Implement Binary Golden Test Framework for C++ Backend

[EricRahm]

This work introduces a new testing framework for the C++ backend that verifies the correctness of binary serialization and deserialization. This framework, called "Binary Golden Tests," allows for the rapid development of tests that cover a wide range of Emboss features and data types. The goal is to make it easy to expand these tests to alternate language backends.

Core Components

1. Test Asset Format: A new test is defined by a set of three files:
   • An .emb file defining the structure.
   • A .json file containing the expected field values and types.
   • A .bin file containing the "golden" raw binary data.
2. Test Generator Script (generate_cpp_gtest.py): A Python script that reads the test asset files and automatically generates a C++ GTest source file. The generated test includes two primary test cases:
   • Read Test: Verifies that reading from the golden .bin file produces the expected values from the .json file.
   • Write Test: Verifies that creating a new view, writing the expected values, and serializing it produces a binary identical to the golden .bin file.
3. Bazel Macro (binary_golden_test): A new Bazel macro that encapsulates the test generation and compilation logic. This allows a new binary golden test to be added to the build with a single, clean rule, significantly improving the maintainability of the BUILD files.

Initial Test Coverage

As a starting point for coverage of the C++ backend, the following test cases were implemented using the new framework:

• simple_structure: A basic test to validate the initial infrastructure.
• integer_types: Verifies integers of all supported sizes, from 8 to 64 bits.
• big_endian: Ensures correct handling of non-default (big-endian) byte ordering.
• floating_point: Tests Float(32) and Float(64) data types.
• enums_and_bools: Verifies Flag (boolean) and enum types.
• bitfields: Tests bits structures with non-byte-aligned fields.
• fixed_arrays: Verifies support for fixed-size arrays