mstc is a minimal mustache-like template processor. It supports objects, arrays and can even call functions on them - all without using eval or new Function! Placeholders can contain any character except curly braces (that means you can have templates that render into templates).
-
-
Save tkissing/1347239 to your computer and use it in GitHub Desktop.
function(a,b) { // a: template string, b: Object or Array with values to fill in | |
// coerce a to a string | |
return(a+'').replace( | |
// search for anything with a mustache around it | |
/\{\{([^{}]+)}}/g, | |
function(c,d) { // c will be the complete placeholder, d the part between the mustaches | |
// b is optional (but if passed, must be a type that is a valid operand for "in" | |
return d in (b||{}) | |
// if d is a method of b, call it, otherwise return its value | |
? (/^f/.test(typeof b[d]) ? b[d]() :b[d] ) | |
// if d is not a member of b, don't replace anything to allow nested calls like | |
// mstc(mstc(tmpl, obj), anotherObj) | |
: c; | |
} | |
); | |
} |
function(a,b){return(a+'').replace(/\{\{([^{}]+)}}/g,function(c,d){return d in(b||{})?(/^f/.test(typeof b[d])?b[d]():b[d]):c})} |
DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE | |
Version 2, December 2004 | |
Copyright (C) 2011 Timo Kissing http://kissing.name | |
Everyone is permitted to copy and distribute verbatim or modified | |
copies of this license document, and changing it is allowed as long | |
as the name is changed. | |
DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE | |
TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION | |
0. You just DO WHAT THE FUCK YOU WANT TO. |
{ | |
"name": "mustacheStyleTemplatesCompressed", | |
"description": "Very minimal implementation of mustache style templating in JavaScript", | |
"keywords": [ | |
"140bytes", | |
"mustache", | |
"template" | |
] | |
} |
<!DOCTYPE html> | |
<head> | |
<title>mstc: mustache style templates - compressed </title> | |
</head> | |
<body> | |
<div>Expected value: <b>Hello @140bytes! How is the weather?</b></div> | |
<div>Actual value: <b id="ret1"></b></div> | |
<div>Expected value: <b>The array has a length of 3 and the first element is "foo". Ah, what the heck, here is all of it: foo,bar,bam</b></div> | |
<div>Actual value: <b id="ret2"></b></div> | |
<script> | |
var mstc = function(a,b){return(a+'').replace(/\{\{([^{}]+)}}/g,function(c,d){return d in(b||{})?(/^f/.test(typeof b[d])?b[d]():b[d]):c})}; | |
document.getElementById( "ret1" ).innerHTML = mstc('Hello {{w}}! How is the {{x}}?',{w:'@140bytes', x: function() {return 'weather';}}); | |
document.getElementById( "ret2" ).innerHTML = mstc('The array has a length of {{length}} and the first element is "{{0}}". Ah, what the heck, here is all of it: {{join}}', ['foo', 'bar', 'bam']); | |
</script> | |
</body> | |
</html> |
- You are right.
- I think the regex actually should actually be one byte longer:
/\{\{([^{}]+)}}/g
mstc('This is a {{{{test}}}}', {test: 'template'});
is handled badly right now, adding that one byte would make it nicer and more useable I think - I can't think of any.
'call' in b[d]
doesn't work because strings and numbers are invalid operands. I had a version that assigned the inner function to a variable and then didtypeof b[d] == typeof h
but with the additional,h
andh=
required elsewhere that is actually 4 bytes more. (If you ever need to check for function type twice though, that will save you some bytes)
The regular expression depends on what you want to fetch and what not. Another possibility is to use /\{\{(\w+)}}/g
. This is a shortcut for [0-9A-Za-z_]
and fetches word characters (hence the \w
) and the underscore. {{{{test}}}}
will work as expected but {{test-this}}
will not work. I checked and yes, these shortcuts work in all web browsers including IE6.
I went with /\{\{([^{}]+)}}/g
- I don't want to be more restrictive as I technically required for the function to work and be predictable. I might change my mind if I could add another feature that would need those bytes :) Allowing dot-notation in placeholders would be awesome, but I don't see how that could work without adding another function and the keywords alone would bring it above 140 again.
Haven't been able to test in older IE, but /^f/.test(typeof x)
works in IE9 at least...
I had an other idea: (typeof b[d]).length==8
. But it's the same length. Your approach is very cool. I checked. Works in IE6.
b[d] && b[d].call
seems to work as well - compared to the original typeof b[d] == 'function'
that's 8 bytes. Need to test in IEs though, so probably won't update the gist today.
I added this to my tests suite (abusing website visitors for compatibility tests is fun). From what I see it works as expected in all web browsers including IE6. We need to add this to Byte-saving Techniques. But don't forget this will lead to wrong results if an object contains a call
property or method.
Yeah, that limitation is important to note. I think I will go with the /^f/.test approach unless I find really good use for those additional bytes ;)
@tkissing, I'm surprised but it seems it's possible to omit all \
because {
is not followed by a number. I will add this to my tests suite.
Results from my tests suit (very few data, unfortunately): Not escaping {
fails in Chrome 8 (but works in Chrome 12+). Also fails in older versions of Safari. Works in all other browsers (IE6+, Firefox 3.6+).
Update (more data): Not escaping {
works in IE 6+ and Opera 10.6+, it works in Firefox 3.0+ but fails in 2, works in Safari 5.1+ but fails in 5.0, works in Chrome 14+ but may fail (inconsistent data) in older versions. Oh, and I added /^f/.test(typeof a)
and a&&a.call
to my tests suite. Both work as expected in all browsers (checking for call
returns true for functions and objects containing a call
method, false for numbers, strings and all other objects).
What about
function f(a,b,c){return++c?b in(c=f.z||{})?c[b].call?c[b]():c[b]:a:a.replace(/{{([^{}]+)}}/g,f,f.z=b)}
?
maettig: How did you get that idea? Would have never thought of that...
tsaniel: That is pretty darn clever! Should read :c[b]:a:a.replace
in the middle there to stay compatible with previous versions (if placeholder is not found in data, leave it untouched; your version removes the mustache), but still very nice work. Adding back the coercion of a
to a string and the /^f/
test for the function it's still 4 bytes shorter! Congratulations!
@tkissing: You're right! That was my bad.
@tsaniel: no worries, that's why I have a bunch of additonal test cases locally ;)
Will update the gist later, your little trick gave me an idea I want to test...
Some ideas:
b=
.[^}]+
with.*?
saves 2 bytes but makes the regular expression slower. I do not recommend this. It will be really slow if applied to very large templates.typeof b[d]=='function'
shorter? To bad(typeof b[d])[0]=='f'
does not work in Internet Explorer.d in b
always returns true, no matter ifd
is a function or not.