Skip to content

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
Newer
Older
100755 730 lines (688 sloc) 40.492 kb
ca120a7 @douglascrockford first commit
authored
1 <html>
2 <head>
3 <title>JSLint: The JavaScript Code Quality Tool</title>
4 <link rel="icon" type="image/gif" href="http://www.JSLint.com/favicon.gif">
5 <style>
6 body {
7 background-color: linen;
8 margin-left: 8%;
9 margin-right: 8%;
10 }
11 pre {margin-left: 40px;}
12 table {margin: 10px; border: 0px;}
13 th, td {border: black solid 1pt; padding-left: 10px;
14 padding-right: 10px; vertical-align: top;}
15 th {
16 background-color: thistle;
17 }
18 td {
19 background-color: white;
20 }
21 #top table {margin: 0px;}
22 #top td {background-color: linen; border: 0pt; vertical-align: middle;}
23 ul {list-style-type: square;}
24 input[type="button"] {border: 2px solid black;}
25 a:link {color: darkblue;}
26 a:visited {color: purple;}
27 a:hover {color: blue; text-decoration: underline;}
28 a:active {color: red;}
29 </style>
30 </head>
31 <body bgcolor="gainsboro">
32 <table id="top" border="0">
33 <tr>
34 <td><img src="jslint.gif" width="383" height="120" alt="JSLint"> </td>
35 <td>
36 <p><big><code>JSLint</code>: The
37 <a href="http://javascript.crockford.com/">JavaScript</a> Code Quality Tool</big></p>
38 <p><a href="http://www.crockford.com/" target="_top">&copy;2002 Douglas Crockford</a></p>
39 </td>
40 </tr>
41 </table>
42 <br clear="all">
43 <h2 id=warning>Warning!</h2>
44 <p><a href="http://www.JSLint.com" target="_blank"><code>JSLint</code></a>
45 will hurt your feelings.</p>
46 <h2 id=what>What is <code>JSLint</code>?</h2>
47
48 <p><a href="http://www.JSLint.com" target="_blank"><code>JSLint</code></a>
49 is a JavaScript program that looks for problems in JavaScript programs.
50 It is a code quality tool.</p>
51
52 <p>When <a href="http://en.wikipedia.org/wiki/C_programming_language">C</a>
53 was a <a href="http://cm.bell-labs.com/cm/cs/who/dmr/chist.html">young</a>
54 programming language, there were several common programming errors that
55 were not caught by the primitive compilers, so an accessory program called
56 <code><a href="http://en.wikipedia.org/wiki/Lint_programming_tool">lint</a></code>
57 was developed that would scan a source file, looking for problems.</p>
58
59 <p>As the language matured, the definition of the language was
60 strengthened to eliminate some insecurities, and compilers got better
61 at issuing warnings. <code>lint</code> is no longer needed.</p>
62
63 <p><a href="http://javascript.crockford.com/">JavaScript</a> is a young-for-its-age
64 language. It was originally intended to do small tasks in webpages, tasks
65 for which Java was too heavy and clumsy. But JavaScript is a very capable
66 language, and it is now being used in larger projects. Many of the features
6735394 @douglascrockford Cleanup.
authored
67 that were intended to make the language easy to use are troublesome when projects become complicated. A <code>lint</code> for JavaScript is needed: <a href="http://www.JSLint.com/"><code>JSLint</code></a>,
ca120a7 @douglascrockford first commit
authored
68 a JavaScript syntax checker and validator.</p>
69
70 <p><code>JSLint</code> takes a JavaScript source and scans it. If it finds
71 a problem, it returns a message describing the problem and an approximate
72 location within the source. The problem is not necessarily a syntax error,
73 although it often is. <code>JSLint</code> looks at some style conventions
74 as well as structural problems. It does not prove that your program is
75 correct. It just provides another set of eyes to help spot problems.</p>
76
77 <p><code>JSLint</code> defines a professional subset of JavaScript, a stricter
78 language than that defined by <a href="http://www.ecma-international.org/publications/standards/Ecma-262.htm" target="ecma">Third
79 Edition of the <i>ECMAScript Programming Language Standard</i></a>. The
80 subset is related to recommendations found in <a href="http://javascript.crockford.com/code.html" target="sun"><i>Code
81 Conventions for the JavaScript Programming Language</i></a>. </p>
82 <p>JavaScript is a sloppy language, but inside it there is an elegant, better
83 language. <code>JSLint</code> helps you to program in that better language
6735394 @douglascrockford Cleanup.
authored
84 and to avoid most of the slop. JSLint will reject programs that browsers will accept because JSLint is concerned with the quality of your code and browsers are not. You should accept all of JSLint's advice.</p>
85 <p><code>JSLint</code> can operate on JavaScript source, HTML source, CSS source, or <a href="http://www.JSON.org/">JSON</a>
ca120a7 @douglascrockford first commit
authored
86 text.</p>
87 <h2 id=global>Global Variables</h2>
88 <p>JavaScript's <a href="http://yuiblog.com/blog/2006/06/01/global-domination/">biggest
89 problem</a> is its dependence on global variables, particularly implied
90 global variables. If a variable is not explicitly declared (usually with
91 the <code>var</code> statement), then JavaScript assumes that the variable
92 was global. This can mask misspelled names and other problems.</p>
93 <p><code>JSLint</code> expects that all variables and functions are declared
94 before they are used or invoked. This allows it to detect implied global
95 variables. It is also good practice because it makes programs easier to
96 read.</p>
97 <p>Sometimes a file is dependent on global variables and functions that
98 are defined elsewhere. You can identify these to <code>JSLint</code> with a <code>var</code> statement that lists the global functions and objects
99 that your program depends on. </p>
100 <p>A global declaration can look like this:</p>
101 <pre>var getElementByAttribute, breakCycles, hanoi;</pre>
9f0303c @douglascrockford appear
authored
102 <p>The declaration should appear near the top of the file. It must appear before the use of the variables
ca120a7 @douglascrockford first commit
authored
103 it declares. </p>
104 <p>It is necessary to use a <code>var</code> statement to declare a variable before that variable is assigned to. </p>
44eec74 @douglascrockford Adrian
authored
105 <p><code>JSLint</code> also recognizes a <code>/*global */</code> directive that can indicate to <code>JSLint</code> that variables used in this file were defined in other files. The comment can contain a comma separated list of names. Each name can optionally be followed by a colon and either <code>true</code> or <code>false</code>, <code>true</code> indicating that the variable may be assigned to by this file, and <code>false</code> indicating that assignment is not allowed (which is the default). The directive respects function scope.</p>
ca120a7 @douglascrockford first commit
authored
106 <p id=browser>Some globals can be predefined for you. Select the <i>Assume
107 a browser</i> (<code>browser</code>) <a href="#options">option</a> to
108 predefine the standard global properties that are supplied by web browsers,
109 such as <code>document</code> and <code>addEventListener</code>. It has the same
110 effect as this comment:</p>
111 <blockquote>
df39101 @douglascrockford option.browser
authored
112 <code>/*global
e752a55 @douglascrockford option.node
authored
113 clearInterval: false, clearTimeout: false, document: false, event: false, frames: false, history: false, Image: false, location: false, name: false, navigator: false, Option: false, parent: false, screen: false, setInterval: false, setTimeout: false, window: false, XMLHttpRequest: false
df39101 @douglascrockford option.browser
authored
114 */</code></blockquote>
e752a55 @douglascrockford option.node
authored
115 <p>Select the
116 <em>Assume console, alert, ...</em>
ca120a7 @douglascrockford first commit
authored
117 (<code>devel</code>) <a href="#options">option</a> to predefine globals that are useful in development but that should be avoided in production, such as <code>console</code> and <code>alert</code>. It has the same
118 effect as this comment:</p>
119 <pre>/*global alert: false, confirm: false, console: false, Debug: false, opera: false, prompt: false */</pre>
706c485 @douglascrockford option.node
authored
120 <p id=node>Select the
e752a55 @douglascrockford option.node
authored
121 <em>Assume Node.js</em>
706c485 @douglascrockford option.node
authored
122 (<code>node</code>) <a href="#options">option</a> to predefine globals that are used in the Node.js environment<code></code>. It has the same
123 effect as this comment:</p>
8d035d3 @douglascrockford e
authored
124 <pre>/*global Buffer: false, clearInterval: false, clearTimeout: false, console: false, global: false, module: false, process: false, querystring: false, require: false, setInterval: false, setTimeout: false, util: false, __filename: false, __dirname: false */</pre>
ca120a7 @douglascrockford first commit
authored
125 <p id=rhino>Select the <i>Assume Rhino</i> (<code>rhino</code>) <a href="#options">option</a>
706c485 @douglascrockford option.node
authored
126 to predefine the global properties provided by the Rhino environment.
ca120a7 @douglascrockford first commit
authored
127 It has the same effect as this statement:</p>
128 <blockquote>
129 <code>/*global defineClass: false, deserialize: false, gc: false, help: false, load: false, loadClass: false, print: false, quit: false, readFile: false, readUrl: false, runCommand: false, seal: false, serialize: false, spawn: false, sync: false, toint32: false, version: false */ </code>
130 </blockquote>
131 <p id=widget>Select the <i>Assume a Yahoo Widget</i> (<code>widget</code>)
132 <a href="#options">option</a> to predefine the global properties provided
133 by the Yahoo! Widgets environment. It has the same effect as this statement:</p>
134 <blockquote>
135 <code>/*global alert: true, animator: true, appleScript: true, beep: true, bytesToUIString: true, Canvas: true, chooseColor: true, chooseFile: true, chooseFolder: true, closeWidget: true, COM: true, convertPathToHFS: true, convertPathToPlatform: true, CustomAnimation: true, escape: true, FadeAnimation: true, filesystem: true, Flash: true, focusWidget: true, form: true, FormField: true, Frame: true, HotKey: true, Image: true, include: true, isApplicationRunning: true, iTunes: true, konfabulatorVersion: true, log: true, md5: true, MenuItem: true, MoveAnimation: true, openURL: true, play: true, Point: true, popupMenu: true, preferenceGroups: true, preferences: true, print: true, prompt: true, random: true, Rectangle: true, reloadWidget: true, ResizeAnimation: true, resolvePath: true, resumeUpdates: true, RotateAnimation: true, runCommand: true, runCommandInBg: true, saveAs: true, savePreferences: true, screen: true, ScrollBar: true, showWidgetPreferences: true, sleep: true, speak: true, Style: true, suppressUpdates: true, system: true, tellWidget: true, Text: true, TextArea: true, Timer: true, unescape: true, updateNow: true, URL: true, Web: true, widget: true, Window: true, XMLDOM: true, XMLHttpRequest: true, yahooCheckLogin: true, yahooLogin: true, yahooLogout: true */</code>
136 </blockquote>
137 <p id=windows>Select the <i>Assume Windows</i> (<code>windows</code>)
138 <a href="#options">option</a> to predefine the global properties provided by Microsoft Windows. It has the same effect as this statement:</p>
139 <blockquote>
140 <p><code>/*global ActiveXObject: false, CScript: false, Debug: false, Enumerator: false, System: false, VBArray: false, WScript: false */</code></p>
141 </blockquote>
142 <h2 id=semicolon>Semicolon</h2>
6735394 @douglascrockford Cleanup.
authored
143 <p>JavaScript uses a C-like syntax which requires the use of semicolons to delimit certain
144 statements. JavaScript attempts to make those semicolons optional with a semicolon
145 insertion mechanism. This is dangerous because it can mask errors.</p>
ca120a7 @douglascrockford first commit
authored
146 <p>Like C, JavaScript has <code>++</code> and <code>--</code> and <code>(</code> operators
147 which can be prefixes or suffixes. The disambiguation is done by the semicolon.</p>
148 <p>In JavaScript, a linefeed can be whitespace or it can act as a semicolon.
149 This replaces one ambiguity with another. </p>
150 <p><code>JSLint</code> expects that every statement be followed by <code>;</code> except
151 for <code>for</code>, <code>function</code>, <code>if</code>, <code>switch</code>, <code>try</code>, and
152 <code>while</code>. <code>JSLint</code> does not expect to see unnecessary semicolons or the
153 empty statement.</p>
154 <h2 id=comma>Comma</h2>
155 <p>The comma operator can lead to excessively tricky expressions. It can also
156 mask some programming errors.</p>
157 <p><code>JSLint</code> expects to see the comma used as a separator, but not as an
158 operator (except in the initialization and incrementation parts of the <code>for</code>
159 statement). It does not expect to see elided elements in array literals. Extra
160 commas should not be used. A comma should not appear after the last element
161 of an array literal or object literal because it can be misinterpreted by some
162 browsers. </p>
163 <h2 id=scope>Scope</h2>
164
165 <p>In many languages, a block introduces a scope. Variables introduced in
166 a block are not visible outside of the block.</p>
167
168 <p>In JavaScript, blocks do not introduce a scope. There is only function-scope.
169 A variable introduced anywhere in a function is visible everywhere in
170 the function. JavaScript's blocks confuse experienced programmers and
171 lead to errors because the familiar syntax makes a false promise.</p>
172
173 <p><code>JSLint</code> expects blocks with <code>function</code>, <code>if</code>,
174 <code>switch</code>, <code>while</code>, <code>for</code>, <code>do</code>,
175 and <code>try</code> statements and nowhere else. </p>
176 <p>In languages with block scope, it is usually recommended that variables
177 be declared at the site of first use. But because JavaScript does not
178 have block scope, it is wiser to declare all of a function's variables
179 at the top of the function. It is recommended that a single <code>var</code>
a27d17c @douglascrockford Tolerate
authored
180 statement be used per function. This can be declined with the <code>vars</code>
ca120a7 @douglascrockford first commit
authored
181 <a href="#options">option</a>.</p>
182
183 <h2 id=required>Required Blocks</h2>
184
185 <p><code>JSLint</code> expects that <code>if</code>, <code>while</code>,
186 <code>do</code> and <code>for</code> statements will be made with blocks
187 <code>{</code>that is, with statements enclosed in braces<code>}</code>.</p>
188
189 <p>JavaScript allows an <code>if</code> to be written like this:</p>
190
191 <pre>if (<i>condition</i><code>)
192 </code><i>statement</i>;</pre>
193
194 <p>That form is known to contribute to mistakes in projects where many programmers
195 are working on the same code. That is why <code>JSLint</code> expects the use of
196 a block:</p>
197
198 <pre>if (<i>condition</i>) {
199 <i>statements</i>;
200 }</pre>
201
202 <p>Experience shows that this form is more resilient.</p>
203
204 <h2 id=expression>Expression Statements</h2>
205 <p>An expression statement is expected to be an assignment or a function/method
206 call or <code>delete</code>. All other expression statements are considered
207 to be errors.</p>
65313e2 @douglascrockford option.confusion
authored
208 <h2 id="confusion">Type Confusion</h2>
209 <p>JSLint can do type inference. It can report cases were variables and properties are
210 used to house multiple types. The warning is <code>Type confusion:</code> {a} <code>and</code> {b}<code>.</code> where the {a} and {b} will be
211 replaced with the names of types.</p>
212 <p> It is usually easy to see what caused the
213 warning. In some cases, it can be very puzzling. In the puzzling cases, try
214 initializing your vars with typed values. For example, if you expect that <code>n</code> will
215 contain numbers, then write</p>
216 <pre>var n = 0;</pre>
217 <p>That should produce clearer warnings. </p>
218 <p>Type confusion is not necessarily an error, particularly in a language that
219 provides as much type freedom as this one does. But some inconsistencies are
220 errors, so type discipline might be something to consider adding to your
221 programming style. Also, the fastest JavaScript engines will slow down in the
222 presence of type confusion.
223
224 To turn off these warnings, turn on the <i>Tolerate type confusion</i> <a href="#options">option</a>. </p>
ca120a7 @douglascrockford first commit
authored
225 <h2 id=forin><code>for</code> <code>in</code></h2>
226 <p>The <code>for</code> <code>in</code> statement allows for looping through
227 the names of all of the properties of an object. <a href="http://yuiblog.com/blog/2006/09/26/for-in-intrigue/">Unfortunately,
d4189ff @douglascrockford properties
authored
228 it also loops through all of the properties that were inherited through
ca120a7 @douglascrockford first commit
authored
229 the prototype chain.</a> This has the bad side effect of serving up method
d4189ff @douglascrockford properties
authored
230 functions when the interest is in data properties. If a program is written without awareness of this situation, then it can fail.</p>
ca120a7 @douglascrockford first commit
authored
231 <p>The body of every <code>for</code> <code>in</code> statement should be
232 wrapped in an <code>if</code> statement that does filtering. It can select
233 for a particular type or range of values, or it can exclude functions,
234 or it can exclude properties from the prototype. For example,</p>
235 <pre>for (name in object) {
236 if (object.hasOwnProperty(name)) {
237 ....
238 }
239
240 }</pre>
241
242 <h2 id=switch><code>switch</code></h2>
243 <p>A <a href="http://yuiblog.com/blog/2007/04/25/id-rather-switch-than-fight/">common
244 error</a> in <code>switch</code> statements is to forget to place a <code>break</code>
245 statement after each case, resulting in unintended fall-through. <code>JSLint</code>
246 expects that the statement before the next <code>case</code> or <code>default</code>
247 is one of these: <code>break</code>, <code>return</code>, or <code>throw</code>.
248 </p>
249 <h2 id=var><code>var</code></h2>
250
251 <p>JavaScript allows <code>var</code> definitions to occur anywhere
252 within a function. <code>JSLint</code> is more strict.</p>
253
254 <p><code>JSLint</code> expects that a <code>var</code> will be declared
255 only once, and that it will be declared before it is used.</p>
256 <p><code></code><code>JSLint</code> expects that a <code>function</code>
257 will be declared before it is used.</p>
258 <p><code>JSLint</code> expects that parameters will not also be declared
259 as vars. </p>
260
261 <p><code>JSLint</code> does not expect the <code>arguments</code> array to be declared
262 as a <code>var</code>.</p>
263 <p><code>JSLint</code> does not expect that a var will be defined in a block.
264 This is because JavaScript blocks do not have block scope. This can have
265 unexpected consequences. Define all variables at the top of the function.</p>
266
267 <h2 id=with><code>with</code></h2>
268
269 <p>The <code>with</code> statement was intended to provide a shorthand in accessing
d4189ff @douglascrockford properties
authored
270 properties in deeply nested objects. Unfortunately, it behaves <a href="http://yuiblog.com/blog/2006/04/11/with-statement-considered-harmful/">very
271 badly</a> when setting new properties. Never use the <code>with</code> statement. Use
ca120a7 @douglascrockford first commit
authored
272 a <code>var</code> instead.</p>
273
274 <p><code>JSLint</code> does not expect to see a <code>with</code> statement.</p>
275
276 <h2 id=assignment>=</h2>
277 <p><code>JSLint</code> does not expect to see an assignment statement in
278 the condition part of an <code>if</code> or <code>for</code> or <code>while</code>
279 <code></code> or <code>do</code> statement. This is because it is more
280 likely that </p>
281 <pre>if (a = b) {
282 ...
283 }</pre>
284 <p>was intended to be </p>
285 <pre>if (a == b) {
286 ...
287 }</pre>
288 <p>It is difficult to write correct programs while using idioms that are
9d80d80 @douglascrockford asignment unwrapped
authored
289 hard to distinguish from obvious errors.</p>
ba257de @douglascrockford eqeq
authored
290 <h2 id=eqeq>== and !=</h2>
ca120a7 @douglascrockford first commit
authored
291 <p>The <code>==</code> and <code>!=</code> operators do type coercion before
292 comparing. This is bad because it causes <code>' \t\r\n' == 0</code> to
ca4cfb5 @douglascrockford Adrian
authored
293 be <code>true</code>. This can mask type errors. JSLint cannot reliably determine if == is being used correctly, so it is best to not use <code>==</code> and != and to always use the more reliable <code>===</code> and <code>!==</code> operators instead. </p>
ca120a7 @douglascrockford first commit
authored
294 <p align="left">If you only care that a value is <i>truthy</i> or <i>falsy</i>,
295 then use the short form. Instead of </p>
296 <pre align="left">(foo != 0)</pre>
297 <p align="left">just say </p>
298 <pre align="left">(foo)</pre>
299 <p align="left">and instead of</p>
300 <pre align="left">(foo == 0)</pre>
301 <p align="left"> say</p>
302 <pre align="left">(!foo)</pre>
ba257de @douglascrockford eqeq
authored
303 <p>There is an <code>eqeq</code> <a href="#options">option</a> that allows the use of <code>==</code> and <code>!=</code>.</p>
ca120a7 @douglascrockford first commit
authored
304 <h2 id=labels>Labels</h2>
305 <p>JavaScript allows any statement to have a label, and labels have a
306 separate name space. <code>JSLint</code> is more strict.</p>
307
308 <p><code>JSLint</code> expects labels only on statements that interact
309 with <code>break</code>: <code>switch</code>, <code>while</code>,
310 <code>do</code>, and <code>for</code>. <code>JSLint</code> expects that labels
311 will be distinct from vars and parameters.</p>
312
313 <h2 id=unreachable>Unreachable Code</h2>
314 <p><code>JSLint</code> expects that
315 a <code>return</code>, <code>break</code>, <code>continue</code>,
316 or <code>throw</code> statement will be followed by
317 a <code>}</code> or <code>case</code> or <code>default</code>.</p>
318
319 <h2 id=pluses>Confusing Pluses and Minuses</h2>
320
321 <p><code>JSLint</code> expects that <code>+</code> will not be followed by
322 <code>+</code> or <code>++</code>, and that <code>-</code> will not be followed
323 by <code>-</code> or <code>--</code>. A misplaced space can turn <code>+ +</code> into <code>++</code>, an error that is difficult to see. Use parens to avoid confusion..</p>
324 <h2 id=inc><code>++</code> and <code>--</code></h2>
325 <p>The <code>++</code> <small>(increment)</small> and <code>--</code> <small>(decrement)</small>
326 operators have been known to contribute to bad code by encouraging excessive
327 trickiness. They are second only to faulty architecture in enabling to
44eec74 @douglascrockford Adrian
authored
328 viruses and other security menaces. Also, preincrement/postincrement confusion can produce off-by-one errors that are extremely difficult to diagnose. There is a <code>plusplus</code> <a href="#options">option</a>
a27d17c @douglascrockford Tolerate
authored
329 that allows the use of these operators.</p>
ca120a7 @douglascrockford first commit
authored
330 <h2 id=bitwise>Bitwise Operators</h2>
331 <p>JavaScript does not have an integer type, but it does have bitwise operators.
332 The bitwise operators convert their operands from floating point to integers
333 and back, so they are not as efficient as in C or other languages. They
334 are rarely useful in browser applications. The similarity to the logical
335 operators can mask some programming errors. The <code>bitwise</code> <a href="#options">option</a>
a27d17c @douglascrockford Tolerate
authored
336 allows the use of these operators: <code>&lt;&lt; &gt;&gt; &gt;&gt;&gt;
ca120a7 @douglascrockford first commit
authored
337 ~ &amp; |</code>.</p>
338 <h2 id=evil><code>eval</code> is evil</h2>
339 <p>The <code>eval</code> function (and its relatives, <code>Function</code>,
340 <code>setTimeout</code>, and <code>setInterval</code>) provide access
341 to the JavaScript compiler. This is sometimes necessary, but in most cases
342 it indicates the presence of extremely bad coding. The <code>eval</code>
343 function is the most misused feature of JavaScript.</p>
344
345 <h2 id=void><code>void</code></h2>
346 <p>In most C-like languages, <code>void</code> is a type. In
347 JavaScript, <code>void</code> is a prefix operator that always
348 returns <code>undefined</code>. <code>JSLint</code> does not expect to
349 see <code>void</code> because it is confusing and not very useful.</p>
350
351 <h2 id=regexp>Regular Expressions</h2>
352 <p>Regular expressions are written in a terse and cryptic notation. <code>JSLint</code>
353 looks for problems that may cause portability problems. It also attempts
354 to resolve visual ambiguities by recommending explicit escapement.</p>
355 <p>JavaScript's syntax for regular expression literals overloads the <code>/</code>
356 character. To avoid ambiguity, <code>JSLint</code> expects that the character
357 preceding a regular expression literal is a <code>(</code> or <code>=</code>
358 or <code>:</code> or <code>,</code> character. </p>
359 <h2 id=new>Constructors and <code>new</code></h2>
360 <p>Constructors are functions that are designed to be used with the <code>new</code>
361 prefix. The <code>new</code> prefix creates a new object based on the
362 function's <code>prototype</code>, and binds that object to the function's
363 implied <code>this</code> parameter. If you neglect to use the <code>new</code>
364 prefix, no new object will be made and <code>this</code> will be bound
365 to the global object. This is a <a href="http://yuiblog.com/blog/2006/11/13/javascript-we-hardly-new-ya/">serious
366 mistake</a>.</p>
367 <p><code>JSLint</code> enforces the convention that constructor functions
368 be given names with initial uppercase. <code>JSLint</code> does not expect
369 to see a function invocation with an initial uppercase name unless it
370 has the <code>new</code> prefix. <code>JSLint</code> does not expect to
371 see the <code>new</code> prefix used with functions whose names do not
a27d17c @douglascrockford Tolerate
authored
372 start with initial uppercase. This can be disabled with the <code>newcap</code>
ca120a7 @douglascrockford first commit
authored
373 <a href="#options">option</a>.</p>
374 <p><code>JSLint</code> does not expect to see the wrapper forms <code>new Number</code>,
375 <code>new String</code>, <code>new Boolean</code>. </p>
376 <p><code>JSLint</code> does not expect to see <code>new Object</code> (use <code>{}</code>
377 instead). </p>
378 <p><code>JSLint</code> does not expect to see <code>new Array</code> (use <code>[]</code>
379 instead).</p>
380 <h2 id=unsafe>Unsafe Characters</h2>
381 <p> There are characters that are handled inconsistently in browsers, and
382 so must be escaped when placed in strings. </p>
383 <pre>\u0000-\u001f
384 \u007f-\u009f
385 \u00ad
386 \u0600-\u0604
387 \u070f
388 \u17b4
389 \u17b5
390 \u200c-\u200f
391 \u2028-\u202f
392 \u2060-\u206f
393 \ufeff
394 \ufff0-\uffff</pre>
395 <h2 id=not>Not Looked For</h2>
396
397 <p><code>JSLint</code> does not do flow analysis to determine that variables are assigned
398 values before used. This is because variables are given a value (<code>undefined</code>)
6735394 @douglascrockford Cleanup.
authored
399 that is a reasonable default for many applications.</p>
ca120a7 @douglascrockford first commit
authored
400
401 <p><code>JSLint</code> does not do any kind of global analysis. It does
402 not attempt to determine that functions used with <code>new</code> are
403 really constructors (<a href="#new">except by enforcing capitalization
d4189ff @douglascrockford properties
authored
404 conventions</a>), or that property names are spelled correctly (<a href="#properties">except
405 for matching against the <code>/*properties */</code> comment</a>).</p>
ca120a7 @douglascrockford first commit
authored
406 <h2 id=html>HTML</h2>
407 <p><code>JSLint</code> is able to handle HTML text. It can inspect the JavaScript content
408 contained within <code>&lt;script&gt;</code>...<code>&lt;/script&gt;</code> tags. It
409 also inspects the HTML content, looking for problems that are known to interfere
410 with JavaScript:</p>
411 <ul>
412 <li>All tag names must be in lower case.</li>
413 <li>All tags that can take a close tag (such as <code>&lt;/p&gt;</code>)
414 must have a close tag.</li>
415 <li>All tags are correctly nested.</li>
416 <li>The entity <code>&amp;lt;</code> must be used for literal <code>'&lt;'</code>.</li>
417 </ul>
418 <p><code>JSLint</code> is less anal than the sycophantic conformity demanded
419 by XHTML, but more strict than the popular browsers. </p>
420 <p><code>JSLint</code> also checks for the occurrence of<code> '&lt;/' </code>in
421 string literals. You should always write<code> '&lt;\/' </code>instead.
422 The extra backslash is ignored by the JavaScript compiler but not by the
423 HTML parser. Tricks like this should not be necessary, and yet they are.</p>
424 <p>There is a <code>cap</code> <a href="#options">option</a> that allows
44eec74 @douglascrockford Adrian
authored
425 use of uppercase tag names. There is also an <code>on</code> <a href="#options">option</a>
ca120a7 @douglascrockford first commit
authored
426 that allows the use of inline HTML event handlers.</p>
427 <p>There is a <code>fragment</code> <a href="#options">option</a> that can
428 inspect a well formed HTML fragment. If the <code>adsafe</code> <a href="#options">option</a>
429 is also used, then the fragment must be a <code>&lt;div&gt;</code> that
430 conforms to the <a href="http://www.ADsafe.org/">ADsafe</a> widget rules.</p>
431 <h2 id=css>CSS</h2>
432 <p><code>JSLint</code> can inspect CSS files. It expects the first line
433 of a CSS file to be </p>
434 <pre>@charset &quot;UTF-8&quot;;</pre>
435 <p>This feature is experimental. Please report any problems or limitations.
436 There is a <code>css</code> <a href="#options">option</a> that will tolerate
437 some of the non-standard-but-customary workarounds. <br>
438 </p>
439
440 <h2 id=options>Options</h2>
441 <p><code>JSLint</code> provides several options that control its operation and
442 its sensitivity. In the <a href="http://www.JSLint.com/">web edition</a>, the
a27d17c @douglascrockford Tolerate
authored
443 options are selected with several checkboxes and two fields. </p>
ca120a7 @douglascrockford first commit
authored
444 <p>It also provides assistance in constructing <code>/*jslint*/</code>
445 comments.
446 </p>
447 <p>When <code>JSLINT</code> is called as a function, it accepts an <code>option</code> object
448 parameter that allows you to determine the subset of JavaScript that is
449 acceptable to you. The web page version of <code>JSLint</code> at <a href="http://www.JSLint.com/">http://www.JSLint.com/</a>
450 does this for you. </p>
451 <p>Options can also be specified within a script with a <code>/*jslint */</code>
475768b @douglascrockford /*properties*/ and /*global*/ now respect function scope
authored
452 directive:</p>
ca120a7 @douglascrockford first commit
authored
453 <pre>/*jslint nomen: true, debug: true,
a27d17c @douglascrockford Tolerate
authored
454 evil: false, vars: true */</pre>
ca120a7 @douglascrockford first commit
authored
455 <p>An option specification starts with <code>/*jslint</code>. Notice that
456 there is no space before the <code>j</code>. The specification contains
457 a sequence of name value pairs, where the names are <code>JSLint</code>
458 options, and the values are <code>true</code> or <code>false</code>. The
459 <code>indent</code> <a href="#options">option</a> can take a number. A <code>/*jslint */</code>
475768b @douglascrockford /*properties*/ and /*global*/ now respect function scope
authored
460 comment takes precedence over the <code>option</code> object. The directive respects function scope.</p>
ca120a7 @douglascrockford first commit
authored
461 <table>
462 <tbody>
463 <tr>
464 <th>Description</th>
465 <th><code>option</code></th>
466 <th>Meaning</th>
467 </tr>
468 <tr>
469 <td>ADsafe</td>
470 <td><code>adsafe</code></td>
471 <td><code>true</code> if <a href="http://www.ADsafe.org/">AD<span style="color: blue;">safe</span></a>
472 rules should be enforced. See <a href="http://www.ADsafe.org/">http://www.ADsafe.org/</a>.</td>
473 </tr>
474 <tr>
a27d17c @douglascrockford Tolerate
authored
475 <td>Tolerate bitwise operators </td>
ca120a7 @douglascrockford first commit
authored
476 <td><code>bitwise</code></td>
a27d17c @douglascrockford Tolerate
authored
477 <td><code>true</code> if bitwise operators should be allowed. <a href="#bitwise"><small>(more)</small></a></td>
ca120a7 @douglascrockford first commit
authored
478 </tr>
479 <tr>
480 <td>Assume a browser </td>
481 <td><code>browser</code></td>
482 <td><code>true</code> if the standard browser globals should be predefined.
483 <a href="#browser"><small>(more)</small></a> </td>
484 </tr>
485 <tr>
486 <td>Tolerate HTML case </td>
487 <td><code>cap</code></td>
44eec74 @douglascrockford Adrian
authored
488 <td><code>true</code> if uppercase HTML should be allowed.</td>
ca120a7 @douglascrockford first commit
authored
489 </tr>
490 <tr>
65313e2 @douglascrockford option.confusion
authored
491 <td>Tolerate type confusion<br>
492 </td>
493 <td><code>confusion</code></td>
494 <td><code>true</code> if variables and properties are allowed to contain more than one type of value.</td>
495 </tr>
496 <tr>
3c8016e @douglascrockford option.continue
authored
497 <td>Tolerate <code>continue</code></td>
498 <td><code>continue</code></td>
499 <td><code>true</code> if the <code>continue</code> statement should be allowed.</td>
500 </tr>
501 <tr>
ca120a7 @douglascrockford first commit
authored
502 <td>Tolerate CSS workarounds</td>
503 <td><code>css</code></td>
504 <td><code>true</code> if CSS workarounds should be tolerated. <a href="#css"><small>(more)</small></a></td>
505 </tr>
506 <tr>
507 <td>Tolerate debugger statements</td>
508 <td><code>debug</code></td>
509 <td><code>true</code> if <code>debugger</code> statements should be
510 allowed. Set this option to <code>false</code> before going into production.</td>
511 </tr>
512 <tr>
513 <td>Assume <code>console</code>, <code>alert</code>, ...</td>
514 <td><code>devel</code></td>
515 <td><code>true</code> if browser globals that are useful in development should be
516 predefined. (<a href="#devel">more</a>)</td>
517 </tr>
518 <tr>
ba257de @douglascrockford eqeq
authored
519 <td>Tolerate <code>==</code> and <code>!=</code></td>
520 <td><code>eqeq</code></td>
521 <td><code>true</code> if the <code>==</code> and <code>!=</code> operators should be tolerated. (<a href="#eqeq">more</a>).</td>
522 <tr>
5675d2c @douglascrockford http://tech.groups.yahoo.com/group/jslint_com/message/1730
authored
523 <td>Tolerate ES5 syntax</td>
524 <td><code>es5</code></td>
40e3f73 @douglascrockford It
authored
525 <td><code>true</code> if ES5 syntax should be allowed.</a> It is likely that programs using this option will produce syntax errors on ES3 systems.</td>
ca120a7 @douglascrockford first commit
authored
526 </tr>
527 <tr>
528 <td>Tolerate <code>eval</code> </td>
529 <td><code>evil</code></td>
530 <td><code>true</code> if <code>eval</code> should be allowed. <a href="#evil"><small>(more)</small></a></td>
531 </tr>
532 <tr>
533 <td>Tolerate unfiltered for in </td>
534 <td><code>forin</code></td>
535 <td><code>true</code> if unfiltered <code>for</code> <code>in</code>
536 statements should be allowed. <a href="#forin"><small>(more)</small></a></td>
537 </tr>
538 <tr>
539 <td>Tolerate HTML fragments </td>
540 <td><code>fragment</code></td>
541 <td><code>true</code> if HTML fragments should be allowed. <a href="#html"><small>(more)</small></a></td>
542 </tr>
543 <tr>
5675d2c @douglascrockford http://tech.groups.yahoo.com/group/jslint_com/message/1730
authored
544 <td>Strict white space indentation</td>
545 <td><code>indent</code></td>
dedfd85 @douglascrockford option.indent
authored
546 <td>The number of spaces used for indentation (default is 4). If 0, then no indentation checking takes place.</td>
ca120a7 @douglascrockford first commit
authored
547 </tr>
548 <tr>
5675d2c @douglascrockford http://tech.groups.yahoo.com/group/jslint_com/message/1730
authored
549 <td>Maximum number of errors</td>
550 <td><code>maxerr</code></td>
0c4a9d6 @douglascrockford .
authored
551 <td>The maximum number of warnings reported. (default is 50)</td>
ca120a7 @douglascrockford first commit
authored
552 </tr>
553 <tr>
554 <td>Maximum line length</td>
555 <td><code>maxlen</code></td>
0c4a9d6 @douglascrockford .
authored
556 <td>The maximum number of characters in a line.</td>
ca120a7 @douglascrockford first commit
authored
557 </tr>
558 <tr>
a27d17c @douglascrockford Tolerate
authored
559 <td>Tolerate uncapitalized constructors</td>
ca120a7 @douglascrockford first commit
authored
560 <td><code>newcap</code></td>
a27d17c @douglascrockford Tolerate
authored
561 <td><code>true</code> if Initial Caps with constructor
562 functions is optional. <a href="#new"><small>(more)</small></a></td>
ca120a7 @douglascrockford first commit
authored
563 </tr>
564 <tr>
706c485 @douglascrockford option.node
authored
565 <td>Assume Node.js</td>
566 <td><code>node</code></td>
0c4a9d6 @douglascrockford .
authored
567 <td><code>true</code> if Node.js globals should be predefined. <a href="#node"><small>(more)</small></a></td>
706c485 @douglascrockford option.node
authored
568 </tr>
569 <tr>
a27d17c @douglascrockford Tolerate
authored
570 <td>Tolerate dangling _ in identifiers </td>
706c485 @douglascrockford option.node
authored
571 <td><code>nomen</code></td>
0c4a9d6 @douglascrockford .
authored
572 <td><code>true</code> if names should not be checked for initial or trailing underbars.</td>
706c485 @douglascrockford option.node
authored
573 </tr>
574 <tr>
ca120a7 @douglascrockford first commit
authored
575 <td>Tolerate HTML event handlers </td>
576 <td><code>on</code></td>
577 <td><code>true</code> if HTML event handlers should be allowed. <a href="#html"><small>(more)</small></a></td>
578 </tr>
579 <tr>
580 <td>Stop on first error </td>
581 <td><code>passfail</code></td>
582 <td><code>true</code> if the scan should stop on first error.</td>
583 </tr>
584 <tr>
a27d17c @douglascrockford Tolerate
authored
585 <td>Tolerate <code>++</code> and <code>--</code> </td>
ca120a7 @douglascrockford first commit
authored
586 <td><code>plusplus</code></td>
587 <td><code>true</code> if <code>++</code> and <code>--</code> should
a27d17c @douglascrockford Tolerate
authored
588 be allowed. <a href="#inc"><small>(more)</small></a></td>
ca120a7 @douglascrockford first commit
authored
589 </tr>
590 <tr>
591 <td>Predefined <small>( , separated)</small></td>
592 <td><code>predef</code></td>
6735394 @douglascrockford Cleanup.
authored
593 <td>An array of strings, the names of predefined global variables, or an object whose keys are global variable names, and whose values are booleans that determine if each variable is assignable (also see <a href="#global">global</a>). <code>predef</code> is used with the <code>option</code> object, but not
00d8d1f @douglascrockford option.predef
authored
594 with the <code>/*jslint */</code> comment. You can also use the <code>var</code>
ca120a7 @douglascrockford first commit
authored
595 statement to declare global variables in a script file.</td>
596 </tr>
597 <tr>
a27d17c @douglascrockford Tolerate
authored
598 <td>Tolerate <code>.</code> and <code>[^</code>...<code>]</code>. in /RegExp/ </td>
ca120a7 @douglascrockford first commit
authored
599 <td><code>regexp</code></td>
a27d17c @douglascrockford Tolerate
authored
600 <td><code>true</code> if <code>.</code> and <code>[^</code>...<code>]</code> should be allowed in RegExp
8d1c4eb @douglascrockford clarification
authored
601 literals. They match more material than might be expected, allowing attackers to confuse applications. These forms should not be used when validating in secure applications. </td>
ca120a7 @douglascrockford first commit
authored
602 </tr>
603 <tr>
604 <td>Assume Rhino </td>
605 <td><code>rhino</code></td>
606 <td><code>true</code> if the <a href="http://www.mozilla.org/rhino/">Rhino</a>
607 environment globals should be predefined. <a href="#rhino"><small>(more)</small></a></td>
608 </tr>
609 <tr>
610 <td>Safe Subset </td>
611 <td><code>safe</code></td>
612 <td><code>true</code> if the safe subset rules are enforced. These rules
613 are used by <a href="http://www.ADsafe.org/">ADsafe</a>. It enforces
614 the safe subset rules but not the widget structure rules.</td>
615 </tr>
616 <tr>
a27d17c @douglascrockford Tolerate
authored
617 <td> Tolerate missing&nbsp;<code>'use strict'</code>&nbsp;pragma </td>
618 <td><code>sloppy</code></td>
619 <td><code>true</code> if the ES5 <code><a href="http://www.yuiblog.com/blog/2010/12/14/strict-mode-is-coming-to-town/">'use strict';</a></code> pragma
0c4a9d6 @douglascrockford .
authored
620 is not required. Do not use this pragma unless you know what you are doing.</td>
ca120a7 @douglascrockford first commit
authored
621 </tr>
622 <tr>
623 <td>Tolerate inefficient subscripting<br>
624 </td>
625 <td><code>sub</code></td>
626 <td><code>true</code> if subscript notation may be used for expressions
627 better expressed in dot notation.</td>
628 </tr>
629 <tr>
a27d17c @douglascrockford Tolerate
authored
630 <td> Tolerate misordered definitions </td>
ca120a7 @douglascrockford first commit
authored
631 <td><code>undef</code></td>
a27d17c @douglascrockford Tolerate
authored
632 <td><code>true</code> if variables and functions need not be declared before used. <a href="#undefined"><small>(more)</small></a></td>
ca120a7 @douglascrockford first commit
authored
633 </tr>
634 <tr>
0b060c1 @douglascrockford unparam
authored
635 <td> Tolerate unused parameters</td>
636 <td><code>unparam</code></td>
0c4a9d6 @douglascrockford .
authored
637 <td><code>true</code> if warnings should not be given for unused parameters.</td>
0b060c1 @douglascrockford unparam
authored
638 </tr>
639 <tr>
640 <td>Tolerate many&nbsp;<tt>var</tt>&nbsp;statements per function</td>
641 <td><code>vars</code></td>
642 <td><code>true</code> if multiple <code>var</code> statement per function
643 should be allowed. <a href="#scope"><small>(more)</small></a></td>
644 </tr>
645 <tr>
a27d17c @douglascrockford Tolerate
authored
646 <td> Tolerate messy white space</td>
ca120a7 @douglascrockford first commit
authored
647 <td><code>white</code></td>
0b060c1 @douglascrockford unparam
authored
648 <td><code>true</code> if strict whitespace rules should be ignored.</td>
ca120a7 @douglascrockford first commit
authored
649 </tr>
650 <tr>
651 <td>Assume a Yahoo Widget </td>
652 <td><code>widget</code></td>
653 <td><code>true</code> if the <a href="http://widgets.yahoo.com/gallery/view.php?widget=37484">Yahoo
654 Widgets</a> globals should be predefined. <a href="#widget"><small>(more)</small></a></td>
655 </tr>
656 <tr>
0b060c1 @douglascrockford unparam
authored
657 <td>Assume Windows</td>
ca120a7 @douglascrockford first commit
authored
658 <td><code>windows</code></td>
659 <td><code>true</code> if the Windows</a> globals should be predefined. <a href="#windows"><small>(more)</small></a></td>
660 </tr>
661 </tbody>
662 </table>
d4189ff @douglascrockford properties
authored
663 <h2 id=properties>Properties</h2>
ca120a7 @douglascrockford first commit
authored
664 <p>Since JavaScript is a loosely-typed, dynamic-object language, it is not
665 possible to determine at compile time if property names are spelled correctly.
666 <code>JSLint</code> provides some assistance with this.</p>
475768b @douglascrockford /*properties*/ and /*global*/ now respect function scope
authored
667 <p>At the bottom of its report, <code>JSLint</code> displays a <code>/*properties*/</code>
ca120a7 @douglascrockford first commit
authored
668 comment. It contains all of the names and string literals that were used
669 with dot notation, subscript notation, and object literals to name the
d4189ff @douglascrockford properties
authored
670 properties of objects. You can look through the list for misspellings. Property
ca120a7 @douglascrockford first commit
authored
671 names that were only used once are shown in italics. This is to make misspellings
672 easier to spot.</p>
475768b @douglascrockford /*properties*/ and /*global*/ now respect function scope
authored
673 <p>You can copy the <code>/*properties*/</code> comment into your script file.
ca120a7 @douglascrockford first commit
authored
674 <code>JSLint</code> will check the spelling of all property names against
675 the list. That way, you can have <code>JSLint</code> look for misspellings
475768b @douglascrockford /*properties*/ and /*global*/ now respect function scope
authored
676 for you. The directive respects function scope.</p>
ca120a7 @douglascrockford first commit
authored
677 <h2 id=report>Report</h2>
678
679 <p>If <code>JSLint</code> is able to complete its scan, it generates a function
680 report. It lists for each function:</p>
681
682 <ul>
683 <li>The line number on which it starts.</li>
684 <li>Its name. In the case of anonymous functions, <code>JSLint</code>
0c4a9d6 @douglascrockford .
authored
685 will 'guess' the name.</li>
ca120a7 @douglascrockford first commit
authored
686 <li>The parameters.</li>
687 <li><i>Closure</i>: The variables and parameters that are declared in
688 the function that are used by its inner functions.</li>
689 <li><i>Variables</i>: The variables that are declared in the function
690 that are used only by the function.</li>
691 <li><i>Exceptions</i>: The variables that are declared by try statements.</li>
692 <li><i>Unused</i>: The variables that are declared in the function that
693 are not used. This may be an indication of an error.</li>
694 <li><i>Outer</i>: Variables used by this function that are declared in
695 another function.</li>
696 <li><i>Global</i>: Global variables that are used by this function. Keep
697 these to a minimum.</li>
698 <li><i>Label</i>: Statement labels that are used by this function.</li>
699 </ul>
d4189ff @douglascrockford properties
authored
700 <p>The report will also include a list of all of the <a href="#properties">property
ca120a7 @douglascrockford first commit
authored
701 names</a> that were used. There is a <a href="msgs.html">list of <code>JSLint</code>
702 messages</a>.</p>
703 <h2 id=feedback>Feedback</h2>
704 <p>Please let me know if <code>JSLint</code> is useful for you. Is it too
705 strict? Is there a check or a report that could help you to improve the
706 quality of your programs? <a href="mailto:douglas@crockford.com">douglas@crockford.com</a></p>
707
708 <p>I intend to continue to adapt <code>JSLint</code> based on your comments.
709 Keep watching for improvements. Updates are announced at <a href="http://tech.groups.yahoo.com/group/jslint_com/">http://tech.groups.yahoo.com/group/jslint_com/</a>.</p>
710
711 <h2 id=try>Try it</h2>
712
713 <p><a href="http://www.JSLint.com" target="_blank">Try it.</a> Paste your script
714 into the window and click the
715 <a href="http://www.JSLint.com" target=jslint><input type="button" value="JSLint"></a>
716 button. The analysis is done by a script running on your machine.
717 Your script is not sent over the network. You can set the options used.
65d4238 @douglascrockford It's all good.
authored
718 </p>
bdd3576 @douglascrockford k
authored
719 <p>
720 JSLint is written entirely in JavaScript, so it can run anywhere that JavaScript can run. See for example <a href="http://tech.groups.yahoo.com/group/jslint_com/database?method=reportRows&tbl=1">http://tech.groups.yahoo.com/group/jslint_com/database?method=reportRows&amp;tbl=1</a>.</p>
ca120a7 @douglascrockford first commit
authored
721 <h2 id=implementation>Implementation</h2>
722 <p><code>JSLint</code> uses a <a href="http://javascript.crockford.com/tdop/tdop.html">Pratt
723 Parser (Top Down Operator Precedence)</a>. It is written in JavaScript.
35ec4a5 @douglascrockford groove
authored
724 The full source code is available: <a href="https://github.com/douglascrockford/JSLint">https://github.com/douglascrockford/JSLint</a>.</p>
ca120a7 @douglascrockford first commit
authored
725 <a href="http://www.JSLint.com/"><img src="jslintpill.gif" width="36" height="17" border="0"></a>
726 <a href="http://www.ADsafe.org/"><img src="adsafepill.gif" width="36" height="17" border="0"></a>
727 <a href="http://www.JSON.org/"><img src="jsonpill.gif" width="36" height="17" border="0"></a>
728 </body>
729 </html>
Something went wrong with that request. Please try again.