regex pro needed

Anon

hello, im trying to write a function that converts ubb encoding to html. ubb is a basic syntax that lets users mark up their posts. using something like this is safer than allowing them to use html. there was an article on this a while ago. (using formatted user input)

heres a quick example:

hello -> hello

get the point?

the problem im having is this:

hello ive been looking at regex all day and im going blind!

the pattern im using matches the first and the LAST so the whole sentence becomes bold. also the [/b] and in the middle are left in the post.

here is the pattern:
$string = eregi_replace("([b])(.+)([/b])", "\2", $string);

here is the whole function:

function ubb2html($string)
{
$string = eregi_replace("([b])(.+)([/b])", "\2", $string);
$string = eregi_replace("([u])(.+)([/u])", "\2", $string);
$string = eregi_replace("([i])(.+)([/i])", "\2", $string);
$string = eregi_replace("([p])(.+)([/p])", "\2", $string);

$string = eregi_replace("(\[code\])(.+)(\[/code\])", "<blockquote><pre>\\2</pre></blockquote>", $string);
$string = eregi_replace("(\[quote\])(.+)(\[/quote\])", "<blockquote>\\2</blockquote>", $string);

$string = eregi_replace("(\[list\])(.+)(\[/list\])", "<ul>\\2</ul>", $string);
$string = eregi_replace("(\[list=)(a|1)(\])(.+)(\[/list\])", "<ol type=\\2>\\4", $string);
$string = eregi_replace("\[\*\]", "<li>", $string);

$string = eregi_replace("(^|[[:space:]])(http://[[:graph:]]+)", "<a href=\"\\2\">\\2</a>", $string);
$string = eregi_replace("(^|[[:space:]])(www\\.[[:graph:]]+)", "<a href=\"http://\\2\">\\2</a>", $string);

$string = eregi_replace("(\[url\])(http|https|ftp)(://[[:graph:]]+)(\[/url\])",
                        "<a href=\"\\2\\3\" target=\"_blank\">\\2\\3</A>", $string);
$string = eregi_replace("(\[url\])([[:graph:]]+)(\[/url\])",
                        "<a href=\"http://\\2\" target=\"_blank\">\\2</A>", $string);
$string = eregi_replace("(\[url=)(http|https|ftp)(://[[:graph:]]+)(\])(.+)(\[/url\])",
                        "<a href=\"\\2\\3\" target=\"_blank\">\\5</A>", $string);
$string = eregi_replace("(\[url=)([[:graph:]]+)(\])(.+)(\[/url\])",
                        "<a href=\"http://\\2\" target=\"_blank\">\\4</A>", $string);

$string = eregi_replace("(\[email\])([[:graph:]]+@[[:graph:]]+)(\[/email\])",
                        "<a href=\"mailto:\\2\">\\2</A>", $string);
$string = eregi_replace("(\[img\])([[:graph:]]+)(\[/img\])",
                        "<img src=\"\\2\" border=0 alt=\"\\2\">", $string);
return $string;

}

everything else seems to work ok except for a similar url problem that ill deal with later 🙂 any help would be GREATLY appreciated

thanks!

-paul
http://xhawk.net

comstar

you need to tell it NOT to match a [/b] while grabbing all the other characters. Regexes are "greedy"... they take as much possible while still matching.

SO... this:
$string = eregi_replace("([b])(.+)([/b])", "\2", $string);

SHOULD BE this:
$string = eregi_replace("([b])(.[^{[/b]]+)([/b])",} "\2", $string);

Note that I don't have time to test this right this minute but I am 95% sure it is correct. 😉 Enjoy.

Oh yeah... did you REALLY have to use [ and ] to denote a tag?

-Philip

Anon

argh, i hate it when the solution is that easy... 🙂

i had been trying
$string = eregi_replace("([b])(.+[^{[/b]]+)([/b])",} "\2", $string);

and other variations.

[ and ] dont seem to be good choices from a regex point of view, i agree. i didnt design the markup language though im just using it.

thanks for your help

-paul
http://xhawk.net

Anon

hrmmm...
sorry for being annoying, but why does the bold line work perfectly and the italics line does not work at all?

$string = eregi_replace("(\[i\])(.[^\[/i\]]+)(\[/i\])", "<i>\\2</i>", $string);
$string = eregi_replace("(\[b\])(.[^\[/b\]]+)(\[/b\])", "<b>\\2</b>", $string);

am i missing something?

-p