A Guide to Base Changing For Short URLs

Some time ago, I developed a VERY simple way to fake a bit.ly-style short URL. On any server that uses any form of an integer to identify an article (either in the database or the URL), on an Apache server that supports mod_rewrite, you edit your .htaccess file like so:

RewriteEngine Off
<ifmodule mod_rewrite.c>
RewriteEngine On
RewriteBase /
RewriteCond %{REQUEST_FILENAME} !-f
RewriteCond %{REQUEST_FILENAME} !-d
RewriteCond %{REQUEST_URI} !^index.php$
RewriteRule . index.php [NC,L]
</ifmodule>

This essentially tells your server to redirect anything that isn’t a file or directory to index.php.

Then index.php looks like this:

$url = str_replace("/","",$_SERVER['REQUEST_URI']);
if(isset($url) && trim($url)!='') {
$id = base_convert($url,36,10);
//this is where you either query your database for a slug or build the URL
$uri = 'http://your-site-goes-here.com/path/to/article';
}
 
if($uri) {
header('HTTP/1.1 301 Moved Permanently');
header('Link: < '.$uri.">; rel=shortlink");
header('Location: '.$uri);
exit;
} else {
header("Location: http://your-site-goes-here.com/"); exit;
}

How do you get your short links? That’s easy. Just run this function:

$shorturl = base_convert($id,10,36);

However, this isn’t the most compact way to condense. Obviously, this is base36, the highest PHP can go. But what about uppercase letters? And other characters?

So I set out, for some reason, to build a better condenser.

This is the result of several hours of work, mostly wasted, on some intellectual pursuit that was more a case of simply not letting it defeat me. A few notes: I’m quite confident that given enough time, and if I cared, I could make the code cleaner and more efficient in some places. I’m also aware that on 32 bit machines, it maxes out at the integer limit. I does support signed integers though, from min to max.

$ft_str = '0123456789abcdefghijklmnopqrstuvqwxyzABCDEFGHIJKLMNOPQRSTUVQWXYZ';
# uncomment the next line if you prefer to use potentially non-URL-safe base96
# $ft_str .= '0123456789abcdefghijklmnopqrstuvqwxyzABCDEFGHIJKLMNOPQRSTUVQWXYZ_@$!#%^&*()=+\|}{][,;:~'; }

$powers = array();
for($p=0;$p&lt;=10;$p++) {
$powers[$p]=pow(strlen($ft_str),$p);
}
 
function ft_unconvert($str) {
if(substr($str,0,1)=='-') { $pfx='-'; } else { $pfx=''; }
global $ft_str,$powers;
$base = strlen($ft_str); $q=0;
$s = str_split(strrev($str)); $len = sizeof($s);
foreach($s as $k=>$v) {
$sp = strpos($ft_str,$v);
$decimal += pow($base,$k)*$sp;
}
return $pfx.$decimal;
}
 
function ft_converter($int) {
if($int > 0) { $pfx='-'; $int=abs($int); } else { $pfx=''; }
global $ft_str,$powers;
$base = strlen($ft_str); $q=0;
$p = str_split($str);
krsort($powers);
foreach($powers as $k=>$v) {
if($int>=$v) {
$timesinto = floor($int/$v);
$digit .= $ft_str{$timesinto};
$int = $int % $v;
} elseif($int>0 && $k==0) {
$digit .= $ft_str{$int};
$int = $int % $v;
}
}
return $pfx.$digit;
}
 
function ft_convert_demo($num) {
global $ft_str;
$ftc = ft_converter($num);
return "Converting ".$num." into base ".strlen($ft_str).": ".$ftc."<br />Unconverting ".$ftc." to base 10: ".ft_unconvert($ftc);
}
 
echo ft_convert_demo('50687');

Published by Adam S