Help with this regular expression

LiamBailey

(Added [noparse]

...

[/noparse] tags around code. -- MOD)[/i][/color]

I am trying to write my own sitemap generator, but I keep getting errors trying to find the links.

<? 
if ($submit)
   {
   $xml = "<?xml version='1.0' encoding='UTF-8'?>
   <urlset xmlns='http://www.sitemaps.org/schemas/sitemap/0.9'>";
   $contents = file_get_contents("http://www.poetry-occasions.co.uk", "r"); 
   $links[] = preg_match('/^[url]http://www.poetry-occasions.co.uk+[/url][/][a-Z0-9_-][^"]*',$contents);
   fclose($file);
      foreach ($links as $link)
            {
             $xml .= "<url><loc>$link</loc></url>";
            }
       $xml .= "</urlset>";
       $sitemap = fopen("/home/files/public_html/test.xml", "w");
       fwrite($sitemap,"$xml");
       fclose($sitemap);
    }
?>

The error is:

Warning: preg_match() [function.preg-match]: Unknown modifier '/' in /home/files/public_html/sitemap_create.php on line 6

Horizon88

/^http://www.poetry-occasions.co.uk+[/][a-Z0-9_-][^"]*

The section [/] is invalid, from the looks of it. Try doing [\/] instead, maybe?

NogDog

If you use "/" as your regular expression delimiter, then you must escape all literal "/" characters as "[FONT="Courier New"]\/[/FONT]". (Also, you are missing the closing regexp delimiter at the end.)

LiamBailey

I have tried this:

$links[] = preg_match('/^http://www.poetry-occasions.co.uk+[\/][a-Z0-9_-][^"]*/',$contents)

And THis:

$links[] = preg_match('/^http://www.poetry-occasions.co.uk+[\/][a-Z0-9_-][^"]*$/',$contents)

Neither works, the error code is the same:

Warning: preg_match() [function.preg-match]: Unknown modifier '/' in /home/bai11814/public_html/sitemap_create.php on line 6

LiamBailey

It seems to be the delimiter, I tried:

$links[] = preg_match('/^http://www.poetry-occasions.co.uk[.^"]*$/',$contents);

and still the same error code that '/' is an uknown modifier.

What regex would you use to find all links in a page?

Horizon88

You need to escape the // in [url]http://,[/url] otherwise it's registering that first "/" after the ":" as an ending delimiter. Either change your delimiter to something like "#" or escape those slashes.

preg_match('#http://www.poetry-occasions.co.uk...#', $contents);
preg_match('/http:\/\/www.poetry-occasions.co.uk.../', $contents);

LiamBailey

Thanks Horizon, but I finally spotted that for myself, forehead red from that slap believe me.

Now I have no error codes whatsoever, but my sitemap has one line and it has a zero as the one link, so obviously my regex is not finding the links.

$links[] = preg_match('/^http:\/\/www.poetry-occasions.co.uk[\/][A-z0-9-_]*/',$contents);

Doesn't work, and neither does:

$links[] = preg_match('/^http:\/\/www.poetry-occasions.co.uk[.*^"]/',$contents)

Both output the following into test.xml:

<?xml version='1.0' encoding='UTF-8'?>
<urlset xmlns='http://www.sitemaps.org/schemas/sitemap/0.9'><url><loc>0</loc></url></urlset>

NogDog

At a guess, you probably want to use preg_match_all() instead of preg_match(). Also, you'll want to use the 3rd parameter to that function to specify the array variable that stores the results, as both functions only return the number of matches, not the actual matches.

preg_match_all('/your_regular_expression/', $contents, $links);

Then when you loop on it, all the full matches will be a sub-array in $links[0], so your foreach would be:

foreach($links[0] as $link) {

LiamBailey

Thanks, that is very helpful. But the fact that my original attempt only put one row into the sitemap, means my regular expression is wrong.

Can anyone tell me what regular expression I should use to find all the link hrefs on my page.

This is easier in java, but I failed there trying to ajax the results through to my PHP script to create the xml file. So if you can help there instead I can post my code.

LiamBailey

function sitemap()
{
sitemap = '
<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">';
link = "";
links = document.anchors.length-1;
for (loop = 0; loop < links; loop++)
{
link = document.anchors[loop].href;
sitemap += '<url><loc>' + link + '</loc></url>';
}
sitemap += '</urlset>';
if (window.XMLHttpRequest) // Object of the current windows
{
xhr = new XMLHttpRequest(); // Firefox, Safari, ...
}
else
if (window.ActiveXObject) // ActiveX version
{
xhr = new ActiveXObject("Microsoft.XMLHTTP"); // Internet Explorer
}

WHat next???

bradgrafelman

LiamBailey wrote:
But the fact that my original attempt only put one row into the sitemap, means my regular expression is wrong.

No, that means that [man]preg_match/man did its job - it found one single match and then stopped.

LiamBailey

I have now got a regex that works, but the problem is now that:

$contents = file_get_contents("/home/files/public_html/index.phtml", "r");

When I echo contents it is not the html source code, it is the text of the page, so my hyperlinks are not registering. Is there any way I can have it scan the html of the file.

whisher06

Hi.
If you are on a php5 cluster you could exploit
the DOM functions so you haven't to worry about the regex at all.

Bye.