Javascript heredoc_问答_开发者_运维开发者技术经验分享

I ne开发者_运维百科ed something like heredoc in JavaScript. Do you have any ideas for this? I need cross-browser functionality.

I found this:

heredoc = '\
<div>\
    <ul>\
        <li><a href="#zzz">zzz</a></li>\
    </ul>\
</div>';

I think it will work for me. :)

Try ES6 String Template, you can do something like

var hereDoc = `
This
is
a
Multiple
Line
String
`.trim()


hereDoc == 'This\nis\na\nMultiple\nLine\nString'

=> true

You can use this great feature even in older browsers with TypeScript

No, unfortunately JavaScript does not support anything like heredoc.

How about this:

function MyHereDoc(){
/*HERE
<div>
   <p>
      This is written in the HEREDOC, notice the multilines :D.
   </p>
   <p>
      HERE
   </p>
   <p>
      And Here
   </p>
</div>
HERE*/
    var here = "HERE";
    var reobj = new RegExp("/\\*"+here+"\\n[\\s\\S]*?\\n"+here+"\\*/", "m");
    str = reobj.exec(MyHereDoc).toString();
    str = str.replace(new RegExp("/\\*"+here+"\\n",'m'),'').toString();
    return str.replace(new RegExp("\\n"+here+"\\*/",'m'),'').toString();
}

//Usage 
document.write(MyHereDoc());

Just replace "/*HERE" and "HERE*/" with word of choice.

Building on Zv_oDD's answer, I created a similar function for easier reuse.

Warning: This is a non-standard feature of many JS interpreters, and will probably be removed at some point, but as I'm building a script to be only used in Chrome, I am using it! Do not ever rely on this for client-facing websites!

// Multiline Function String - Nate Ferrero - Public Domain
function heredoc(fn) {
  return fn.toString().match(/\/\*\s*([\s\S]*?)\s*\*\//m)[1];
};

Use:

var txt = heredoc(function () {/*
A test of horrible
Multi-line strings!
*/});

Returns:

"A test of horrible
Multi-line strings!"

Notes:

Text is trimmed on both ends, so any extra whitespace on either end is OK.

Edits:

2/2/2014 - changed to not mess with the Function prototype at all and use the name heredoc instead.

5/26/2017 - updated whitespace to reflect modern coding standards.

Depending on what flavour of JS/JS engine you're running (SpiderMonkey, AS3) you can simply write inline XML, into which you can place text on multiple lines, like heredoc:

var xml = <xml>
    Here 
    is 
    some 
    multiline 
    text!
</xml>

console.log(xml.toXMLString())
console.log(xml.toString()) // just gets the content

ES6 Template Strings has heredoc feature.

You can declare strings enclosed by back-tick (` `) and can be expanded through multiple lines.

var str = `This is my template string...
and is working across lines`;

You can also include expressions inside Template Strings. These are indicated by the Dollar sign and curly braces (${expression}).

var js = "Java Script";
var des = `Template strings can now be used in ${js} with lot of additional features`;

console.log(des); //"Template strings can now be used in Java Script with lot of additional features"

There are in fact more features such as Tagged Temple Strings and Raw Strings in it. Please find the documentation at

https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/template_strings

ES5 and earlier versions

(function(){/**
some random
multi line
text here
**/}).toString().slice(15,-5);

ES6 and later versions

`some random
multi line
text here`

result

some random
multi line
text here

I feel bad writing a separate answer for merely an extension to @NateFerrero's answer, but I don't feel editing his answer is appropriate either, so please upvote @NateFerrero if this answer was useful to you.

tl;dr—For those who wish to use block comments inside their heredoc...

I mainly needed Javascript heredocs to store a block of CSS, e.g.

var css = heredoc(function() {/*
    /**
     * Nuke rounded corners.
     */
    body div {
        border-top-left-radius: 0 !important;
        border-top-right-radius: 0 !important;
        border-bottom-right-radius: 0 !important;
        border-bottom-left-radius: 0 !important;
    }
*/});

As you can see however, I like to comment my CSS, and unfortunately (as hinted by the syntax highlighting) the first */ ends the overall comment, breaking the heredoc.

For this specific purpose (CSS), my workaround was to add

.replace(/(\/\*[\s\S]*?\*) \//g, '$1/')

to the chain inside @NateFerrero's heredoc; in complete form:

function heredoc (f) {
    return f.toString().match(/\/\*\s*([\s\S]*?)\s*\*\//m)[1].replace(/(\/\*[\s\S]*?\*) \//g, '$1/');
};

and use it by adding a space between the * and / for "inner" block comments, like so:

var css = heredoc(function() {/*
    /**
     * Nuke rounded corners.
     * /
    body div {
        border-top-left-radius: 0 !important;
        border-top-right-radius: 0 !important;
        border-bottom-right-radius: 0 !important;
        border-bottom-left-radius: 0 !important;
    }
*/});

The replace simply finds /* ... * / and removes the space to make /* ... */, thereby preserving the heredoc until called.

You can of course remove the comments altogether using

.replace(/\/\*[\s\S]*?\* \//g, '')

You can also support // comments if you add them to the chain:

.replace(/^\s*\/\/.*$/mg, '')

Also, you can do something other than the single space between * and /, like a -:

    /**
     * Nuke rounded corners.
     *-/

if you just update the regex appropriately:

.replace(/(\/\*[\s\S]*?\*)-\//g, '$1/')
                          ^

Or maybe you'd like an arbitrary amount of whitespace instead of a single space?

.replace(/(\/\*[\s\S]*?\*)\s+\//g, '$1/')
                          ^^^

You could use CoffeeScript, a language that compiles down to JavaScript. The code compiles one-to-one into the equivalent JS, and there is no interpretation at runtime.

And of course, it has heredocs :)

As others have said, ES6 template strings give you most of what traditional heredocs provide.

If you want to go a step further and use a tagged template string, theredoc is a nice utility function that lets you do this:

if (yourCodeIsIndented) {
  console.log(theredoc`
    Theredoc will strip the
    same amount of indentation
    from each line.

      You can still indent
      further if you want.

    It will also chop off the
    whitespace-only first and
    last lines.
  `)
}

You can use Sweet.js Macros to add it like so, as created by Tim Disney in this post

Note that this approach uses backticks as the string delimiters instead:

let str = macro {
    case {_ $template } => {
        var temp = #{$template}[0];
        var tempString = temp.token.value.raw;
        letstx $newTemp = [makeValue(tempString, #{here})];
        return #{$newTemp}
    }
}

str `foo bar baz`

If you have some html and jQuery at hand and the string is valid HTML, this may be useful:

<div id="heredoc"><!--heredoc content
with multiple lines, even 'quotes' or "double quotes",
beware not to leave any tag open--></div>
<script>
var str = (function() {
   var div = jQuery('#heredoc');
   var str = div.html();
   str = str.replace(/^<\!--/, "").toString();
   str = str.replace(/-->$/, "").toString();
   return str;
})();
</script>

If text have comments "" in between, it works as well, but a part of the text may be visible. Here's the fiddle: https://jsfiddle.net/hr6ar152/1/

// js heredoc - http://stackoverflow.com/a/32915549/466363
// a function with comment with eval-able string, use it just like regular string

function extractFuncCommentString(func,comments) {
  var matches = func.toString().match(/function\s*\(\)\s*\{\s*\/\*\!?\s*([\s\S]+?)\s*\*\/\s*\}/);
  if (!matches) return undefined;
  var str=matches[1];

   // i have made few flavors of comment removal add yours if you need something special, copy replacement lines from examples below, mix them
  if(comments===1 )
  {
   // keep comments, in order to keep comments  you need to convert /**/ to / * * / to be able to put them inside /**/ like /*    / * * /    */
   return (
    str
   .replace(/\/\s\*([\s\S]*?)\*\s\//g,"/*$1*/") //       change   / * text * /  to   /* text */ 
   )
  }
  else if(comments===2)
  {
   // keep comments and replace singleline comment to multiline comment
   return (
    str
   .replace(/\/\s\*([\s\S]*?)\*\s\//g,"/*$1*/") //       change   / * text * /  to   /* text */ 
   .replace(/\/\/(.*)/g,"/*$1*/")          //           change   //abc to  /*abc*/
   )
  }
  else if(comments===3)
  {
   // remove comments
   return (
      str
      .replace(/\/\s\*([\s\S]*?)\*\s\//g,"") //       match / * abc * /
      .replace(/\/\/(.*)/g,"")             // match //abc
     )
  }
  else if(comments===4)
  {
   // remove comments and trim and replace new lines with escape codes
   return (
      str
      .replace(/\/\s\*([\s\S]*?)\*\s\//g,"") //       match / * abc * /
      .replace(/\/\/(.*)/g,"")             // match //abc
      .trim() // after removing comments trim and:
      .replace(/\n/g,'\\n').replace(/\r/g,'\\r') // replace new lines with escape codes. allows further eval() of the string, you put in the comment function: a quoted text but with new lines
     )
  }
  else if(comments===5)
  {
   // keep comments comments and replace strings, might not suit when there are spaces or comments before and after quotes 
   // no comments allowed before quotes of the string
   return (
      str
      .replace(/\/\s\*([\s\S]*?)\*\s\//g,"/*$1*/") //       change   / * text * /  to   /* text */
      .replace(/\/\/(.*)/g,"/*$1*/")          //           change   //abc to  /*abc*/
      .trim() // trim space around quotes to not escape it and:
      .replace(/\n/g,'\\n').replace(/\r/g,'\\r') // replace new lines with escape codes. allows further eval() of the string, you put in the comment function: a quoted text but with new lines
     )
  }
  else 
  return str
}

example

var week=true,b=123;
var q = eval(extractFuncCommentString(function(){/*!

// this is a comment     


'select 

/ * this
is a multiline 
comment * /

 a
,b  // this is a comment  
,c
from `table`
where b='+b+' and monthweek="'+(week?'w':'m')+'" 
//+' where  a=124
order by a asc
'
*/},4));

with cache: - make a simple template function, and save the function:(second time works fast)

var myfunction_sql1;
function myfunction(week,a){


    if(!myfunction_sql1) eval('myfunction_sql1=function(week,a){return ('+extractFuncCommentString(function(){/*!
'select 

/ * this
is a multiline 
comment * /

 a
,b  // this is a comment  
,c
from `table`
where b='+b+' and monthweek="'+(week?'w':'m')+'" 
//+' where  a=124
order by a asc

'*/},4)+')}');
    q=myfunction_sql1(week,a);
    console.log(q)
}
myfunction(true,1234)

I'm posting this version as it avoids the use of regex for something so trivial.

IMHO regex is an obfuscation that was created as a practical joke amongst perl developers. the rest of the community took them seriously and we now pay the price, decades later. don't use regex, except for backward compatabilty with legacy code. there is no excuse these days to write code that is not immediately human readable and understandable. regex violates this principle on every level.

I've also added a way to add the result to the current page, not that this was asked for.

function pretty_css () {
/*
    pre { color: blue; }

*/
}
function css_src (css_fn) {
   var css = css_fn.toString();
   css = css.substr(css.indexOf("/*")+2);
   return css.substr(0,css.lastIndexOf("*/")).trim();
}

function addCss(rule) {
  let css = document.createElement('style');
  css.type = 'text/css';
  if (css.styleSheet) css.styleSheet.cssText = rule; // Support for IE
  else css.appendChild(document.createTextNode(rule)); // Support for the rest
  document.getElementsByTagName("head")[0].appendChild(css);
}

addCss(css_src(pretty_css));

document.querySelector("pre").innerHTML=css_src(pretty_css);