gulp-strip-comments

decomment - removes comments from JSON, JavaScript, CSS, HTML, etc.

Features

• Removes both single and multi-line comments from JSON, JavaScript and CSS/Text
• Automatically recognizes HTML and removes all <!-- comments --> from it
• Does not change layout / formatting of the original document
• Removes lines that have only comments on them
• Compatible with CSS3, JSON5 and ECMAScript 6

The library does not support mixed content - HTML with JavaScript or CSS in it. Once the input code is recognized as HTML, only the HTML comments will be removed from it.

Performance

For JSON and JavaScript this library uses [esprima] to guarantee correct processing for regular expressions.

As an example, it can process AngularJS 1.5 Core in under 100ms, which is 1.1MB ~ 30,000 lines of JavaScript.

Install

$ npm install --save-dev gulp-strip-comments

Usage

var gulp = require("gulp");
var strip = require("gulp-strip-comments");

gulp.task("default", function () {
  return gulp.src("template.js").pipe(strip()).pipe(gulp.dest("dist"));
});

API

See decomment for examples and more information.

let strip = require('gulp-strip-comments') // == decomment

strip(options)

This method first checks if the code starts with <, as an HTML, and if so, all <!-- comment --> entries are removed, according to the options.

When the code is not recognized as HTML, it is assumed to be either JSON or JavaScript. It is then parsed through [esprima] for ECMAScript 6 compliance, and to extract details about regular expressions.

If [esprima] fails to validate the code, it will throw a parsing error. When successful, this method will remove // and /**/ comments according to the options (see below).

options.safe ⇒ Boolean

• false (default) - remove all multi-line comments
• true - keep special multi-line comments that begin with:
  • <!--[if - for conditional comments in HTML
  • /*! - for everything else (other than HTML)

options.ignore ⇒ RegExp | [RegExp,...]

Takes either a single or an array of regular expressions to match against. All matching blocks are then skipped, as well as any comment-like content inside them.

Examples:

• CSS may contain Base64-encoded strings with comment-like symbols:

src: url(data:font/woff;base64,d09GRg//ABAAAAAAZ);

You can isolate all url(*) blocks by using:

{
  ignore: /url\([\w\s:\/=\-\+;,]*\)/g;
}

• If you want to isolate jsDoc blocks (start with /**, followed by a line break, end with */), you can use the following:

{
  ignore: /\/\*\*\s*\n([^\*]|(\*(?!\/)))*\*\//g;
}

options.space ⇒ Boolean

• false (default) - remove comment blocks entirely
• true - replace comment blocks with white spaces where needed, in order to preserve the original line + column position of every code element.

NOTE: When this option is enabled, option trim is ignored.

options.trim ⇒ Boolean

• false (default) - do not trim comments
• true - remove empty lines that follow removed full-line comments

NOTE: This option has no effect when option space is enabled.

options.tolerant ⇒ Boolean

• false (default) - perform strict JavaScript parsing (parser throws on invalid JavaScript)
• true - pass tolerant flag to [esprima] parser (the parser may choose to continue parsing and produce a syntax tree). Usefull for parsing Angular/TypeScript code, for example.

strip.text(options)

Unlike the default strip method, it instructs the library that text is not a JSON, JavaScript or HTML, rather a plain text that needs no parsing or validation, only to remove // and /**/ comments from it according to the options.

This method is good for any text file that uses syntax // and /**/ for comments, such as: .CSS, .CPP, .H, etc.

Please note that while the same rules apply for the text blocks ('', "" and ``), you should not use this method for JSON or JavaScript, as it can break your regular expressions.

strip.html(options)

Unlike the default strip method, it instructs the library not to parse or validate the input in any way, rather assume it to be HTML, and remove all <!-- comment --> entries from it according to the options.

strip.getEOL()

Returns End-of-Line string used within the text, based on the occurrence frequency:

• \n - for Unix-encoded text
• \r\n - for Windows-encoded text

When impossible to conclude (the same or 0 occurrence), it returns the default End-of-Line for the current OS.