Web 2.1

If you are obfuscating source code to use with the Prototype library, be careful. Prototype actually interrogates parts of your source code literally!!! The only example I am aware of so far is the use of $super as a function argument. This cannot be renamed because Prototype searches for it as a string literal.

To solve this problem, use the –protect option of jsobf, and don’t forget to escape your “$” symbols ;)

jsobf –protect=\$super

Further comments about the –protect switch.

I ended up adding the identifiers that I wish to protect in a “special” comment syntax. Then I used a script to scan the file, collect all the identifiers to protect and then invoke jsobf.

I used the following format:
//#: MyClassName
function MyClassName()
{
…
}
MyClassName.prototype = {
…
}

Note the line starting with the //#:
It would be great if you could implement a similar feature.

Thanks, keep up the great work !!!

This is a great tool.
But I can’t remove the ‘-i’ switch without problems:

Example:
function ModalDisplay()
{
…
}
ModalDisplay.prototype = {
…
clearFavourites : function()
{
var i;
for (i=0; i<this.thumbNails.length; i++)
{
this.thumbNails[i].setSelected(false);
}
},
…
}

Example after jsobf and then jsfmt on the above:
clearFavourites : function ( ) {
var $3a;
for ( $3a = 0; $3a < this.thumbNails.length; $3a ++ ) {
this.thumbNails[i].setSelected ( false );
}
}

NOTE: the code reads this.thumbNails[i] and not this.thumbNails[$3a]

There are several places in my code where this happens and I have simply had to manually weed them out using the –protect switch. But I am still puzzled. Am I doing something wrong?

Reg

SOLVED – with workaround
I think I have isolated the situation in which the above bug occurs. In any situation as follows:
> myClass.myMember[index]
‘index’ appears to get ignored. A quick work around is to –protect one or two variable names which you can then re-use in these situations.

eg:
–protect ix,…
> for(var ix; i myClass.myMember[ix] = whatever;
> }

So yea one little bug, but still a great tool!

Thanks for this Reg. I guess the part of the parse tree that deals with bracket access isn’t carrying the obfuscated variables into its scope. I’ll take a look when I get the chance.

I didn’t realise anyone was actually using it ;)