[cvsnt] Problems managing unicode files

[Tony Hoyle]

On Fri, 12 Sep 2003 09:42:59 +0200, Daniele_Cortecchia at sacmi.it wrote:

>I'm experiencing some problems with cvsnt when i'm trying to add some 
>unicode text files (starting bytes: FF FE, chinese language) in the 
>repository.
>First I added the files as unicode but there were a character in some 
>files that caused cvsnt server to store uncorrectly the file: when I tried 
>to retrieve the file via update it was displayed uncorrectly after the 
>occurrence of that character (probably because  of that character is 
>formed by three bytes I think).

If you have any characters formed by three bytes then this is not a
unicode file - unicode is always two bytes.

>Later I tried to add files as binary. When I made some modification  and 
>committed them the server gave me,  after an update command, the error 
>"pos > nvec->lines" or sometimes files appeared to be damaged.

How are you accessing the repository?

Tony