Something like bbCode...

rIkLuNd

In my forum that is a work in progressm I want the users to be able to write php codes, and the code should be higlighted, just like on this forum. I know how to turn into <b>, but how should do to make text within

 and [ /php] to be higtlighted?

suntra

Take the code between

and

, put it in a variable, for example $phpcode. Then, use highlight_string(). (ex: highlight_string($phpcode)).

planetsim

Actually i made my own.. I havent got the PHP or code working yet.. But theres some good Tutorials on Regex. A forum i visited had a good post that i thought helped me alot to understand regex..

http://tutorialforums.com/showthread.php?s=&threadid=16817&highlight=bbcode

This really isnt that hard once you understand it..

if you want url you need //2 //3 2 for address 3 for the output words.. instead of just to.. The only thing i need is to get Javascript help to get them to work once i press a button..

suntra

Oops... !!! Take 2 !

Take the code between [ PHP ] and [ /PHP ], put it in a variable, for example $phpcode. Then, use highlight_string(). (ex: highlight_string($phpcode)).

planetsim

suntra that wont work... as hes gonna be getting this from the same var.. eg $message.. so it can have /n so that if theres a new line etc.

then his bbcode

which will change a bit.. Then of course he may have similes.. So we cant keep making vars out of nothing.. But its similar to what needs to be done

suntra

Well, in my head, that's absolutely possible...

Cut the message in three parts...
The first, all before [ PHP ] (suppose $first_part)
The second, all between [ PHP ] and [ /PHP ] (suppose $second_part)
The third, all after [ /PHP ] (suppose $third_part)

Then just gather them like this :
$message = $first_part . highlight($second_part) . $third_part;

This won't work ?!?! Hé bien, t'es drôle toi !

suntra

Well... there's effectively a little problem... highlight_string() automatically echoes the string... but you can use the output buffering option...

<?
ob_start();
$var ="<?\n \$var1 = \"value 1\";\n \$var2 = Function(\"9090\", \"AAA\", 1);\n \$1 = 1 + 1;\n?>";
highlight_string($var);
$var_h = ob_get_contents();
ob_end_clean();
echo $var_h;
?>

suntra

Could also make that :

ighlight_string ($string_to_highlight, TRUE);

planetsim

Forums are big enough as it is extra code to split things up is just a waste of time.. Thats what regex is for.. If he doesnt use that then hes gonna have lines and lines of code when he could just have maybe 5-6lines.. In fact with italics,bold,url,email i got it down to 4lines.. Quite impressive...

You should be smart and code smart.. That just sounds silly to cut them up..

suntra

I think he'll need a lot of lines to correctly highlight PHP code... the must find string, function, parameters, .............. ! There's already a function for that ! C'mon, don't make your vieux garçon !

planetsim

Regex will do it

$php= '/(\[php\])(.+)(\[\/php\])/';
$message=preg_replace($php,"highlight the stuff and do what you need remember to use //2",$message);

Look one line code thats needed.. You could change the $php and put that code there..

rIkLuNd

Your last code did not work. It said:

Unknown modifier 'p' on line 13 in ...

Line 13:

$message=preg_replace($php...

Duey

you could also write what evers between [PHP/] [/PHP/] to a file and use show_source(filename)

rIkLuNd

yes, but then there would be lots of files... but how do I get the stuff between

 and [ /php]?

Duey

this is how vbulitin does it:

<?
function phphighlite($code) {
  //PHP 4 only

  if (floor(phpversion())<4) {
    $buffer=$code;
  } else {
		$code = str_replace("<br>", "", $code);
		$code = str_replace("<br />", "", $code);
		$code = str_replace("&gt;", ">", $code);
		$code = str_replace("&lt;", "<", $code);

	$code = str_replace("&amp;", "&", $code);
	$code = str_replace('$', '\$', $code);
	$code = str_replace('\n', '\\\\n', $code);
	$code = str_replace('\r', '\\\\r', $code);
	$code = str_replace('\t', '\\\\t', $code);

	$code = stripslashes($code);

	if (!strpos($code,"<?") and substr($code,0,2)!="<?") {
		$code="<?\n".trim($code)."\n?>";
		$addedtags=1;
	}
	ob_start();
	$oldlevel=error_reporting(0);
	highlight_string($code);
	error_reporting($oldlevel);
	$buffer = ob_get_contents();
	ob_end_clean();
	if ($addedtags) {
	  $openingpos = strpos($buffer,'&lt;?');
	  $closingpos = strrpos($buffer, '?');
	  $buffer=substr($buffer, 0, $openingpos).substr($buffer, $openingpos+5, $closingpos-($openingpos+5)).substr($buffer, $closingpos+5);
	}
	$buffer = str_replace("&quot;", "\"", $buffer);
  }

  return "</normalfont><blockquote><pre><smallfont>PHP:</smallfont><hr>$buffer<hr></pre></blockquote><normalfont>";
}?>

Weedpacket

Just a little something to throw in here:

$string = preg_replace('/\[php\](.*?)\[\/php\]/se',
          "highlight_string('<?php\n$1\n?>',true)",
          $string);

Not that I'm saying it's bulletproof or anything... I do think breaking the text into pieces is safer...

rIkLuNd

How do I do to just highlight the stuff between

 and [ /php]? To do that I have to select ONLY that stuff! Plz help me!

Weedpacket

Doesn't that regexp I dashed off work? I mean, it even replaces [pseudotags] with <?php mode ones ?> ...

planetsim

Weedpacket what you did should work.. Maybe he didnt try it.. Or he didnt copy it all..

I didnt test it but that looks pretty much what is needed

Weedpacket

A browse through the Bugzilla source, and a bit of porting, turned up an idea.

Keep arrays of all the various bits that need to be worked on. One array for each type. In this case, there is just one, for $phpcodeblocks.

First, I'm going to "escape" # marks, so that I can use them as delimiters.

$text = str_replace('#','%#',$text);

Next, go through, finding bits we'll need to work on.

$count=0;
$placeholders = array();
$phpcodeblocks = array();
while(preg_match("/pattern to match things/", $text, $matches))
{	$placeholders[$count]="##$count##";
	$text = preg_replace("/pattern to match things/",
	        $placeholders[$count],
	        $text,
	        1); // Not all at once!
	$phpcodeblocks[$count++]=$matches[0];
}

(Bugzilla is written in Perl, which allows capturing bits of matched strings even as they're being replaced. PHP don't do that, though, which is why the pattern matching is done twice.) In this case, the pattern woul be much like the one I gave earlier. A faster one would be

#\[php\]((?!\[/php\]).)*\[/php\]#s

(The e modifier indicated that the replacement string is to be executed; it has no use in preg_match(), and we're not using it here in preg_replace() either.)

Now $phpcodeblocks is an array of all the bits in the text that we want to work over; in the original text they've been replaced by placeholders of the form ##\d+## (which is why we started by making sure such a pattern was otherwise impossible). We can work on each in splendid isolation.

When we're done, we copy them back in to the original. The reformatted $phpcodeblocks[0] will replace "##0##", $phpcodeblocks[1] will replace "##1##" and so on.

$text=str_replace($placeholders, $phpcodeblocks, $text);

And unescape the placeholder characters.

$text = str_replace('%#','#',$text);

I might go and try this out myself.

Of course, this was intended for situations where there was more than one type of block that needed to be reformatted - the idea being that the various modifications each one carried out didn't conflict with each other. WIth just one type of block, you can wrap all the reformatting code into a single function, and use an e-modified preg_replace() (like that of my previous post) to call it.