a php parser written in php ?

athinker

Hi, All:

I am looking for a php parser written in php.

I would like to use it to fetch the chunk of the code of every individual function in a php file and stroe the code in an array.

It seems that it is easier to find out the starting point of a function since we can simply search the keywork "function". However to find out the end point of a function is not such easy.

If someone has or knows some code, please help me out.

I really appreciate for your help in advance.

rachel2004

Two points regarding this:

Do your PHP files have so many PHP functions that work is warranted to build a parser? It would seem to be faster to do things manually where functions are concerned and just remove them by hand from a file.
If you do as you plan, wouldn't you have so many function snippets in your array that you'd be consuming quite a deal of memory?

I wasn't sure if you've thought about this yet.

athinker

many thanks for your kind reply and suggestions.

There are lots of reasons that I would like to achieve the goal listed in my original post.

So please help me to reach that goal.

Many thanks indeed again.

rachel2004

Off the bat, you will need to write some regular expressions code, assuming, of course, that you already know what strings you are looking for.

I have no futher suggestions beyond that.

LordShryku

Write comments in your functions to signify the end and this wouldn't be hard.

function blah() {
  $this = do_something();
  return $this;
} // End blah

As far as a "parser", eval should suffice.

athinker

The problem is we have lots of php files written by someone else.

To add such comment/tag at the end of each function is one of our task.

So we need a script to do the job rather than using a person.

We know that the main things needed are some regular expressions.

I may be able to write my own. However I wonder whether there are some fully tested code existing that I can pick it and reuse it.

Any further suggestions?

Thanks alot.

rachel2004

Originally posted by LordShryku

As far as a "parser", eval should suffice.

I'm not trying to challenge your knowledge, so my following question is just to help me understand... if he uses eval(), wouldn't he still have the chore of escaping stuff within the functions in order for eval() to work correctly? Again, I'm not trying to argue with you... I'm just curious where you were going with it.

LordShryku

I think escaping certain quotes and symbols would be easier than building a full on php parser. From the sounds of it, this is something he has to read in, not include. Unless I mis-understand, and he has no intention of executing this code...

Arc

I would simply load the file into a string and use explode to break it up into an array of functions based on the word "Function"

<?php
$TheString="function html_header($title){
//bunch of code
}
function html_footer($title){
//bunch of code
}
function html_side($title){
//bunch of code
}";

$TheArray=explode("function",$TheString);
for($Counter=1;$Counter<count($TheArray);$Counter++){
print "function".$TheArray[$Counter]."//Your Comments at the end<br>";
}
?>

of course you'll have to make it so the word "function" is not case sensitive since not everyone uses all lower case.

Also, when printing this out to a php page as the example does, the variables in the function, "$title" for example will not show up as the variable, but as the variable value. But from what I can guess you are not printing this out to a browser, but are just I/O with afile so that shouldn't be a problem.

athinker

Thank you very much, Arc.

Yes, it IS a solution providing there is no other code between two functions.

If we have something like:

function foo1(){
blah blah;
}

blah blah;

function foo2(){
blah blah;
}

then ?

An interesting problem, isn't it?

Any further suggestion, please ?

Arc

An algorithm can only do so much. If your pages are that diverse and vary that much I think spending this much time writing an algorithm is probably inefficient. This is something you only have to do once right? You could probably have been half done by now.

Weedpacket

The [man]token_get_all[/man] function can be used to at least get started with parsing (before parsing comes tokenising)....

crimaniak

You need full parser to reach this goal. PHP is not the language to write parsers. Too slooooow... Better to write separate application (using boost::spirit) and run it.
I think fast and dirty solution is to run Doxigen with [x] Source included and strip functions from output. 🙂 preg_match is enought for you in this case.