Pattern help!

persa

I use this pattern in a preg_replace_callback() function, and it works fine:

$pattern_html = "#(http|https|ftp|ftps)://([^<>\s]+)#i";

But I want exclude if it stands href=" i before the word, or just =" would be better..

How can that be done in a simple way??

Thanks.🙂

bubblenut

$pattern_html="#[^(=\")](http|https|ftp|ftps):\\/\\/([^<>\\s]+)#i";

Is that what you mean? It will return the path if it is not prepended by a ="
HTH
Bubble

persa

It doenst work exactly like i planned..
Now it only takes those with =" before..

$pattern_html = "#(http|https|ftp|ftps)://([^<>\s]+)#i";

on a other side, i can make the pattern to only transform if the word starts with http(s)/ftp(s).. like http://www.cnn.com .. and dont act to ">http://www.cnn.com

You see?

bubblenut

Originally posted by persa
It doenst work exactly like i planned..
Now it only takes those with =" before..

No it doesn't, it only takes those without a =" before

Originally posted by persa
on a other side, i can make the pattern to only transform if the word starts with http(s)/ftp(s).. like http://www.cnn.com .. and dont act to ">http://www.cnn.com

The board changes your links so I can;t really see what you've written, try putting it in code blocks.
Bubble

persa

How would it be, if the word should have "blank" space before?

like http://www.cnn.com

and not ..>http://www.cnn.com

persa

I tried to use

$pattern_html = "#\b(http|https|ftp|ftps)://([^<>\s]+)#i";

to get words blank space infront..

\b = word boundary

but that would work either.. now it transform all

OhLordy

\b denotes a word boundary, this includes any character that cannot be in a word. What you are looking for is \s (whitespace).

persa

ok that avoid transforming the ones that not has a blank space in front.. but now it doesnt transform the others, that are "ok"..

The might some other thinks in the script that fucks up..

doesnt seems that the "if()" runs in the CheckImage function...

function transform($text){ 

$pattern_html = "#\b(http|https|ftp|ftps)://([^<>\s]+)#i";   

$text = preg_replace_callback($pattern_html,'Check_if_Image',$text); 

return $text; 
} 


function Check_if_Image($matches) { 
//print "<pre>"; var_dump($matches); print "</pre><hr>\n"; 
$dst = $matches[0]; 
if(@fopen($dst, 'r')){ 
$suffix = strtolower(substr($matches[0],-4));  // last 4 chars of match 
if ( $suffix == '.jpg' or $suffix == '.gif') { 
   $dsn = $matches[0]; // $matches[1].'://'.$matches[2]; 
   if (list($width, $height, $type, $attr) = getimagesize($dsn)) { 
     if ( ($width > 150) ) { 
         $koef = $width / 150; 
         $width = 150; 
         $height /= $koef; 
         } 
     if ( ($height > 150) ) { 
         $koef = $height / 150; 
         $height = 150; 
         $width /= $koef; 
         } 
     } 
     else { 
         $height=150; 
         $width=150; 
     } 

 $ret = "<img src=\"$dsn\""; 
 $ret .= "\" border=0 width=\"$width\" height=\"$height\" style=\"display:block;\">"; 
   } else { 
     $ret = '<a href="'.$matches[0].'" target="_blank">'.$matches[0].'</a>'; 
   } 
   return $ret; 
} 
else { 
  return ''; 
} 
}

OhLordy

Haha, I get what you're trying to do now (sorry I'm a little slow today).
try this as your regex
#\s((https?|ftps?)://[^<>\s]+)\s#i
and then access by using $matches[1] rather than $matches[0].
This should match all words begining with either http, https, ftp or ftps etc and surrounded by a whitespace character at the begining and at the end.
HTH
Bubble

persa

Thanks, that worked.. 🙂

Iam using fopen() to check if a image exits.. even if the images are small, or the orthers file.. it takes so much bandwitch.. It seems so for me....

Are there an alternative function, that would do the same? file_exits() are only for local, so.. any others, or some edits that would speed up??

persa

The pattern that you recommended, works fine.. but in some occasions, it doesnt work, even if its a white space before.. I Can't really tell when it works, and when its not.. nothing seems logic for me now..

If you take this string i.e:

$text = "Tekst vil komme...  :)  [url]http://nsmh.net/design/pictures/logo-top2.jpg[/url] 
Kom tilbake senere for en beskrivelse av NSMH, og
 [url]http://www.nsmh.net[/url]";

😕

persa

The script turned the script wrong out.. its like

$text = "Tekst vil komme... 🙂 http://nsmh.net/design/pictures/logo-top2.jpg
Kom tilbake senere for en beskrivelse av NSMH, og
http://www.nsmh.net";

bubblenut

There aren't spaces after it.

Originally posted by OhLordy
and surrounded by a whitespace character at the begining and at the end.

HTH
Bubble

persa

thanks.. didn't read carfully enough, are there someway to make an exception if it stands last/first in a string, then even accept is as a pattern, as my example shows...

dave420

Change the last \s to [\s$]

persa

That didnt work..

I tried it on my example..

And the complete pattern look like:

$pattern_html = "#[\\s$]((https?|ftps?)://[^<>\\s]+)[\\s$]#i";

persa

that script function screwed my script.. it like:

$pattern_html = "#\s$[\s$]#i";