Regex for adding a versionnumber to .js and .css references in a HTML document

I have a HTML document where I want to automatically have a version number added to all script and style sheet references.

the following should be replaced

 <link ... href="file.css" media="all" />
 <script ... src="file.js"></script>
 <script ... src="http://myhost.com/file.js"></script>

with that

<link ... href="file.v123456.css" media="all" />
<script ... src="file.v123456.js"></script>
<script ... src="http://开发者_JAVA技巧myhost.com/file.v123456.js"></script>

where 123456 is my dynamic version number.

But it shoud ONLY do that on local files

<link ... href="http://otherhost.com/file.css" media="all" />

should be left untouched.

So far I have the following regex:

$html = preg_replace('#(src|href)=("|\')(?!http)(?!("|\'| |\+))(.*)\.(css|js|swf)("|\')#Ui', '\\1=\\2\\3\\4.v'.$version .'.\\5\\6', $html);

But it does not work 100% and I am sure that there is a better way in doing this. How would you write it?

Edit:

I have it now using DOMDocument, but it turns out that it's pretty slow!

    <?php
//------- snip --------------
            $host       = (isset($_SERVER['HTTPS']) && $_SERVER['HTTPS'] == 'on' ? 'https' : 'http') . '://' . $_SERVER['HTTP_HOST'];
            $refs       = array();
            $version    = "v". $version;
            $doc        = new DOMDocument();        
            $tmpDoc     = $html;

            $doc->loadHTML($tmpDoc);        
            $xpath = new DOMXPath($doc);
            foreach($xpath->query('/html/head/link/@href') as $href) {
                $ref = $href->value;
                if(
                    !preg_match('/^https?:/', $ref) || 
                    strpos($ref, $host) === 0               
                ) {
                    $refs[$ref] = preg_replace('/\.css$/', '.'.$version.'$0', $ref);
                }
            }
            foreach ($xpath->query('//script/@src') as $src) {
                $ref = $src->value;
                if(
                    !preg_match('/^https?:/', $ref) ||
                    strpos($ref, $host) === 0
                ) {
                    $refs[$ref] = preg_replace('/\.js$/', '.'.$version.'$0', $ref);
                }
            }       
            $html = str_replace(
                        array_keys($refs), 
                        array_values($refs), 
                        $tmpDoc
                    );
//------- snip --------------
    ?>

---

Don't use regex, at least not to find the src values. Use the DOM.

var allScriptTags = document.getElementsByTagName("script");
var allLinkTags = document.getElementsByTagName("link");

for(var i=0; i<allScriptTags.length; i++) {
    var srcAttribute = allScriptTags[i].getAttribute("src");

    // ... do something with srcAttribute ...

    // replace it with the modified value
    allScriptTags[i].setAttribute("src", srcAttribute);
}

---

You should better use a HTML parser like Simple HTML DOM Parser or DOMDocument to find the elements/attributes. Then you can use regular expressions to modify the attribute values.

Here’s an example for DOMDocument:

$version = 'v123456';
$doc = new DOMDocument();
$doc->loadHTML($code);
$xpath = new DOMXPath($doc);
foreach ($xpath->query('/html/head/link/@href') as $href) {
    if (!preg_match('/^https?:/', $href->value)) {
        $href->value = preg_replace('/\.css$/', '.'.$version.'$0', $href->value);
    }
}
foreach ($xpath->query('//script/@src') as $src) {
    if (!preg_match('/^https?:/', $src->value)) {
        $src->value = preg_replace('/\.js$/', '.'.$version.'$0', $src->value);
    }
}

---

why not something a lot more simple like:

$version = 12345;
function print_local_css($base_filename)
{
    global $version;
    print '<link rel="stylesheet" type="text/css" href="'.$base_filename.'.v'.$version.'.css" />';
}

function print_local_js($base_filename)
{
    global $version;
    print '<script type="text/javascript" src="'.$base_filename.'.v'.$version.'.js"></script>';
}

It'll be quick, one location to change the version and any files you don't want to be versioned, just manually put the HTML in.