Rebrand for IvorySQL distribution: ivy_mooncake

[NotHimmel]

The repo is renamed "ivy_mooncake" (GitHub fork name, README title); the extension, cdylib, Cargo package, and SQL-level identifiers all stay "pg_mooncake" so SQL written against pg_mooncake keeps working. Mirrors the ivy_duckdb pattern (the IvorySQL pg_duckdb fork installs as the pg_duckdb extension).

Changes vs upstream Mooncake-Labs/pg_mooncake (b7c67e0):

Submodules - all three forked to IvorySQL with ivy_* path names:

• pg_duckdb -> ivy_duckdb (IvorySQL/ivy_duckdb)
• moonlink -> ivy_moonlink (IvorySQL/ivy_moonlink)
• duckdb_mooncake -> ivy_duckdb_mooncake (IvorySQL/ivy_duckdb_mooncake; pinned at the same upstream SHA bd9e5f3a)

Path references following the renamed submodules:

• Cargo.toml: moonlink_*.path = "ivy_moonlink/..."
• Makefile: targets ivy_duckdb, ivy_duckdb_mooncake
• .vscode/c_cpp_properties.json: ${workspaceFolder}/ivy_duckdb/**
• Dockerfile: COPY ivy_moonlink ivy_moonlink

Repo identity:

• README rewritten as "IvorySQL distribution of pg_mooncake" framing, clone URL IvorySQL/ivy_mooncake.git; SQL examples and shared_preload_libraries stay pg_mooncake

Toolchain:

• rust-toolchain 1.88.0 -> 1.91.1 (required by aws-sdk-* MSRV in moonlink_service via the bgworker feature)

Intentionally NOT renamed (preserved at upstream values for SQL/ABI compatibility):

• Extension name, cdylib library, Cargo package -> pg_mooncake
• module_pathname='pg_mooncake', requires='pg_duckdb' in control file
• DuckDB-side names: mooncake schema, mooncake access method, pgmooncake_* C symbols, INSTALL mooncake FROM community

Verified end-to-end: make ivy_duckdb installs pg_duckdb.so; cargo pgrx regress --resetdb passes 3/3 (setup, partitioned_table, sanity) under rust 1.91.1 with
shared_preload_libraries='pg_duckdb,pg_mooncake'.