Re: Re: Bug #1683 Updated: Failing to parse/compile a *? pattern.

Howdy. Sure. A tiny bit of background. Phorum has this regexp: $body=eregi_replace("&lt;([^:]*://[^ >\n\t]*)&gt;", "<a href=\"\\1\">\\1</a>", $body); where the 2nd * being greedy means two links on the same line are treated like a single long link. I want the second * to be non-greedy. So in theory I just do this: $body=preg_replace("/&lt;([^:]*:\/\/[^ >\n\t]*?)&gt;/i", "<a href=\"\\1\">\\1</a>", $body); however, that tells me "illegal escape sequence \/". Trying to debug my regexp, I tried this line: $body=preg_replace("/\//", "", $body); same error. However, the stand-alone program: <?PHP $body=preg_replace("/\//", "", $body); ?> creates no error message. If you want to reproduce this, and happen to have phorum yourself, put my small preg_replace command in around line 183 or so of post.php3, after the closing } and before the two egrei_replace commands. This is with Apache 1.3.6, php 3.0.11, on a linux redhat 5 machine. At least, it'll reproduce it on my machine/setup. -Jonathan *********** REPLY SEPARATOR *********** On 7/9/99 at 8:53 AM Andrey Zmievski wrote: >Jonathan Roy wrote: >> But no matter. preg with non-greedy matching will do it, if I can ever >> make preg work. In my stand alone test program, the preg regexp "/\//" is >> fine and works. In the actual program I'm trying to replace an egreg with >> it, it tells me "illegal escape sequence \/". I'm about to give up and let >> someone else fix this Phorum bug. :) > >Can you give me the exact regexp you're having problems with? > >-Andrey >* The future has arrived, it's just not evenly distributed. -W. Gibson * --- Jonathan Roy - roy@idle.com - Idle Communications, Inc.