Running Tasks

Once your tasks are configured via .unirtm.toml or the .unirtm/tasks directory, you can unleash their power using the unirtm run command.

Basic Usage

# Run a single task
unirtm run build

# Run multiple tasks (sequentially or in parallel, depending on the DAG)
unirtm run lint test build

If you execute unirtm run without any arguments, it smartly parses all available tasks and presents an interactive list alongside their short descriptions (similar to a dynamic make help).

Parallel Execution & Dependency Resolution (DAG)

UniRTM features an internal concurrent Directed Acyclic Graph (DAG) engine. If you run a task C that defines prerequisites depends_on = ["A", "B"]:

1. UniRTM recursively checks if A and B have their own dependencies.
2. UniRTM executes A and B simultaneously in parallel.
3. Task C will only begin execution once both A and B return successfully with an exit code of 0.
4. If either A or B fails, the entire execution chain is immediately aborted with an error.

This architecture allows you to squeeze maximum performance out of your multi-core CPU!

Terminal UI & Pterm Integration

UniRTM utilizes a high-performance terminal UI library (Pterm) to manage task output.

• Stream Isolation: During parallel execution, the output of different tasks is smartly isolated and prefixed (e.g., [build] ..., [test] ...) to prevent log interleaving and race conditions on stdout.
• Dynamic Status Bar: A dynamic progress indicator is displayed at the bottom, providing real-time visibility into which tasks are currently Running and which are Done.
• Plaintext Degradation: When a CI environment (like GitHub Actions) is detected, UniRTM automatically disables the dynamic UI and colors, degrading gracefully to a plaintext streaming output to ensure log file readability.

Advanced Run Flags

1. Passing Arguments

If you need to pass additional command-line arguments to the underlying script, use -- after the task name:

# The --fix argument will be passed directly to the lint script
unirtm run lint -- --fix

2. Dry-run Mode

When debugging a complex task graph, you might only want to verify the execution order without actually running dangerous commands. You can utilize the dry-run mode:

unirtm run deploy --dry-run

This prints the sequentially ordered execution plan alongside the actual shell commands that would be run.

3. Concurrency Limits

By default, UniRTM determines the maximum number of parallel threads based on your machine's CPU core count. You can forcefully override this behavior:

unirtm run build --jobs 2