Using preg_replace to convert '<?php*?>' to nothing (removing all PHP code)

[bissatch]

Hi,

I am trying to use preg_replace to take out all occurances of PHP code
after reading (fread()) the contents of a PHP file. The code I am using
is:

$html = preg_replace("< ?php*?>", "", $html);

Unfortuntely it is not working and giving the following error:
Warning: preg_replace(): Nothing to repeat for offset 0 in
C:/apache/...
Can someone please try this out and see what they get. How can I get
this to work?

Cheers

Burnsy

[Andy Hassall]

On 2 Sep 2005 13:57:49 -0700, bi******@yahoo. co.uk wrote:

I am trying to use preg_replace to take out all occurances of PHP code
after reading (fread()) the contents of a PHP file. The code I am using
is:

$html = preg_replace("< ?php*?>", "", $html);

OK, what do you think that expression should match, and why?

--
Andy Hassall :: an**@andyh.co.u k :: http://www.andyh.co.uk
http://www.andyhsoftware.co.uk/space :: disk and FTP usage analysis tool

[IWP506]

Do you get the same error if you switch out <?php*?> with something
simple, as a test?

[bissatch]

> $html = preg_replace("< ?php*?>", "", $html);

To my basic knowledge, that expression should match a string with
'<?php' at the start, and '?>' at the end with any collection of
characters inbetween (represented by the asterix - * - character).

Do you get the same error if you switch out <?php*?> with something

simple, as a test?

I did try and switch the '<?php*?>' part with just something like
'<h1>' and it seemed to work.

[Chung Leong]

First, PCRE is interpreting the brackets as brackets for the
expression. Second, the question mark is a special character in RegEx
thus needs to be escaped. Third, to match any character zero or many
times you need ".*" and not just the asterisk.

The correct expression is "/<\?php.*?\?>/s". The s modifier is
necessary for matching across lines.

If this is intended as a security check of sort, then you need a lot
more code than a call to preg_replace().

[Gordon Burditt]

>I am trying to use preg_replace to take out all occurances of PHP code

after reading (fread()) the contents of a PHP file. The code I am using
is:

$html = preg_replace("< ?php*?>", "", $html);

What characters when they are intended to be literal have to be
backslashed? This at least includes ? and *. Not sure about <>.

What does the * character mean in a regular expression? 0 or more
repetitions of what preceeds it. This is *NOT* the same as a
filename wildcard. p* matches a whole bunch of p's. As a filename
wildcard, it matches something that begins with p followed by
something else.

Offhand I'm having trouble finding documentation on the particular
flavor of regular expressions PHP uses.

Gordon L. Burditt

[Andy Hassall]

On 2 Sep 2005 14:46:41 -0700, bi******@yahoo. co.uk wrote:

$html = preg_replace("< ?php*?>", "", $html);
To my basic knowledge, that expression should match a string with
'<?php' at the start, and '?>' at the end with any collection of
characters inbetween (represented by the asterix - * - character).

It doesn't match that, it's an invalid pattern. Have you read:

http://uk2.php.net/manual/en/ref.pcre.php
http://uk2.php.net/manual/en/referen...ern.syntax.php
http://uk2.php.net/manual/en/referen....modifiers.php

Even if you'd surrounded it with delimiters, e.g. you'd used

/<?php*?>/

that would match (courtesy of the YAPE::Regex::Ex plain Perl module):

----------------------------------------------------------------------
<? '<' (optional (matching the most amount
possible))
----------------------------------------------------------------------
ph 'ph'
----------------------------------------------------------------------
p*? 'p' (0 or more times (matching the least
amount possible))
---------------------------------------------------------------------- '>'

----------------------------------------------------------------------

Do you get the same error if you switch out <?php*?> with something

simple, as a test?

I did try and switch the '<?php*?>' part with just something like
'<h1>' and it seemed to work.

But this doesn't match "<h1>" which it appears you might be expecting, it
matches "h1". Read http://uk2.php.net/manual/en/ref.pcre.php for why.

You may want to look at http://www.weitz.de/regex-coach/ which I've seen
mentioned a few times, and looks pretty useful for learning regexes from the
description and screenshot.

--
Andy Hassall :: an**@andyh.co.u k :: http://www.andyh.co.uk
http://www.andyhsoftware.co.uk/space :: disk and FTP usage analysis tool

[John Dunlop]

bi******@yahoo. co.uk wrote:

I am trying to use preg_replace to take out all occurances of PHP code
after reading (fread()) the contents of a PHP file. The code I am using
is:

$html = preg_replace("< ?php*?>", "", $html);

<?php

// ?>

?>

You're fired.

--
Jock

[John Dunlop]

John Dunlop wrote:

<?php

// ?>

?>

Make that

<?php

/* ?> */

?>

You're still fired.

--
Jock

[Chung Leong]

The block would be removed correctly if the greedy matching isn't
turned off.