refactor handling of DLC input validation

[conduition]

• Removed trailing whitespace (my editor does this automatically so i separated it in the first commit)
• Refactor handling of inputs for DLC registration:
  • Raise TransactionErrors instead of returning false, to give callers better feedback and to mirror the behavior of other input validation code
  • Avoid duplicating SCT validation logic in mint/dlc.py. Instead rely on the recursive SCT secret verification logic in mint/conditions.py.
• Rename DLCWitness to SCTWitness, reflecting the fact that this data structure is the witness to an SCT secret (including those which may not be related to DLCs at all)

[lollerfirst]

lgtm

[conduition]

Tests are failing, i need some additional time to get unit test runners working on my machine

[conduition]

I found the cause of my woes: cashubtc#607

[conduition]

I've updated this PR to target the latest commit 6b1d04c in cashubtc#576. The overall changes are the same:

• Whitespace formatting fixes c58cf9b
• Rename DLCWitness to SCTWitness 97684a6
• Raise errors instead of returning false. 42964e7
• Refactor how DLC input validation is handled 0b5c191
• [new] Fix a bunch of linter errors 1f287d3 (validated with make check)

I've confirmed tests pass on my machine with make test, though only if I apply cashubtc#607 on top of this PR.

[lollerfirst]

Good job. I'm gonna merge this. Unfortunately some tests fail because I fucked up. I changed the migration for the DLC table to have the column settled of type BIT instead of BOOL, as I had read that BOOL is not supported across different DBs while BIT is. Apparently it doesn't like the 0 default value for it.