| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107 |
- /*
- * tex_highlight_rules.js
- *
- * Copyright (C) 2009-11 by RStudio, Inc.
- *
- * The Initial Developer of the Original Code is
- * Ajax.org B.V.
- * Portions created by the Initial Developer are Copyright (C) 2010
- * the Initial Developer. All Rights Reserved.
- *
- * This program is licensed to you under the terms of version 3 of the
- * GNU Affero General Public License. This program is distributed WITHOUT
- * ANY EXPRESS OR IMPLIED WARRANTY, INCLUDING THOSE OF NON-INFRINGEMENT,
- * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Please refer to the
- * AGPL (http://www.gnu.org/licenses/agpl-3.0.txt) for more details.
- *
- */
- define(function(require, exports, module) {
- "use strict";
- var oop = require("../lib/oop");
- var lang = require("../lib/lang");
- var TextHighlightRules = require("./text_highlight_rules").TextHighlightRules;
- var TexHighlightRules = function(textClass) {
- if (!textClass)
- textClass = "text";
- // regexp must not have capturing parentheses. Use (?:) instead.
- // regexps are ordered -> the first match is used
- this.$rules = {
- "start" : [
- {
- token : "comment",
- regex : "%.*$"
- }, {
- token : textClass, // non-command
- regex : "\\\\[$&%#\\{\\}]"
- }, {
- token : "keyword", // command
- regex : "\\\\(?:documentclass|usepackage|newcounter|setcounter|addtocounter|value|arabic|stepcounter|newenvironment|renewenvironment|ref|vref|eqref|pageref|label|cite[a-zA-Z]*|tag|begin|end|bibitem)\\b",
- next : "nospell"
- }, {
- token : "keyword", // command
- regex : "\\\\(?:[a-zA-z0-9]+|[^a-zA-z0-9])"
- }, {
- // Obviously these are neither keywords nor operators, but
- // labelling them as such was the easiest way to get them
- // to be colored distinctly from regular text
- token : "paren.keyword.operator",
- regex : "[[({]"
- }, {
- // Obviously these are neither keywords nor operators, but
- // labelling them as such was the easiest way to get them
- // to be colored distinctly from regular text
- token : "paren.keyword.operator",
- regex : "[\\])}]"
- }, {
- token : textClass,
- regex : "\\s+"
- }
- ],
- // This mode is necessary to prevent spell checking, but to keep the
- // same syntax highlighting behavior. The list of commands comes from
- // Texlipse.
- "nospell" : [
- {
- token : "comment",
- regex : "%.*$",
- next : "start"
- }, {
- token : "nospell." + textClass, // non-command
- regex : "\\\\[$&%#\\{\\}]"
- }, {
- token : "keyword", // command
- regex : "\\\\(?:documentclass|usepackage|newcounter|setcounter|addtocounter|value|arabic|stepcounter|newenvironment|renewenvironment|ref|vref|eqref|pageref|label|cite[a-zA-Z]*|tag|begin|end|bibitem)\\b"
- }, {
- token : "keyword", // command
- regex : "\\\\(?:[a-zA-z0-9]+|[^a-zA-z0-9])",
- next : "start"
- }, {
- token : "paren.keyword.operator",
- regex : "[[({]"
- }, {
- token : "paren.keyword.operator",
- regex : "[\\])]"
- }, {
- token : "paren.keyword.operator",
- regex : "}",
- next : "start"
- }, {
- token : "nospell." + textClass,
- regex : "\\s+"
- }, {
- token : "nospell." + textClass,
- regex : "\\w+"
- }
- ]
- };
- };
- oop.inherits(TexHighlightRules, TextHighlightRules);
- exports.TexHighlightRules = TexHighlightRules;
- });
|