One thing to fix: for 'fossil commit' there should be some hook to convert a file after it leaves EDITOR and before it get into Fossil database. EDITOR (a program with command line got from 'EDITOR' environment variable) most often using default 8-bit console charset (for Cyrillic it is 'Windows-1251' under Windows and 'KOI8-R' under Unix), but in the Fossil repo there is 'UTF-8' encoding. When I do commit with a Cyrillic (Russian) text in comment, Fossil put it into db w/o conversion (as Windows-1251), and when displaing it in UTF-8 UI it becomes unreadable. So it should be hook for convertor or some internal option or command line for iconv.

Something like these GIT's options (from .gitconfig):

[i18n]
	commitencoding = windows-1251
	logoutputencoding = cp866

anonymous added on 2009-09-15 18:17:06:
Or just make charset in this line (cgi.c, line 322):

  fprintf(g.httpOut, "Content-Type: %s; charset=utf-8\r\n", zContentType);

anonymous claiming to be Dmitry Chestnykh added on 2009-10-11 20:11:34:
Maybe you don't need my opinion on this, but I think it's time drop old extended ASCII encodings everywhere and stick to UTF-8. Unicode standard has been available since 1991. Every major operating system and many editors support it. If your editor doesn't produce UTF-8 encoded output, pick another editor. Encoding conversion only makes code more complex (there are dozens of non-Unicode encodings) and doesn't add much value in 2009.

Major Linux distributions, Mac OS X come with UTF-8 terminals by default, and Windows has it as an option, AFAIK.