Hi Kyle,

Came across your blog post when searching for "golang xor encryption" and was testing it. Unfortunately, your function isn't Unicode safe. I've created a few test cases in https://play.golang.org/p/UdCubbEZECx to show the issue.

Might be worth putting a warning into the code/blog post, as string handling in Go has its gotchas unfortunately.

Thanks,
tsak

I had encountered a bug with the c version code when I had tried to decrypt the encrypted string.

void encryptDecrypt(char *input, char *output) {
	char key[] = {'K', 'C', 'Q'}; //Can be any chars, and any size array
	
	int i;
	for(i = 0; i < strlen(input); i++) {
		output[i] = input[i] ^ key[i % (sizeof(key)/sizeof(char))];
	}
}

There is a bug when do the encryption and decryption in one function above.
As the strlen(input) may be smaller than the original string length.So is the below:

void encryptDecrypt(char *input, char *output, int size) {
	char key[] = {'K', 'C', 'Q'}; //Can be any chars, and any size array
	
	int i;
	for(i = 0; i < size; i++) {
		output[i] = input[i] ^ key[i % (sizeof(key)/sizeof(char))];
	}
}

Here's a Lua version

function M.obfuscate(input, key)
	key = key or M.obfuscation_key
	local output = ""
	local key_iterator = 1
	
	local input_length = #input
	local key_length = #key
	
	for i=1, input_length do
		local character = string.byte(input:sub(i,i))
		if key_iterator >= key_length then key_iterator = 1 end -- cycle
		local key_byte = string.byte(key:sub(key_iterator,key_iterator))
		output = output .. string.char(bit.bxor( character , key_byte))

		key_iterator = key_iterator + 1
		
	end
	return output
end

I am trying to encrypt something on Cpp and decrypt it in PHP but they won't work. Any tips, they are the same key.

Please add a license to this project, I'd go with MIT: http://choosealicense.com/licenses/mit/.

See https://help.github.com/articles/open-source-licensing/ for help on this topic.

On the C version I made some modifications for my own needs, however now it seems to not work properly on windows. (the encrypted string does not decrypt to be the same as the original and I test printing them all out as well, I started with making the key be passed to the function itself, and thinking if it is because output has to be larger than the input string or something.

void encryptDecrypt(char *input, char *output, char *key) {
	size_t i;
	for (i = 0; i < strlen(input); i++)
		output[i] = input[i] ^ key[i % sizeof(key)];
}

On this I am taking a key of size 134 characters, and a input string of 114 characters and I expect and output (encrypted string of about the same size of larger, however I am not sure the actual size and in some places it returns multiple characters with the int value of -52. It does that value however on the decryption though after 52 characters into the string decryption. I am officially left wondering why this is.

https://cdn.discordapp.com/attachments/184740186436009984/367877048011718656/ConsoleApplication1.c

This code above demonstrates the valid problems I got with all of this.

Also why are you dividing by 1 because sizeof(char) is always 1 no matter what the platform is.

std::string EncryptDecrypt(std::string toEncrypt) { char key[4] = { 'j', 'T', 'J','d' }; std::string output = toEncrypt; for (int i = 0; i < toEncrypt.size(); i++) { output[i] = toEncrypt[i] ^ key[i % (sizeof(key) / sizeof(char))]; if (output[i] == 0) { output[i] = toEncrypt[i]; } } output[toEncrypt.size()] = 0; return output; }

Source code: https://github.com/KyleBanks/XOREncryption/blob/master/Go/xor.go

The string concat operator output += will make this XOR function run heavily. I tested input with ~500k characters and it took about 22s to complete. Here is my suggestion:

func Xor(input, key string) (output string) {
	var tmp []string

	for i := 0; i < len(input); i++ {
		tmp = append(tmp, string(input[i]^key[i%len(key)]))
	}
	output = strings.Join(tmp, "")
	return output
}

By doing this, input with ~500k characters took ~0.1s to complete. Much better.

Could you tag your releases and make the changelog directly visible in your GitHub releases? This would increase transparency for potential users and distribution maintainers. Something like https://skywinder.github.io/github-changelog-generator/ could be used to generate the changelog based on merged PRs and closed issues since the last release tag.