I solved the problem using only browser resources (no node.js); I also found some official sources for explanations, test data, field decoding (see at the end of the answer):
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
<meta name="generator" content="PSPad editor, www.pspad.com">
<title></title>
</head>
<body>
<textarea id="encoded" name="encoded" cols=100 rows=10></textarea><br>
<textarea id="decoded" name="decoded" cols=100 rows=10 value = "prova">hello</textarea><br>
<span id="plain1" name="plain1">-</span> -->
<span id ="cod" name="cod">-</span> -->
<span id="plain2" name="plain2">-</span><br>
<button onclick="test()">Test</button><br>
<button onclick="vai()">Vai</button><br>
<script src="pako.min.js"></script>
<script src="my_base45_2.js"></script>
<script src="cbor.js" type="text/javascript"></script>
<script>
// https://dev.to/lmillucci/javascript-how-to-decode-the-greenpass-qr-code-3dh0
function convert(decoded) {
text="";
for (var i = 0; i < decoded.length; i++) {
text += String.fromCharCode(decoded[i]);
}
console.log("Text=",text);
values = text.split(",");
console.log("values: " ,values);
final = "";
for (var i=0; i<values.length; i++) {
final += String.fromCharCode(values[i])
}
return final;
}
function test() {
source = document.getElementById("decoded").value;
console.log("source=",source);
document.getElementById("plain1").innerHTML = source;
var enc = new TextEncoder();
buff= enc.encode(source);
console.log("buff:" , buff);
encoded2 = encode(buff);
console.log("encoded2:" , encoded2);
document.getElementById("cod").innerHTML = encoded2;
source = encoded2;
console.log("Encoded source=",source);
decoded = decode(source).enc;
converted = convert(decoded);
console.log("Decoded=", decoded);
console.log("converted=", converted);
document.getElementById("plain2").innerHTML = converted;
}
function buf2hex(buffer) {
// https://stackoverflow.com/questions/34309988/byte-array-to-hex-string-conversion-in-javascript
var u = new Uint8Array(buffer),
a = new Array(u.length),
i = u.length;
while (i--) // map to hex
a[i] = (u[i] < 16 ? '0' : '') + u[i].toString(16);
u = null; // free memory
return a.join('');
};
function typedArrayToBuffer(array) {
// https://stackoverflow.com/questions/37228285/uint8array-to-arraybuffer
return array.buffer.slice(array.byteOffset, array.byteLength + array.byteOffset)
}
function vai() {
source = document.getElementById("encoded").value;
console.log("Encoded source=",source);
// Decode BASE45:
decoded = decode(source).enc;
// Unzip the decoded:
COSEbin = pako.inflate(decode(source).raw);
COSE = buf2hex(COSEbin);
var typedArray = new Uint8Array(COSE.match(/[\da-f]{2}/gi).map(function (h) {
return parseInt(h, 16)
})) // https://stackoverflow.com/questions/43131242/how-to-convert-a-hexadecimal-string-of-data-to-an-arraybuffer-in-javascript
var unzipped = typedArray.buffer;
[headers1, headers2, cbor_data, signature] = CBOR.decode(unzipped);
cbor_dataArr = typedArrayToBuffer(cbor_data);
greenpassData = CBOR.decode(cbor_dataArr);
console.log(greenpassData);
}
</script>
</body>
</html>
The key is the "double CBOR decoding": once is not enough:
CBOR decoding:
[headers1, headers2, cbor_data, signature] = CBOR.decode(unzipped);
Conversion to Array:
cbor_dataArr = typedArrayToBuffer(cbor_data);
Further CBOR conversion:
greenpassData = CBOR.decode(cbor_dataArr);
My version of base45 decoder ("my_base45_2.js"), adapted from a node.js version:
function encode(uint8array) {
var output = [];
for (var i = 0, length = uint8array.length; i < length; i+=2) {
if (uint8array.length -i > 1) {
var x = (uint8array[i]<<8)+ uint8array[i+1]
var [ e, x ] = divmod(x, 45*45)
var [ d, c ] = divmod(x, 45)
output.push(fromCharCode(c) + fromCharCode(d) + fromCharCode(e))
} else {
var x = uint8array[i]
var [ d, c ] = divmod(x, 45)
output.push(fromCharCode(c) + fromCharCode(d))
}
}
return output.join('')
};
var divmod = function divmod(a,b) {
var remainder = a
var quotient = 0
if (a >= b) {
remainder = a % b
quotient = (a - remainder) / b
}
return [ quotient, remainder ]
}
const BASE45_CHARSET = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ $%*+-./:"
var fromCharCode = function fromCharCode(c) {
return BASE45_CHARSET.charAt(c);
};
function decode(str) {
var output = []
var buf = []
for(var i = 0, length=str.length; i < length; i++) {
//console.log(i);
var j = BASE45_CHARSET.indexOf(str[i])
if (j < 0)
console.log('Base45 decode: unknown character n.', i, j);
//throw new Error('Base45 decode: unknown character');
buf.push(j)
}
for(var i = 0, length=buf.length; i < length; i+=3) {
var x = buf[i] + buf[i + 1] * 45
if (length - i >= 3) {
var [d, c] = divmod(x + buf[i + 2] * 45 * 45,256)
output.push(d)
output.push(c)
} else {
output.push(x)
}
}
console.log("output",output);
var enc = new TextEncoder();
return {"enc" : enc.encode(output), "raw" : output};
//return Buffer.from(output);
};
These are the decoding steps:
QR code --> QR DECODER --> RAW QR-decoded string --> BASE45 decoder --> zlib compressed string --> pako library --> COSE string --> CBOR decoder --> CBOR string --> CBOR decoder --> final JSON file
My reference guides for decoding greenpass:
