[RESOLVED] preg help please, should be easy

[bfuzze_98]

I'm abissmal at reg exp. I'm learning, but slowly. Been trying to put together what should be a fairly easy one, but to no not having much luck so far. I just need to scan a text file for links submitted in the same format as most of these forums (HTTP,/HTTP each within brackets), so I can reformat it for posting to a web page.

Heres what I've tried so far:

$path = "...";
$filename = "file.txt";
$handle = fopen($path.$filename, "r");
$contents = fread($handle, 8096);
[B] while (preg_match('/[HTTP]/', $contents))	{
 $split1 = preg_split('/[HTTP]/', $contents); 
  $beforelink = $split1[0];
  $afterlink = preg_split('%[/HTTP]%', $split1[1]);

  $raw_url1 = preg_match('%[HTTP](.*?)[/HTTP]%si', $contents);
  $stripped1 = preg_replace('%[HTTP]%', '', $raw_url);
  $subject = preg_replace('%[/HTTP]%', '', $stripped1);
  $newsub = '<a href="'.$subject.'" target="_blank">'.$subject.'</a>';
  $contents = $beforelink.$newsub.$afterlink;
}[/B]

For some reason it keeps splitting at the letter 'T'. I suck at reg exp :mad:

[robche]

something like this?

<?

$replace = array(
	'/\[http\](.+)\[\/http\]/mi'	=> '<a href="$1">$1</a>'
	);


$text_from_some_flat_file = "
	hello world my website can be viewed at [http]http://blog.ubrio.us[/http] 
	i hope you will come visit me! or you can [http]mailto:[e-mail][/http] email me";


$filtered_text = preg_replace(array_keys($replace), array_values($replace), $text_from_some_flat_file);

echo $filtered_text;

?>

sidenote:

it might be easier to gather the entire file (depending on filesize) into a string using

$some_file_text = file_get_contents('file.txt');

[bfuzze_98]

Just wanted to thank you for the assistance. Ran into another problem extracting the text file from the server where posted, but this did/will come in handy in some other applications I'm working on.

I understand what you were doing escaping the '[' and fslashes, but I never would have guessed these:
'(.+)' -- what does this mean? I see why you need '+' (1 or more matches), but why '.'(any char)?
'/mi' -- what does this mean?
'$1' -- I assume this references anything in between the pattern limits (start/end) but not included with the pattern?

I recently came accross this tutorial:
http://www.phpbuilder.com/columns/dario19990616.php3
and I have regexBuddy, which has come in pretty handy. Any other suggestions?

Thanks again for the assistance...

Have a good weekend!

[Weedpacket]

'(.+)' -- what does this mean? I see why you need '+' (1 or more matches), but why '.'(any char)?

Well, you have one or more matches, but one or more matches of what? In this case, anything. If it's a character, it matches.

Not incidentally, the fact that [ is a character, and so are h, t, p, and /, then the pattern "(.+)" will match the "[/http]" that was supposed to indicate the end of the URL. So if there were two URLs in the line, the entire pattern would only see one (starting with the first [http] and ending with the last [/http). If you don't say otherwise, quantifiers like "+" will attempt to match as much as possible, whereas in this case you only want them to match as much as necessary.

Because of that, the "(.+)" needs to be modified slightly, to "(.+?)", where the "?" makes the necessary change to the behaviour.

'/mi' -- what does this mean?

Multiline matching, case insensitive. Multiline so that ^ and $ anchors match the start and end of every line (instead of just the start and end of the entire string); unnecessary here, since ^ and $ are not being used. Case insensitive, so that the pattern will match [HTTP] as well as [http].

These and other pattern modifiers are covered on this page in the manual.

'$1' -- I assume this references anything in between the pattern limits (start/end) but not included with the pattern?

Everything matched by the part of the pattern between the first set of "()" - in this case, the ".+".

[bfuzze_98]

Thanks for the assistance guys!