When commit messages contain umlauts (ä,ö,ü,ß,...) the tags in the tag cloud are split at these points into two or more single tags.
Example: Log message "ausgängen" -> Tag "ausg" + Tag "ngen"
In other parts of StatSVN the umlauts are handled correctly, i.e. in section "Commit Logs".
I observed that behaviour as well and posted it in the user-forum. I want to tell my further findings here:
- The switch -charset=utf-8 is not working. It only switches the character-set of the html
- In MS Windows batch-script, setting the LC_MESSAGE environment variable (SET LC_MESSAGES=de_DE) doesn't work either.