
|
||
w += runeDisplayWidth(r)
|
||
}
|
||
return w
|
||
}
|
||
|
||
const (
|
||
ellipsisWidth = 3
|
||
minWidthForEllipsis = 5
|
||
)
|
||
|
||
// Truncate shortens a string to fit the maximum display width
|
||
func Truncate(max int, s string) string {
|
||
w := DisplayWidth(s)
|
||
if w <= max {
|
||
return s
|
||
}
|
||
|
||
useEllipsis := false
|
||
if max >= minWidthForEllipsis {
|
||
useEllipsis = true
|
||
max -= ellipsisWidth
|
||
}
|
||
|
||
cw := 0
|
||
ri := 0
|
||
for _, r := range s {
|
||
rw := runeDisplayWidth(r)
|
||
if cw+rw > max {
|
||
break
|
||
}
|
||
cw += rw
|
||
ri++
|
||
}
|
||
|
||
res := string([]rune(s)[:ri])
|
||
if useEllipsis {
|
||
res += "..."
|
||
}
|
||
if cw < max {
|
||
// compensate if truncating a wide character left an odd space
|
||
res += " "
|
||
}
|
||
return res
|
||
}
|
||
|
||
var runeDisplayWidthOverrides = map[rune]int{
|
||
'“': 1,
|
||
'”': 1,
|
||
'‘': 1,
|
||
'’': 1,
|
||
'–': 1, // en dash
|
||
'—': 1, // em dash
|
||
'→': 1,
|
||
'…': 1,
|
||
'•': 1, // bullet
|
||
'·': 1, // middle dot
|
||
}
|
||
|
||
func runeDisplayWidth(r rune) int {
|
||
if w, ok := runeDisplayWidthOverrides[r]; ok {
|
||
return w
|
||
}
|
||
|
||
switch width.LookupRune(r).Kind() {
|
||
case width.EastAsianWide, width.EastAsianAmbiguous, width.EastAsianFullwidth:
|
||
return 2
|
||
default:
|
||
return 1
|
||
}
|
||
}
|