Nadya Morozova said the following on 31.01.2008 19:42:
!ALERT! Help needed!I got the latest version of the site, all looks nice with the new encoding. However, if I edit any Russian page and build, Anakia tampers with all Russian files and the encoding somehow changes (offset of certain symbols). I use Windows, VS 2005 as editor, and it recognizes the encoding, so it's not the problem with the editor. the file velocity.properties is up to date and specifies UTF-8 as output and input encoding. On my system, only 2 letters in Cyrillic are not converted properly ("я" "с") in both cases. Can anybody reproduce this strange thing? Any Anakia gurus to help me out?
It was me who switched the site to Unicode and on Linux I had no such problems. Now that I tried to process XML documents on Windows I get some similar problems with Cyrillic letter 'И', it is replaced with D0 3F UTF-8 sequence instead of correct D0 98. I tried setting file.encoding property to UTF-8 (as it is set for me on Linux by default) but it didn't help.
A funny thing is that when I set JAVA_HOME to point to Harmony, I get correct output, although Anakia indeed processes all of the Cyrillic pages, the result has no difference with the currently stored in svn.
-- Gregory
