Files

36 lines
901 B
Go

package linknorm
import (
"strings"
"unicode"
)
// SanitizeShareText strips BOM, bidi isolates and zero-width chars that break
// scheme detection (common when pasting from messengers / Amnezia share).
func SanitizeShareText(raw string) string {
raw = strings.TrimPrefix(raw, "\ufeff")
raw = strings.TrimSpace(raw)
if raw == "" {
return ""
}
var b strings.Builder
b.Grow(len(raw))
for _, r := range raw {
switch r {
case '\u200b', '\u200c', '\u200d', '\u2060', '\ufeff', // zero-width / BOM
'\u2066', '\u2067', '\u2068', '\u2069', // LRI/RLI/FSI/PDI
'\u200e', '\u200f', // LRM/RLM
'\u202a', '\u202b', '\u202c', '\u202d', '\u202e': // embedding overrides
continue
}
if r < 0x20 && r != '\n' && r != '\r' && r != '\t' {
continue
}
if unicode.Is(unicode.Cf, r) { // other format controls
continue
}
b.WriteRune(r)
}
return strings.TrimSpace(b.String())
}