Develop Reference

Base64 encoding of String in Node.js

I try to encode a string in both browser and server but I have different base64 encoding for a specific string
Here is my string: "£aº©S=³hPó c¨¸" (Hexa: 00a3006100ba00a900940053003d00b30068005000f300900020006300a800b8 )
Client-side: I encode this String using btoa() and I have : o2G6qZRTPbNoUPOQIGOouA== and this is the result I expect.
Server-side: I code this String using Buffer according to this answer in Node.js i have :
var ciphertext = ... // myString
console.log(ciphertext.hexEncode()); // 00a3006100ba00a900940053003d00b30068005000f300900020006300a800b8
console.log(Buffer.from(ciphertext, 'utf8').toString('base64')) // wqNhwrrCqcKUUz3Cs2hQw7PCkCBjwqjCuA==
console.log(Buffer.from(ciphertext, 'ucs2').toString('base64')) // owBhALoAqQCUAFMAPQCzAGgAUADzAJAAIABjAKgAuAA=

I managed to obtain the base64 encoding I expect using the node-package base-64
I still don't know why, so if someone has a clue
var base64 = require('base-64');
var ciphertext = ...; //myString
var encoded = base64.encode(bytes);
console.log(encoded); // o2G6qZRTPbNoUPOQIGOouA==

How to replace / except in http://

I have a string and I need to replace %2f with /, except in http://.
Example:
var str = "http://locahost%2f";
str = str.replace(/%2f/g, "/");
Somehow I got str to output http:/locahost/.
Thanks in advance!

What should be used
You should use decodeURIComponent or decodeURI functions to decode encoded strings like yours. For example, decodeURIComponent("http://locahost%2f") will return http://localhost/.
That being said, decodeURIComponent should be used to decode components of a URI and not a full URI like http://locahost/.
You should look at this answer, it explains what is the difference between decodeURIComponent and decodeURI and when each should be used.
A workaround
As the first / are not encoded, it makes it difficult to find a Javascript functions doing exactly what you want. A working solution to decode your %2f after http:// would be to use String.prototype.split.
Working example:
var encoded = "http://localhost%2f";
var encodedArray = str.split("://");
var decoded = encodedArray[0] + "://" + encodedArray[1].replace(/%2f/g, "/");

This should work:
str = str.replace("%2f", "/");

CryptoJS not decrypting non-Latin characters faithfully

I am trying to use CryptoJS AES, like so:
var msg = "café";
var key = "something";
var c = CryptoJS.AES.encrypt(msg, key).toString();
CryptoJS.AES.decrypt(c, key).toString(CryptoJS.enc.Latin1);
Unfortunately this returns café, not café. Clearly Latin1 is not the right encoding to use, but I can't find a better one. Is there a solution?
Thanks.

You are just missing the format
The proper way is using CryptoJS.enc.Utf8
So, Please try:
CryptoJS.AES.decrypt(c, key).toString(CryptoJS.enc.Utf8);

https://code.google.com/p/crypto-js/#The_Hasher_Input
The hash algorithms accept either strings or instances of CryptoJS.lib.WordArray [...] an array of 32-bit words. When you pass a string, it's automatically converted to a WordArray encoded as UTF-8.
So, when you pass a string (and don't use CryptoJS.enc.* to generate a WordArray) it automatically converts the string (message) to a utf8 WordArray.
See here for sample roundtrip encrypt/decrypt:
https://code.google.com/p/crypto-js/#The_Cipher_Output
Here's a jsfiddle to play with CryptoJS
https://jsfiddle.net/8qbf4746/4/
var message = "café";
var key = "something";
var encrypted = CryptoJS.AES.encrypt(message, key);
//equivalent to CryptoJS.AES.encrypt(CryptoJS.enc.Utf8.parse(message), key);
var decrypted = CryptoJS.AES.decrypt(encrypted, key);
$('#1').text("Encrypted: "+encrypted);
$('#2').text("Decrypted: "+decrypted.toString(CryptoJS.enc.Utf8));
To emphasize my point here is the same thing using Latin1 encoding:
https://jsfiddle.net/3a8tf48f/2/
var message = "café";
var key = "something";
var encrypted = CryptoJS.AES.encrypt(CryptoJS.enc.Latin1.parse(message), key);
var decrypted = CryptoJS.AES.decrypt(encrypted, key);
$('#1').text("Encrypted: " + encrypted);
$('#2').text("Decrypted: " + decrypted.toString(CryptoJS.enc.Latin1));
On a side note, the API would probably be better if it only accepted WordArray and didn't overload the toString method (which is just a convenience interface to CryptoJS.enc.*.stringify). The string conversion magic is a little misleading.

You are trying to decrypt your data as a Latin1 string, even though your input string is not in Latin1. The encoding used by CryptoJS internally is not the same as the encoding you use to write the input file.
You need to specify the same encoding both when encrypting (for the string -> byte array conversion) and when decrypting (for the byte array -> string conversion).

Escape Base64 String in URL?

For some reason I need to put a Base64 encoded string in the URL, i.e.
<script>
var URL = "localhost:8080/MyApp/order?z=AAAAAAAEJs4"
</script>
The URL is generated by Java, problem here is I can only use java to make the Base 64 encoded string to be URL friendly, but not javascript friendly, so fire bug give me the following error:
SyntaxError: unterminated string literal
Apparently there are charactors in the Base64 string requires escaping. However I cannot use escape() or URLencoding() method as the request will directly deliver to the controller and manipulated by java code, so there is no "next page" in this situation.
So how to make this work then?

If your're trying to convert that z url attribute into an understandable string/variable, you have to use a library to convert that. Below is a link to a base64 library for Javascript that you must load in.
http://www.webtoolkit.info/javascript-base64.html
You maybe also need a library to access the z attribute in your url. I would recommend this:
https://github.com/allmarkedup/purl

Just in case it helps, here is a short JS code requiring no dependency:
String.prototype.base64EncodeUrl = function () {
var str = this;
str = window.btoa(unescape(encodeURIComponent( str )));
return str.replace(/\+/g, '-').replace(/\//g, '_').replace(/\=+$/, '');
};
String.prototype.base64DecodeUrl = function () {
var str = this, str_pad = (str + '===');
str = str_pad.slice(0, str.length + (str.length % 4));
str = str.replace(/-/g, '+').replace(/_/g, '/');
return decodeURIComponent(escape(window.atob( str )));
};

What is DigestUtils sha256 encoding equivalent from JavaScript?

I call following and create the password hash.
ByteString password = ByteString.copyFrom(DigestUtils.sha256("mypassword"));
But now I need to send the sha256 converted password message from client (JavaScript). I tired to use CryptoJS as following
var pass = CryptoJS.SHA256(document.getElementById('password').value);
var passhash = pass.toString(CryptoJS.enc.Latin1)
login(passhash);
I tried all Base64, Latin1, and Hex types to get the string. But it will not produce the same password as the one in Java

Problem was with character encoding. Following fixed the problem.
in JS:
var password = pass.toString(CryptoJS.enc.Utf16);
In Java:
byte[] passhash = jsCryptoString.getBytes("UTF-16BE");