UglifyJS is a JavaScript parser, minifier, compressor or beautifier toolkit.
This page documents the command line utility. For API and internals documentation see my website. There's also an in-browser online demo (for Firefox, Chrome and probably Safari).
uglify-js only supports ECMAScript 5 (ES5).
Support for const is present but incomplete, and may not be
transformed properly.
Those wishing to minify ES2015+ (ES6+) should use the npm package uglify-es.
First make sure you have installed the latest version of node.js (You may need to restart your computer after this step).
From NPM for use as a command line app:
From NPM for programmatic use:
UglifyJS2 can take multiple input files. It's recommended that you pass the input files first, then pass the options. UglifyJS will parse input files in sequence and apply any compression options. The files are parsed in the same global scope, that is, a reference from a file to some variable/function declared in another file will be matched properly.
If you want to read from STDIN instead, pass a single dash instead of input files.
If you wish to pass your options before the input files, separate the two with a double dash to prevent input files being used as option arguments:
The available options are:
Specify --output (-o) to declare the output file. Otherwise the output goes to STDOUT.
UglifyJS2 can generate a source map file, which is highly useful for debugging your compressed JavaScript. To get a source map, pass --source-map output.js.map (full path to the file where you want the source map dumped).
Additionally you might need --source-map-root to pass the URL where the original files can be found. In case you are passing full paths to input files to UglifyJS, you can use --prefix (-p) to specify the number of directories to drop from the path prefix when declaring files in the source map.
For example:
The above will compress and mangle file1.js and file2.js, will drop the output in foo.min.js and the source map in foo.min.js.map. The source mapping will refer to http://foo.com/src/js/file1.js and http://foo.com/src/js/file2.js (in fact it will list http://foo.com/src as the source map root, and the original files as js/file1.js and js/file2.js).
When you're compressing JS code that was output by a compiler such as CoffeeScript, mapping to the JS code won't be too helpful. Instead, you'd like to map back to the original code (i.e. CoffeeScript). UglifyJS has an option to take an input source map. Assuming you have a mapping from CoffeeScript → compiled JS, UglifyJS can generate a map from CoffeeScript → compressed JS by mapping every token in the compiled JS to its original location.
To use this feature you need to pass --in-source-map /path/to/input/source.map or --in-source-map inline if the source map is included inline with the sources. Normally the input source map should also point to the file containing the generated JS, so if that's correct you can omit input files from the command line.
To enable the mangler you need to pass --mangle (-m). The following (comma-separated) options are supported:
toplevel — mangle names declared in the toplevel scope (disabled by default).
eval — mangle names visible in scopes where eval or with are used (disabled by default).
When mangling is enabled but you want to prevent certain names from being mangled, you can declare those names with --reserved (-r) — pass a comma-separated list of names. For example:
to prevent the require, exports and $ names from being changed.
--mangle-props)Note: this will probably break your code. Mangling property names is a separate step, different from variable name mangling. Pass --mangle-props. It will mangle all properties that are seen in some object literal, or that are assigned to. For example:
In the above code, foo, bar, baz, moo and boo will be replaced with single characters, while something() will be left as is.
In order for this to be of any use, we should avoid mangling standard JS names. For instance, if your code would contain x.length = 10, then length becomes a candidate for mangling and it will be mangled throughout the code, regardless if it's being used as part of your own objects or accessing an array's length. To avoid that, you can use --reserved-file to pass a filename that should contain the names to be excluded from mangling. This file can be used both for excluding variable names and property names. It could look like this, for example:
--reserved-file can be an array of file names (either a single comma-separated argument, or you can pass multiple --reserved-file arguments) — in this case it will exclude names from all those files.
A default exclusion file is provided in tools/domprops.json which should cover most standard JS and DOM properties defined in various browsers. Pass --reserve-domprops to read that in.
You can also use a regular expression to define which property names should be mangled. For example, --mangle-regex="/^_/" will only mangle property names that start with an underscore.
When you compress multiple files using this option, in order for them to work together in the end we need to ensure somehow that one property gets mangled to the same name in all of them. For this, pass --name-cache filename.json and UglifyJS will maintain these mappings in a file which can then be reused. It should be initially empty. Example:
Now, part1.js and part2.js will be consistent with each other in terms of mangled property names.
Using the name cache is not necessary if you compress all your files in a single call to UglifyJS.
--mangle-props=unquoted or --mangle-props=2)Using quoted property name (o["foo"]) reserves the property name (foo) so that it is not mangled throughout the entire script even when used in an unquoted style (o.foo). Example:
You can also pass --mangle-props-debug in order to mangle property names without completely obscuring them. For example the property o.foo would mangle to o._$foo$_ with this option. This allows property mangling of a large codebase while still being able to debug the code and identify where mangling is breaking things.
You can also pass a custom suffix using --mangle-props-debug=XYZ. This would then mangle o.foo to o._$foo$XYZ_. You can change this each time you compile a script to identify how a property got mangled. One technique is to pass a random number on every compile to simulate mangling changing with different inputs (e.g. as you update the input script with new properties), and to help identify mistakes like writing mangled keys to storage.
You need to pass --compress (-c) to enable the compressor. Optionally you can pass a comma-separated list of options. Options are in the form foo=bar, or just foo (the latter implies a boolean option that you want to set true; it's effectively a shortcut for foo=true).
sequences (default: true) -- join consecutive simple statements using the comma operator. May be set to a positive integer to specify the maximum number of consecutive comma sequences that will be generated. If this option is set to true then the default sequences limit is 200. Set option to false or 0 to disable. The smallest sequences length is 2. A sequences value of 1 is grandfathered to be equivalent to true and as such means 200. On rare occasions the default sequences limit leads to very slow compress times in which case a value of 20 or less is recommended.
properties -- rewrite property access using the dot notation, for example foo["bar"] → foo.bar
dead_code -- remove unreachable code
drop_debugger -- remove debugger; statements
unsafe (default: false) -- apply "unsafe" transformations (discussion below)
unsafe_comps (default: false) -- Reverse < and <= to > and >= to allow improved compression. This might be unsafe when an at least one of two operands is an object with computed values due the use of methods like get, or valueOf. This could cause change in execution order after operands in the comparison are switching. Compression only works if both comparisons and unsafe_comps are both set to true.
unsafe_math (default: false) -- optimize numerical expressions like 2 * x * 3 into 6 * x, which may give imprecise floating point results.
unsafe_proto (default: false) -- optimize expressions like Array.prototype.slice.call(a) into [].slice.call(a)
unsafe_regexp (default: false) -- enable substitutions of variables with RegExp values the same way as if they are constants.
conditionals -- apply optimizations for if-s and conditional expressions
comparisons -- apply certain optimizations to binary nodes, for example: !(a <= b) → a > b (only when unsafe_comps), attempts to negate binary nodes, e.g. a = !b && !c && !d && !e → a=!(b||c||d||e) etc.
evaluate -- attempt to evaluate constant expressions
booleans -- various optimizations for boolean context, for example !!a ? b : c → a ? b : c
loops -- optimizations for do, while and for loops when we can statically determine the condition
unused -- drop unreferenced functions and variables (simple direct variable assignments do not count as references unless set to "keep_assign")
toplevel -- drop unreferenced functions ("funcs") and/or variables ("vars") in the toplevel scope (false by default, true to drop both unreferenced functions and variables)
top_retain -- prevent specific toplevel functions and variables from unused removal (can be array, comma-separated, RegExp or function. Implies toplevel)
hoist_funs -- hoist function declarations
hoist_vars (default: false) -- hoist var declarations (this is false by default because it seems to increase the size of the output in general)
if_return -- optimizations for if/return and if/continue
join_vars -- join consecutive var statements
cascade -- small optimization for sequences, transform x, x into x and x = something(), x into x = something()
collapse_vars -- Collapse single-use var and const definitions when possible.
reduce_vars -- Improve optimization on variables assigned with and used as constant values.
warnings -- display warnings when dropping unreachable code or unused declarations etc.
negate_iife -- negate "Immediately-Called Function Expressions" where the return value is discarded, to avoid the parens that the code generator would insert.
pure_getters -- the default is false. If you pass true for this, UglifyJS will assume that object property access (e.g. foo.bar or foo["bar"]) doesn't have any side effects. Specify "strict" to treat foo.bar as side-effect-free only when foo is certain to not throw, i.e. not null or undefined.
pure_funcs -- default null. You can pass an array of names and UglifyJS will assume that those functions do not produce side effects. DANGER: will not check if the name is redefined in scope. An example case here, for instance var q = Math.floor(a/b). If variable q is not used elsewhere, UglifyJS will drop it, but will still keep the Math.floor(a/b), not knowing what it does. You can pass pure_funcs: [ 'Math.floor' ] to let it know that this function won't produce any side effect, in which case the whole statement would get discarded. The current implementation adds some overhead (compression will be slower).
drop_console -- default false. Pass true to discard calls to console.* functions. If you wish to drop a specific function call such as console.info and/or retain side effects from function arguments after dropping the function call then use pure_funcs instead.
expression -- default false. Pass true to preserve completion values from terminal statements without return, e.g. in bookmarklets.
keep_fargs -- default true. Prevents the compressor from discarding unused function arguments. You need this for code which relies on Function.length.
keep_fnames -- default false. Pass true to prevent the compressor from discarding function names. Useful for code relying on Function.prototype.name. See also: the keep_fnames mangle option.
passes -- default 1. Number of times to run compress with a maximum of 3. In some cases more than one pass leads to further compressed code. Keep in mind more passes will take more time.
keep_infinity -- default false. Pass true to prevent Infinity from being compressed into 1/0, which may cause performance issues on Chrome.
side_effects -- default true. Pass false to disable potentially dropping functions marked as "pure". A function call is marked as "pure" if a comment annotation /*@__PURE__*/ or /*#__PURE__*/ immediately precedes the call. For example: /*@__PURE__*/foo();
unsafe optionIt enables some transformations that might break code logic in certain contrived cases, but should be fine for most code. You might want to try it on your own code, it should reduce the minified size. Here's what happens when this flag is on:
new Array(1, 2, 3) or Array(1, 2, 3) → [ 1, 2, 3 ]
new Object() → {}
String(exp) or exp.toString() → "" + exp
new Object/RegExp/Function/Error/Array (...) → we discard the new
typeof foo == "undefined" → foo === void 0
void 0 → undefined (if there is a variable named "undefined" in
scope; we do it because the variable name will be mangled, typically
reduced to a single character)
You can use the --define (-d) switch in order to declare global variables that UglifyJS will assume to be constants (unless defined in scope). For example if you pass --define DEBUG=false then, coupled with dead code removal UglifyJS will discard the following from the output:
You can specify nested constants in the form of --define env.DEBUG=false.
UglifyJS will warn about the condition being always false and about dropping unreachable code; for now there is no option to turn off only this specific warning, you can pass warnings=false to turn off all warnings.
Another way of doing that is to declare your globals as constants in a separate file and include it into the build. For example you can have a build/defines.js file with the following:
and build your code like this:
UglifyJS will notice the constants and, since they cannot be altered, it will evaluate references to them to the value itself and drop unreachable code as usual. The build will contain the const declarations if you use them. If you are targeting < ES6 environments which does not support const, using var with reduce_vars (enabled by default) should suffice.
You can also use conditional compilation via the programmatic API. With the difference that the property name is global_defs and is a compressor property:
The code generator tries to output shortest code possible by default. In case you want beautified output, pass --beautify (-b). Optionally you can pass additional arguments that control the code output:
beautify (default true) -- whether to actually beautify the output.
Passing -b will set this to true, but you might need to pass -b even
when you want to generate minified code, in order to specify additional
arguments, so you can use -b beautify=false to override it.
indent-level (default 4)
indent-start (default 0) -- prefix all lines by that many spaces
quote-keys (default false) -- pass true to quote all keys in literal
objects
space-colon (default true) -- insert a space after the colon signs
ascii-only (default false) -- escape Unicode characters in strings and
regexps (affects directives with non-ascii characters becoming invalid)
inline-script (default false) -- escape the slash in occurrences of
</script in strings
width (default 80) -- only takes effect when beautification is on, this
specifies an (orientative) line width that the beautifier will try to
obey. It refers to the width of the line text (excluding indentation).
It doesn't work very well currently, but it does make the code generated
by UglifyJS more readable.
max-line-len (default 32000) -- maximum line length (for uglified code)
bracketize (default false) -- always insert brackets in if, for,
do, while or with statements, even if their body is a single
statement.
semicolons (default true) -- separate statements with semicolons. If
you pass false then whenever possible we will use a newline instead of a
semicolon, leading to more readable output of uglified code (size before
gzip could be smaller; size after gzip insignificantly larger).
preamble (default null) -- when passed it must be a string and
it will be prepended to the output literally. The source map will
adjust for this text. Can be used to insert a comment containing
licensing information, for example.
quote_style (default 0) -- preferred quote style for strings (affects
quoted property names and directives as well):
0 -- prefers double quotes, switches to single quotes when there are
more double quotes in the string itself.
1 -- always use single quotes
2 -- always use double quotes
3 -- always use the original quotes
keep_quoted_props (default false) -- when turned on, prevents stripping
quotes from property names in object literals.
You can pass --comments to retain certain comments in the output. By default it will keep JSDoc-style comments that contain "@preserve", "@license" or "@cc_on" (conditional compilation for IE). You can pass --comments all to keep all the comments, or a valid JavaScript regexp to keep only comments that match this regexp. For example --comments '/foo|bar/' will keep only comments that contain "foo" or "bar".
Note, however, that there might be situations where comments are lost. For example:
Even though it has "@preserve", the comment will be lost because the inner function g (which is the AST node to which the comment is attached to) is discarded by the compressor as not referenced.
The safest comments where to place copyright information (or other info that needs to be kept in the output) are comments attached to toplevel nodes.
UglifyJS2 has its own abstract syntax tree format; for practical reasons we can't easily change to using the SpiderMonkey AST internally. However, UglifyJS now has a converter which can import a SpiderMonkey AST.
For example Acorn is a super-fast parser that produces a SpiderMonkey AST. It has a small CLI utility that parses one file and dumps the AST in JSON on the standard output. To use UglifyJS to mangle and compress that:
The --spidermonkey option tells UglifyJS that all input files are not JavaScript, but JS code described in SpiderMonkey AST in JSON. Therefore we don't use our own parser in this case, but just transform that AST into our internal AST.
More for fun, I added the --acorn option which will use Acorn to do all the parsing. If you pass this option, UglifyJS will require("acorn").
Acorn is really fast (e.g. 250ms instead of 380ms on some 650K code), but converting the SpiderMonkey tree that Acorn produces takes another 150ms so in total it's a bit more than just using UglifyJS's own parser.
Now you can use UglifyJS as any other intermediate tool for transforming JavaScript ASTs in SpiderMonkey format.
Example:
Check out original blog post for details.
Assuming installation via NPM, you can load UglifyJS in your application like this:
It exports a lot of names, but I'll discuss here the basics that are needed for parsing, mangling and compressing a piece of code. The sequence is (1) parse, (2) compress, (3) mangle, (4) generate output code.
There's a single toplevel function which combines all the steps. If you don't need additional customization, you might want to go with minify. Example:
You can also compress multiple files:
To generate a source map:
To generate a source map with the fromString option, you can also use an object:
Note that the source map is not saved in a file, it's just returned in result.map. The value passed for outSourceMap is only used to set //# sourceMappingURL=out.js.map in result.code. The value of outFileName is only used to set file attribute in source map file.
The file attribute in the source map (see the spec) will use outFileName firstly, if it's falsy, then will be deduced from outSourceMap (by removing '.map').
You can set option sourceMapInline to be true and source map will be appended to code.
You can also specify sourceRoot property to be included in source map:
If you're compressing compiled JavaScript and have a source map for it, you can use the inSourceMap argument:
If your input source map is not in a file, you can pass it in as an object using the inSourceMap argument:
The inSourceMap is only used if you also request outSourceMap (it makes no sense otherwise).
To set the source map url, use the sourceMapUrl option. If you're using the X-SourceMap header instead, you can just set the sourceMapUrl option to false. Defaults to outSourceMap:
Other options:
warnings (default false) — pass true to display compressor warnings.
fromString (default false) — if you pass true then you can pass JavaScript source code, rather than file names.
mangle (default true) — pass false to skip mangling names, or pass an object to specify mangling options (see below).
mangleProperties (default false) — pass an object to specify custom mangle property options.
output (default null) — pass an object if you wish to specify additional output options. The defaults are optimized for best compression.
compress (default {}) — pass false to skip compressing entirely. Pass an object to specify custom compressor options.
parse (default {}) — pass an object if you wish to specify some additional parser options. (not all options available... see below)
mangle
except - pass an array of identifiers that should be excluded from mangling
toplevel — mangle names declared in the toplevel scope (disabled by default).
eval — mangle names visible in scopes where eval or with are used (disabled by default).
keep_fnames -- default false. Pass true to not mangle function names. Useful for code relying on Function.prototype.name. See also: the keep_fnames compress option.
Examples:
mangleProperties options
regex — Pass a RegExp to only mangle certain names (maps to the --mangle-regex CLI arguments option)
ignore_quoted – Only mangle unquoted property names (maps to the --mangle-props 2 CLI arguments option)
debug – Mangle names with the original name still present (maps to the --mangle-props-debug CLI arguments option). Defaults to false. Pass an empty string to enable, or a non-empty string to set the suffix.
We could add more options to UglifyJS.minify — if you need additional functionality please suggest!
Following there's more detailed API info, in case the minify function is too simple for your needs.
options is optional and if present it must be an object. The following properties are available:
strict — disable automatic semicolon insertion and support for trailing
comma in arrays and objects
bare_returns — Allow return outside of functions. (maps to the
--bare-returns CLI arguments option and available to minify parse
other options object)
filename — the name of the file where this code is coming from
toplevel — a toplevel node (as returned by a previous invocation of
parse)
The last two options are useful when you'd like to minify multiple files and get a single file as the output and a proper source map. Our CLI tool does something like this:
After this, we have in toplevel a big AST containing all our files, with each token having proper information about where it came from.
UglifyJS contains a scope analyzer that you need to call manually before compressing or mangling. Basically it augments various nodes in the AST with information about where is a name defined, how many times is a name referenced, if it is a global or not, if a function is using eval or the with statement etc. I will discuss this some place else, for now what's important to know is that you need to call the following before doing anything with the tree:
Like this:
The options can be missing. Available options are discussed above in “Compressor options”. Defaults should lead to best compression in most scripts.
The compressor is destructive, so don't rely that toplevel remains the original tree.
After compression it is a good idea to call again figure_out_scope (since the compressor might drop unused variables / unreachable code and this might change the number of identifiers or their position). Optionally, you can call a trick that helps after Gzip (counting character frequency in non-mangleable words). Example:
AST nodes have a print method that takes an output stream. Essentially, to generate code you do this:
or, for a shortcut you can do:
As usual, options is optional. The output stream accepts a lot of options, most of them documented above in section “Beautifier options”. The two which we care about here are source_map and comments.
In order to keep certain comments in the output you need to pass the comments option. Pass a RegExp (as string starting and closing with / or pass a RegExp object), a boolean or a function. Stringified options all and some can be passed too, where some behaves like it's cli equivalent --comments without passing a value. If you pass a RegExp, only those comments whose body matches the RegExp will be kept. Note that body means without the initial // or /*. If you pass a function, it will be called for every comment in the tree and will receive two arguments: the node that the comment is attached to, and the comment token itself.
The comment token has these properties:
type: "comment1" for single-line comments or "comment2" for multi-line
comments
value: the comment body
pos and endpos: the start/end positions (zero-based indexes) in the
original code where this comment appears
line and col: the line and column where this comment appears in the
original code
file — the file name of the original file
nlb — true if there was a newline before this comment in the original
code, or if this comment contains a newline.
Your function should return true to keep the comment, or a falsy value otherwise.
You need to pass the source_map argument when calling print. It needs to be a SourceMap object (which is a thin wrapper on top of the source-map library).
Example:
The source_map_options (optional) can contain the following properties:
file: the name of the JavaScript output file that this mapping refers to
root: the sourceRoot property (see the spec)
orig: the "original source map", handy when you compress generated JS
and want to map the minified output back to the original code where it
came from. It can be simply a string in JSON, or a JSON object containing
the original source map.
constconst in uglify-js@2.x has function scope and as such behaves much like var - unlike const in ES2015 (ES6) which has block scope. It is recommended to avoid using const for this reason as it will have undefined behavior when run on an ES2015 compatible browser.