Skip to content


Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
100755 713 lines (672 sloc) 39.37 kb
ca120a7 @douglascrockford first commit
1 <html>
2 <head>
3 <title>JSLint: The JavaScript Code Quality Tool</title>
4 <link rel="icon" type="image/gif" href="">
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="">JavaScript</a> Code Quality Tool</big></p>
38 <p><a href="" 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="" target="_blank"><code>JSLint</code></a>
45 will hurt your feelings.</p>
46 <h2 id=what>What is <code>JSLint</code>?</h2>
48 <p><a href="" 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>
52 <p>When <a href="">C</a>
53 was a <a href="">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="">lint</a></code>
57 was developed that would scan a source file, looking for problems.</p>
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>
63 <p><a href="">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.
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=""><code>JSLint</code></a>,
ca120a7 @douglascrockford first commit
68 a JavaScript syntax checker and validator.</p>
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>
77 <p><code>JSLint</code> defines a professional subset of JavaScript, a stricter
78 language than that defined by <a href="" 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="" 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.
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="">JSON</a>
ca120a7 @douglascrockford first commit
86 text.</p>
87 <h2 id=global>Global Variables</h2>
88 <p>JavaScript's <a href="">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
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
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
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
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
112 <code>/*global
e752a55 @douglascrockford option.node
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
114 */</code></blockquote>
e752a55 @douglascrockford option.node
115 <p>Select the
116 <em>Assume console, alert, ...</em>
ca120a7 @douglascrockford first commit
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
120 <p id=node>Select the
e752a55 @douglascrockford option.node
121 <em>Assume Node.js</em>
706c485 @douglascrockford option.node
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
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
125 <p id=rhino>Select the <i>Assume Rhino</i> (<code>rhino</code>) <a href="#options">option</a>
706c485 @douglascrockford option.node
126 to predefine the global properties provided by the Rhino environment.
ca120a7 @douglascrockford first commit
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.
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
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>
165 <p>In many languages, a block introduces a scope. Variables introduced in
166 a block are not visible outside of the block.</p>
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>
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
180 statement be used per function. This can be declined with the <code>vars</code>
ca120a7 @douglascrockford first commit
181 <a href="#options">option</a>.</p>
183 <h2 id=required>Required Blocks</h2>
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>
189 <p>JavaScript allows an <code>if</code> to be written like this:</p>
191 <pre>if (<i>condition</i><code>)
192 </code><i>statement</i>;</pre>
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>
198 <pre>if (<i>condition</i>) {
199 <i>statements</i>;
200 }</pre>
202 <p>Experience shows that this form is more resilient.</p>
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>
208 <h2 id=forin><code>for</code> <code>in</code></h2>
209 <p>The <code>for</code> <code>in</code> statement allows for looping through
210 the names of all of the properties of an object. <a href="">Unfortunately,
d4189ff @douglascrockford properties
211 it also loops through all of the properties that were inherited through
ca120a7 @douglascrockford first commit
212 the prototype chain.</a> This has the bad side effect of serving up method
d4189ff @douglascrockford properties
213 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
214 <p>The body of every <code>for</code> <code>in</code> statement should be
215 wrapped in an <code>if</code> statement that does filtering. It can select
216 for a particular type or range of values, or it can exclude functions,
217 or it can exclude properties from the prototype. For example,</p>
218 <pre>for (name in object) {
219 if (object.hasOwnProperty(name)) {
220 ....
221 }
223 }</pre>
225 <h2 id=switch><code>switch</code></h2>
226 <p>A <a href="">common
227 error</a> in <code>switch</code> statements is to forget to place a <code>break</code>
228 statement after each case, resulting in unintended fall-through. <code>JSLint</code>
229 expects that the statement before the next <code>case</code> or <code>default</code>
230 is one of these: <code>break</code>, <code>return</code>, or <code>throw</code>.
231 </p>
232 <h2 id=var><code>var</code></h2>
234 <p>JavaScript allows <code>var</code> definitions to occur anywhere
235 within a function. <code>JSLint</code> is more strict.</p>
237 <p><code>JSLint</code> expects that a <code>var</code> will be declared
238 only once, and that it will be declared before it is used.</p>
239 <p><code></code><code>JSLint</code> expects that a <code>function</code>
240 will be declared before it is used.</p>
241 <p><code>JSLint</code> expects that parameters will not also be declared
242 as vars. </p>
244 <p><code>JSLint</code> does not expect the <code>arguments</code> array to be declared
245 as a <code>var</code>.</p>
246 <p><code>JSLint</code> does not expect that a var will be defined in a block.
247 This is because JavaScript blocks do not have block scope. This can have
248 unexpected consequences. Define all variables at the top of the function.</p>
250 <h2 id=with><code>with</code></h2>
252 <p>The <code>with</code> statement was intended to provide a shorthand in accessing
d4189ff @douglascrockford properties
253 properties in deeply nested objects. Unfortunately, it behaves <a href="">very
254 badly</a> when setting new properties. Never use the <code>with</code> statement. Use
ca120a7 @douglascrockford first commit
255 a <code>var</code> instead.</p>
257 <p><code>JSLint</code> does not expect to see a <code>with</code> statement.</p>
259 <h2 id=assignment>=</h2>
260 <p><code>JSLint</code> does not expect to see an assignment statement in
261 the condition part of an <code>if</code> or <code>for</code> or <code>while</code>
262 <code></code> or <code>do</code> statement. This is because it is more
263 likely that </p>
264 <pre>if (a = b) {
265 ...
266 }</pre>
267 <p>was intended to be </p>
268 <pre>if (a == b) {
269 ...
270 }</pre>
271 <p>It is difficult to write correct programs while using idioms that are
9d80d80 @douglascrockford asignment unwrapped
272 hard to distinguish from obvious errors.</p>
ba257de @douglascrockford eqeq
273 <h2 id=eqeq>== and !=</h2>
ca120a7 @douglascrockford first commit
274 <p>The <code>==</code> and <code>!=</code> operators do type coercion before
275 comparing. This is bad because it causes <code>' \t\r\n' == 0</code> to
44eec74 @douglascrockford Adrian
276 be <code>true</code>. This can mask type errors. JSLint cannot reliably determine if == is being used correctly, to it is best to not use <code>==</code> and != and always use the more reliable <code>===</code> and <code>!==</code> operators instead. </p>
ca120a7 @douglascrockford first commit
277 <p align="left">If you only care that a value is <i>truthy</i> or <i>falsy</i>,
278 then use the short form. Instead of </p>
279 <pre align="left">(foo != 0)</pre>
280 <p align="left">just say </p>
281 <pre align="left">(foo)</pre>
282 <p align="left">and instead of</p>
283 <pre align="left">(foo == 0)</pre>
284 <p align="left"> say</p>
285 <pre align="left">(!foo)</pre>
ba257de @douglascrockford eqeq
286 <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
287 <h2 id=labels>Labels</h2>
288 <p>JavaScript allows any statement to have a label, and labels have a
289 separate name space. <code>JSLint</code> is more strict.</p>
291 <p><code>JSLint</code> expects labels only on statements that interact
292 with <code>break</code>: <code>switch</code>, <code>while</code>,
293 <code>do</code>, and <code>for</code>. <code>JSLint</code> expects that labels
294 will be distinct from vars and parameters.</p>
296 <h2 id=unreachable>Unreachable Code</h2>
297 <p><code>JSLint</code> expects that
298 a <code>return</code>, <code>break</code>, <code>continue</code>,
299 or <code>throw</code> statement will be followed by
300 a <code>}</code> or <code>case</code> or <code>default</code>.</p>
302 <h2 id=pluses>Confusing Pluses and Minuses</h2>
304 <p><code>JSLint</code> expects that <code>+</code> will not be followed by
305 <code>+</code> or <code>++</code>, and that <code>-</code> will not be followed
306 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>
307 <h2 id=inc><code>++</code> and <code>--</code></h2>
308 <p>The <code>++</code> <small>(increment)</small> and <code>--</code> <small>(decrement)</small>
309 operators have been known to contribute to bad code by encouraging excessive
310 trickiness. They are second only to faulty architecture in enabling to
44eec74 @douglascrockford Adrian
311 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
312 that allows the use of these operators.</p>
ca120a7 @douglascrockford first commit
313 <h2 id=bitwise>Bitwise Operators</h2>
314 <p>JavaScript does not have an integer type, but it does have bitwise operators.
315 The bitwise operators convert their operands from floating point to integers
316 and back, so they are not as efficient as in C or other languages. They
317 are rarely useful in browser applications. The similarity to the logical
318 operators can mask some programming errors. The <code>bitwise</code> <a href="#options">option</a>
a27d17c @douglascrockford Tolerate
319 allows the use of these operators: <code>&lt;&lt; &gt;&gt; &gt;&gt;&gt;
ca120a7 @douglascrockford first commit
320 ~ &amp; |</code>.</p>
321 <h2 id=evil><code>eval</code> is evil</h2>
322 <p>The <code>eval</code> function (and its relatives, <code>Function</code>,
323 <code>setTimeout</code>, and <code>setInterval</code>) provide access
324 to the JavaScript compiler. This is sometimes necessary, but in most cases
325 it indicates the presence of extremely bad coding. The <code>eval</code>
326 function is the most misused feature of JavaScript.</p>
328 <h2 id=void><code>void</code></h2>
329 <p>In most C-like languages, <code>void</code> is a type. In
330 JavaScript, <code>void</code> is a prefix operator that always
331 returns <code>undefined</code>. <code>JSLint</code> does not expect to
332 see <code>void</code> because it is confusing and not very useful.</p>
334 <h2 id=regexp>Regular Expressions</h2>
335 <p>Regular expressions are written in a terse and cryptic notation. <code>JSLint</code>
336 looks for problems that may cause portability problems. It also attempts
337 to resolve visual ambiguities by recommending explicit escapement.</p>
338 <p>JavaScript's syntax for regular expression literals overloads the <code>/</code>
339 character. To avoid ambiguity, <code>JSLint</code> expects that the character
340 preceding a regular expression literal is a <code>(</code> or <code>=</code>
341 or <code>:</code> or <code>,</code> character. </p>
342 <h2 id=new>Constructors and <code>new</code></h2>
343 <p>Constructors are functions that are designed to be used with the <code>new</code>
344 prefix. The <code>new</code> prefix creates a new object based on the
345 function's <code>prototype</code>, and binds that object to the function's
346 implied <code>this</code> parameter. If you neglect to use the <code>new</code>
347 prefix, no new object will be made and <code>this</code> will be bound
348 to the global object. This is a <a href="">serious
349 mistake</a>.</p>
350 <p><code>JSLint</code> enforces the convention that constructor functions
351 be given names with initial uppercase. <code>JSLint</code> does not expect
352 to see a function invocation with an initial uppercase name unless it
353 has the <code>new</code> prefix. <code>JSLint</code> does not expect to
354 see the <code>new</code> prefix used with functions whose names do not
a27d17c @douglascrockford Tolerate
355 start with initial uppercase. This can be disabled with the <code>newcap</code>
ca120a7 @douglascrockford first commit
356 <a href="#options">option</a>.</p>
357 <p><code>JSLint</code> does not expect to see the wrapper forms <code>new Number</code>,
358 <code>new String</code>, <code>new Boolean</code>. </p>
359 <p><code>JSLint</code> does not expect to see <code>new Object</code> (use <code>{}</code>
360 instead). </p>
361 <p><code>JSLint</code> does not expect to see <code>new Array</code> (use <code>[]</code>
362 instead).</p>
363 <h2 id=unsafe>Unsafe Characters</h2>
364 <p> There are characters that are handled inconsistently in browsers, and
365 so must be escaped when placed in strings. </p>
366 <pre>\u0000-\u001f
367 \u007f-\u009f
368 \u00ad
369 \u0600-\u0604
370 \u070f
371 \u17b4
372 \u17b5
373 \u200c-\u200f
374 \u2028-\u202f
375 \u2060-\u206f
376 \ufeff
377 \ufff0-\uffff</pre>
378 <h2 id=not>Not Looked For</h2>
380 <p><code>JSLint</code> does not do flow analysis to determine that variables are assigned
381 values before used. This is because variables are given a value (<code>undefined</code>)
6735394 @douglascrockford Cleanup.
382 that is a reasonable default for many applications.</p>
ca120a7 @douglascrockford first commit
384 <p><code>JSLint</code> does not do any kind of global analysis. It does
385 not attempt to determine that functions used with <code>new</code> are
386 really constructors (<a href="#new">except by enforcing capitalization
d4189ff @douglascrockford properties
387 conventions</a>), or that property names are spelled correctly (<a href="#properties">except
388 for matching against the <code>/*properties */</code> comment</a>).</p>
ca120a7 @douglascrockford first commit
389 <h2 id=html>HTML</h2>
390 <p><code>JSLint</code> is able to handle HTML text. It can inspect the JavaScript content
391 contained within <code>&lt;script&gt;</code>...<code>&lt;/script&gt;</code> tags. It
392 also inspects the HTML content, looking for problems that are known to interfere
393 with JavaScript:</p>
394 <ul>
395 <li>All tag names must be in lower case.</li>
396 <li>All tags that can take a close tag (such as <code>&lt;/p&gt;</code>)
397 must have a close tag.</li>
398 <li>All tags are correctly nested.</li>
399 <li>The entity <code>&amp;lt;</code> must be used for literal <code>'&lt;'</code>.</li>
400 </ul>
401 <p><code>JSLint</code> is less anal than the sycophantic conformity demanded
402 by XHTML, but more strict than the popular browsers. </p>
403 <p><code>JSLint</code> also checks for the occurrence of<code> '&lt;/' </code>in
404 string literals. You should always write<code> '&lt;\/' </code>instead.
405 The extra backslash is ignored by the JavaScript compiler but not by the
406 HTML parser. Tricks like this should not be necessary, and yet they are.</p>
407 <p>There is a <code>cap</code> <a href="#options">option</a> that allows
44eec74 @douglascrockford Adrian
408 use of uppercase tag names. There is also an <code>on</code> <a href="#options">option</a>
ca120a7 @douglascrockford first commit
409 that allows the use of inline HTML event handlers.</p>
410 <p>There is a <code>fragment</code> <a href="#options">option</a> that can
411 inspect a well formed HTML fragment. If the <code>adsafe</code> <a href="#options">option</a>
412 is also used, then the fragment must be a <code>&lt;div&gt;</code> that
413 conforms to the <a href="">ADsafe</a> widget rules.</p>
414 <h2 id=css>CSS</h2>
415 <p><code>JSLint</code> can inspect CSS files. It expects the first line
416 of a CSS file to be </p>
417 <pre>@charset &quot;UTF-8&quot;;</pre>
418 <p>This feature is experimental. Please report any problems or limitations.
419 There is a <code>css</code> <a href="#options">option</a> that will tolerate
420 some of the non-standard-but-customary workarounds. <br>
421 </p>
423 <h2 id=options>Options</h2>
424 <p><code>JSLint</code> provides several options that control its operation and
425 its sensitivity. In the <a href="">web edition</a>, the
a27d17c @douglascrockford Tolerate
426 options are selected with several checkboxes and two fields. </p>
ca120a7 @douglascrockford first commit
427 <p>It also provides assistance in constructing <code>/*jslint*/</code>
429 </p>
430 <p>When <code>JSLINT</code> is called as a function, it accepts an <code>option</code> object
431 parameter that allows you to determine the subset of JavaScript that is
432 acceptable to you. The web page version of <code>JSLint</code> at <a href=""></a>
433 does this for you. </p>
434 <p>Options can also be specified within a script with a <code>/*jslint */</code>
475768b @douglascrockford /*properties*/ and /*global*/ now respect function scope
435 directive:</p>
ca120a7 @douglascrockford first commit
436 <pre>/*jslint nomen: true, debug: true,
a27d17c @douglascrockford Tolerate
437 evil: false, vars: true */</pre>
ca120a7 @douglascrockford first commit
438 <p>An option specification starts with <code>/*jslint</code>. Notice that
439 there is no space before the <code>j</code>. The specification contains
440 a sequence of name value pairs, where the names are <code>JSLint</code>
441 options, and the values are <code>true</code> or <code>false</code>. The
442 <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
443 comment takes precedence over the <code>option</code> object. The directive respects function scope.</p>
ca120a7 @douglascrockford first commit
444 <table>
445 <tbody>
446 <tr>
447 <th>Description</th>
448 <th><code>option</code></th>
449 <th>Meaning</th>
450 </tr>
451 <tr>
452 <td>ADsafe</td>
453 <td><code>adsafe</code></td>
454 <td><code>true</code> if <a href="">AD<span style="color: blue;">safe</span></a>
455 rules should be enforced. See <a href=""></a>.</td>
456 </tr>
457 <tr>
a27d17c @douglascrockford Tolerate
458 <td>Tolerate bitwise operators </td>
ca120a7 @douglascrockford first commit
459 <td><code>bitwise</code></td>
a27d17c @douglascrockford Tolerate
460 <td><code>true</code> if bitwise operators should be allowed. <a href="#bitwise"><small>(more)</small></a></td>
ca120a7 @douglascrockford first commit
461 </tr>
462 <tr>
463 <td>Assume a browser </td>
464 <td><code>browser</code></td>
465 <td><code>true</code> if the standard browser globals should be predefined.
466 <a href="#browser"><small>(more)</small></a> </td>
467 </tr>
468 <tr>
469 <td>Tolerate HTML case </td>
470 <td><code>cap</code></td>
44eec74 @douglascrockford Adrian
471 <td><code>true</code> if uppercase HTML should be allowed.</td>
ca120a7 @douglascrockford first commit
472 </tr>
473 <tr>
3c8016e @douglascrockford option.continue
474 <td>Tolerate <code>continue</code></td>
475 <td><code>continue</code></td>
476 <td><code>true</code> if the <code>continue</code> statement should be allowed.</td>
477 </tr>
478 <tr>
ca120a7 @douglascrockford first commit
479 <td>Tolerate CSS workarounds</td>
480 <td><code>css</code></td>
481 <td><code>true</code> if CSS workarounds should be tolerated. <a href="#css"><small>(more)</small></a></td>
482 </tr>
483 <tr>
484 <td>Tolerate debugger statements</td>
485 <td><code>debug</code></td>
486 <td><code>true</code> if <code>debugger</code> statements should be
487 allowed. Set this option to <code>false</code> before going into production.</td>
488 </tr>
489 <tr>
490 <td>Assume <code>console</code>, <code>alert</code>, ...</td>
491 <td><code>devel</code></td>
492 <td><code>true</code> if browser globals that are useful in development should be
493 predefined. (<a href="#devel">more</a>)</td>
494 </tr>
495 <tr>
ba257de @douglascrockford eqeq
496 <td>Tolerate <code>==</code> and <code>!=</code></td>
497 <td><code>eqeq</code></td>
498 <td><code>true</code> if the <code>==</code> and <code>!=</code> operators should be tolerated. (<a href="#eqeq">more</a>).</td>
499 <tr>
5675d2c @douglascrockford
500 <td>Tolerate ES5 syntax</td>
501 <td><code>es5</code></td>
40e3f73 @douglascrockford It
502 <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
503 </tr>
504 <tr>
505 <td>Tolerate <code>eval</code> </td>
506 <td><code>evil</code></td>
507 <td><code>true</code> if <code>eval</code> should be allowed. <a href="#evil"><small>(more)</small></a></td>
508 </tr>
509 <tr>
510 <td>Tolerate unfiltered for in </td>
511 <td><code>forin</code></td>
512 <td><code>true</code> if unfiltered <code>for</code> <code>in</code>
513 statements should be allowed. <a href="#forin"><small>(more)</small></a></td>
514 </tr>
515 <tr>
516 <td>Tolerate HTML fragments </td>
517 <td><code>fragment</code></td>
518 <td><code>true</code> if HTML fragments should be allowed. <a href="#html"><small>(more)</small></a></td>
519 </tr>
520 <tr>
5675d2c @douglascrockford
521 <td>Strict white space indentation</td>
522 <td><code>indent</code></td>
dedfd85 @douglascrockford option.indent
523 <td>The number of spaces used for indentation (default is 4). If 0, then no indentation checking takes place.</td>
ca120a7 @douglascrockford first commit
524 </tr>
525 <tr>
5675d2c @douglascrockford
526 <td>Maximum number of errors</td>
527 <td><code>maxerr</code></td>
0c4a9d6 @douglascrockford .
528 <td>The maximum number of warnings reported. (default is 50)</td>
ca120a7 @douglascrockford first commit
529 </tr>
530 <tr>
531 <td>Maximum line length</td>
532 <td><code>maxlen</code></td>
0c4a9d6 @douglascrockford .
533 <td>The maximum number of characters in a line.</td>
ca120a7 @douglascrockford first commit
534 </tr>
535 <tr>
a27d17c @douglascrockford Tolerate
536 <td>Tolerate uncapitalized constructors</td>
ca120a7 @douglascrockford first commit
537 <td><code>newcap</code></td>
a27d17c @douglascrockford Tolerate
538 <td><code>true</code> if Initial Caps with constructor
539 functions is optional. <a href="#new"><small>(more)</small></a></td>
ca120a7 @douglascrockford first commit
540 </tr>
541 <tr>
706c485 @douglascrockford option.node
542 <td>Assume Node.js</td>
543 <td><code>node</code></td>
0c4a9d6 @douglascrockford .
544 <td><code>true</code> if Node.js globals should be predefined. <a href="#node"><small>(more)</small></a></td>
706c485 @douglascrockford option.node
545 </tr>
546 <tr>
a27d17c @douglascrockford Tolerate
547 <td>Tolerate dangling _ in identifiers </td>
706c485 @douglascrockford option.node
548 <td><code>nomen</code></td>
0c4a9d6 @douglascrockford .
549 <td><code>true</code> if names should not be checked for initial or trailing underbars.</td>
706c485 @douglascrockford option.node
550 </tr>
551 <tr>
ca120a7 @douglascrockford first commit
552 <td>Tolerate HTML event handlers </td>
553 <td><code>on</code></td>
554 <td><code>true</code> if HTML event handlers should be allowed. <a href="#html"><small>(more)</small></a></td>
555 </tr>
556 <tr>
557 <td>Stop on first error </td>
558 <td><code>passfail</code></td>
559 <td><code>true</code> if the scan should stop on first error.</td>
560 </tr>
561 <tr>
a27d17c @douglascrockford Tolerate
562 <td>Tolerate <code>++</code> and <code>--</code> </td>
ca120a7 @douglascrockford first commit
563 <td><code>plusplus</code></td>
564 <td><code>true</code> if <code>++</code> and <code>--</code> should
a27d17c @douglascrockford Tolerate
565 be allowed. <a href="#inc"><small>(more)</small></a></td>
ca120a7 @douglascrockford first commit
566 </tr>
567 <tr>
568 <td>Predefined <small>( , separated)</small></td>
569 <td><code>predef</code></td>
6735394 @douglascrockford Cleanup.
570 <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
571 with the <code>/*jslint */</code> comment. You can also use the <code>var</code>
ca120a7 @douglascrockford first commit
572 statement to declare global variables in a script file.</td>
573 </tr>
574 <tr>
a27d17c @douglascrockford Tolerate
575 <td>Tolerate <code>.</code> and <code>[^</code>...<code>]</code>. in /RegExp/ </td>
ca120a7 @douglascrockford first commit
576 <td><code>regexp</code></td>
a27d17c @douglascrockford Tolerate
577 <td><code>true</code> if <code>.</code> and <code>[^</code>...<code>]</code> should be allowed in RegExp
8d1c4eb @douglascrockford clarification
578 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
579 </tr>
580 <tr>
581 <td>Assume Rhino </td>
582 <td><code>rhino</code></td>
583 <td><code>true</code> if the <a href="">Rhino</a>
584 environment globals should be predefined. <a href="#rhino"><small>(more)</small></a></td>
585 </tr>
586 <tr>
587 <td>Safe Subset </td>
588 <td><code>safe</code></td>
589 <td><code>true</code> if the safe subset rules are enforced. These rules
590 are used by <a href="">ADsafe</a>. It enforces
591 the safe subset rules but not the widget structure rules.</td>
592 </tr>
593 <tr>
a27d17c @douglascrockford Tolerate
594 <td> Tolerate missing&nbsp;<code>'use strict'</code>&nbsp;pragma </td>
595 <td><code>sloppy</code></td>
596 <td><code>true</code> if the ES5 <code><a href="">'use strict';</a></code> pragma
0c4a9d6 @douglascrockford .
597 is not required. Do not use this pragma unless you know what you are doing.</td>
ca120a7 @douglascrockford first commit
598 </tr>
599 <tr>
600 <td>Tolerate inefficient subscripting<br>
601 </td>
602 <td><code>sub</code></td>
603 <td><code>true</code> if subscript notation may be used for expressions
604 better expressed in dot notation.</td>
605 </tr>
606 <tr>
ae6a9f8 @douglascrockford option.type
607 <td>Tolerate type inconsistency<br>
608 </td>
609 <td><code>type</code></td>
610 <td><code>true</code> if variables are allowed to contain more than one type of value.</td>
611 </tr>
612 <tr>
a27d17c @douglascrockford Tolerate
613 <td> Tolerate misordered definitions </td>
ca120a7 @douglascrockford first commit
614 <td><code>undef</code></td>
a27d17c @douglascrockford Tolerate
615 <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
616 </tr>
617 <tr>
0b060c1 @douglascrockford unparam
618 <td> Tolerate unused parameters</td>
619 <td><code>unparam</code></td>
0c4a9d6 @douglascrockford .
620 <td><code>true</code> if warnings should not be given for unused parameters.</td>
0b060c1 @douglascrockford unparam
621 </tr>
622 <tr>
623 <td>Tolerate many&nbsp;<tt>var</tt>&nbsp;statements per function</td>
624 <td><code>vars</code></td>
625 <td><code>true</code> if multiple <code>var</code> statement per function
626 should be allowed. <a href="#scope"><small>(more)</small></a></td>
627 </tr>
628 <tr>
a27d17c @douglascrockford Tolerate
629 <td> Tolerate messy white space</td>
ca120a7 @douglascrockford first commit
630 <td><code>white</code></td>
0b060c1 @douglascrockford unparam
631 <td><code>true</code> if strict whitespace rules should be ignored.</td>
ca120a7 @douglascrockford first commit
632 </tr>
633 <tr>
634 <td>Assume a Yahoo Widget </td>
635 <td><code>widget</code></td>
636 <td><code>true</code> if the <a href="">Yahoo
637 Widgets</a> globals should be predefined. <a href="#widget"><small>(more)</small></a></td>
638 </tr>
639 <tr>
0b060c1 @douglascrockford unparam
640 <td>Assume Windows</td>
ca120a7 @douglascrockford first commit
641 <td><code>windows</code></td>
642 <td><code>true</code> if the Windows</a> globals should be predefined. <a href="#windows"><small>(more)</small></a></td>
643 </tr>
644 </tbody>
645 </table>
d4189ff @douglascrockford properties
646 <h2 id=properties>Properties</h2>
ca120a7 @douglascrockford first commit
647 <p>Since JavaScript is a loosely-typed, dynamic-object language, it is not
648 possible to determine at compile time if property names are spelled correctly.
649 <code>JSLint</code> provides some assistance with this.</p>
475768b @douglascrockford /*properties*/ and /*global*/ now respect function scope
650 <p>At the bottom of its report, <code>JSLint</code> displays a <code>/*properties*/</code>
ca120a7 @douglascrockford first commit
651 comment. It contains all of the names and string literals that were used
652 with dot notation, subscript notation, and object literals to name the
d4189ff @douglascrockford properties
653 properties of objects. You can look through the list for misspellings. Property
ca120a7 @douglascrockford first commit
654 names that were only used once are shown in italics. This is to make misspellings
655 easier to spot.</p>
475768b @douglascrockford /*properties*/ and /*global*/ now respect function scope
656 <p>You can copy the <code>/*properties*/</code> comment into your script file.
ca120a7 @douglascrockford first commit
657 <code>JSLint</code> will check the spelling of all property names against
658 the list. That way, you can have <code>JSLint</code> look for misspellings
475768b @douglascrockford /*properties*/ and /*global*/ now respect function scope
659 for you. The directive respects function scope.</p>
ca120a7 @douglascrockford first commit
660 <h2 id=report>Report</h2>
662 <p>If <code>JSLint</code> is able to complete its scan, it generates a function
663 report. It lists for each function:</p>
665 <ul>
666 <li>The line number on which it starts.</li>
667 <li>Its name. In the case of anonymous functions, <code>JSLint</code>
0c4a9d6 @douglascrockford .
668 will 'guess' the name.</li>
ca120a7 @douglascrockford first commit
669 <li>The parameters.</li>
670 <li><i>Closure</i>: The variables and parameters that are declared in
671 the function that are used by its inner functions.</li>
672 <li><i>Variables</i>: The variables that are declared in the function
673 that are used only by the function.</li>
674 <li><i>Exceptions</i>: The variables that are declared by try statements.</li>
675 <li><i>Unused</i>: The variables that are declared in the function that
676 are not used. This may be an indication of an error.</li>
677 <li><i>Outer</i>: Variables used by this function that are declared in
678 another function.</li>
679 <li><i>Global</i>: Global variables that are used by this function. Keep
680 these to a minimum.</li>
681 <li><i>Label</i>: Statement labels that are used by this function.</li>
682 </ul>
d4189ff @douglascrockford properties
683 <p>The report will also include a list of all of the <a href="#properties">property
ca120a7 @douglascrockford first commit
684 names</a> that were used. There is a <a href="msgs.html">list of <code>JSLint</code>
685 messages</a>.</p>
686 <h2 id=feedback>Feedback</h2>
687 <p>Please let me know if <code>JSLint</code> is useful for you. Is it too
688 strict? Is there a check or a report that could help you to improve the
689 quality of your programs? <a href=""></a></p>
691 <p>I intend to continue to adapt <code>JSLint</code> based on your comments.
692 Keep watching for improvements. Updates are announced at <a href=""></a>.</p>
694 <h2 id=try>Try it</h2>
696 <p><a href="" target="_blank">Try it.</a> Paste your script
697 into the window and click the
698 <a href="" target=jslint><input type="button" value="JSLint"></a>
699 button. The analysis is done by a script running on your machine.
700 Your script is not sent over the network. You can set the options used.
65d4238 @douglascrockford It's all good.
701 </p>
bdd3576 @douglascrockford k
702 <p>
703 JSLint is written entirely in JavaScript, so it can run anywhere that JavaScript can run. See for example <a href="">;tbl=1</a>.</p>
ca120a7 @douglascrockford first commit
704 <h2 id=implementation>Implementation</h2>
705 <p><code>JSLint</code> uses a <a href="">Pratt
706 Parser (Top Down Operator Precedence)</a>. It is written in JavaScript.
35ec4a5 @douglascrockford groove
707 The full source code is available: <a href=""></a>.</p>
ca120a7 @douglascrockford first commit
708 <a href=""><img src="jslintpill.gif" width="36" height="17" border="0"></a>
709 <a href=""><img src="adsafepill.gif" width="36" height="17" border="0"></a>
710 <a href=""><img src="jsonpill.gif" width="36" height="17" border="0"></a>
711 </body>
712 </html>
Something went wrong with that request. Please try again.