Linter

Roughly includes a built-in linter to catch errors and enforce consistent coding practices in your R code.

Usage

Lint your R files using the command line:

roughly check              # Check all files in the current directory
roughly check <path>       # Check all files in <path>

Current Checks

The linter is divided into two main categories of checks:

Syntax Checks

Syntax checks identify structural problems in your code that would cause R to raise errors:

• Missing delimiters: Detects unclosed parentheses, brackets, and braces
• Missing operands: Identifies binary operators missing their right-hand side operand
• Missing function bodies: Finds function definitions without an implementation
• Unexpected delimiters: Warns about unexpected closing delimiters like }, ), or ]
• General syntax errors: Catches other syntax issues that would prevent code execution

Example:

# Missing closing bracket
foo <- function(x {
  x + 1
}

# Missing right-hand side
y <-

Semantics Checks

Semantics checks enforce coding conventions and best practices that won’t necessarily cause errors but may lead to bugs or reduce code quality:

• Variable naming convention: Enforces consistent naming style (snake_case or camelCase) according to your configuration
• Parameter naming convention: Ensures function parameters follow the same naming convention
• Assignment operator: Recommends using <- rather than = for variable assignment
• Trailing commas: Flags unnecessary trailing commas in function calls
• Booleans values: Use TRUE and FALSE over T and F

Example:

# Using = instead of <- for assignment (warning)
x = 10

# Inconsistent naming convention (warning)
calculate_mean <- function(dataSet) {
  # ...
}

# Trailing comma (warning)
result <- sum(1, 2, 3,)

Configuration

Linter behavior can be customized via the roughly.toml configuration file:

# Control the naming convention for variables and parameters
case = "snake_case"  # or "camelCase"

Roadmap

Roughly’s static analysis capabilities will expand in future versions to include:

• Argument validation: Checking that function calls provide the correct number of arguments
• Type checking: Basic type inference and validation for function arguments and returns
• Unused variables: Detection of unused variable declarations
• Unreachable code: Identification of code that will never be executed
• Control flow analysis: Detecting potential infinite loops or missing return statements
• Package-specific rules: Special rules for popular packages like dplyr, ggplot2, and data.table

These advanced static analysis features are currently in development and will be added in future releases.

Integration

Roughly’s linter integrates seamlessly with its language server, providing real-time feedback in compatible editors like VS Code. Diagnostics appear as you type, with appropriate highlighting and hover information.