Removing white space and line breaks from source using PHP - php

Question

I have the following function that I use in a PHP application to remove white space and line breaks from the source of a page.
It's based on some examples I have read on Stack Overflow, with some amends to handle JS and HTML comments. Note: I've not used an exisiting library because I wanted something simple without all the additional features that others include and with this code I have fine-grained control over what is stripped and what is not.
protected function MinifyHTML($str) {
$str = preg_replace("/(?<!\S)\/\/\s*[^\r\n]*/", "", $str); // strip JS/CSS comments
$str = preg_replace("/<!--(.*)-->/Uis", "", $str); // strip HTML comments
$protected_parts = array('<pre>,</pre>','<textarea>,</textarea>','<,>');
$extracted_values = array();
$i = 0;
foreach ($protected_parts as $part) {
$finished = false;
$search_offset = $first_offset = 0;
$end_offset = 1;
$startend = explode(',', $part);
if (count($startend) === 1) $startend[1] = $startend[0];
$len0 = strlen($startend[0]); $len1 = strlen($startend[1]);
while ($finished === false) {
$first_offset = strpos($str, $startend[0], $search_offset);
if ($first_offset === false) $finished = true;
else {
$search_offset = strpos($str, $startend[1], $first_offset + $len0);
$extracted_values[$i] = substr($str, $first_offset + $len0, $search_offset - $first_offset - $len0);
$str = substr($str, 0, $first_offset + $len0).'$$#'.$i.'$$'.substr($str, $search_offset);
$search_offset += $len1 + strlen((string)$i) + 5 - strlen($extracted_values[$i]);
++$i;
}
}
}
$str = preg_replace("/\s/", " ", $str);
$str = preg_replace("/\s{2,}/", " ", $str);
$replace = array('> <'=>'><', ' >'=>'>','< '=>'<','</ '=>'</');
$str = str_replace(array_keys($replace), array_values($replace), $str);
for ($d = 0; $d < $i; ++$d)
$str = str_replace('$$#'.$d.'$$', $extracted_values[$d], $str);
return $str;
}
However if I get a scenario like:
Link Link
It will remove that space between the two anchor tags.
I've added '</a> <a' to my $protected_parts in an attempt to stop this, but it still strips out the space between them. So I end up with LinkLink in the source which isn't what I want.
The same also happens with:
<p>This is <span class="">some</span> <span class="">styled</span> text.</p>
Also it seems the protected_parts arn't working as my textareas are being minified too so all the content inside them is compressed down into one line...
Any ideas on the fixes? I've also not been able to find alternatives to use instead that don't implement caching, gzipping and other features I don't want. I purely want a simple solution that strips spaces, line breaks and comments and that's it.

Answer 1

UPDATED 2014/02/25 (late):
Here's another workaround. Instead of touching $protected_parts I'm just adding another replace operation at the end that adds a space after every </a> -- again a workaround, but this shouldn't screw up any of your original operability, and the penalty this time is only one space character after every anchor tag, not bad. Here it is: http://phpfiddle.org/main/code/5qj-13z
UPDATED 2014/02/25:
I added '</a> ' to $protected_parts and it does not strip the space. I threw it into phpfiddle over here, http://phpfiddle.org/lite/code/dms-cud. This is only a workaround for a few lines of synethetic-emulated HTML... I'm not sure what kind of organic code you're running through your function. Obviously this workaround is not a universal fix either.
Original
I added '</a>',' <a ', to $protected_parts and it does not strip the space. I threw it into phpfiddle over here, http://phpfiddle.org/lite/code/ztz-5hf.
Your function is scary to me, but I like some of the basic functionality, like stripping HTML, JS and CSS comments. I'd still recommend using an apache extension or library. Using other people's open source code is the most powerful witchcraft a programmer can yield. :)