Question

Similar to How replace all spaces inside HTML elements with   using preg_replace?

Except I only want to modify spaces found between PRE tags. For example:

<table atrr="zxzx"><tr>
<td>adfa a   adfadfaf></td><td><br /> dfa  dfa</td>
</tr></table>
<pre class="abc" id="abc">abc abc</pre>
<pre>123 123</pre>

would be converted to (note the pre tag may contain attributes, or may not):

<table atrr="zxzx"><tr>
<td>adfa a   adfadfaf></td><td><br /> dfa  dfa</td>
</tr></table>
<pre class="abc" id="abc">abc&nbsp;abc</pre>
<pre>123&nbsp;123</pre>

Answer 1

$html = preg_replace(
      '#(\<pre[^>]*>)(.*)(</pre>)#Umie'
    , "'$1'.str_replace(' ', '&nbsp;', '$2').'$3'"
    , $html);

Has been tested, works with the sample string you provided. It’s ungreedy, you don’t want to replace spaces between </pre> and <pre>. Also works if the <pre></pre> section spans several lines.

Note: this will fail if you have nested situations like <pre> <pre> </pre> </pre>. If you want to be able to parse that, you need to parse the (X)HTML using the Document Object Model.

Update:
I have done some benchmarking and it turns out the callback version is faster by about 1 second per 100,000 iterations, so I think I should also mention that option.

$html = preg_replace_callback(
      '#(\<pre[^>]*>)(.*)(</pre>)#Uim'
    , function($matches){
          return $matches[1].str_replace(' ', '&nbsp;', $matches[2]).$matches[3];
      }
    , $html);

This requires PHP 5.3 or newer, earlier versions do not support anonymous functions.