[NeilHanlon]

Try this:

	function word_limiter($str, $limit = 100, $end_char = '…')
	{
		if (trim($str) == '')
		{
			return $str;
		}

		preg_match('/^\s*+(?:\S++\s*+){1,'.(int) $limit.'}/', $str, $matches);

		if (strlen($str) == strlen($matches[0]))
		{
			$end_char = '';
		}

		return rtrim($matches[0]).$end_char;
	}

Edit:

After reading Kyek's post, I'll admit I didn't think about what if you're in the middle of a tag.

[Cyril]

I'd suggest cutting it after a punctuation mark (a dot, interrogation / exclamation mark), and even better, if it has a line break after it. So basically, look through your post until you find something like this:

.<br>

Get the index of it, and split the string there. I'll code up something a little later.

[Kyek]

Interesting issue! This DEFINITELY warrants more thought than I've given it, but here's the logical process I'd follow:

• Determine the character position at which you'd like to cut the string (since this contains HTML, this might be an entire topic itself since HTML characters probably shouldn't be counted)
  
• Call strrpos() to search backward from that position, looking for the last occurrence of '<'. Then do it again, looking for the last occurrence of '>'.
  
• If both return false, you're safe to cut.
  
• If the index of the last '<' is greater than the index of the last '>', use strpos (with an offset) to find the next index of >. Move your cut point after that.
  
• Now you need to make sure you're not cutting in between an opening tag and a closing tag -- for example, this would be bad:

  <strong>How to code [CUT] in PHP</strong>

  . You have a couple different options here.
  • Use an HTML parsing library. This is easiest.
    
  • Cut the string at your cut point. Use preg_match to count the number of <tags> WITHOUT a '/' character after the <, then use it again to find the number of </tags> WITH a slash character after the <. Make sure to ignore tags that self <close/>. If the numbers aren't equal, you're probably cutting before a closing tag.
  
  
• If you used an HTML parser to do the above, you may be able to find out what tags weren't closed and close them automatically. Otherwise, you're stuck doing another really ugly round of preg_match to figure that out on your own.

All in all, this is a really tough one. I've never looked for or used an HTML parser in PHP, but it sounds like that might be the better way to go if you can find one.

Edit: Just read Cyril's post above. That's probably the best idea so far, but *only* if you can guarantee that you won't have opened tags that span <br/>s.

[Cyril]

Kyek, on 13 July 2012 - 08:50 AM, said:

Interesting issue! This DEFINITELY warrants more thought than I've given it

Kyek, on 13 July 2012 - 08:50 AM, said:

Edit: Just read Cyril's post above. That's probably the best idea so far, but *only* if you can guarantee that you won't have opened tags that span <br/>s.

Well, that's just a question of doing a check; as you said, simply search backward from the position where it's going to cut to see if there are any open tags. (Also, you could simply close the tag instead of trying to find another cutting point. Shouldn't be an issue, as you'd usually find punctuation and line breaks in paragraph tags.)

[callumacrae]

Use this to convert to HTML, then this to convert to tree. Cycle through elements counting characters, cut text from the element that the text ends on, and then delete all nodes after that.

You'll want to cache that.

[Rob]

Another, perhaps cleaner way would be to render the text, including BBCODE in memory and then strip all of the tags, determine where you want to insert the break. Then you could optionally re-insert the BBCODE tags or leave them out.

Rob

[callumacrae]

That's not as easy as you make it sound, though.