Related
I'm looking for a good JavaScript equivalent of the C/PHP printf() or for C#/Java programmers, String.Format() (IFormatProvider for .NET).
My basic requirement is a thousand separator format for numbers for now, but something that handles lots of combinations (including dates) would be good.
I realize Microsoft's Ajax library provides a version of String.Format(), but we don't want the entire overhead of that framework.
Current JavaScript
From ES6 on you could use template strings:
let soMany = 10;
console.log(`This is ${soMany} times easier!`);
// "This is 10 times easier!"
See Kim's answer below for details.
Older answer
Try sprintf() for JavaScript.
If you really want to do a simple format method on your own, don’t do the replacements successively but do them simultaneously.
Because most of the other proposals that are mentioned fail when a replace string of previous replacement does also contain a format sequence like this:
"{0}{1}".format("{1}", "{0}")
Normally you would expect the output to be {1}{0} but the actual output is {1}{1}. So do a simultaneous replacement instead like in fearphage’s suggestion.
Building on the previously suggested solutions:
// First, checks if it isn't implemented yet.
if (!String.prototype.format) {
String.prototype.format = function() {
var args = arguments;
return this.replace(/{(\d+)}/g, function(match, number) {
return typeof args[number] != 'undefined'
? args[number]
: match
;
});
};
}
"{0} is dead, but {1} is alive! {0} {2}".format("ASP", "ASP.NET")
outputs
ASP is dead, but ASP.NET is alive! ASP {2}
If you prefer not to modify String's prototype:
if (!String.format) {
String.format = function(format) {
var args = Array.prototype.slice.call(arguments, 1);
return format.replace(/{(\d+)}/g, function(match, number) {
return typeof args[number] != 'undefined'
? args[number]
: match
;
});
};
}
Gives you the much more familiar:
String.format('{0} is dead, but {1} is alive! {0} {2}', 'ASP', 'ASP.NET');
with the same result:
ASP is dead, but ASP.NET is alive! ASP {2}
It's funny because Stack Overflow actually has their own formatting function for the String prototype called formatUnicorn. Try it! Go into the console and type something like:
"Hello, {name}, are you feeling {adjective}?".formatUnicorn({name:"Gabriel", adjective: "OK"});
You get this output:
Hello, Gabriel, are you feeling OK?
You can use objects, arrays, and strings as arguments! I got its code and reworked it to produce a new version of String.prototype.format:
String.prototype.formatUnicorn = String.prototype.formatUnicorn ||
function () {
"use strict";
var str = this.toString();
if (arguments.length) {
var t = typeof arguments[0];
var key;
var args = ("string" === t || "number" === t) ?
Array.prototype.slice.call(arguments)
: arguments[0];
for (key in args) {
str = str.replace(new RegExp("\\{" + key + "\\}", "gi"), args[key]);
}
}
return str;
};
Note the clever Array.prototype.slice.call(arguments) call -- that means if you throw in arguments that are strings or numbers, not a single JSON-style object, you get C#'s String.Format behavior almost exactly.
"a{0}bcd{1}ef".formatUnicorn("FOO", "BAR"); // yields "aFOObcdBARef"
That's because Array's slice will force whatever's in arguments into an Array, whether it was originally or not, and the key will be the index (0, 1, 2...) of each array element coerced into a string (eg, "0", so "\\{0\\}" for your first regexp pattern).
Neat.
Number Formatting in JavaScript
I got to this question page hoping to find how to format numbers in JavaScript, without introducing yet another library. Here's what I've found:
Rounding floating-point numbers
The equivalent of sprintf("%.2f", num) in JavaScript seems to be num.toFixed(2), which formats num to 2 decimal places, with rounding (but see #ars265's comment about Math.round below).
(12.345).toFixed(2); // returns "12.35" (rounding!)
(12.3).toFixed(2); // returns "12.30" (zero padding)
Exponential form
The equivalent of sprintf("%.2e", num) is num.toExponential(2).
(33333).toExponential(2); // "3.33e+4"
Hexadecimal and other bases
To print numbers in base B, try num.toString(B). JavaScript supports automatic conversion to and from bases 2 through 36 (in addition, some browsers have limited support for base64 encoding).
(3735928559).toString(16); // to base 16: "deadbeef"
parseInt("deadbeef", 16); // from base 16: 3735928559
Reference Pages
Quick tutorial on JS number formatting
Mozilla reference page for toFixed() (with links to toPrecision(), toExponential(), toLocaleString(), ...)
From ES6 on you could use template strings:
let soMany = 10;
console.log(`This is ${soMany} times easier!`);
// "This is 10 times easier!"
Be aware that template strings are surrounded by backticks ` instead of (single) quotes.
For further information:
https://developers.google.com/web/updates/2015/01/ES6-Template-Strings
https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/template_strings
Note:
Check the mozilla-site to find a list of supported browsers.
jsxt, Zippo
This option fits better.
String.prototype.format = function() {
var formatted = this;
for (var i = 0; i < arguments.length; i++) {
var regexp = new RegExp('\\{'+i+'\\}', 'gi');
formatted = formatted.replace(regexp, arguments[i]);
}
return formatted;
};
With this option I can replace strings like these:
'The {0} is dead. Don\'t code {0}. Code {1} that is open source!'.format('ASP', 'PHP');
With your code the second {0} wouldn't be replaced. ;)
I use this simple function:
String.prototype.format = function() {
var formatted = this;
for( var arg in arguments ) {
formatted = formatted.replace("{" + arg + "}", arguments[arg]);
}
return formatted;
};
That's very similar to string.format:
"{0} is dead, but {1} is alive!".format("ASP", "ASP.NET")
For Node.js users there is util.format which has printf-like functionality:
util.format("%s world", "Hello")
I'm surprised no one used reduce, this is a native concise and powerful JavaScript function.
ES6 (EcmaScript2015)
String.prototype.format = function() {
return [...arguments].reduce((p,c) => p.replace(/%s/,c), this);
};
console.log('Is that a %s or a %s?... No, it\'s %s!'.format('plane', 'bird', 'SOman'));
< ES6
function interpolate(theString, argumentArray) {
var regex = /%s/;
var _r=function(p,c){return p.replace(regex,c);}
return argumentArray.reduce(_r, theString);
}
interpolate("%s, %s and %s", ["Me", "myself", "I"]); // "Me, myself and I"
How it works:
reduce applies a function against an accumulator and each element in the array (from left to right) to reduce it to a single value.
var _r= function(p,c){return p.replace(/%s/,c)};
console.log(
["a", "b", "c"].reduce(_r, "[%s], [%s] and [%s]") + '\n',
[1, 2, 3].reduce(_r, "%s+%s=%s") + '\n',
["cool", 1337, "stuff"].reduce(_r, "%s %s %s")
);
Here's a minimal implementation of sprintf in JavaScript: it only does "%s" and "%d", but I have left space for it to be extended. It is useless to the OP, but other people who stumble across this thread coming from Google might benefit from it.
function sprintf() {
var args = arguments,
string = args[0],
i = 1;
return string.replace(/%((%)|s|d)/g, function (m) {
// m is the matched format, e.g. %s, %d
var val = null;
if (m[2]) {
val = m[2];
} else {
val = args[i];
// A switch statement so that the formatter can be extended. Default is %s
switch (m) {
case '%d':
val = parseFloat(val);
if (isNaN(val)) {
val = 0;
}
break;
}
i++;
}
return val;
});
}
Example:
alert(sprintf('Latitude: %s, Longitude: %s, Count: %d', 41.847, -87.661, 'two'));
// Expected output: Latitude: 41.847, Longitude: -87.661, Count: 0
In contrast with similar solutions in previous replies, this one does all substitutions in one go, so it will not replace parts of previously replaced values.
3 different ways to format javascript string
There are 3 different ways to format a string by replacing placeholders with the variable value.
Using template literal (backticks ``)
let name = 'John';
let age = 30;
// using backticks
console.log(`${name} is ${age} years old.`);
// John is 30 years old.
Using concatenation
let name = 'John';
let age = 30;
// using concatenation
console.log(name + ' is ' + age + ' years old.');
// John is 30 years old.
Creating own format function
String.prototype.format = function () {
var args = arguments;
return this.replace(/{([0-9]+)}/g, function (match, index) {
// check if the argument is there
return typeof args[index] == 'undefined' ? match : args[index];
});
};
console.log('{0} is {1} years old.'.format('John', 30));
JavaScript programmers can use String.prototype.sprintf at https://github.com/ildar-shaimordanov/jsxt/blob/master/js/String.js. Below is example:
var d = new Date();
var dateStr = '%02d:%02d:%02d'.sprintf(
d.getHours(),
d.getMinutes(),
d.getSeconds());
I want to share my solution for the 'problem'. I haven't re-invented the wheel but tries to find a solution based on what JavaScript already does. The advantage is, that you get all implicit conversions for free. Setting the prototype property $ of String gives a very nice and compact syntax (see examples below). It is maybe not the most efficient way, but in most cases dealing with output it does not have to be super optimized.
String.form = function(str, arr) {
var i = -1;
function callback(exp, p0, p1, p2, p3, p4) {
if (exp=='%%') return '%';
if (arr[++i]===undefined) return undefined;
exp = p2 ? parseInt(p2.substr(1)) : undefined;
var base = p3 ? parseInt(p3.substr(1)) : undefined;
var val;
switch (p4) {
case 's': val = arr[i]; break;
case 'c': val = arr[i][0]; break;
case 'f': val = parseFloat(arr[i]).toFixed(exp); break;
case 'p': val = parseFloat(arr[i]).toPrecision(exp); break;
case 'e': val = parseFloat(arr[i]).toExponential(exp); break;
case 'x': val = parseInt(arr[i]).toString(base?base:16); break;
case 'd': val = parseFloat(parseInt(arr[i], base?base:10).toPrecision(exp)).toFixed(0); break;
}
val = typeof(val)=='object' ? JSON.stringify(val) : val.toString(base);
var sz = parseInt(p1); /* padding size */
var ch = p1 && p1[0]=='0' ? '0' : ' '; /* isnull? */
while (val.length<sz) val = p0 !== undefined ? val+ch : ch+val; /* isminus? */
return val;
}
var regex = /%(-)?(0?[0-9]+)?([.][0-9]+)?([#][0-9]+)?([scfpexd%])/g;
return str.replace(regex, callback);
}
String.prototype.$ = function() {
return String.form(this, Array.prototype.slice.call(arguments));
}
Here are a few examples:
String.format("%s %s", [ "This is a string", 11 ])
console.log("%s %s".$("This is a string", 11))
var arr = [ "12.3", 13.6 ]; console.log("Array: %s".$(arr));
var obj = { test:"test", id:12 }; console.log("Object: %s".$(obj));
console.log("%c", "Test");
console.log("%5d".$(12)); // ' 12'
console.log("%05d".$(12)); // '00012'
console.log("%-5d".$(12)); // '12 '
console.log("%5.2d".$(123)); // ' 120'
console.log("%5.2f".$(1.1)); // ' 1.10'
console.log("%10.2e".$(1.1)); // ' 1.10e+0'
console.log("%5.3p".$(1.12345)); // ' 1.12'
console.log("%5x".$(45054)); // ' affe'
console.log("%20#2x".$("45054")); // ' 1010111111111110'
console.log("%6#2d".$("111")); // ' 7'
console.log("%6#16d".$("affe")); // ' 45054'
Adding to zippoxer's answer, I use this function:
String.prototype.format = function () {
var a = this, b;
for (b in arguments) {
a = a.replace(/%[a-z]/, arguments[b]);
}
return a; // Make chainable
};
var s = 'Hello %s The magic number is %d.';
s.format('world!', 12); // Hello World! The magic number is 12.
I also have a non-prototype version which I use more often for its Java-like syntax:
function format() {
var a, b, c;
a = arguments[0];
b = [];
for(c = 1; c < arguments.length; c++){
b.push(arguments[c]);
}
for (c in b) {
a = a.replace(/%[a-z]/, b[c]);
}
return a;
}
format('%d ducks, 55 %s', 12, 'cats'); // 12 ducks, 55 cats
ES 2015 update
All the cool new stuff in ES 2015 makes this a lot easier:
function format(fmt, ...args){
return fmt
.split("%%")
.reduce((aggregate, chunk, i) =>
aggregate + chunk + (args[i] || ""), "");
}
format("Hello %%! I ate %% apples today.", "World", 44);
// "Hello World, I ate 44 apples today."
I figured that since this, like the older ones, doesn't actually parse the letters, it might as well just use a single token %%. This has the benefit of being obvious and not making it difficult to use a single %. However, if you need %% for some reason, you would need to replace it with itself:
format("I love percentage signs! %%", "%%");
// "I love percentage signs! %%"
+1 Zippo with the exception that the function body needs to be as below or otherwise it appends the current string on every iteration:
String.prototype.format = function() {
var formatted = this;
for (var arg in arguments) {
formatted = formatted.replace("{" + arg + "}", arguments[arg]);
}
return formatted;
};
I use a small library called String.format for JavaScript which supports most of the format string capabilities (including format of numbers and dates), and uses the .NET syntax. The script itself is smaller than 4 kB, so it doesn't create much of overhead.
I'll add my own discoveries which I've found since I asked:
number_format (for thousand separator/currency formatting)
sprintf (same author as above)
Sadly it seems sprintf doesn't handle thousand separator formatting like .NET's string format.
If you are looking to handle the thousands separator, you should really use toLocaleString() from the JavaScript Number class since it will format the string for the user's region.
The JavaScript Date class can format localized dates and times.
Very elegant:
String.prototype.format = function (){
var args = arguments;
return this.replace(/\{\{|\}\}|\{(\d+)\}/g, function (curlyBrack, index) {
return ((curlyBrack == "{{") ? "{" : ((curlyBrack == "}}") ? "}" : args[index]));
});
};
// Usage:
"{0}{1}".format("{1}", "{0}")
Credit goes to (broken link) https://gist.github.com/0i0/1519811
There is "sprintf" for JavaScript which you can find at http://www.webtoolkit.info/javascript-sprintf.html.
The PHPJS project has written JavaScript implementations for many of PHP's functions. Since PHP's sprintf() function is basically the same as C's printf(), their JavaScript implementation of it should satisfy your needs.
I use this one:
String.prototype.format = function() {
var newStr = this, i = 0;
while (/%s/.test(newStr))
newStr = newStr.replace("%s", arguments[i++])
return newStr;
}
Then I call it:
"<h1>%s</h1><p>%s</p>".format("Header", "Just a test!");
I have a solution very close to Peter's, but it deals with number and object case.
if (!String.prototype.format) {
String.prototype.format = function() {
var args;
args = arguments;
if (args.length === 1 && args[0] !== null && typeof args[0] === 'object') {
args = args[0];
}
return this.replace(/{([^}]*)}/g, function(match, key) {
return (typeof args[key] !== "undefined" ? args[key] : match);
});
};
}
Maybe it could be even better to deal with the all deeps cases, but for my needs this is just fine.
"This is an example from {name}".format({name:"Blaine"});
"This is an example from {0}".format("Blaine");
PS: This function is very cool if you are using translations in templates frameworks like AngularJS:
<h1> {{('hello-message'|translate).format(user)}} <h1>
<h1> {{('hello-by-name'|translate).format( user ? user.name : 'You' )}} <h1>
Where the en.json is something like
{
"hello-message": "Hello {name}, welcome.",
"hello-by-name": "Hello {0}, welcome."
}
One very slightly different version, the one I prefer (this one uses {xxx} tokens rather than {0} numbered arguments, this is much more self-documenting and suits localization much better):
String.prototype.format = function(tokens) {
var formatted = this;
for (var token in tokens)
if (tokens.hasOwnProperty(token))
formatted = formatted.replace(RegExp("{" + token + "}", "g"), tokens[token]);
return formatted;
};
A variation would be:
var formatted = l(this);
that calls an l() localization function first.
For basic formatting:
var template = jQuery.validator.format("{0} is not a valid value");
var result = template("abc");
We can use a simple lightweight String.Format string operation library for Typescript.
String.Format():
var id = image.GetId()
String.Format("image_{0}.jpg", id)
output: "image_2db5da20-1c5d-4f1a-8fd4-b41e34c8c5b5.jpg";
String Format for specifiers:
var value = String.Format("{0:L}", "APPLE"); //output "apple"
value = String.Format("{0:U}", "apple"); // output "APPLE"
value = String.Format("{0:d}", "2017-01-23 00:00"); //output "23.01.2017"
value = String.Format("{0:s}", "21.03.2017 22:15:01") //output "2017-03-21T22:15:01"
value = String.Format("{0:n}", 1000000);
//output "1.000.000"
value = String.Format("{0:00}", 1);
//output "01"
String Format for Objects including specifiers:
var fruit = new Fruit();
fruit.type = "apple";
fruit.color = "RED";
fruit.shippingDate = new Date(2018, 1, 1);
fruit.amount = 10000;
String.Format("the {type:U} is {color:L} shipped on {shippingDate:s} with an amount of {amount:n}", fruit);
// output: the APPLE is red shipped on 2018-01-01 with an amount of 10.000
Just in case someone needs a function to prevent polluting global scope, here is the function that does the same:
function _format (str, arr) {
return str.replace(/{(\d+)}/g, function (match, number) {
return typeof arr[number] != 'undefined' ? arr[number] : match;
});
};
For those who like Node.JS and its util.format feature, I've just extracted it out into its vanilla JavaScript form (with only functions that util.format uses):
exports = {};
function isString(arg) {
return typeof arg === 'string';
}
function isNull(arg) {
return arg === null;
}
function isObject(arg) {
return typeof arg === 'object' && arg !== null;
}
function isBoolean(arg) {
return typeof arg === 'boolean';
}
function isUndefined(arg) {
return arg === void 0;
}
function stylizeNoColor(str, styleType) {
return str;
}
function stylizeWithColor(str, styleType) {
var style = inspect.styles[styleType];
if (style) {
return '\u001b[' + inspect.colors[style][0] + 'm' + str +
'\u001b[' + inspect.colors[style][3] + 'm';
} else {
return str;
}
}
function isFunction(arg) {
return typeof arg === 'function';
}
function isNumber(arg) {
return typeof arg === 'number';
}
function isSymbol(arg) {
return typeof arg === 'symbol';
}
function formatPrimitive(ctx, value) {
if (isUndefined(value))
return ctx.stylize('undefined', 'undefined');
if (isString(value)) {
var simple = '\'' + JSON.stringify(value).replace(/^"|"$/g, '')
.replace(/'/g, "\\'")
.replace(/\\"/g, '"') + '\'';
return ctx.stylize(simple, 'string');
}
if (isNumber(value)) {
// Format -0 as '-0'. Strict equality won't distinguish 0 from -0,
// so instead we use the fact that 1 / -0 < 0 whereas 1 / 0 > 0 .
if (value === 0 && 1 / value < 0)
return ctx.stylize('-0', 'number');
return ctx.stylize('' + value, 'number');
}
if (isBoolean(value))
return ctx.stylize('' + value, 'boolean');
// For some reason typeof null is "object", so special case here.
if (isNull(value))
return ctx.stylize('null', 'null');
// es6 symbol primitive
if (isSymbol(value))
return ctx.stylize(value.toString(), 'symbol');
}
function arrayToHash(array) {
var hash = {};
array.forEach(function (val, idx) {
hash[val] = true;
});
return hash;
}
function objectToString(o) {
return Object.prototype.toString.call(o);
}
function isDate(d) {
return isObject(d) && objectToString(d) === '[object Date]';
}
function isError(e) {
return isObject(e) &&
(objectToString(e) === '[object Error]' || e instanceof Error);
}
function isRegExp(re) {
return isObject(re) && objectToString(re) === '[object RegExp]';
}
function formatError(value) {
return '[' + Error.prototype.toString.call(value) + ']';
}
function formatPrimitiveNoColor(ctx, value) {
var stylize = ctx.stylize;
ctx.stylize = stylizeNoColor;
var str = formatPrimitive(ctx, value);
ctx.stylize = stylize;
return str;
}
function isArray(ar) {
return Array.isArray(ar);
}
function hasOwnProperty(obj, prop) {
return Object.prototype.hasOwnProperty.call(obj, prop);
}
function formatProperty(ctx, value, recurseTimes, visibleKeys, key, array) {
var name, str, desc;
desc = Object.getOwnPropertyDescriptor(value, key) || {value: value[key]};
if (desc.get) {
if (desc.set) {
str = ctx.stylize('[Getter/Setter]', 'special');
} else {
str = ctx.stylize('[Getter]', 'special');
}
} else {
if (desc.set) {
str = ctx.stylize('[Setter]', 'special');
}
}
if (!hasOwnProperty(visibleKeys, key)) {
name = '[' + key + ']';
}
if (!str) {
if (ctx.seen.indexOf(desc.value) < 0) {
if (isNull(recurseTimes)) {
str = formatValue(ctx, desc.value, null);
} else {
str = formatValue(ctx, desc.value, recurseTimes - 1);
}
if (str.indexOf('\n') > -1) {
if (array) {
str = str.split('\n').map(function (line) {
return ' ' + line;
}).join('\n').substr(2);
} else {
str = '\n' + str.split('\n').map(function (line) {
return ' ' + line;
}).join('\n');
}
}
} else {
str = ctx.stylize('[Circular]', 'special');
}
}
if (isUndefined(name)) {
if (array && key.match(/^\d+$/)) {
return str;
}
name = JSON.stringify('' + key);
if (name.match(/^"([a-zA-Z_][a-zA-Z_0-9]*)"$/)) {
name = name.substr(1, name.length - 2);
name = ctx.stylize(name, 'name');
} else {
name = name.replace(/'/g, "\\'")
.replace(/\\"/g, '"')
.replace(/(^"|"$)/g, "'")
.replace(/\\\\/g, '\\');
name = ctx.stylize(name, 'string');
}
}
return name + ': ' + str;
}
function formatArray(ctx, value, recurseTimes, visibleKeys, keys) {
var output = [];
for (var i = 0, l = value.length; i < l; ++i) {
if (hasOwnProperty(value, String(i))) {
output.push(formatProperty(ctx, value, recurseTimes, visibleKeys,
String(i), true));
} else {
output.push('');
}
}
keys.forEach(function (key) {
if (!key.match(/^\d+$/)) {
output.push(formatProperty(ctx, value, recurseTimes, visibleKeys,
key, true));
}
});
return output;
}
function reduceToSingleString(output, base, braces) {
var length = output.reduce(function (prev, cur) {
return prev + cur.replace(/\u001b\[\d\d?m/g, '').length + 1;
}, 0);
if (length > 60) {
return braces[0] +
(base === '' ? '' : base + '\n ') +
' ' +
output.join(',\n ') +
' ' +
braces[1];
}
return braces[0] + base + ' ' + output.join(', ') + ' ' + braces[1];
}
function formatValue(ctx, value, recurseTimes) {
// Provide a hook for user-specified inspect functions.
// Check that value is an object with an inspect function on it
if (ctx.customInspect &&
value &&
isFunction(value.inspect) &&
// Filter out the util module, it's inspect function is special
value.inspect !== exports.inspect &&
// Also filter out any prototype objects using the circular check.
!(value.constructor && value.constructor.prototype === value)) {
var ret = value.inspect(recurseTimes, ctx);
if (!isString(ret)) {
ret = formatValue(ctx, ret, recurseTimes);
}
return ret;
}
// Primitive types cannot have properties
var primitive = formatPrimitive(ctx, value);
if (primitive) {
return primitive;
}
// Look up the keys of the object.
var keys = Object.keys(value);
var visibleKeys = arrayToHash(keys);
if (ctx.showHidden) {
keys = Object.getOwnPropertyNames(value);
}
// This could be a boxed primitive (new String(), etc.), check valueOf()
// NOTE: Avoid calling `valueOf` on `Date` instance because it will return
// a number which, when object has some additional user-stored `keys`,
// will be printed out.
var formatted;
var raw = value;
try {
// the .valueOf() call can fail for a multitude of reasons
if (!isDate(value))
raw = value.valueOf();
} catch (e) {
// ignore...
}
if (isString(raw)) {
// for boxed Strings, we have to remove the 0-n indexed entries,
// since they just noisey up the output and are redundant
keys = keys.filter(function (key) {
return !(key >= 0 && key < raw.length);
});
}
// Some type of object without properties can be shortcutted.
if (keys.length === 0) {
if (isFunction(value)) {
var name = value.name ? ': ' + value.name : '';
return ctx.stylize('[Function' + name + ']', 'special');
}
if (isRegExp(value)) {
return ctx.stylize(RegExp.prototype.toString.call(value), 'regexp');
}
if (isDate(value)) {
return ctx.stylize(Date.prototype.toString.call(value), 'date');
}
if (isError(value)) {
return formatError(value);
}
// now check the `raw` value to handle boxed primitives
if (isString(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
return ctx.stylize('[String: ' + formatted + ']', 'string');
}
if (isNumber(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
return ctx.stylize('[Number: ' + formatted + ']', 'number');
}
if (isBoolean(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
return ctx.stylize('[Boolean: ' + formatted + ']', 'boolean');
}
}
var base = '', array = false, braces = ['{', '}'];
// Make Array say that they are Array
if (isArray(value)) {
array = true;
braces = ['[', ']'];
}
// Make functions say that they are functions
if (isFunction(value)) {
var n = value.name ? ': ' + value.name : '';
base = ' [Function' + n + ']';
}
// Make RegExps say that they are RegExps
if (isRegExp(value)) {
base = ' ' + RegExp.prototype.toString.call(value);
}
// Make dates with properties first say the date
if (isDate(value)) {
base = ' ' + Date.prototype.toUTCString.call(value);
}
// Make error with message first say the error
if (isError(value)) {
base = ' ' + formatError(value);
}
// Make boxed primitive Strings look like such
if (isString(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
base = ' ' + '[String: ' + formatted + ']';
}
// Make boxed primitive Numbers look like such
if (isNumber(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
base = ' ' + '[Number: ' + formatted + ']';
}
// Make boxed primitive Booleans look like such
if (isBoolean(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
base = ' ' + '[Boolean: ' + formatted + ']';
}
if (keys.length === 0 && (!array || value.length === 0)) {
return braces[0] + base + braces[1];
}
if (recurseTimes < 0) {
if (isRegExp(value)) {
return ctx.stylize(RegExp.prototype.toString.call(value), 'regexp');
} else {
return ctx.stylize('[Object]', 'special');
}
}
ctx.seen.push(value);
var output;
if (array) {
output = formatArray(ctx, value, recurseTimes, visibleKeys, keys);
} else {
output = keys.map(function (key) {
return formatProperty(ctx, value, recurseTimes, visibleKeys, key, array);
});
}
ctx.seen.pop();
return reduceToSingleString(output, base, braces);
}
function inspect(obj, opts) {
// default options
var ctx = {
seen: [],
stylize: stylizeNoColor
};
// legacy...
if (arguments.length >= 3) ctx.depth = arguments[2];
if (arguments.length >= 4) ctx.colors = arguments[3];
if (isBoolean(opts)) {
// legacy...
ctx.showHidden = opts;
} else if (opts) {
// got an "options" object
exports._extend(ctx, opts);
}
// set default options
if (isUndefined(ctx.showHidden)) ctx.showHidden = false;
if (isUndefined(ctx.depth)) ctx.depth = 2;
if (isUndefined(ctx.colors)) ctx.colors = false;
if (isUndefined(ctx.customInspect)) ctx.customInspect = true;
if (ctx.colors) ctx.stylize = stylizeWithColor;
return formatValue(ctx, obj, ctx.depth);
}
exports.inspect = inspect;
// http://en.wikipedia.org/wiki/ANSI_escape_code#graphics
inspect.colors = {
'bold': [1, 22],
'italic': [3, 23],
'underline': [4, 24],
'inverse': [7, 27],
'white': [37, 39],
'grey': [90, 39],
'black': [30, 39],
'blue': [34, 39],
'cyan': [36, 39],
'green': [32, 39],
'magenta': [35, 39],
'red': [31, 39],
'yellow': [33, 39]
};
// Don't use 'blue' not visible on cmd.exe
inspect.styles = {
'special': 'cyan',
'number': 'yellow',
'boolean': 'yellow',
'undefined': 'grey',
'null': 'bold',
'string': 'green',
'symbol': 'green',
'date': 'magenta',
// "name": intentionally not styling
'regexp': 'red'
};
var formatRegExp = /%[sdj%]/g;
exports.format = function (f) {
if (!isString(f)) {
var objects = [];
for (var j = 0; j < arguments.length; j++) {
objects.push(inspect(arguments[j]));
}
return objects.join(' ');
}
var i = 1;
var args = arguments;
var len = args.length;
var str = String(f).replace(formatRegExp, function (x) {
if (x === '%%') return '%';
if (i >= len) return x;
switch (x) {
case '%s':
return String(args[i++]);
case '%d':
return Number(args[i++]);
case '%j':
try {
return JSON.stringify(args[i++]);
} catch (_) {
return '[Circular]';
}
default:
return x;
}
});
for (var x = args[i]; i < len; x = args[++i]) {
if (isNull(x) || !isObject(x)) {
str += ' ' + x;
} else {
str += ' ' + inspect(x);
}
}
return str;
};
Harvested from: https://github.com/joyent/node/blob/master/lib/util.js
I have a slightly longer formatter for JavaScript here...
You can do formatting several ways:
String.format(input, args0, arg1, ...)
String.format(input, obj)
"literal".format(arg0, arg1, ...)
"literal".format(obj)
Also, if you have say a ObjectBase.prototype.format (such as with DateJS) it will use that.
Examples...
var input = "numbered args ({0}-{1}-{2}-{3})";
console.log(String.format(input, "first", 2, new Date()));
//Outputs "numbered args (first-2-Thu May 31 2012...Time)-{3})"
console.log(input.format("first", 2, new Date()));
//Outputs "numbered args(first-2-Thu May 31 2012...Time)-{3})"
console.log(input.format(
"object properties ({first}-{second}-{third:yyyy-MM-dd}-{fourth})"
,{
'first':'first'
,'second':2
,'third':new Date() //assumes Date.prototype.format method
}
));
//Outputs "object properties (first-2-2012-05-31-{3})"
I've also aliased with .asFormat and have some detection in place in case there's already a string.format (such as with MS Ajax Toolkit (I hate that library).
Using Lodash you can get template functionality:
Use the ES template literal delimiter as an "interpolate" delimiter.
Disable support by replacing the "interpolate" delimiter.
var compiled = _.template('hello ${ user }!');
compiled({ 'user': 'pebbles' });
// => 'hello pebbles!
I'm looking for a good JavaScript equivalent of the C/PHP printf() or for C#/Java programmers, String.Format() (IFormatProvider for .NET).
My basic requirement is a thousand separator format for numbers for now, but something that handles lots of combinations (including dates) would be good.
I realize Microsoft's Ajax library provides a version of String.Format(), but we don't want the entire overhead of that framework.
Current JavaScript
From ES6 on you could use template strings:
let soMany = 10;
console.log(`This is ${soMany} times easier!`);
// "This is 10 times easier!"
See Kim's answer below for details.
Older answer
Try sprintf() for JavaScript.
If you really want to do a simple format method on your own, don’t do the replacements successively but do them simultaneously.
Because most of the other proposals that are mentioned fail when a replace string of previous replacement does also contain a format sequence like this:
"{0}{1}".format("{1}", "{0}")
Normally you would expect the output to be {1}{0} but the actual output is {1}{1}. So do a simultaneous replacement instead like in fearphage’s suggestion.
Building on the previously suggested solutions:
// First, checks if it isn't implemented yet.
if (!String.prototype.format) {
String.prototype.format = function() {
var args = arguments;
return this.replace(/{(\d+)}/g, function(match, number) {
return typeof args[number] != 'undefined'
? args[number]
: match
;
});
};
}
"{0} is dead, but {1} is alive! {0} {2}".format("ASP", "ASP.NET")
outputs
ASP is dead, but ASP.NET is alive! ASP {2}
If you prefer not to modify String's prototype:
if (!String.format) {
String.format = function(format) {
var args = Array.prototype.slice.call(arguments, 1);
return format.replace(/{(\d+)}/g, function(match, number) {
return typeof args[number] != 'undefined'
? args[number]
: match
;
});
};
}
Gives you the much more familiar:
String.format('{0} is dead, but {1} is alive! {0} {2}', 'ASP', 'ASP.NET');
with the same result:
ASP is dead, but ASP.NET is alive! ASP {2}
It's funny because Stack Overflow actually has their own formatting function for the String prototype called formatUnicorn. Try it! Go into the console and type something like:
"Hello, {name}, are you feeling {adjective}?".formatUnicorn({name:"Gabriel", adjective: "OK"});
You get this output:
Hello, Gabriel, are you feeling OK?
You can use objects, arrays, and strings as arguments! I got its code and reworked it to produce a new version of String.prototype.format:
String.prototype.formatUnicorn = String.prototype.formatUnicorn ||
function () {
"use strict";
var str = this.toString();
if (arguments.length) {
var t = typeof arguments[0];
var key;
var args = ("string" === t || "number" === t) ?
Array.prototype.slice.call(arguments)
: arguments[0];
for (key in args) {
str = str.replace(new RegExp("\\{" + key + "\\}", "gi"), args[key]);
}
}
return str;
};
Note the clever Array.prototype.slice.call(arguments) call -- that means if you throw in arguments that are strings or numbers, not a single JSON-style object, you get C#'s String.Format behavior almost exactly.
"a{0}bcd{1}ef".formatUnicorn("FOO", "BAR"); // yields "aFOObcdBARef"
That's because Array's slice will force whatever's in arguments into an Array, whether it was originally or not, and the key will be the index (0, 1, 2...) of each array element coerced into a string (eg, "0", so "\\{0\\}" for your first regexp pattern).
Neat.
Number Formatting in JavaScript
I got to this question page hoping to find how to format numbers in JavaScript, without introducing yet another library. Here's what I've found:
Rounding floating-point numbers
The equivalent of sprintf("%.2f", num) in JavaScript seems to be num.toFixed(2), which formats num to 2 decimal places, with rounding (but see #ars265's comment about Math.round below).
(12.345).toFixed(2); // returns "12.35" (rounding!)
(12.3).toFixed(2); // returns "12.30" (zero padding)
Exponential form
The equivalent of sprintf("%.2e", num) is num.toExponential(2).
(33333).toExponential(2); // "3.33e+4"
Hexadecimal and other bases
To print numbers in base B, try num.toString(B). JavaScript supports automatic conversion to and from bases 2 through 36 (in addition, some browsers have limited support for base64 encoding).
(3735928559).toString(16); // to base 16: "deadbeef"
parseInt("deadbeef", 16); // from base 16: 3735928559
Reference Pages
Quick tutorial on JS number formatting
Mozilla reference page for toFixed() (with links to toPrecision(), toExponential(), toLocaleString(), ...)
From ES6 on you could use template strings:
let soMany = 10;
console.log(`This is ${soMany} times easier!`);
// "This is 10 times easier!"
Be aware that template strings are surrounded by backticks ` instead of (single) quotes.
For further information:
https://developers.google.com/web/updates/2015/01/ES6-Template-Strings
https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/template_strings
Note:
Check the mozilla-site to find a list of supported browsers.
jsxt, Zippo
This option fits better.
String.prototype.format = function() {
var formatted = this;
for (var i = 0; i < arguments.length; i++) {
var regexp = new RegExp('\\{'+i+'\\}', 'gi');
formatted = formatted.replace(regexp, arguments[i]);
}
return formatted;
};
With this option I can replace strings like these:
'The {0} is dead. Don\'t code {0}. Code {1} that is open source!'.format('ASP', 'PHP');
With your code the second {0} wouldn't be replaced. ;)
I use this simple function:
String.prototype.format = function() {
var formatted = this;
for( var arg in arguments ) {
formatted = formatted.replace("{" + arg + "}", arguments[arg]);
}
return formatted;
};
That's very similar to string.format:
"{0} is dead, but {1} is alive!".format("ASP", "ASP.NET")
For Node.js users there is util.format which has printf-like functionality:
util.format("%s world", "Hello")
I'm surprised no one used reduce, this is a native concise and powerful JavaScript function.
ES6 (EcmaScript2015)
String.prototype.format = function() {
return [...arguments].reduce((p,c) => p.replace(/%s/,c), this);
};
console.log('Is that a %s or a %s?... No, it\'s %s!'.format('plane', 'bird', 'SOman'));
< ES6
function interpolate(theString, argumentArray) {
var regex = /%s/;
var _r=function(p,c){return p.replace(regex,c);}
return argumentArray.reduce(_r, theString);
}
interpolate("%s, %s and %s", ["Me", "myself", "I"]); // "Me, myself and I"
How it works:
reduce applies a function against an accumulator and each element in the array (from left to right) to reduce it to a single value.
var _r= function(p,c){return p.replace(/%s/,c)};
console.log(
["a", "b", "c"].reduce(_r, "[%s], [%s] and [%s]") + '\n',
[1, 2, 3].reduce(_r, "%s+%s=%s") + '\n',
["cool", 1337, "stuff"].reduce(_r, "%s %s %s")
);
Here's a minimal implementation of sprintf in JavaScript: it only does "%s" and "%d", but I have left space for it to be extended. It is useless to the OP, but other people who stumble across this thread coming from Google might benefit from it.
function sprintf() {
var args = arguments,
string = args[0],
i = 1;
return string.replace(/%((%)|s|d)/g, function (m) {
// m is the matched format, e.g. %s, %d
var val = null;
if (m[2]) {
val = m[2];
} else {
val = args[i];
// A switch statement so that the formatter can be extended. Default is %s
switch (m) {
case '%d':
val = parseFloat(val);
if (isNaN(val)) {
val = 0;
}
break;
}
i++;
}
return val;
});
}
Example:
alert(sprintf('Latitude: %s, Longitude: %s, Count: %d', 41.847, -87.661, 'two'));
// Expected output: Latitude: 41.847, Longitude: -87.661, Count: 0
In contrast with similar solutions in previous replies, this one does all substitutions in one go, so it will not replace parts of previously replaced values.
3 different ways to format javascript string
There are 3 different ways to format a string by replacing placeholders with the variable value.
Using template literal (backticks ``)
let name = 'John';
let age = 30;
// using backticks
console.log(`${name} is ${age} years old.`);
// John is 30 years old.
Using concatenation
let name = 'John';
let age = 30;
// using concatenation
console.log(name + ' is ' + age + ' years old.');
// John is 30 years old.
Creating own format function
String.prototype.format = function () {
var args = arguments;
return this.replace(/{([0-9]+)}/g, function (match, index) {
// check if the argument is there
return typeof args[index] == 'undefined' ? match : args[index];
});
};
console.log('{0} is {1} years old.'.format('John', 30));
JavaScript programmers can use String.prototype.sprintf at https://github.com/ildar-shaimordanov/jsxt/blob/master/js/String.js. Below is example:
var d = new Date();
var dateStr = '%02d:%02d:%02d'.sprintf(
d.getHours(),
d.getMinutes(),
d.getSeconds());
I want to share my solution for the 'problem'. I haven't re-invented the wheel but tries to find a solution based on what JavaScript already does. The advantage is, that you get all implicit conversions for free. Setting the prototype property $ of String gives a very nice and compact syntax (see examples below). It is maybe not the most efficient way, but in most cases dealing with output it does not have to be super optimized.
String.form = function(str, arr) {
var i = -1;
function callback(exp, p0, p1, p2, p3, p4) {
if (exp=='%%') return '%';
if (arr[++i]===undefined) return undefined;
exp = p2 ? parseInt(p2.substr(1)) : undefined;
var base = p3 ? parseInt(p3.substr(1)) : undefined;
var val;
switch (p4) {
case 's': val = arr[i]; break;
case 'c': val = arr[i][0]; break;
case 'f': val = parseFloat(arr[i]).toFixed(exp); break;
case 'p': val = parseFloat(arr[i]).toPrecision(exp); break;
case 'e': val = parseFloat(arr[i]).toExponential(exp); break;
case 'x': val = parseInt(arr[i]).toString(base?base:16); break;
case 'd': val = parseFloat(parseInt(arr[i], base?base:10).toPrecision(exp)).toFixed(0); break;
}
val = typeof(val)=='object' ? JSON.stringify(val) : val.toString(base);
var sz = parseInt(p1); /* padding size */
var ch = p1 && p1[0]=='0' ? '0' : ' '; /* isnull? */
while (val.length<sz) val = p0 !== undefined ? val+ch : ch+val; /* isminus? */
return val;
}
var regex = /%(-)?(0?[0-9]+)?([.][0-9]+)?([#][0-9]+)?([scfpexd%])/g;
return str.replace(regex, callback);
}
String.prototype.$ = function() {
return String.form(this, Array.prototype.slice.call(arguments));
}
Here are a few examples:
String.format("%s %s", [ "This is a string", 11 ])
console.log("%s %s".$("This is a string", 11))
var arr = [ "12.3", 13.6 ]; console.log("Array: %s".$(arr));
var obj = { test:"test", id:12 }; console.log("Object: %s".$(obj));
console.log("%c", "Test");
console.log("%5d".$(12)); // ' 12'
console.log("%05d".$(12)); // '00012'
console.log("%-5d".$(12)); // '12 '
console.log("%5.2d".$(123)); // ' 120'
console.log("%5.2f".$(1.1)); // ' 1.10'
console.log("%10.2e".$(1.1)); // ' 1.10e+0'
console.log("%5.3p".$(1.12345)); // ' 1.12'
console.log("%5x".$(45054)); // ' affe'
console.log("%20#2x".$("45054")); // ' 1010111111111110'
console.log("%6#2d".$("111")); // ' 7'
console.log("%6#16d".$("affe")); // ' 45054'
Adding to zippoxer's answer, I use this function:
String.prototype.format = function () {
var a = this, b;
for (b in arguments) {
a = a.replace(/%[a-z]/, arguments[b]);
}
return a; // Make chainable
};
var s = 'Hello %s The magic number is %d.';
s.format('world!', 12); // Hello World! The magic number is 12.
I also have a non-prototype version which I use more often for its Java-like syntax:
function format() {
var a, b, c;
a = arguments[0];
b = [];
for(c = 1; c < arguments.length; c++){
b.push(arguments[c]);
}
for (c in b) {
a = a.replace(/%[a-z]/, b[c]);
}
return a;
}
format('%d ducks, 55 %s', 12, 'cats'); // 12 ducks, 55 cats
ES 2015 update
All the cool new stuff in ES 2015 makes this a lot easier:
function format(fmt, ...args){
return fmt
.split("%%")
.reduce((aggregate, chunk, i) =>
aggregate + chunk + (args[i] || ""), "");
}
format("Hello %%! I ate %% apples today.", "World", 44);
// "Hello World, I ate 44 apples today."
I figured that since this, like the older ones, doesn't actually parse the letters, it might as well just use a single token %%. This has the benefit of being obvious and not making it difficult to use a single %. However, if you need %% for some reason, you would need to replace it with itself:
format("I love percentage signs! %%", "%%");
// "I love percentage signs! %%"
+1 Zippo with the exception that the function body needs to be as below or otherwise it appends the current string on every iteration:
String.prototype.format = function() {
var formatted = this;
for (var arg in arguments) {
formatted = formatted.replace("{" + arg + "}", arguments[arg]);
}
return formatted;
};
I use a small library called String.format for JavaScript which supports most of the format string capabilities (including format of numbers and dates), and uses the .NET syntax. The script itself is smaller than 4 kB, so it doesn't create much of overhead.
I'll add my own discoveries which I've found since I asked:
number_format (for thousand separator/currency formatting)
sprintf (same author as above)
Sadly it seems sprintf doesn't handle thousand separator formatting like .NET's string format.
If you are looking to handle the thousands separator, you should really use toLocaleString() from the JavaScript Number class since it will format the string for the user's region.
The JavaScript Date class can format localized dates and times.
Very elegant:
String.prototype.format = function (){
var args = arguments;
return this.replace(/\{\{|\}\}|\{(\d+)\}/g, function (curlyBrack, index) {
return ((curlyBrack == "{{") ? "{" : ((curlyBrack == "}}") ? "}" : args[index]));
});
};
// Usage:
"{0}{1}".format("{1}", "{0}")
Credit goes to (broken link) https://gist.github.com/0i0/1519811
There is "sprintf" for JavaScript which you can find at http://www.webtoolkit.info/javascript-sprintf.html.
The PHPJS project has written JavaScript implementations for many of PHP's functions. Since PHP's sprintf() function is basically the same as C's printf(), their JavaScript implementation of it should satisfy your needs.
I use this one:
String.prototype.format = function() {
var newStr = this, i = 0;
while (/%s/.test(newStr))
newStr = newStr.replace("%s", arguments[i++])
return newStr;
}
Then I call it:
"<h1>%s</h1><p>%s</p>".format("Header", "Just a test!");
I have a solution very close to Peter's, but it deals with number and object case.
if (!String.prototype.format) {
String.prototype.format = function() {
var args;
args = arguments;
if (args.length === 1 && args[0] !== null && typeof args[0] === 'object') {
args = args[0];
}
return this.replace(/{([^}]*)}/g, function(match, key) {
return (typeof args[key] !== "undefined" ? args[key] : match);
});
};
}
Maybe it could be even better to deal with the all deeps cases, but for my needs this is just fine.
"This is an example from {name}".format({name:"Blaine"});
"This is an example from {0}".format("Blaine");
PS: This function is very cool if you are using translations in templates frameworks like AngularJS:
<h1> {{('hello-message'|translate).format(user)}} <h1>
<h1> {{('hello-by-name'|translate).format( user ? user.name : 'You' )}} <h1>
Where the en.json is something like
{
"hello-message": "Hello {name}, welcome.",
"hello-by-name": "Hello {0}, welcome."
}
One very slightly different version, the one I prefer (this one uses {xxx} tokens rather than {0} numbered arguments, this is much more self-documenting and suits localization much better):
String.prototype.format = function(tokens) {
var formatted = this;
for (var token in tokens)
if (tokens.hasOwnProperty(token))
formatted = formatted.replace(RegExp("{" + token + "}", "g"), tokens[token]);
return formatted;
};
A variation would be:
var formatted = l(this);
that calls an l() localization function first.
For basic formatting:
var template = jQuery.validator.format("{0} is not a valid value");
var result = template("abc");
We can use a simple lightweight String.Format string operation library for Typescript.
String.Format():
var id = image.GetId()
String.Format("image_{0}.jpg", id)
output: "image_2db5da20-1c5d-4f1a-8fd4-b41e34c8c5b5.jpg";
String Format for specifiers:
var value = String.Format("{0:L}", "APPLE"); //output "apple"
value = String.Format("{0:U}", "apple"); // output "APPLE"
value = String.Format("{0:d}", "2017-01-23 00:00"); //output "23.01.2017"
value = String.Format("{0:s}", "21.03.2017 22:15:01") //output "2017-03-21T22:15:01"
value = String.Format("{0:n}", 1000000);
//output "1.000.000"
value = String.Format("{0:00}", 1);
//output "01"
String Format for Objects including specifiers:
var fruit = new Fruit();
fruit.type = "apple";
fruit.color = "RED";
fruit.shippingDate = new Date(2018, 1, 1);
fruit.amount = 10000;
String.Format("the {type:U} is {color:L} shipped on {shippingDate:s} with an amount of {amount:n}", fruit);
// output: the APPLE is red shipped on 2018-01-01 with an amount of 10.000
Just in case someone needs a function to prevent polluting global scope, here is the function that does the same:
function _format (str, arr) {
return str.replace(/{(\d+)}/g, function (match, number) {
return typeof arr[number] != 'undefined' ? arr[number] : match;
});
};
For those who like Node.JS and its util.format feature, I've just extracted it out into its vanilla JavaScript form (with only functions that util.format uses):
exports = {};
function isString(arg) {
return typeof arg === 'string';
}
function isNull(arg) {
return arg === null;
}
function isObject(arg) {
return typeof arg === 'object' && arg !== null;
}
function isBoolean(arg) {
return typeof arg === 'boolean';
}
function isUndefined(arg) {
return arg === void 0;
}
function stylizeNoColor(str, styleType) {
return str;
}
function stylizeWithColor(str, styleType) {
var style = inspect.styles[styleType];
if (style) {
return '\u001b[' + inspect.colors[style][0] + 'm' + str +
'\u001b[' + inspect.colors[style][3] + 'm';
} else {
return str;
}
}
function isFunction(arg) {
return typeof arg === 'function';
}
function isNumber(arg) {
return typeof arg === 'number';
}
function isSymbol(arg) {
return typeof arg === 'symbol';
}
function formatPrimitive(ctx, value) {
if (isUndefined(value))
return ctx.stylize('undefined', 'undefined');
if (isString(value)) {
var simple = '\'' + JSON.stringify(value).replace(/^"|"$/g, '')
.replace(/'/g, "\\'")
.replace(/\\"/g, '"') + '\'';
return ctx.stylize(simple, 'string');
}
if (isNumber(value)) {
// Format -0 as '-0'. Strict equality won't distinguish 0 from -0,
// so instead we use the fact that 1 / -0 < 0 whereas 1 / 0 > 0 .
if (value === 0 && 1 / value < 0)
return ctx.stylize('-0', 'number');
return ctx.stylize('' + value, 'number');
}
if (isBoolean(value))
return ctx.stylize('' + value, 'boolean');
// For some reason typeof null is "object", so special case here.
if (isNull(value))
return ctx.stylize('null', 'null');
// es6 symbol primitive
if (isSymbol(value))
return ctx.stylize(value.toString(), 'symbol');
}
function arrayToHash(array) {
var hash = {};
array.forEach(function (val, idx) {
hash[val] = true;
});
return hash;
}
function objectToString(o) {
return Object.prototype.toString.call(o);
}
function isDate(d) {
return isObject(d) && objectToString(d) === '[object Date]';
}
function isError(e) {
return isObject(e) &&
(objectToString(e) === '[object Error]' || e instanceof Error);
}
function isRegExp(re) {
return isObject(re) && objectToString(re) === '[object RegExp]';
}
function formatError(value) {
return '[' + Error.prototype.toString.call(value) + ']';
}
function formatPrimitiveNoColor(ctx, value) {
var stylize = ctx.stylize;
ctx.stylize = stylizeNoColor;
var str = formatPrimitive(ctx, value);
ctx.stylize = stylize;
return str;
}
function isArray(ar) {
return Array.isArray(ar);
}
function hasOwnProperty(obj, prop) {
return Object.prototype.hasOwnProperty.call(obj, prop);
}
function formatProperty(ctx, value, recurseTimes, visibleKeys, key, array) {
var name, str, desc;
desc = Object.getOwnPropertyDescriptor(value, key) || {value: value[key]};
if (desc.get) {
if (desc.set) {
str = ctx.stylize('[Getter/Setter]', 'special');
} else {
str = ctx.stylize('[Getter]', 'special');
}
} else {
if (desc.set) {
str = ctx.stylize('[Setter]', 'special');
}
}
if (!hasOwnProperty(visibleKeys, key)) {
name = '[' + key + ']';
}
if (!str) {
if (ctx.seen.indexOf(desc.value) < 0) {
if (isNull(recurseTimes)) {
str = formatValue(ctx, desc.value, null);
} else {
str = formatValue(ctx, desc.value, recurseTimes - 1);
}
if (str.indexOf('\n') > -1) {
if (array) {
str = str.split('\n').map(function (line) {
return ' ' + line;
}).join('\n').substr(2);
} else {
str = '\n' + str.split('\n').map(function (line) {
return ' ' + line;
}).join('\n');
}
}
} else {
str = ctx.stylize('[Circular]', 'special');
}
}
if (isUndefined(name)) {
if (array && key.match(/^\d+$/)) {
return str;
}
name = JSON.stringify('' + key);
if (name.match(/^"([a-zA-Z_][a-zA-Z_0-9]*)"$/)) {
name = name.substr(1, name.length - 2);
name = ctx.stylize(name, 'name');
} else {
name = name.replace(/'/g, "\\'")
.replace(/\\"/g, '"')
.replace(/(^"|"$)/g, "'")
.replace(/\\\\/g, '\\');
name = ctx.stylize(name, 'string');
}
}
return name + ': ' + str;
}
function formatArray(ctx, value, recurseTimes, visibleKeys, keys) {
var output = [];
for (var i = 0, l = value.length; i < l; ++i) {
if (hasOwnProperty(value, String(i))) {
output.push(formatProperty(ctx, value, recurseTimes, visibleKeys,
String(i), true));
} else {
output.push('');
}
}
keys.forEach(function (key) {
if (!key.match(/^\d+$/)) {
output.push(formatProperty(ctx, value, recurseTimes, visibleKeys,
key, true));
}
});
return output;
}
function reduceToSingleString(output, base, braces) {
var length = output.reduce(function (prev, cur) {
return prev + cur.replace(/\u001b\[\d\d?m/g, '').length + 1;
}, 0);
if (length > 60) {
return braces[0] +
(base === '' ? '' : base + '\n ') +
' ' +
output.join(',\n ') +
' ' +
braces[1];
}
return braces[0] + base + ' ' + output.join(', ') + ' ' + braces[1];
}
function formatValue(ctx, value, recurseTimes) {
// Provide a hook for user-specified inspect functions.
// Check that value is an object with an inspect function on it
if (ctx.customInspect &&
value &&
isFunction(value.inspect) &&
// Filter out the util module, it's inspect function is special
value.inspect !== exports.inspect &&
// Also filter out any prototype objects using the circular check.
!(value.constructor && value.constructor.prototype === value)) {
var ret = value.inspect(recurseTimes, ctx);
if (!isString(ret)) {
ret = formatValue(ctx, ret, recurseTimes);
}
return ret;
}
// Primitive types cannot have properties
var primitive = formatPrimitive(ctx, value);
if (primitive) {
return primitive;
}
// Look up the keys of the object.
var keys = Object.keys(value);
var visibleKeys = arrayToHash(keys);
if (ctx.showHidden) {
keys = Object.getOwnPropertyNames(value);
}
// This could be a boxed primitive (new String(), etc.), check valueOf()
// NOTE: Avoid calling `valueOf` on `Date` instance because it will return
// a number which, when object has some additional user-stored `keys`,
// will be printed out.
var formatted;
var raw = value;
try {
// the .valueOf() call can fail for a multitude of reasons
if (!isDate(value))
raw = value.valueOf();
} catch (e) {
// ignore...
}
if (isString(raw)) {
// for boxed Strings, we have to remove the 0-n indexed entries,
// since they just noisey up the output and are redundant
keys = keys.filter(function (key) {
return !(key >= 0 && key < raw.length);
});
}
// Some type of object without properties can be shortcutted.
if (keys.length === 0) {
if (isFunction(value)) {
var name = value.name ? ': ' + value.name : '';
return ctx.stylize('[Function' + name + ']', 'special');
}
if (isRegExp(value)) {
return ctx.stylize(RegExp.prototype.toString.call(value), 'regexp');
}
if (isDate(value)) {
return ctx.stylize(Date.prototype.toString.call(value), 'date');
}
if (isError(value)) {
return formatError(value);
}
// now check the `raw` value to handle boxed primitives
if (isString(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
return ctx.stylize('[String: ' + formatted + ']', 'string');
}
if (isNumber(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
return ctx.stylize('[Number: ' + formatted + ']', 'number');
}
if (isBoolean(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
return ctx.stylize('[Boolean: ' + formatted + ']', 'boolean');
}
}
var base = '', array = false, braces = ['{', '}'];
// Make Array say that they are Array
if (isArray(value)) {
array = true;
braces = ['[', ']'];
}
// Make functions say that they are functions
if (isFunction(value)) {
var n = value.name ? ': ' + value.name : '';
base = ' [Function' + n + ']';
}
// Make RegExps say that they are RegExps
if (isRegExp(value)) {
base = ' ' + RegExp.prototype.toString.call(value);
}
// Make dates with properties first say the date
if (isDate(value)) {
base = ' ' + Date.prototype.toUTCString.call(value);
}
// Make error with message first say the error
if (isError(value)) {
base = ' ' + formatError(value);
}
// Make boxed primitive Strings look like such
if (isString(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
base = ' ' + '[String: ' + formatted + ']';
}
// Make boxed primitive Numbers look like such
if (isNumber(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
base = ' ' + '[Number: ' + formatted + ']';
}
// Make boxed primitive Booleans look like such
if (isBoolean(raw)) {
formatted = formatPrimitiveNoColor(ctx, raw);
base = ' ' + '[Boolean: ' + formatted + ']';
}
if (keys.length === 0 && (!array || value.length === 0)) {
return braces[0] + base + braces[1];
}
if (recurseTimes < 0) {
if (isRegExp(value)) {
return ctx.stylize(RegExp.prototype.toString.call(value), 'regexp');
} else {
return ctx.stylize('[Object]', 'special');
}
}
ctx.seen.push(value);
var output;
if (array) {
output = formatArray(ctx, value, recurseTimes, visibleKeys, keys);
} else {
output = keys.map(function (key) {
return formatProperty(ctx, value, recurseTimes, visibleKeys, key, array);
});
}
ctx.seen.pop();
return reduceToSingleString(output, base, braces);
}
function inspect(obj, opts) {
// default options
var ctx = {
seen: [],
stylize: stylizeNoColor
};
// legacy...
if (arguments.length >= 3) ctx.depth = arguments[2];
if (arguments.length >= 4) ctx.colors = arguments[3];
if (isBoolean(opts)) {
// legacy...
ctx.showHidden = opts;
} else if (opts) {
// got an "options" object
exports._extend(ctx, opts);
}
// set default options
if (isUndefined(ctx.showHidden)) ctx.showHidden = false;
if (isUndefined(ctx.depth)) ctx.depth = 2;
if (isUndefined(ctx.colors)) ctx.colors = false;
if (isUndefined(ctx.customInspect)) ctx.customInspect = true;
if (ctx.colors) ctx.stylize = stylizeWithColor;
return formatValue(ctx, obj, ctx.depth);
}
exports.inspect = inspect;
// http://en.wikipedia.org/wiki/ANSI_escape_code#graphics
inspect.colors = {
'bold': [1, 22],
'italic': [3, 23],
'underline': [4, 24],
'inverse': [7, 27],
'white': [37, 39],
'grey': [90, 39],
'black': [30, 39],
'blue': [34, 39],
'cyan': [36, 39],
'green': [32, 39],
'magenta': [35, 39],
'red': [31, 39],
'yellow': [33, 39]
};
// Don't use 'blue' not visible on cmd.exe
inspect.styles = {
'special': 'cyan',
'number': 'yellow',
'boolean': 'yellow',
'undefined': 'grey',
'null': 'bold',
'string': 'green',
'symbol': 'green',
'date': 'magenta',
// "name": intentionally not styling
'regexp': 'red'
};
var formatRegExp = /%[sdj%]/g;
exports.format = function (f) {
if (!isString(f)) {
var objects = [];
for (var j = 0; j < arguments.length; j++) {
objects.push(inspect(arguments[j]));
}
return objects.join(' ');
}
var i = 1;
var args = arguments;
var len = args.length;
var str = String(f).replace(formatRegExp, function (x) {
if (x === '%%') return '%';
if (i >= len) return x;
switch (x) {
case '%s':
return String(args[i++]);
case '%d':
return Number(args[i++]);
case '%j':
try {
return JSON.stringify(args[i++]);
} catch (_) {
return '[Circular]';
}
default:
return x;
}
});
for (var x = args[i]; i < len; x = args[++i]) {
if (isNull(x) || !isObject(x)) {
str += ' ' + x;
} else {
str += ' ' + inspect(x);
}
}
return str;
};
Harvested from: https://github.com/joyent/node/blob/master/lib/util.js
I have a slightly longer formatter for JavaScript here...
You can do formatting several ways:
String.format(input, args0, arg1, ...)
String.format(input, obj)
"literal".format(arg0, arg1, ...)
"literal".format(obj)
Also, if you have say a ObjectBase.prototype.format (such as with DateJS) it will use that.
Examples...
var input = "numbered args ({0}-{1}-{2}-{3})";
console.log(String.format(input, "first", 2, new Date()));
//Outputs "numbered args (first-2-Thu May 31 2012...Time)-{3})"
console.log(input.format("first", 2, new Date()));
//Outputs "numbered args(first-2-Thu May 31 2012...Time)-{3})"
console.log(input.format(
"object properties ({first}-{second}-{third:yyyy-MM-dd}-{fourth})"
,{
'first':'first'
,'second':2
,'third':new Date() //assumes Date.prototype.format method
}
));
//Outputs "object properties (first-2-2012-05-31-{3})"
I've also aliased with .asFormat and have some detection in place in case there's already a string.format (such as with MS Ajax Toolkit (I hate that library).
Using Lodash you can get template functionality:
Use the ES template literal delimiter as an "interpolate" delimiter.
Disable support by replacing the "interpolate" delimiter.
var compiled = _.template('hello ${ user }!');
compiled({ 'user': 'pebbles' });
// => 'hello pebbles!
How would you write a function that is like this
f()()
f('it') == fit
f()('x') == fox
I have
function f(s){
return "f"+s;
}
I had to discern what you're looking for not only from your question, but also from your comments. It looks like every string begins with 'f', and each empty bracket-pair appends an 'o'. Finally, a non-empty bracket-pair appends its argument.
I actually think this is a cool metaprogramming challenge.
This should work:
let f = (str, depth=0) => str
? `f${'o'.repeat(depth)}${str}` // If given param, terminate
: str => f(str, depth + 1); // If no param, return func
// "fit"
console.log(f('it'));
// "fox"
console.log(f()('x'));
// "fortress"
console.log(f()('rtress'));
// "football"
console.log(f()()('tball'));
// "foooooool!!!"
console.log(f()()()()()()()('l!!!'));
You have only 3 required outputs and no additional behavior requirements explained. Hence, the simplest approach would be:
let f = arg =>
arg ? 'f' + arg : arg => arg ? 'fo' + arg : '';
console.log(f()()); // ''
console.log(f('it')); // 'fit'
console.log(f()('x')); // 'fox'
You could return a function for more than one call of the function and implement a toString method to get the final string.
function f(v) {
var s = 'f' + (v || 'o');
function g(v) {
s += v || 'o';
return g;
};
g.toString = function () { return s; };
return g;
}
console.log(f()()); // foo
console.log(f('it')); // fit
console.log(f()('x')); // fox
console.log(f()()('b')('a')('r')); // foobar
That's a very strange function to want, but you can do it by having your named function return an anonymous function if you didn't get a parameter the first time:
function f(s, str) {
if(typeof str === "undefined") {
str = "f"
}
if(typeof s === "undefined") {
return function(t) {
return f(t, str + "o")
}
} else {
return str + s;
}
}
Haha I saw this problem the other day, I used function currying:
var f = s => s ? `f${s}` : (g=o=>e=>e?`f${o+e}`:g(o+'o')) && g('o')
console.log(f('x'))
console.log(f()()('l'))
console.log(f()()()('bar'))
I am POSTing JSON from my Javascript application to this API that behaves differently if you send it doubles vs. ints. I just use JSON.stringify to build my request bodies. I always want to send doubles. But when I have Javascript numbers that are integral, JSON.stringify does this:
> JSON.stringify({x: 1.0});
'{"x":1}'
That unadorned 1 goes to the API where the other server thinks it's an int, and I get the integer behavior, which I don't want. Is there any known hack to get a double in my JSON in Javascript?
JSON (and JavaScript in general) don't distinguish between different numeric types; there's only "number". As such this is a bit outside the scope of JSON.stringify.
Here's a pretty awful workaround:
JSON.stringify({x: 1.0}).replace(
/("(?:[^"\\]|\\.)*")|[\[:,]\s*-?\d+(?![\d.eE])/g,
function (m0, m1) { return m1 || m0 + '.0'; }
)
This (rather big hammer) uses a regex to add .0 to all numbers that don't already have a . and are not part of a quoted string.
Another approach: build a custom version of JSON.stringify which will always output numbers with a decimal point.
This simply takes an example implementation from the Mozilla JSON documentation and modifies the line of code that handles numbers:
window.JSON.stringifyDecimal = (function(){
var toString = Object.prototype.toString;
var isArray = Array.isArray || function (a) { return toString.call(a) === '[object Array]'; };
var escMap = {'"': '\\"', '\\': '\\\\', '\b': '\\b', '\f': '\\f', '\n': '\\n', '\r': '\\r', '\t': '\\t'};
var escFunc = function (m) { return escMap[m] || '\\u' + (m.charCodeAt(0) + 0x10000).toString(16).substr(1); };
var escRE = /[\\"\u0000-\u001F\u2028\u2029]/g;
return function stringify(value) {
if (value == null) {
return 'null';
} else if (typeof value === 'number') {
// Ensure decimal point is present
return isFinite(value) ? value.toString() + ( Number.isInteger( value ) ? '.0' : '' ) : 'null';
} else if (typeof value === 'boolean') {
return value.toString();
} else if (typeof value === 'object') {
if (typeof value.toJSON === 'function') {
return stringify(value.toJSON());
} else if (isArray(value)) {
var res = '[';
for (var i = 0; i < value.length; i++)
res += (i ? ', ' : '') + stringify(value[i]);
return res + ']';
} else if (toString.call(value) === '[object Object]') {
var tmp = [];
for (var k in value) {
if (value.hasOwnProperty(k))
tmp.push(stringify(k) + ': ' + stringify(value[k]));
}
return '{' + tmp.join(', ') + '}';
}
}
return '"' + value.toString().replace(escRE, escFunc) + '"';
};
})();
With JSON, you have two parties: The writer and the reader. In your case, the writer is your request and you're using JSON.stringify to generate the JSON string, and the reader is the API (which I don't know what it uses to parse the string back to a JSON object). You need both sides to be able to understand that 1.0 should stay as is because it's a double not an integer.
JSON string itself is just text and keeps whatever you give it as is. It is the JSON.stringify that is converting 1.0 to 1, so you can easily create your JSON string manually instead of using JSON.stringify:
var text = JSON.stringify({x: 1.0}); //JSON string will be {"x": 1}
var text = '{"x": 1.0}'; //JSON string will be exactly {"x": 1.0}
However, even if you do that, you need the other side (i.e. the API) to parse the value 1.0 as double, not an integer. Is that under your control? For example, the JavaScript JSON.parse function will parse the 1.0 as 1 exactly like JSON.stringify does:
var obj = JSON.parse('{"x": 1.0}');
alert(obj.x); //It prints 1, even though we gave it 1.0
For example, say I have a function defined as follows:
function foo() {
return "Hello, serialized world!";
}
I want to be able to serialize that function and store it using localStorage. How can I go about doing that?
Most browsers (Chrome, Safari, Firefox, possibly others) return the definition of functions from the .toString() method:
> function foo() { return 42; }
> foo.toString()
"function foo() { return 42; }"
Just be careful because native functions won't serialize properly. For example:
> alert.toString()
"function alert() { [native code] }"
function foo() {
alert('native function');
return 'Hello, serialised world!';
}
Serializing
var storedFunction = foo.toString();
Deserializing
var actualFunction = new Function('return ' + foo.toString())()
Explanation
foo.toString() will be string version of the function foo
"function foo() { ... return 'Hello, serialised world!';}"
But new Function takes the body of a function and not the function itself.
See MDN: Function
So we can create a function that returns us back this function and assign it to some variable.
"return function foo() { ... return 'Hello, serialised world!';}"
So now when we pass this string to the constructor we get a function and we immediately execute it to get back our original function. :)
I made this answer to address some pretty big flaws with the existing answers: .toString()/eval() and new Function() on their own wont work at all if your function uses this or named arguments (function (named, arg) {}), respectively.
Using toJSON() below, all you need to do is call JSON.stringify() as usual on the function, and use Function.deserialise when parse()ing.
The following wont work for concise functions (hello => 'there'), but for standard ES5 fat functions it'll return it as it was defined, closures notwithstanding of course. My other answer will work with all that ES6 goodness.
Function.prototype.toJSON = function() {
var parts = this
.toString()
.match(/^\s*function[^(]*\(([^)]*)\)\s*{(.*)}\s*$/)
;
if (parts == null)
throw 'Function form not supported';
return [
'window.Function',
parts[1].trim().split(/\s*,\s*/),
parts[2]
];
};
Function.deserialise = function(key, data) {
return (data instanceof Array && data[0] == 'window.Function') ?
new (Function.bind.apply(Function, [Function].concat(data[1], [data[2]]))) :
data
;
};
Take a look at the DEMO
At it's simplest:
var test = function(where) { return 'hello ' + where; };
test = JSON.parse(JSON.stringify(test), Function.deserialise);
console.log(test('there'));
//prints 'hello there'
More usefully, you can serialise entire objects containing functions and pull them back out:
test = {
a : 2,
run : function(x, y, z) { return this.a + x + y + z; }
};
var serialised = JSON.stringify(test);
console.log(serialised);
console.log(typeof serialised);
var tester = JSON.parse(serialised, Function.deserialise);
console.log(tester.run(3, 4, 5));
Outputs:
{"a":2,"run":["window.Function",["x","y","z"]," return this.a + x + y + z; "]}
string
14
I didn't test older IE's, but it works on IE11, FF, Chrome, Edge.
NB, the name of the function is lost, if you use that property then there's nothing you can do, really.
You can change it to not use prototype easily, but that's for you to do if that's what you need.
If you needed a way to serialise Arrow Functions in ES6 I have written a serialiser that makes everything work.
All you need to do is call JSON.stringify() as usual on the function or object containing the function, and call Function.deserialise on the other side for the magic to work.
Obviously you shouldn't expect closures to work, it is serialisation after all, but defaults, destructuring, this, arguments, class member functions, it'll all be preserved.
If you're only using ES5 notations please just use my other answer. This one really is above and beyond
Here's the demonstration
Working in Chrome/Firefox/Edge.
Bellow is the output from the demo; a few functions, the serialised string, then calling the new function created after deserialisation.
test = {
//make the function
run : function name(x, y, z) { return this.a + x + y + z; },
a : 2
};
//serialise it, see what it looks like
test = JSON.stringify(test) //{"run":["window.Function",["x","y","z"],"return this.a + x + y + z;"],"a":2}
test = JSON.parse(test, Function.deserialise)
//see if `this` worked, should be 2+3+4+5 : 14
test.run(3, 4, 5) //14
test = () => 7
test = JSON.stringify(test) //["window.Function",[""],"return 7"]
JSON.parse(test, Function.deserialise)() //7
test = material => material.length
test = JSON.stringify(test) //["window.Function",["material"],"return material.length"]
JSON.parse(test, Function.deserialise)([1, 2, 3]) //3
test = ([a, b] = [1, 2], {x: c} = {x: a + b}) => a + b + c
test = JSON.stringify(test) //["window.Function",["[a, b] = [1, 2]","{ x: c } = { x: a + b }"],"return a + b + c"]
JSON.parse(test, Function.deserialise)([3, 4]) //14
class Bob {
constructor(bob) { this.bob = bob; }
//a fat function with no `function` keyword!!
test() { return this.bob; }
toJSON() { return {bob:this.bob, test:this.test} }
}
test = new Bob(7);
test.test(); //7
test = JSON.stringify(test); //{"bob":7,"test":["window.Function",[""],"return this.bob;"]}
test = JSON.parse(test, Function.deserialise);
test.test(); //7
And finally, the magic
Function.deserialise = function(key, data) {
return (data instanceof Array && data[0] == 'window.Function') ?
new (Function.bind.apply(Function, [Function].concat(data[1], [data[2]]))) :
data
;
};
Function.prototype.toJSON = function() {
var whitespace = /\s/;
var pair = /\(\)|\[\]|\{\}/;
var args = new Array();
var string = this.toString();
var fat = (new RegExp(
'^\s*(' +
((this.name) ? this.name + '|' : '') +
'function' +
')[^)]*\\('
)).test(string);
var state = 'start';
var depth = new Array();
var tmp;
for (var index = 0; index < string.length; ++index) {
var ch = string[index];
switch (state) {
case 'start':
if (whitespace.test(ch) || (fat && ch != '('))
continue;
if (ch == '(') {
state = 'arg';
tmp = index + 1;
}
else {
state = 'singleArg';
tmp = index;
}
break;
case 'arg':
case 'singleArg':
var escaped = depth.length > 0 && depth[depth.length - 1] == '\\';
if (escaped) {
depth.pop();
continue;
}
if (whitespace.test(ch))
continue;
switch (ch) {
case '\\':
depth.push(ch);
break;
case ']':
case '}':
case ')':
if (depth.length > 0) {
if (pair.test(depth[depth.length - 1] + ch))
depth.pop();
continue;
}
if (state == 'singleArg')
throw '';
args.push(string.substring(tmp, index).trim());
state = (fat) ? 'body' : 'arrow';
break;
case ',':
if (depth.length > 0)
continue;
if (state == 'singleArg')
throw '';
args.push(string.substring(tmp, index).trim());
tmp = index + 1;
break;
case '>':
if (depth.length > 0)
continue;
if (string[index - 1] != '=')
continue;
if (state == 'arg')
throw '';
args.push(string.substring(tmp, index - 1).trim());
state = 'body';
break;
case '{':
case '[':
case '(':
if (
depth.length < 1 ||
!(depth[depth.length - 1] == '"' || depth[depth.length - 1] == '\'')
)
depth.push(ch);
break;
case '"':
if (depth.length < 1)
depth.push(ch);
else if (depth[depth.length - 1] == '"')
depth.pop();
break;
case '\'':
if (depth.length < 1)
depth.push(ch);
else if (depth[depth.length - 1] == '\'')
depth.pop();
break;
}
break;
case 'arrow':
if (whitespace.test(ch))
continue;
if (ch != '=')
throw '';
if (string[++index] != '>')
throw '';
state = 'body';
break;
case 'body':
if (whitespace.test(ch))
continue;
string = string.substring(index);
if (ch == '{')
string = string.replace(/^{\s*(.*)\s*}\s*$/, '$1');
else
string = 'return ' + string.trim();
index = string.length;
break;
default:
throw '';
}
}
return ['window.Function', args, string];
};
Don't serialize the call, instead try serializing the info, allowing to repeat the call, which can include things like class & method names, arguments passed into the call or just a call scenario name.
w = (function(x){
return function(y){
return x+y;
};
});""+w returns "function(x){
return function(y){
return x+y;
};
}" but ""+w(3) returns "function(y){
return x+y;
}"
which is not the same as w(3) which somehow still remembers to add 3.