Cardano Updates

* Make `Builtins.unsafeUncons` use builtin list casing instead of head, tail calls

* correct according to PR feedback by @SeungheonOh

* Add PIR definition to unsafeCaseList, add tests

* Change log

* Silly mistake

* Update golden

---------

Co-authored-by: Sohang Chopra <[email protected]>

Updated valueContains cost model parameters based on benchmark run with expanded above-diagonal coverage (3→20 samples).

Cost parameter changes:

- Above-diagonal constant: 233391 → 246085 (+5.4%)

- Below-diagonal intercept: 658617 → 686640 (+4.3%)

- Slope parameters adjusted for better fit

Remove redundant try wrappers:
- symbol is auto-backtracking since Megaparsec 4.4.0
- Removed 10 redundant try wrappers from keyword and delimiter parsing
- Fixed duplicate "constr" entry bug

Add strategic labels for better error messages:
- Use <?> operator for two-level labeling (delimiters + keywords)
- Replace explicit fail with <?> in type parser (more idiomatic)
- Labels provide clearer context in parse errors

Refactor to use between combinator:
- Cleaner, more declarative code structure
- Standard Megaparsec pattern for delimited parsing
- Eliminates intermediate bindings

Expand test coverage:
- Add 6 new error message tests covering common error scenarios
- Extract testParseError helper to reduce duplication
- Organize error tests into dedicated test group
- Verify labels appear correctly in error messages

All 2,973 tests pass (781 UPLC + 1,864 PLC + 328 PlutusIR)

Kenneth requested more data points above the diagonal for a better
average of the constant-cost region.

Changes:
- Increased from 3 to 20 above-diagonal samples
- Added variety in container/contained size ratios:
  * Small containers (4-16 entries) with larger contained
  * Medium containers (25-100 entries) with larger contained
  * Large containers (150-1024 entries) with much larger contained
- Total benchmark points: ~570 (550 below + 20 above diagonal)

Co-authored-by: Ana Pantilie <[email protected]>

Co-authored-by: Ana Pantilie <[email protected]>

Co-authored-by: Ana Pantilie <[email protected]>

Co-authored-by: Ana Pantilie <[email protected]>

Co-authored-by: Yura <[email protected]>

Co-authored-by: Yura <[email protected]>

Two different "diagonal" concepts were confused:
1. Value structure diagonal (numPolicies vs tokensPerPolicy)
2. Costing diagonal (container totalSize vs contained totalSize)

The previous `aboveDiagonalCombos` tested #1 but the cost model needs #2.

Changes:
- Remove misleading `aboveDiagonalCombos` variable
- Add true above-diagonal samples where contained size > container size
- Update comments to clarify both diagonal concepts
- Generate 3 independent (container, contained) pairs for the constant-cost region

Closes #7525