Question

So basically my regex is not working as I expect & I don’t know why.

I am working in a fairly regulated environment so this should not be too much of a problem – all the html tags are generated by a script & follow this pattern: only li, p and h(3-6) tags are present. all text is between tags and there are no spaces between tags.

I ‘need’ to write something to surround the lis with ul tags. here is what i got:

preg_replace('#(<li>[^<p|<h]+</li>)(?!<li>)#', '<ul>$1</ul>', $html)

however it only matches the last li pair in a set for some reason. Anyone can tell me why … please?

Answer 1

[^<p|<h] doesn’t do what you expect. It matches a single character that is not any of the characters <p|h. If your HTML really is as constrained as you say, and you cannot have an <li> nested inside another <li>, then the following should work:

preg_replace('#(<li>.*?</li>)+#', '<ul>$0</ul>', $html)

The sequence .*? is just like .* except the trailing ? is the non-greedy modifier. By default .* is greedy – it will consume as many characters as it can, then backtrack if the rest of the pattern doesn’t match. The non-greedy modifier inverts this. It consumes as few characters as it can and advances if the rest of the pattern cannot match. As the rest of the pattern is simply </li>, this effectively captures all text up to, but not including, the first sequence </li>. This pattern is then nested inside a capture which is then repeated with +, meaning it will match one or more sequences of <li> tags.