2011-11-30 20:52 AEST by Arthur Barrett - I suspect that the SJIS client did not translate 'log messages' to Unicode, so a utility that parses the RCS 
file and rewrites all the log messages in Unicode is required.  

RCS files from unix would usually be in unicode anyway, but if the unix codepage was not set to UTF8 
then this same feature could be used for CVS 1.x conversions too.

Basically the 'standalone migrate' would need to detect the current codepage, version of CVS etc and 
maybe sneak a look at the RCS files and if so needs to put a flag in the 'package' that the RCS files need 
converting (or convert them 'in place' if it is safe).

Comments from user:


I tried many options but all Japanese characters (in the comment and log) in converted CVSNT Suite 
repository are garbled characters.