Text Parasing

[guest]

I am trying to make a script that would be able to read only the parts i specify from a list of like 300 rows

here is an example of one of the rows in the text file

<VirtualHost 000.1.000.000>
ServerAlias www.google.com google.com
ServerAdmin [e-mail]
DocumentRoot /home/google/
BytesLog domlogs/google.com-bytes_log
<IfModule mod_php4.c>
php_admin_value open_basedir "/home/fontmob/:/usr/lib/php:/usr/local/lib/php:/tmp"
</IfModule>
ServerName www.google.com
ScriptAlias /cgi-bin/ /home/gogole/public_html/cgi-bin/
User google
Group google
</VirtualHost>

would it be possible for the php to grab "ServerName" and put that in a url so i could just click it and put "DocumentRoot" just as plain text, i have a text file that has like 300 of these and i dont need all the other info so if you could help me out it would be awsome!

[bradgrafelman]

I'm sure there's an easier, possibly faster way to use Perl's regular expression syntax to achieve this, but I would do this as such:

$start = '<VirtualHost 000.1.000.000>';
$start_pos = strpos($file_contents, $start);
if($start_pos === FALSE) die('VirtualHost not found in file.');

$end = '</VirtualHost>';
$end_pos = strpos($file_contents, $end, $start_pos);
if($end_pos === FALSE) die('File incorrectly formatted; VirtualHost tag is not terminated.');

$config_data = substr($file_contents, $start_pos,($end_pos - $start_pos));
if($config_data == "") die('Error parsing file.');

$document_root_keyword = 'DocumentRoot ';
$document_root_start = strpos($config_data, $document_root_keyword);
$document_root_end = strpos($config_data, "\n", $document_root_start);
$document_root = substr($config_data, $document_root_start + strlen($document_root_keyword), ($document_root_end - $document_root_start - strlen($document_root_keyword)));

$server_name_keyword = 'ServerName ';
$server_name_start = strpos($config_data, $server_name_keyword);
$server_name_end = strpos($config_data, "\n", $server_name_start);
$server_name = substr($config_data, $server_name_start + strlen($server_name_keyword), ($server_name_end - $server_name_start - strlen($server_name_keyword)));

echo "doc root is: " . $document_root;

assuming you've read the file into a variable called $file_contents (by using [man]fread/man, etc.).

This is (and even looks it) overkill; I'm sure regular expression syntax guru's could simplify this for you. Perhaps I could even write some POSIX regular expression, but at the moment I'm very tired, and wanted at least to give you one solution to think about, if no one else supplies you with one.

[wilku]

I am new to regular expressions, but to learn something I tried to solve this problem. This is what I came to:

<?php
$input = '
<VirtualHost 000.1.000.000>
ServerAlias [url]www.google.com[/url] google.com
ServerAdmin [email][e-mail][/email]
DocumentRoot /home/google/
BytesLog domlogs/google.com-bytes_log
<IfModule mod_php4.c>
php_admin_value open_basedir "/home/fontmob/:/usr/lib/php:/usr/local/lib/php:/tmp"
</IfModule>
ServerName [url]www.google.com[/url]
ScriptAlias /cgi-bin/ /home/gogole/public_html/cgi-bin/
User google
Group google
</VirtualHost>';

// make sure special chars are escaped in the start and end strings (  .?*\()  )
$DocumentRoot = FindTextBetween("DocumentRoot","\nBytesLog",$input);
$ServerName = FindTextBetween("ServerName","\nScriptAlias",$input);

echo "$ServerName - $DocumentRoot";

function FindTextBetween($start_phrase,$end_phrase,$input)
{
$regex = "/$start_phrase(.*)$end_phrase/";
preg_match($regex,$input,$matches);
return trim($matches[1]);
}
?>

There are some limitations though:
- it will work only if start and end phrase exist in string only once
- for some reason "\n" is required before end_phrase
So my questions are:
- how to look for more than one occurence in a string of the same pair of start-end (using regular expressions)?
- how to make regex look for the smallest string matching?