preg replace problems processing XML data

[gnetcon]

Hello, all!

I apologize if this has been asked and answered. I couldn't find a good solution anywhere. Thought I'd ask the experts. : )

I'm parsing an XML file, and I have run across a problem in the output. The SimpleXML construct is having "issues" because some the data is in the form of "&", "<", ">", etc.

What I'm looking for is a regex pattern to replace all of the HTML characters INSIDE the XML tags, and not within the tags (or XML tags). I tried htmlentities() on it, but that screws up all of the XML.

So, I need:

<tag name="tag">Replace the &lt; and &amp; signs but not the ...</tag>

Any guidance? TIA!

[eelixduppy]

How are you getting the XML file? Are you making it yourself? If you can change the way the XML is created you can avoid this problem altogether.

[gnetcon]

It's provided by a 3rd party program that cannot be changed. I just load the contents of the file into SimpleXML.

[gnetcon]

An update:

If this helps, this is the exact problem I'm having:

<proto name="packet">This is the <&> text that needs escaped.</proto>

So, anything inside the <proto name="packet"></proto> needs to be escaped. Everything else is fine.

[gnetcon]

Update #2:

Well, the regex was really super simple. The following gathers what I need:

preg_match_all("/<proto name=\"packet\">(.*?)<\/proto>/is", $xml, $matches);

From this I get my array of matches. Now my problem is updating the data. I tried this:

foreach ($matches[1] as $match) {
 $replaced = str_replace(array("&", "<", ">"), array("&amp;", "&lt;", "&gt;"), $match);
 $xml = str_replace($match, $replaced, $xml);
}

I just get a blank page.

What would be the fastest way to update these items? This file is currently 3 MB's, but it could MUCH larger.

TIA!

[idfer]

Considering the size of your text, you may be better off rolling your own search and replace with strpos() and substr(). Here's an example:

$tagStart = '<proto name="packet">';
$tagEnd = '</proto>';
$tagStartLen = strlen($tagStart);
$tagEndLen = strlen($tagEnd);
 
$corrected = '';
$ptr = 0;
while(($posStart = strpos($xml, $tagStart, $ptr)) !== false) {
 // Append text up to and including tagStart, advance ptr.
 $corrected .= substr($xml, $ptr, $posStart-$ptr+$tagStartLen);
 $ptr = $posStart + $tagStartLen;
 
 // Find tagEnd.
 $posEnd = strpos($xml, $tagEnd, $ptr);
 if($posEnd === false) // No tagEnd!
$posEnd = strlen($xml) - 1;
 
 // Append corrected tag content, advance ptr to beginning of tagEnd.
 $corrected .= htmlentities(substr($xml, $ptr, $posEnd-$ptr));
 $ptr = $posEnd;
}
 
// Append rest of text.
$corrected .= substr($xml, $ptr);

Hope this helps.

[gnetcon]

idfer:

That code worked well. I ended up working with the developer to make sure the code is now escaped, but this should help anyone else who might have the same issue.