Dragon-Runner

A test runner for CMPUT 415 Compiler Design that services both student testing and automated grading.

Installation

Requirements: Python ≥ 3.8

git clone https://github.com/cmput415/Dragon-Runner.git
cd Dragon-Runner
pip install .

Some newer versions of python prevent system-wide package installations by default. To get around this use a virtual environment or --break-system-packages. If dragon-runneris not found in your $PATH after install, ensure ~/.local/bin is added.

Quick Start

# Run tests normally
dragon-runner config.json

# Run in tournament mode (for grading)
dragon-runner tournament config.json

# Check for memory leaks (see Valgrind Config in `/tests`)
dragon-runner memcheck valgrindConfig.json

# Start HTTP server for explorer
dragon-runner serve /path/to/configs

Configuration

Dragon-Runner uses JSON configuration files to define test packages, executables, and toolchains.

Basic Example

{
  "testDir": "../packages/CPackage", 
  "testedExecutablePaths": {
    "gcc": "/usr/bin/gcc"
  },
  "toolchains": {
    "compile-and-run": [
      {
        "stepName": "compile",
        "executablePath": "$EXE", 
        "arguments": ["$INPUT", "-o", "$OUTPUT"],
        "output": "/tmp/test.o",
        "allowError": true
      },
      {
        "stepName": "run",
        "executablePath": "$INPUT",
        "arguments": [],
        "usesInStr": true,
        "allowError": true
      }
    ]
  }
}

Configuration Reference

Top-level Properties

Property	Description	Required
testDir	Path to directory containing test packages	✓
testedExecutablePaths	Map of executable IDs to file paths	✓
toolchains	Map of toolchain names to step lists	✓
runtimes	Map of runtime libraries (optional)
solutionExecutable	Reference solution ID (optional)

Toolchain Steps

Property	Description	Required
stepName	Human-readable step name	✓
executablePath	Path to executable (use $EXE, $INPUT)	✓
arguments	Command arguments list	✓
output	Output file path (optional)
allowError	Allow non-zero exit codes (optional)
usesInStr	Use test input stream as stdin (optional)
usesRuntime	Load runtime library (optional)

Magic Variables

• $EXE - Path to the tested executable
• $INPUT - Input file (testfile for first step, previous output for others)
• $OUTPUT - Output file for next step
• $RT_PATH - Runtime library directory
• $RT_LIB - Runtime library name

Test File Format

Tests support inline directives using comment syntax:

// INPUT: hello world
// CHECK: HELLO WORLD

int main() {
    // Your test code here
    return 0;
}

Directives

• INPUT: - Single line of stdin (no newline)
• INPUT_FILE: - Path to input file
• CHECK: - Expected stdout (no newline)
• CHECK_FILE: - Path to expected output file

Multiple INPUT: and CHECK: directives are supported. INPUT: and INPUT_FILE: cannot be used together.

Command Line Reference

Basic Usage

dragon-runner [mode] config.json [options...]

Modes

• regular (default) - Standard test execution
• tournament - Cross-product testing for grading
• perf - Performance benchmarking
• memcheck - Memory leak detection
• serve - HTTP server mode
• script - Run grading scripts

Options

Option	Description
--timeout SECONDS	Test timeout (default: 2.0)
--fail-log FILE	Log failures to file
--verify	Verify package exists for CCID
--debug-package PATH	Test single package
-t, --time	Show execution times
-v, --verbosity	Increase output verbosity (repeat for more)
-s, --show-testcase	Display test file contents
-o, --output FILE	Output file for results

Examples

# Basic testing with timing
dragon-runner -t config.json

# Verbose tournament mode
dragon-runner tournament -vv config.json

# Performance testing with 5-second timeout
dragon-runner perf --timeout 5.0 config.json

# Serve configs on port 8080
dragon-runner serve --port 8080 /path/to/configs

# Run grading script
dragon-runner script build.py /path/to/submissions build.log 4

Contributing

Contributions welcome! Please file issues for bugs or feature requests, and feel free to submit pull requests.