Bug #115927 for JSON-PP: Reject U+D800 to U+DFFF in strings for encoding

For comparison, the Go language rejects completely U+D800 when converting an array of byte "\xed\xa0\x80" to a string using UTF-8 (it replaces each *byte* with U+FFFD) and the JSON decoder decodes "\uD800" as U+FFFD.

https://play.golang.org/p/2aCCCTluVZ



package main

import (
    "fmt"
    "encoding/json"
)

func main() {
    fmt.Print("Decoding \"\\uD800\": ")
    var s string
    json.Unmarshal([]byte(`"\uD800"`), &s)
    for _, c := range s {
        fmt.Printf("U+%X ", c)
    }
    fmt.Print("\nDecoding \\xed\\xa0\\x80: ")
    json.Unmarshal([]byte{'"', 0xed, 0xa0, 0x80, '"'}, &s)
    for _, c := range s {
        fmt.Printf("U+%X ", c)
    }
    fmt.Println()

    fmt.Print("Converting bytes to UTF-8: ")    
    invalidSurrogate := []byte{0xed, 0xa0, 0x80}
    s = string(invalidSurrogate)
    for _, c := range s {
        fmt.Printf("U+%X ", c)
    }
    fmt.Println()
}





-- 
Olivier Mengué - http://perlresume.org/DOLMEN