Create an index vector defined by two input vectors, one defining the beginnings of one or more index ranges, and the other defining the ends of the ranges.

The difficult part is that for and while loops are disallowed. No looping! [I apologize for the rudimentary regexp test for 'for' 'while' and 'eval'.] My test cases scan for usage of these substrings, so choice of variable names is effectively restricted as well.

So given 2 input vectors x1 and x2, the desired output is a third vector y, containing [x1(1):x2(1) x1(2):x2(2) ... x1(end):x2(end)]

Example:

 x1 = [1 5 12];
 x2 = [2 8 21];

Result:

 y = [1 2 5 6 7 8 12 13 14 15 16 17 18 19 20 21]