Recently, several tools at Toolserver have not been functioning properly, affected by text encoding problems.
Some changes in the PHP configuration may solve this.
The tools below are reported to not function when they receive non-latin texts.
Others tagged with 'UTF-8' https://jira.toolserver.org/secure/IssueNavigator.jspa?sorter/field=updated&sorter/order=DESC
They seem to be affected by the same text encoding problem.
A common behaviour is that the first part of the non-latin text appear as unreadable characters, while the rest is kept as input. (See the title of http://toolserver.org/~vvv/sulutil.php?user=%E5%88%A9%E7%94%A8%E8%80%85 for example, which should be 利用者 but shown as ň�用者)
Since the affected scripts are independent from each other, what's needed to be fixed should not be in the scripts, but in the system level.
I found a possible cause in the PHP configuration of Toolserver. The current configuration (assuming I'm not looking at the wrong one) leaves the value of 'mbstring.detect_order' as 'auto'. The official document says if it's 'auto', PHP prefer to interpret texts as ASCII over JIS and UTF-8. I think UTF-8 should be preferred over ASCII in a multilingual situation like Toolserver.
While it's safe to interpret a ASCII string as UTF-8, the opposite may break the content. I suggest to make PHP prefer UTF-8 over ASCII, by setting 'mbstring.detect_order' to 'UTF-8, ASCII'. I believe this change causes no problem and possibly solves several recent problems across different tools regarding text encoding.