How to clean and simplify this code? - php

Question

After thinking about This Question and giving an answer to it I wanted to do more about that to train myself.
So I wrote a function which will calc the length of an given function. Th given php-file has to start at the beginning of the needed function.
Example: If the function is in a big phpfile with lots of functions, like
/* lots of functions */
function f_interesting($arg) {
/* function */
}
/* lots of other functions */
then $part3 of my function will require to begin like that (after the starting-{ of the interesting function):
/* function */
}
/* lots of other functions */
Now that's not the problem, but I would like to know if there are an cleaner or simplier ways to do this. Here's my function: (I already cleaned a lot of testing-echo-commands)
(The idea behind it is explained here)
function f_analysis ($part3) {
if(isset($part3)) {
$char_array = str_split($part3); //get array of chars
$end_key = false; //length of function
$depth = 0; //How much of unclosed '{'
$in_sstr = false; //is next char inside in ''-String?
$in_dstr = false; //is nect char inside an ""-String?
$in_sl_comment = false; //inside an //-comment?
$in_ml_comment = false; //inside an /* */-comment?
$may_comment = false; //was the last char an '/' which can start a comment?
$may_ml_comment_end = false; //was the last char an '*' which may end a /**/-comment?
foreach($char_array as $key=>$char) {
if($in_sstr) {
if ($char == "'") {
$in_sstr = false;
}
}
else if($in_dstr) {
if($char == '"') {
$in_dstr = false;
}
}
else if($in_sl_comment) {
if($char == "\n") {
$in_sl_comment = false;
}
}
else if($in_ml_comment) {
if($may_ml_comment_end) {
$may_ml_comment_end = false;
if($char == '/') {
$in_ml_comment = false;
}
}
if($char == '*') {
$may_ml_comment_end = true;
}
}
else if ($may_comment) {
if($char == '/') {
$in_sl_comment = true;
}
else if($char == '*') {
$in_ml_comment = true;
}
$may_comment = false;
}
else {
switch ($char) {
case '{':
$depth++;
break;
case '}':
$depth--;
break;
case '/':
$may_comment = true;
break;
case '"':
$in_dstr = true;
break;
case "'":
$in_sstr = true;
break;
}
}
if($depth < 0) {
$last_key = $key;
break;
}
}
} else echo '<br>$part3 of f_analysis not set!';
return ($last_key===false) ? false : $last_key+1; //will be false or the length of the function
}

Answer 1

Tokenizer (Example) - Learn it, love it.

Answer 2

You could probably reduce the number of state variables a little, but truthfully... yes, it will be messy code. I would probably get rid of $may_ml_comment_end and peek ahead for the next character when I encounter an asterisk, for example. You will need to rewrite your foreach loop to a regular for loop be able to do that without creating a bigger mess though.
PS: I don't see you handling the escape character yet. Without the above approach, that would introduce another boolean variable.
Another problem with your current code is that characters immediately following a / don't get interpreted as they should. However unlikely
echo 5/'2'; // NB: no space in between
is valid in PHP and would break your parser.