NoCodeComments

NoCodeComments is a browser-first code comment stripper. Version 2 replaces the original regex-based implementation with a conservative lexical scanner that understands comment delimiters without blindly modifying quoted strings, template literals, or regular-expression literals.

Supported language families

The language registry currently covers C, C++, C#, Java, JavaScript, TypeScript, Go, Rust, Swift, Kotlin, Scala, Dart, Groovy, PHP, CSS/SCSS/Less, HTML/XML/SVG, Python, Ruby, Perl, R, Shell, PowerShell, YAML, Dockerfiles, SQL, Lua, Haskell, Julia, Nim, MATLAB, Fortran, Lisp-family languages, Erlang, Elixir, Prolog, Pascal, D, Solidity, HCL/Terraform, JSONC, Visual Basic, Batch, Assembly and COBOL. A conservative generic fallback is also available.

The exact language profiles and filename mappings are maintained in src/languages.js. A profile defines line-comment markers, block-comment delimiters, string delimiters, and optional regex/triple-string handling.

This is intentionally not marketed as a mathematically universal parser. Every programming language has its own grammar and some languages have context-sensitive, preprocessor-specific, or dialect-specific comment rules. The scanner therefore uses explicit language profiles and avoids destructive heuristics where possible.

Phase 0 baseline

Phase 0 establishes the current parser and CLI baseline before broader reliability work:

• The current product is a browser-first static application. It is not yet a standalone command-line executable.
• The reusable transformation API is stripComments(code, language) in src/comment-stripper.js.
• The scanner removes configured line and block comments while preserving text inside configured string delimiters.
• Where a profile enables it, regex literals and triple-quoted strings are treated as protected regions.
• Multiline block-comment removal preserves the line breaks contained in the removed comment.
• Language selection can be explicit or inferred from a filename/content detector. Unknown languages use the conservative generic profile.
• The generic profile is a fallback, not a guarantee of language correctness.
• The transformation is text-based and does not build a full language AST. Syntax that requires semantic or grammar-aware parsing remains outside the scanner's guarantees.
• Malformed or unterminated strings/comments are handled conservatively by consuming through the end of the available input rather than attempting speculative recovery.
• The tool does not execute, compile, or validate the processed source. Syntax validity remains the responsibility of later validation/testing phases.

See docs/phase-0-parser-cli-baseline.md for the supported transformation matrix, safety assumptions, and parser boundaries.

Features

• Removes line and block comments with a conservative lexical scanner.
• Preserves comment-like text inside strings and template literals.
• Preserves JavaScript/PHP/Perl-style regular-expression literals where they can be identified conservatively.
• Preserves line endings when multiline comments are removed.
• Supports nested block comments for language families that use them.
• Auto-detects languages from uploaded filenames and common source-code content.
• Drag-and-drop file loading.
• Copy input or output with one click.
• Swap the processed output back into the input editor.
• Download stripped code with the original filename when a file was opened.
• Optional blank-line collapsing and long-line wrapping controls.
• Live character and line-count statistics.
• Dark mode for comfortable editing.
• Keyboard shortcuts: Ctrl+Enter reprocesses, Ctrl+Shift+S swaps input/output, Ctrl+D downloads the output, and Ctrl+L clears the editor.
• Keeps the original vanilla JavaScript implementation intact under /legacy/.
• Adds automated Node.js tests covering representative language families and edge cases.

Legacy version

The original implementation is intentionally preserved at /legacy/. The new application exposes a Legacy version button at the bottom of the main page.

Testing

Requires Node.js 20 or newer.

npm test

The test suite focuses on the most failure-prone behavior: comment markers inside strings, regex literals, multiline comments, multiple language syntaxes, and code with no comments.

Static deployment

There is no build step. The project can be served as a static site, including through GitHub Pages.