#19. Kevin on 16 March 2009

@ Jay: I will review your comment soon over at http://phpjs.org . Please bear with me, and use that new site for comments on PHP.JS from now on.

Thanks a lot!!

#18. Jay on 06 March 2009

You should really consider putting this project up on GitHub.com, it would be a perfect fit. That way people could modify and improve the code, and then send you a pull request for you to merge their changes. It's a very nice way to develop open source projects.

#17. Kevin on 15 January 2009

@ Paul: It was actually an issue in the exit function. Should be fixed now, thanks for letting us know!

#16. Paul on 15 January 2009

I downloaded php.namespaced.*.js at 2009-01-14 15:28 HST. There appears to be a space missing which is causing javascript errors. Search string "functionPropagation". I was able to resolve the error in php.namespaced.js and php.namespaced.min.js, but not php.namespaced.packed.js due to the packing operation.

BTW: This package is incredible!

#15. Kevin on 12 January 2009

@ Nile: Hey Nile thanks for your contribution. Unfortunately there are a couple of issues with it:
- depends on an outside function: applyEncode
- does not support associative arrays (in terms of traversing)
- does not support index arrays, numbers, etc (in terms of encoding)

... [more] Please have a look here:
http://www.json.org/json2.js

The code is public domain, so maybe it's possible to modify it and use it in PHP.JS. We could look into that!

#14. Nile on 09 January 2009

json_encode

var arr = new Array('Hello');
var applyEncode = function(varIn){
return '{'+varIn+'}';
}
var json_encode = function(value){
   var data = '';
   for(i=0,endCount='';i<value.length;++i){
   var endCount = (i!=value.length-1) ? ", " : "";
   data += '"'+key(value[i])+'" : "'+value[key(value)]+'"'+endCount;
   }
   return applyEncode(data);
}
document.write(json_encode(arr));

#13. Reena on 08 September 2008

Good

#12. Kevin on 13 April 2008

@ Jonas Raoni: Thanks I'll update the function.
@ Philip: I'll update your name!

#11. Philip on 12 April 2008

Hey, just fyi... Philip and Philip Peterson are both me ;-)

#10. Jonas Raoni on 12 April 2008

A shorter version of this would be:

return Math.abs (n) || 0;

#9. Kevin on 02 April 2008

Philip: Good idea, I'll add it, thanks!

#8. Philip on 31 March 2008

Dunno where this belongs, but how about this code for echo()?

function echo()
{
  for(i=0;i<echo.arguments.length;i++)
  {
    if(document.body && document.body.innerHTML) {
document.body.innerHTML=document.body.innerHTML+echo.arguments[i];
} else {
document.write(echo.arguments[i]);
}
  }
}

#7. Kevin on 31 January 2008

@ _argos: Nice work! Added!

#6. _argos on 31 January 2008

Hi Kevin I'm here again look this ports.

if ( defined ( 'CONSTANTE' ) ) {
 
    console.log ( CONSTANTE );
 
}
 
function defined ( constant_name ) {
 
  return ( ( window [ constant_name ] !== undefined ) ? true : false );
 
}
 
 
 
// -----
 
// [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]
 
console.log ( range ( 0, 12 ) );
 
 
 
// [0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100]
 
console.log ( range( 0, 100, 10 ) );
 
 
 
// ['a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i'];
 
console.log ( range( 'a', 'i' ) );
 
 
 
// ['c', 'b', 'a'];
 
console.log ( range ('c', 'a' ) );
 
 
 
function range ( low, high, step ) {
 
  var matrix = [];
 
  var inival, endval, plus;
 
  var walker = step || 1;
 
  var chars  = false;
 
  
 
  if ( !isNaN ( low ) && !isNaN ( high ) ) {
 
    inival = low;
 
    endval = high;    
 
  } else if ( isNaN ( low ) && isNaN ( high ) ) {
 
    chars = true;
 
    inival = low.charCodeAt ( 0 );
 
    endval = high.charCodeAt ( 0 );
 
  } else {
 
    inival = ( isNaN ( low ) ? 0 : low );
 
    endval = ( isNaN ( high ) ? 0 : high );
 
  }
 
  
 
  plus = ( ( inival > endval ) ? false : true );
 
  
 
  if ( plus ) {
 
    while ( inival <= endval ) {
 
      matrix.push ( ( ( chars ) ? String.fromCharCode ( inival ) : inival ) );
 
      inival += walker;
 
    }
 
  } else {
 
    while ( inival >= endval ) {
 
      matrix.push ( ( ( chars ) ? String.fromCharCode ( inival ) : inival ) );
 
      inival -= walker;
 
    }
 
  }
 
    
 
  return matrix;
 
}
 
 
 
// -----
 
console.log ( strcmp ( 'waldo', 'Waldo' ) );
 
console.log ( strcmp ( 'Waldo', 'waldo' ) );
 
console.log ( strcmp ( 'waldo', 'waldo' ) );
 
 
 
function strcmp ( str1, str2 ) {
 
  var size1 = str1.charCodeAt ( 0 );
 
  var size2 = str2.charCodeAt ( 0 );
 
  
 
  return ( ( size1 == size2 ) ? 0 : ( ( size1 > size2 ) ? 1 : -1 ) );
 
}

#5. _argos on 30 January 2008

@Karol Kowalski : Hi, thanks for take the time to check my port, I make the same tests like you, because I think that native functions always are more slower, so ever I use self hacks :p again thanxs for your time.

PS: Sorry for my badly English, but i have 4 years without use it :p

#4. Kevin on 30 January 2008

@ Karol Kowalski: Awesome work Karol! I've updated the function. Thanks for putting in the extra effort. Greatly appreciated!

#3. Karol Kowalski on 30 January 2008

I run a test and it appeared that my function with Math.abs was around 40% more time consuming than yours. I tried to optimize it and ended up with 20% gain, and the code is still smaller and more readable. I tested it in Firefox, IE7, Safari and Opera with similar results. To see what's causing the overhead I run a function that would just return Math.abs, withoud checking for validity, still it was 10% slower than your function. It seems that Math.abs API is always slower than what can be accomplished with JS hack, a bit sad.

Still, for code readability I would sugget using the code of abs_math2.

Here's the test code:

if (!window['console']) {
 
console={}
console.log=alert
}
 
var start;
 
function abs_math2 ( mixed_number ) {
  return ( ( isNaN ( mixed_number ) ) ? 0 : Math.abs ( mixed_number ) );
}
 
function abs_math( mixed_number )  {
  var abs=Math.abs( mixed_number );
  return ( !isNaN ( abs ) ) ? abs : 0
}
 
function abs_cond( mixed_number )  {
  return ( ( !isNaN ( mixed_number ) ) ? ( ( mixed_number < 0 ) ? ( mixed_number * -1 ) : mixed_number ) : 0 );
}
 
 
start=new Date();
 
for (var i=100000;i;i--) {
 
abs_cond(4.2);
abs_cond(-4.2);
abs_cond(-5);
abs_cond('_argos');
abs_cond(-Infinity);
 
}
 
console.log((new Date())-start)
 
 
start=new Date();
 
for (var i=100000;i;i--) {
 
abs_math(4.2);
abs_math(-4.2);
abs_math(-5);
abs_math('_argos');
abs_math(-Infinity);
 
}
 
console.log((new Date())-start)
 
 
start=new Date();
 
for (var i=100000;i;i--) {
 
abs_math2(4.2);
abs_math2(-4.2);
abs_math2(-5);
abs_math2('_argos');
abs_math2(-Infinity);
 
}
 
console.log((new Date())-start)
 
 
//my results
//Firefox 2.0.11
//1000
//1422
//1219

#2. Kevin on 30 January 2008

@ Karol Kowalski: Using native Javascript for this is probably best so I've updated the function. But I'm not totally certain that this is faster. Doesn't the Math libary have to be loaded or something? Can somebod shed a light on this?

For reference, this was the original by _argos:

function abs( mixed_number )  {
    // +   original by: _argos
    return ( ( !isNaN ( mixed_number ) ) ? ( ( mixed_number < 0 ) ? ( mixed_number * -1 ) : mixed_number ) : 0 );
}

#1. Karol Kowalski on 30 January 2008

Why not use native JS Math.abs for this. I believe this should be faster than conditional expressions.

function abs( mixed_number )  {
    // http://kevin.vanzonneveld.net
    // +   original by: _argos
    // *     example 1: abs(4.2);
    // *     returns 1: 4.2
    // *     example 2: abs(-4.2);
    // *     returns 2: 4.2
    // *     example 3: abs(-5);
    // *     returns 3: 5
    // *     example 4: abs('_argos');
    // *     returns 4: 0
 
  var abs=Math.abs( mixed_number )
  return ( !isNaN ( abs) ) ? abs : 0
}