Update test vectors for sub-sat output sorting

[TheBlueMatt]

See lightning/bolts#876 we already do it right but should add the new test vectors.

[jkczyz]

Hmm... so I'm getting an off-by-one for the to-local output when making the change. The problem seems similar to lightning/bolts#539 (comment). Namely, the outputs plus fees total one less than expected.