CLI Integration Tests#

Rust-based integration tests for the mlf command-line interface. These tests verify end-to-end CLI workflows by invoking the actual binary.

Structure#

tests/cli/
├── README.md (this file)
├── check_command/        # mlf check tests
│   ├── valid_file/
│   │   ├── input.mlf
│   │   └── expected.json
│   └── invalid_syntax/
│       ├── input.mlf
│       └── expected.json
├── generate_command/     # mlf generate tests
│   ├── lexicon_output/
│   └── typescript_output/
└── validate_command/     # mlf validate tests
    └── basic_validation/

Writing CLI Tests#

Each test case is a directory with:

• input.mlf - The MLF source file to test
• expected.json - Expected output (status, stdout, stderr)
• args.txt - CLI arguments (optional, one per line)

Example Test Case#

tests/cli/check_command/valid_file/
├── input.mlf
├── expected.json
└── args.txt

input.mlf:

record post {
    text!: string,
    createdAt!: datetime,
}

expected.json:

{
  "status": "success",
  "exit_code": 0,
  "stdout": "",
  "stderr": ""
}

args.txt:

check
{input}

Running Tests#

# All CLI tests (via just)
just test-cli

# Or via cargo
cargo test -p mlf-integration-tests --test cli_integration -- --nocapture

# Build CLI first if needed
cargo build --release

Test Categories#

check_command#

• ⏳ Valid MLF files pass
• ⏳ Invalid MLF files fail with error
• ⏳ Missing files return proper error
• ⏳ Multiple files are checked

generate_command#

• ⏳ Generate lexicon JSON
• ⏳ Generate TypeScript code
• ⏳ Generate Rust code
• ⏳ Generate Go code
• ⏳ Output to specific directory

validate_command#

• ⏳ Validate JSON against MLF schema
• ⏳ Detect validation errors
• ⏳ Batch validation

fetch_command#

• ⏳ Fetch lexicon from network
• ⏳ Cache downloaded lexicons
• ⏳ Update existing lexicons

init_command#

• ⏳ Initialize new project
• ⏳ Create .mlf directory structure
• ⏳ Don't overwrite existing files

Test Runner Implementation#

The test runner (tests/cli_integration.rs) will:

1. Discover test directories automatically
2. Build the CLI binary if needed
3. Execute commands via std::process::Command
4. Compare output against expected results
5. Report pass/fail with clear error messages

This follows the same pattern as our codegen tests, providing:

• Automatic discovery - No manual test registration
• File-based tests - Easy to add new test cases
• Clear output - ✓/✗ indicators for each test
• Version control friendly - All test data in files
• No shell scripting - Pure Rust test infrastructure