test: Expand script verification tests to cover all output types and flags (continued)

[alexanderwiederin]

Addresses #132.

Summary

• New output types not previously tested: P2SH-multisig, CLTV, CSV, native P2WPKH, native P2WSH, P2TR script-path
• New flags exercised individually: VERIFY_NONE, VERIFY_P2SH, VERIFY_DERSIG, VERIFY_NULLDUMMY, VERIFY_CHECKLOCKTIMEVERIFY, VERIFY_CHECKSEQUENCEVERIFY
• Split script_verify_test into separate tests for each output type

Approach

Subtests are ordered from oldest output types to newest. Each test follows the same pattern:

1. Valid transaction passes with both minimal and full flags
2. An altered transaction fails — the spk stays the same and the transaction itself is modified (corrupted signature, wrong amount, different locktime/sequence)
3. Stripping the relevant flag lets the altered transaction pass, confirming that flag is what catches the issue

For CLTV and CSV, changing locktime/sequence requires a fresh signature, so the invalid case uses a re-signed transaction rather than a byte flip. Unlike the other tests which use real mainnet transactions, the CLTV and CSV test data was created on regtest using a functional test.

Supersedes

This PR continues #137 by @stringintech with the following review feedback addressed:

• Use descriptive variable names instead of comments (e.g. tx_corrupted_sig instead of // same tx but with corrupted signature)
• Expand the flag loop into explicit separate calls for readability

[sedited]

LGTM

We should really expose those granular script error types, would make looking this over much easier.