transfer_array_intrinsic failures

[Paul Thomas]

Jack Howarth ran a couple of tests for me last night. TRANSFER is 
working correctly and it is the testcases that make assumptions about 
endian-ness; or rather the first certainly has that problem, the second 
looks like it does but is just a wee bit more difficult to understand.

I'll try to come up with modified testcases tonight. Otherwise, I will 
remove both temporarily, whilst I progress replacements, but will leave 
the TRANSFER function in place.

Thank you, Jack.

Paul