[RESOLVED] REGEX : ^, $, m modifier, *, GETTING CRAZY !

suntra

Hi everyone !

I'm getting totally crazy !

I'm getting more and more used with REGEX, but tonight, I'm just too confused !

Isn't the "m" modifier supposed to work line by line, so "^" is the start of a line, and "$" the end of the same line... What's I'm trying to do is to use preg_replace_callback() to replace everything after a HTML tag and before the end of a line... so normally, it should be something like this : />[^<]$/m, but it returns a line with a new line character and the content of the other line up to the "<" character !! Even with [^<]? it doesn't work !

I noticed that as soon as I write a "*", it gets mixed up...

Can someone make my hamster run normally ? (funny expression)

Here are some patterns used with preg_replace_callback()... I printed the resultats with print_r() and put them below (some lines were not included because they were empty...)

Here is the string I used :

<html>
 no tags $id
 </html>

And now :
pattern
results...

/^.*$/m

-> <html>
-> no tags $id
-> </html>

/>.*$/m

-> <
-> <

/^.*>$/m

-> <html>
-> </html>

/>[^<]*$/m

-> >
 no tags $id
^^^ There is a /$/?"$?% new line !!!
->

[^<]*?$/m

-> 
 no tags $id
^^^ There is a /$/?"$?% new line !!!
-> /html>

bradgrafelman

Try using the 's' modifier?

EDIT: Nvm, didn't look at question carefully... no clue what to suggest.

Weedpacket

The ^ that appears inside a character class has nothing to do with the ^ that appears outside the character class. The one that appears outside the class means "the start of a line", the one inside means "not these characters".

Basically it's an overloaded operator.

Could you explain what you actually mean by "everything after a HTML tag and before the end of a line" because by the way I understand it, there's nothing in your "<html>...</html>" sample that would match.

suntra

Well I simply wanted to show the error I got...

Example :

<html>
<body>
<table>
  <tr>
    <td> This is what I want to get !
    </td>
  </tr>
</table>
</body>
</html>

So, last ">", then anything EXCEPT ">", then end of line (not of string)...

Weedpacket

Something like...

<?php
$string=<<<EOF
<html>
<body>
<table>
  <tr>
    <td>

 I don't want to get this!
    <td> This is what I want to get !
    <td> I want this <but not this>
    </td>
  </tr>
</table>
</body>
</html>
EOF;

preg_match_all('/>.*?((?=<)|$)/m', $string, $m);
print_r($m[0]);

That is to say, ">", followed either by everything (except a line break) up to (but not including) the next "<" or the end of the line. It could probably be simplified.

In fact, I'm guessing you don't really want the > either, you just want to know it's there.

/(?<=>).*?((?=<)|$)/m

suntra

That's fantastic, you're such an intelligent guy 😃

One day I'll master REGEX perfectly...

I was wondering if it was possible to avoid getting empty lines... like when there's nothing at the end of a line, I don't it to be included in the $m array... is that possible ?

At the moment, I'm using preg_replace_callback() (cause that's what is going to be used), and in that function, I look if if the result is empty / only \n / only \r / only \r\n (even though I converted every \r\n to \n)...

$string=<<<EOF
<html>
<body>
<table>
  <tr>
    <td>
I don't want to get this!
Sexy    <td> This is what I want to get ! <and> ceci
    <td> I want this <but not this>
    </td>
  </tr>
</table>
</body>
</html>
EOF;
function F($arrParties) {
    if(
        ($arrParties[1] <> '')
        && ($arrParties[1] <> "\n")
        && ($arrParties[1] <> "\r")
        && ($arrParties[1] <> "\r\n")
      ) {
      return '(((' . $arrParties[1] . ')))';
    } else {
        return $arrParties[0];
    }
}
$string = str_replace("\r\n", "\n", $string);
$string = preg_replace_callback('/^([^<]*?)((?=<)|$)/m', 'F', $string);
$string = preg_replace_callback('/(?<=>)(.*?)((?=<)|$)/m', 'F', $string);
echo nl2br(htmlentities($string));

suntra

Well looks like I've got it. What I wanted to do is to "extract" and replace any text between tags that contains a variable (like $id), withtout extracting the space... I wanted this to work line by line, so this is what I've got :

<?
 $string='
<html>
<body>
<table>
  <tr>
 $id	<td>
I don\'t want $id to get this!
Sexy    <td> This is what $id I want to get ! <and> ceci $id
    <td> I want this $id<but not this>
    </td>
  </tr>
</table>
</body>
</html>';
function F($arrParties) {
    if(
        ($arrParties[1] <> '')
        && ($arrParties[1] <> "\n")
        && ($arrParties[1] <> "\r")
        && ($arrParties[1] <> "\r\n")
      ) {
      return '(((' . str_replace('$', '_', $arrParties[1]) . ')))';
    } else {
        return $arrParties[0];
    }
}
$string = str_replace("\r\n", "\n", $string);
$string = preg_replace_callback('/^([^<]*\$([a-zA-Z_\x7f-\xff][a-zA-Z0-9_\x7f-\xff]*)[^<]*?)((?=<)|$)/m', 'F', $string);
$string = preg_replace_callback('/(?<=>)([^<]*\$([a-zA-Z_\x7f-\xff][a-zA-Z0-9_\x7f-\xff]*)[^<]*?)((?=<)|$)/m', 'F', $string);
echo nl2br(htmlentities($string));
?>

So thanks Weedpacket 🙂