About character encoding (28May07 Markus)

CompletionSearch supports ISO-8859-1 and the multibyte character encoding UTF-8. UTF-8 is the default encoding with the following consequences:

The $AC->settings->encoding is 'utf-8' unless overriden in autocomplete_config.php
The texts in text.php are saved as UTF-8
The css file uses '@charset "utf-8";'
We use mb_strtolower (instead of strtolower) with parameter $AC->settings->encoding to enable UTF-8

We do the following depending on the defined encoding:

We UTF-8 encode $AC->settings->capitals if $AC->settings->encoding is UTF-8
In ajax.php we UTF-8 encode the query string if $AC->settings->encoding is UTF-8 and the charset of content_type is not UTF-8 (means the request is sent as a non-UTF-8 type)
We set the page encoding of index.php, options.php and change_options.php according to $AC->settings->encoding (<meta http-equiv="content-type" content="text/html;charset=<?php echo $AC->settings->encoding; ?>">

)

Texts from text.php are UTF-8 decoded by $AC->get_text() if $AC->settings->encoding is ISO-8859-1
We url encode the javascript code in function javascript_rhs (in generate_javascript.php) if $AC->settings->encoding is not UTF-8 (this is not necessary if utf-8 is used)

The PHP Apache extension php_mbstring

The use of the mb_strtolower function (and other mb_ functions) requires the extension php_mbstring in php.ini:

In windows:
extension=php_mbstring.dll

or in linux:
extension=php_mbstring.so

(On geek, the mb_... functions were available by default, on Markus' laptop the line above had to be added.)