eslint/lib/rules/no-trailing-spaces.js

/**
 * @fileoverview Disallow trailing spaces at the end of lines.
 * @author Nodeca Team <https://github.com/nodeca>
 * @copyright 2015 Greg Cochard
 */
"use strict";

//------------------------------------------------------------------------------
// Rule Definition
//------------------------------------------------------------------------------

module.exports = function(context) {

    var BLANK_CLASS = "[ \t\u00a0\u2000-\u200b\u2028\u2029\u3000]",
        SKIP_BLANK = "^" + BLANK_CLASS + "*$",
        NONBLANK = BLANK_CLASS + "+$";

    var options = context.options[0] || {},
        skipBlankLines = options.skipBlankLines || false;

    /**
     * Report the error message
     * @param {ASTNode} node node to report
     * @param {int[]} location range information
     * @param {int[]} fixRange Range based on the whole program
     * @returns {void}
     */
    function report(node, location, fixRange) {
        // Passing node is a bit dirty, because message data will contain
        // big text in `source`. But... who cares :) ?
        // One more kludge will not make worse the bloody wizardry of this plugin.
        context.report({
            node: node,
            loc: location,
            message: "Trailing spaces not allowed.",
            fix: function(fixer) {
                return fixer.removeRange(fixRange);
            }
        });
    }


    //--------------------------------------------------------------------------
    // Public
    //--------------------------------------------------------------------------

    return {

        "Program": function checkTrailingSpaces(node) {

            // Let's hack. Since Espree does not return whitespace nodes,
            // fetch the source code and do matching via regexps.

            var src = context.getSource(),
                re = new RegExp(NONBLANK),
                skipMatch = new RegExp(SKIP_BLANK),
                matches, lines = src.split(/\r?\n/),
                linebreaks = context.getSource().match(/\r\n|\r|\n|\u2028|\u2029/g),
                location,
                totalLength = 0,
                fixRange = [];

            for (var i = 0, ii = lines.length; i < ii; i++) {
                matches = re.exec(lines[i]);

                // Always add linebreak length to line length to accommodate for line break (\n or \r\n)
                // Because during the fix time they also reserve one spot in the array.
                // Usually linebreak length is 2 for \r\n (CRLF) and 1 for \n (LF)
                var linebreakLength = linebreaks && linebreaks[i] ? linebreaks[i].length : 1;
                var lineLength = lines[i].length + linebreakLength;

                if (matches) {

                    // If the line has only whitespace, and skipBlankLines
                    // is true, don't report it
                    if (skipBlankLines && skipMatch.test(lines[i])) {
                        continue;
                    }
                    location = {
                        line: i + 1,
                        column: matches.index
                    };

                    fixRange = [totalLength + location.column, totalLength + lineLength - linebreakLength];

                    report(node, location, fixRange);
                }

                totalLength += lineLength;
            }
        }

    };
};

module.exports.schema = [
    {
        "type": "object",
        "properties": {
            "skipBlankLines": {
                "type": "boolean"
            }
        },
        "additionalProperties": false
    }
];