decode error for some chinese words

[orangemi]

Try this:

var utf7 = require('utf7')
var msg = '新建'
console.log(msg, utf7.encode(msg), utf7.decode(utf7.encode(msg)))

Output:

新建 +ZbBe+g- 新帀

If I change the msg to a single word 建, encode & decode is correct. When it msg contains 2 words it got wrong decode.

Any idea ?

[TSlivede]

The problem seems to be, that the char + is not in the regex that searches for possible base64 chars.

line 105 shouldn't be
return str.replace(/\+([A-Za-z0-9\/]*)-?/gi, function(_, chunk) {
but
return str.replace(/\+([A-Za-z0-9+\/]*)-?/gi, function(_, chunk) {

This also applies if you try to decode the utf-7 version of eg. "⏩" (utf7.encode('⏩') -> "+I+k-" -> also contains a + within the base64 part).