Redshift: Support implicit string concatenation using newline

[yoavcloud]

We've encountered a statement of the following form that failed to parse due to a missing comma, that is accepted by Redshift:

SELECT 'ab' IN (
   'a'
   'b'
)

Turns out Redshift considers the 'a' <NEWLINE> 'b' as a concatenation, so it's evaluted as 'ab'. The fix here is to parse this construct as 'ab'.