Posted by yawnmoth on 11/11/08 11:43

yawnmoth wrote:
> Erwin Moller wrote:
> <snip>
>
> Thanks for the suggestion - that helped!
>
> I tried modifyng it such that it'd delete all lines not begining with a
> b but am once again having some difficulty...
>
> First, here's the (new) script:
>
> <?
> $contents = file_get_contents('dummy.txt');
> echo "\r\n";
> echo preg_replace('/^(?!b).*?\n/m','',$contents);
> ?>
>
> The (?!b) should, as I understand it, match everything but b. What
> instead seems to be happening is that, when using the dummy.txt file
> that was posted earlier, all lines begining with a are deleted, but the
> lines with c are still there.
>
> Now, I realize that in this example, I could just as effectively
> replace (?!b) with [^b] but this approach fails if b is anything other
> than a single character. (eg. if I wanted to replace b with
> 192.168.1.1, I couldn't use [^...] - I'd have to use something like
> (?!b))

Never mind - I got it working:

<?
$contents = file_get_contents('dummy.txt');
echo "\r\n";
echo preg_replace('/^(?!b).*?(?:\n|$)/m','',$contents);
?>

The problem was due to the fact that the last line - which in dummy.txt
didn't start with a b - didn't end with a new line character but rather
with a EOF character (or maybe not a character at all? is EOF actually
a character?)

[Back to original message]