plan: add some clarifying docs

[benma]

"The size in bytes of the script sig that" is ambiguous and could refer to the actual scriptSig, exlcuding the varint prefix that indicates its size.

The code comment in witness_size is promoted to the docstring, as its helpful to anyone trying to compute the right size of a transaction.

[apoelstra]

ACK e89e226; successfully ran local tests; nice, good call

[benma]

@apoelstra speaking of, I just noticed that plan.satisfy() returns the scriptSig as a ScriptBuf, which does not include the var-int prefix, so it's a inconsistent that the size of the var-int prefix is included in plan.sigscript_size().

Furthermore, it's not possible to know the size of the scriptsig without the var-int prefix during planning.

I think it would make sense to offer a function to return the size excluding the var-int prefix. WDYT?

[apoelstra]

Oh, yeah, I like this. Definitely agreed that plan.satisfy().len() ought to return the same value as plan.???_size().

I am torn between deprecating the version that has the length prefix, or not. (We definitely should not just silently change the behavior of scriptsig_size()! We need a new name. Maybe satisfy_len?)