Comprehensive Guide to SwiftLint: Working Principles, Configuration, Custom Rules, and Performance Optimization

SwiftLint is a command‑line tool that checks Swift code style and conventions by leveraging static analysis techniques such as lexical, syntactic, control‑flow, and data‑flow analysis. It integrates with the Swift compiler via Hook Clang and SourceKit to generate an Abstract Syntax Tree (AST) for precise diagnostics.

The tool’s workflow consists of several compilation stages: Parse creates a raw AST, Sema performs semantic analysis and injects warnings, SilLGen generates intermediate SIL, and IRGen produces LLVM IR, which is then optimized and emitted as object files.

Configuration is done through a .swiftlint.yml file where developers can enable or disable rules, set severity levels, and define custom rule parameters. Example configuration snippets:

disabled_rules:
  - colon
  - comma
opt_in_rules:
  - empty_count
included:
  - Source
excluded:
  - Pods
  - Carthage

Built‑in rules are divided into default and opt‑in categories. Default rules are active without a config file, while opt‑in rules must be explicitly enabled. Rules can be customized for warning or error severity, and many support parameterized thresholds (e.g., line_length: 110 ).

Custom rules can be added using regular expressions and syntax‑kind filters. A sample custom rule to detect deprecated UIWebView usage:

custom_rules:
  uiwebview_deprecated:
    regex: "(UIWebView)"
    match_kinds:
      - comment
      - identifier
    message: "UIWebView is deprecated, use WKWebView instead."
    severity: error

To avoid full‑project scans on every build, the article proposes a Git‑diff‑based incremental linting script that runs SwiftLint only on staged, unstaged, or newly added Swift files. The script also checks file extensions before invoking linting:

if [[ "${filename##*.}" == "swift" ]]; then
  swiftlint lint "${filename}"
  swiftlint --fix --config .swiftlint.auto.yml "${filename}"
fi

Integrating this script into Xcode’s build phases with input and output file lists dramatically reduces linting time—from ~20 seconds for a 200k‑line codebase to 1–5 seconds—by preventing unnecessary executions.

The article concludes that SwiftLint, when properly configured and optimized, can greatly improve code quality, reduce code‑review effort, and increase development efficiency in iOS projects.