`sane` in the typechecker should be called `signed`

[apoelstra]

Also the doccomment for this field should be changed. It currently says:

true if satisfactions cannot be created by any 3rd party who has not yet seen a satisfaction.

Hash preimages and signature checks are safe; timelocks are not. Affects malleability.

This is inaccurate. Hash preimages are not "safe" (which is why we renamed this term to "signed").

The code is correct, just the doccomment is wrong.

[apoelstra]

Was noticed by @darosior -- see IRC https://gnusha.org/miniscript/2026-02-19.log?__goaway_challenge=meta-refresh&__goaway_id=09e42b3902db88b85b8f91abbdc20684 -- but I guess I didn't notice his noticing.