add tests for script deserialization

[joemphilips]

While I was trying to port this library into C# , I found a bug in script deserialization.

I was trying to fix it by myself, but I'm afraid that the spec will change again. So I will not do it for now (unless we are sure that the spec won't change drastically anymore.)

Here is the failing case, tell me if I'm doing something wrong.

[apoelstra]

Thanks for this report - sorry for letting it drop. I will try to investigate today.

[apoelstra]

offending script:
OP_IF OP_SIZE OP_PUSHBYTES_1 20 OP_EQUALVERIFY OP_HASH256 OP_PUSHBYTES_32 131772552c01444cd81360818376a040b7c3b2b7b0a53550ee3edde216cec61b OP_EQUAL OP_ELSE OP_0 OP_ENDIF OP_VERIFY OP_SIZE OP_PUSHBYTES_1 20 OP_EQUALVERIFY OP_SHA256 OP_PUSHBYTES_32 ec4916dd28fc4c10d78e287ca5d9cc51ee1ae73cbfde08c6b37324cbfaac8bc5 OP_EQUALVERIFY OP_PUSHNUM_1

or 6382012088aa20131772552c01444cd81360818376a040b7c3b2b7b0a53550ee3edde216cec61b876700686982012088a820ec4916dd28fc4c10d78e287ca5d9cc51ee1ae73cbfde08c6b37324cbfaac8bc58851

[apoelstra]

Yep, I see ... the problem is that when we are parsing backward after seeing VERIFY, we parse OP_EQUAL and think we're in a pubkeyhash construction, but we're actually in a hashlock construction. We need to split up Terminal::PkH into a NonTerminal::PkHOrHashLock so that we can differentiate these cases.

Great find!

[apoelstra]

This PR needs an update to rename c:pk to pk, thresh_m to multi, and to use Segwitv0Script rather than Miniscript::<bitcoin::PublicKey>.

It still does not pass by the way, even with these fixes.

[sanket1729]

@apoelstra This test has been merged by #78 / #106 and this can be closed.