fix(sql): treat X'...' as binary literal

[sundy-li]

I hereby agree to the terms of the CLA available at: https://docs.databend.com/dev/policies/cla/

Summary

• parse X'...' as a binary literal instead of routing it through the unsigned hex integer path
• keep MySQL-style 0x... literals as numeric literals
• add parser, function-layer, and sqllogictest regression coverage for the new semantics

Tests

• cargo test -p databend-common-ast
• cargo test -p databend-common-functions --test it type_check::test_pg_hex_literal_type_check -- --nocapture
• cargo test -p databend-common-sql --test it --no-run
• cargo clippy -p databend-common-ast --tests -- -D warnings
• cargo run -p databend-sqllogictests -- --run_file binary_format.test --handlers mysql (requires a running Databend server on 127.0.0.1:3307; connection was refused in this environment)

Fixes #19600

---

This change is 

[github-actions]

At least one type of change must be checked in the PR description.
@sundy-li please update it 🙏.

[github-actions]

🤖 CI Job Analysis

Workflow: 23490482570

📊 Summary

• Total Jobs: 25
• Failed Jobs: 2
• Retryable: 0
• Code Issues: 2

❌ NO RETRY NEEDED

All failures appear to be code/test issues requiring manual fixes.

🔍 Job Details

• ❌ linux / test_unit: Not retryable (Code/Test)
• ❌ linux / check: Not retryable (Code/Test)

---

🤖 About

Automated analysis using job annotations to distinguish infrastructure issues (auto-retried) from code/test issues (manual fixes needed).

[sundy-li]

Blocking issue:

The new Literal::Binary variant is not handled in src/tests/sqlsmith/src/query_fuzzer.rs:554, so the workspace no longer compiles. CI is failing with error[E0004]: non-exhaustive patterns: &databend_common_ast::ast::Literal::Binary(_) not covered in fuzz_literal_bad_value(). Please thread the new literal variant through sqlsmith before merging.