Consider string buffer encoding.

[sc0Vu]

I think string in PHP is utf8 bytes. Consider the following code:

$buffer = new Buffer('abcàê߀abc', 'ascii');
echo $buffer->length(); // 15

if you dump you'll see:

  array(15) {
    [0]=>
    int(97)
    [1]=>
    int(98)
    [2]=>
    int(99)
    [3]=>
    int(195)
    [4]=>
    int(160)
    [5]=>
    int(195)
    [6]=>
    int(170)
    [7]=>
    int(195)
    [8]=>
    int(159)
    [9]=>
    int(226)
    [10]=>
    int(130)
    [11]=>
    int(172)
    [12]=>
    int(97)
    [13]=>
    int(98)
    [14]=>
    int(99)
  }

It should be 10 and looks like:

  array(10) {
    [0]=>
    int(97)
    [1]=>
    int(98)
    [2]=>
    int(99)
    [3]=>
    int(224)
    [4]=>
    int(234)
    [5]=>
    int(223)
    [6]=>
    int(172)
    [7]=>
    int(97)
    [8]=>
    int(98)
    [9]=>
    int(99)
  }

In nodejs:

• Ascii:

new Buffer('abcàê߀abc', 'ascii');
<Buffer 61 62 63 e0 ea df ac 61 62 63> // [97, 98, 99, 224, 234, 223, 172, 97, 98, 99]

• Utf8 (like the array above)

new Buffer('abcàê߀abc', 'utf8');
<Buffer 61 62 63 c3 a0 c3 aa c3 9f e2 82 ac 61 62 63> // [97, 98, 99, 195, 160, 195, 170, 195, 159, 226, 130, 172, 97, 98, 99]

How to fix it?

[sc0Vu]

IMHO buffer usage looks bad (.dirty), so I remove the buffer in the new version of rlp.