Bump acorn version

[bgrins]

Changes for version compatibility:

• Uses tokenizer instead of tokenize
• token.type.type is renamed to token.type.label
• token.startLoc -> token.start.loc and token.endLoc -> token.end.loc
• The order that comments come through is different now (the onComment handler is fired immediately after the token that the comment comes after), so we don't need to use the commentQueue object. However this meant that we didn't know if we needed to add a newline after a token because we didn't know if there was going to be an inline comment next. My solution is to do one pass through the tokenizer and onComment handler and build up an array of tokens+comments, and then loop through that array and do processing. Couldn't think of an easy way to do it without storing all of the tokens in this array - I considered trying to wait and add a newline at the beginning of the loop for the last token, but I was having bugs. I'd say the current implementation is easier to follow, but I could try to go back to the single pass through if this is a performance issues (do we have performance measurements somewhere?).

Anyway, the test is passing locally and I need to test this on m-c (which I'm assuming will simultaneously require a tern update since they share the acorn implementation), but I figure it's worth starting a review here, so r? @fitzgen

[fitzgen]

FYI: I'm on vacation all next week, won't be able to get to this until after that.

[fitzgen]

Thanks for this, Brian. LGTM!