Number: Fix parser to handle empty string

[rxaviers]

Fixes #682
Ref #292
Ref #353

[jzaefferer]

That's a lot of regular expressions that are pretty hard to read. I can't tell where in the tokenizer update you're now handling the empty string. But it seems to be working, so, up to you to add more comments or not.

[rxaviers]

Yeap, the regexps are starting to go crazy. The explanation I have is:

• https://github.com/rxaviers/globalize/blob/b1745f1babf7fb9d5a7bcb8293405c3682b0523c/src/number/parse-properties.js#L101 which corresponds to https://github.com/rxaviers/globalize/blob/b1745f1babf7fb9d5a7bcb8293405c3682b0523c/src/number/parse-properties.js#L138-L139
• The extra part has its own comment https://github.com/rxaviers/globalize/blob/b1745f1babf7fb9d5a7bcb8293405c3682b0523c/src/number/parse-properties.js#L141-L143

Any suggestion how to improve that please?

[jzaefferer]

I've looked at those, but barely understand anything. I think I'd have to spend a few hours to understand the context, probably not quite worth it as this point.

So, good to go :-)

[rxaviers]

Sure thanks for the review