@frankwallis you might find this helpful.

thanks!

Also pinging @teppeis and @basarat for typescript-simple that can use this change.

Is the algorithm provided basically saying:

1. We have a string.
2. We want to consider the UTF8 encoding of that string.
3. We want to take that encoding and convert it to Base64?

If so, i would personally prefer breaking out the two parts of this process. The UTF8 encoding, and the Base64 conversion.

Note: MDN already covers a good way to do this:
https://developer.mozilla.org/en-US/docs/Web/API/WindowBase64/btoa#Unicode_Strings

function utf8_to_b64( str ) {
  return window.btoa(unescape(encodeURIComponent( str )));
}
or
function b64EncodeUnicode(str) {
    return btoa(encodeURIComponent(str).replace(/%([0-9A-F]{2})/g, function(match, p1) {
        return String.fromCharCode('0x' + p1);
    }));
}

Now, if we don't have access to these functions in Node or Chakra, we could still do the above steps, just individually. i.e. break up a string into utf8 bytes. Then convert those bytes to Base64. I think this would be far easier to understand (and verify).

mm. let me try this out then. i remember there was a reason why i did not do that :)

So btoa is not available on node, and escape is not available on chakra :), so there is not an easy way to use these. I did split them into two functions, and added unit tests as well.

I have also handled --inlineSourceMap to be a parallel to --sourceMap instead of a dependent.

The reason covertToBase64 is in utilities is to allow for targeted unit tests. I have limited compiler baselines to limt the noise in .js files as the map is actually included in the .js file.

LGTM

Can you also update transpile function to use inlineSources\inlineSourceMaps?

@mhegazy duplicated lines in wiki https://github.com/Microsoft/TypeScript/wiki/Roadmap/_compare/bfc82f102a462b485ed5fbebf04a457eb76987a7...307473332000299d8fe56b70460deca31b5e1f03#diff-0fa168c44ae96a89e7ba2d227c7c545cR44

thanks @teppeis. fixed