Hi, I asking for help in my PHP code optimization. I try to rewrite my old/slow BBCode conversion function to use regex. The allowed BBCodes are

[ b]text[ /b]
[ i]text[ /i]
[ url]domain.com[ /url] or [ url]domain.com|domainname[ /url]
[ img]filename.ext[ /img] or [ img]filename.ext|description[ /img]

It is possible to construct regex with the following conditions?

[ b] - inside only [ i] and [ url] BBCode allowed, newline NOT allowed
[ i] - as [ b]
[ url] - inside NO BBCode allowed, newline NOT allowed
[ img] - as [ url]

(I try to make difference between two [ url] (or [ img]) syntax with '/\[url\](.+?)\[\/url\]/is' and '/\[url\](.+?)\|(.+?)\[\/url\]/is' patterns, but it generates malfunction.)

Thanks for help in advance 

Are you testing a string to see if it matches the conditions outlined?

If so, if the string matches the following regex it has failed at least 1 of your conditions:

if(preg_match('#(?=.*(?:\[([bi])\](?:(?!\[/\1\]).)*?(?:\[(?:img|b)\]|\n))|(?:\[(url|img)\](?:(?!\[/\2\]).)*?(?:\[[a-z]+\]|\n)))#s',$sourcestring)){

  echo "failed validation";

}

If you are not trying to validate based on the conditions given, possibly explain the question further.

I don't validate the string, just replace tags one-by-one with preg_replace (if matches for the pattern)
for example
preg_replace('/\[b\](.+?)\[\/b\]/is', '<b>$1</b>', $str);
preg_replace('/\[img\](.+?)\[\/img\]/is', '<div style='text-align:center'><img src='$1'></div>', $str);

I guess the (complex) pattern above matches for every conditions I mentioned
To understand the formula, please tell me what will be the pattern only for [ b] (for example)?

Thanks

if(preg_match('#(?=.*(?:\[([bi])\](?:(?!\[/\1\]).)*?(?:\[(?:img|b)\]|\n))|(?:\[(url|img)\](?:(?!\[/\2\]).)*?(?:\[[a-z]+\]|\n)))#s',$bbcode)){
  echo 'failed validation';
} else {
  $bbcode=preg_replace('#\[(/?[bi])\]#i','<$1>',$bbcode);
  $bbcode=preg_replace('#\[img\](.*?)\[/img\]#i',"<div style='text-align:center'><img src='$1'></div>',$bbcode);
  $bbcode=preg_replace('#\[url\](.*?)\[/url\]#i',"<a href='$1'>$1</a>",$bbcode);
  echo $bbcode;
}

Thank you for the fast response.
What is the regex pattern for [ url]domain.com|webpage[ /url] to <a href='domain.com'>webpage</a>
while keep working the '#\[url\](.*?)\[/url\]#i' pattern?
(For example in case [ url]domain1.com[ /url] [ url]domain2.com|webpage2[ /url] [ url]domain3.com[ /url])

And how to modify the
$bbcode=preg_replace('#\[url\](.*?)\[/url\]#i',"<a href='$1'>$1</a>",$bbcode);
patternt to match just in case if no [ b] , [/b] , [  i] , [ /i] , [ img] , [ /img] and newline tags inside [ url] and [ /url]

Let's say that you remove the validation logic and just apply the preg_replace function to the string in steps.

Let's start with [b]:  for [b] tags which do not have [b] or [img] tags within their enclosing [/b] block:

$bbcode=preg_replace('#\[b\]((?:(?!\[/?(?:img|b)\]|\n).)*?)\[/b\]#i','<b>$1</b>',$bbcode);

If that preg_replace function works for you as desired that should give you the framework for [i] and [img].

For [url]:

<?php
$bbcode='[url]http://test.com[/url]
[url]http://test2.com|test2[/url]';
$bbcode=preg_replace('#\[url\]((?:(?!\[(?:b|img|i)\]).)*?)\|(.*)\[/url\]#i',"<a href='$1'>$2</a>",$bbcode);
$bbcode=preg_replace('#\[url\]((?:(?!\[(?:b|img|i)\]).)*?)\[/url\]#i',"<a href='$1'>$1</a>",$bbcode);
echo '<pre>'.$bbcode;
?>

Thanks. It works fine when url tags are separated with newline, but without newline it generates strange output.
Try $bbcode='[ url]http://test.com[ /url] [ url]http://test2.com|test2[ /url]';

Try:

<?php
$bbcode='[url]http://test.com[/url]test[url]http://test2.com|test2[/url]test[url]http://test.com[/url]test[url]http://test2.com|test2[/url]';
$bbcode=preg_replace('#\[url\]((?:(?!/?\[(?:url|b|img|i)\]).)*?)\|(.*?)\[/url\]#i',"<a href='$1'>$2</a>",$bbcode);
$bbcode=preg_replace('#\[url\]((?:(?!/?\[(?:url|b|img|i)\]).)*?)\[/url\]#i',"<a href='$1'>$1</a>",$bbcode);
echo '<pre>'.$bbcode;
?>