As an international company, we have similar problem.

Currently, we solve this problem pretty easy with VIM and iconv library. In vim, we open an encoding file and issue 3 commands:

:e ++enc=original_encoding
:file new_encoding_name.php
:w ++enc=utf-8

The first command loads file in right encoding, the second one assures that we won't overwrite original file, then we change encoding name in the text. Finally, the third command saves file in UTF-8.

The problem with this is that we have to regenerate encodings every time new mantis version appears. So separating translation from encodings could really be a good feature.

I have a suggestion to the feature - translation file could contain "default" encoding, which could be used by default if admin won't set some global encoding used by mantis.

For translating from one encoding to another, maybe this could help:

http://www.gnu.org/software/libiconv/

Thank you all for your comments. Actually, I used iconv from the Linux command line to convert the lang/strings_english file to utf-8. But, since ascii overlaps with utf-8 the first 256 bytes, it probably wasn't necessary.

PHP has native support for the Linux iconv library, so long as you compile it in when you compile PHP. This is what my company uses to serve content in different character sets. What it means is you don't have to store multiple files on your server that have the same content but different character encoding.

This change would be relatively easy to make in Mantis...if only I had the time...

This problem could be fixed with utf8-ization of Mantis.
For now, no conversion of strings_english is required, but you could just change $s_charset = 'windows-1252'; to $s_charset = 'utf-8';.