Added error checking to melt_check to ensure unique measure.vars

[jmacdon]

Got caught by this today - identical measure.vars will silently recycle the first column.

z <- data.frame(A = letters[1:5], this = 1:5, that = 6:10, this = 11:15, that = 16:20, check.names = FALSE)
melt(z)
Using A as id variables
A variable value
1 a this 1
2 b this 2
3 c this 3
4 d this 4
5 e this 5
6 a that 6
7 b that 7
8 c that 8
9 d that 9
10 e that 10
11 a this 1
12 b this 2
13 c this 3
14 d this 4
15 e this 5
16 a that 6
17 b that 7
18 c that 8
19 d that 9
20 e that 10