Regular Expression

Pna

Can anybody help me with a regular expression to split long words in a string which contains html.

Consider this string:

<a href="http://www.loooooong.com">This is a looooooooong wooooooooord</a>

I want to split words above 10 characters, but I do not want to split html tags. The result should be:

<a href="http://www.loooooong.com">This is a looooooooo ng wooooooooo rd</a>

I have tried below regular expression:
$string= preg_replace('/([^{\s<>\&]{10,10})/',} '\1 ', $string);

But that will also split the URL inside the html tag:
<a href="http ://www.loo oooong.com" >This is a looooooooo ng wooooooooo rd</a>

Installer

$txt = strip_tags($str);
$words = explode(' ', $txt);
foreach ($words as & $word) {
    $word = trim(chunk_split($word, 10, ' '));
}
$str = str_replace($txt, implode(' ', $words), $str);

Pna

Installer;10905717 wrote:

$txt = strip_tags($str);
$words = explode(' ', $txt);
foreach ($words as & $word) {
    $word = trim(chunk_split($word, 10, ' '));
}
$str = str_replace($txt, implode(' ', $words), $str);

Thanks. But if the string contains more than one html tag it does not work:

<a href="http://www.loooooong.dk">This is a looooooooong word</a><p>This is also a looooonnnnnnng word</p><form action="" method="get">Search: <input type="text" name="loooongName"><input type="submit" value="Login"/></form>

I need to split all long words in the string, and the string can contain lots of html tags.

scrupul0us

sounds like you want to parse the DOM and with each pass use installers code to render the output you desire

is there a fixed list of html tags you are working with or is it dynamic?

Pna

scrupul0us;10905754 wrote:
sounds like you want to parse the DOM and with each pass use installers code to render the output you desire

is there a fixed list of html tags you are working with or is it dynamic?

It is dynamic.
I have a web forum where users can post from a text editor (fckeditor http://www.fckeditor.net/ )

It is a rich editor where users can create html. And sometimes they post long words which has to be splitted.

scrupul0us

are you looking to split words simply to have them wrap or conform to a fixed width layout box of some sorts?

just trying to get a handle on why you want to split the big words

Pna

scrupul0us;10905768 wrote:
are you looking to split words simply to have them wrap or conform to a fixed width layout box of some sorts?

just trying to get a handle on why you want to split the big words

Yes, that exactly what I am trying.

scrupul0us

What i would do since there isnt any uniform CBC css hack to handle it is this...

First determine if you want to permanently modify the input (e.g your layout will never change)...

in this case... when the user data is submitted, create a new DOM object and feed it the data from the editor... parse each node (html tag) and for each nodes value run the code the break the text apart and write it back to the node... when thats done put it in your database

If your layout may change then youll want to put the input directly into the DB as is and then run the above DOM on the output...

there's ups and downs to both...

the first option will output to the screen faster but you wont be able to reconstruct the user data later on if your layout changes

the second option will output slower as youll have to parse a DOM object on each output, but if you ever change your layout all you have to change is one snippet of code to break the wording

the middle ground would be to store the data twice in the db, once altered once raw... your overhead will be in the db storage and not your process time during the output

Installer

Other things to consider:
- Your users will almost certainly, and frequently, input html that's invalid for all sorts of reasons, breaking any scheme you have.
- If you don't filter (at the least whitelist with strip_tags()), you'll sooner or later be using malicious input.

nrg_alpha

Pna;10905755 wrote:
It is a rich editor where users can create html. And sometimes they post long words which has to be splitted.

I'm not sure I follow... why would long words need to be split up? Aren't you destroying the integrity of the word(s) in question by doing that?

Pna

nrg_alpha;10905794 wrote:
I'm not sure I follow... why would long words need to be split up? Aren't you destroying the integrity of the word(s) in question by doing that?

I need to split up long words because sometimes users post long word which will destroy the layout of the web page. Example:

XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX

In IE the table will wrap the word above, but if you open this page in Firefox you will see a horizontal scroller because of the very long word above. I do not want that.

nrg_alpha

Well, I would be very hard-pressed to find a word that long!

In either case, could not take the input that user has entered and use word_wrap()?

Pna

nrg_alpha;10905820 wrote:
Well, I would be very hard-pressed to find a word that long!

In either case, could not take the input that user has entered and use word_wrap()?

I agree that words that long should never be there. But sometimes users post a linie like this:

I can not just use word_wrap() because the post contains html tags. There might be a <a> tag with a long URL:

And word_wrap() will break the link.

Pna

I made this very simple script to solve my problem.

<?php


$txt = '
<table border="0"> 
   <tbody> 
      <tr> 
         <td><a href="http://www.loooooooong.com" title="Link to loooooooong page">Very loooooooong word</a></td>
         <td>This is a loooooooong word.</td> 
      </tr> 
   </tbody> 
</table>
'; 

echo $txt;
echo "<br/><br/>";

// Length of the text
$length = strlen($txt);

// Initialize
$tag=0;
$word_length=0;
$txt_out="";

// Run through all chars
for($i=0; $i<$length; $i++)
{
	$char=substr($txt,$i,1);

// <tag> begins
if($char=="<")
{
	$tag=1;
}

// If NOT inside <tag>
if($tag==0)
{
	// If new word
	if($char==" ")
	{
		$word_length=0;
	}
	else
	{
		$word_length++;
	}

	// If word to long
	if($word_length==10)
	{
		// Reset word length
		$word_length=0;

		// Put in space in the long word
		$char=$char . " ";
	}
}

// Append
$txt_out = $txt_out . $char;

// <tag> ends
if($char==">")
{
	$tag=0;	
}
}

echo $txt_out;

?>

scrupul0us

what is someone enters:

<<reaaaly long word>>

since you arent doing any filtering it will break

you really should be using the DOM to parse the actual valid tags and their values

Pna

Can somebody show me a simple example how to use the DOM to run through all the tags and their values?

scrupul0us

what element and id/class is your user posted data wrapped in?

Pna

scrupul0us;10905942 wrote:
what element and id/class is your user posted data wrapped in?

When users post their posts with the fckeditor they can use different elements:

<ul>
    <li>Hi</li>
</ul>

<p>
hi
</p>

So I need to loop through all tags and check the values. But I do not know how.