Re: Regarding check-in of unicode characters.

[Rahul]

Hi Maninder -

You can use UTF-8 encoded files.

UTF-8 is a method of representing Unicode text with a stream of
8-bit bytes. The resulting stream is both ASCII-compatible and
reverse-ASCII-compatible. A single character can occupy from 1 to 4
bytes.

Most IDE and editors like emacs these days, will let you switch the
default charset encoding from ASCII to UTF-8.

More information on Unicode and UTF-8 is here -
http://www.unicode.org/faq/utf_bom.html#14

A UTF-8 encoded source file with text in UTF-8 can be
handled by CVS (on Linux etc)  without any problems. Most IDEs and
Editors supporting CVS can be configured for UTF-8 encoding by default
to get Unicode support in CVS to work seamlessly.


With regards to CVSNT, the big difference is CVSNT supports a -ku
keyword
flag to let you have the file in UTF-16 be automatically be converted
to
UTF-8.

Internally CVSNT also stores the files in UTF-8. Since most IDE,
editors and tools will do UTF-8 encoding, this may be good enough.

For command-line cvs usage, tools like iconv (on UNIX, Cygwin)
can be used to go from any format to any other.  So just use UTF-8
and you should be home. You could have a script do iconv before and
after any ci/co to make it seamless.

Rahul Bhargava,
CTO, WANdisco
Mountain View, CA
http://www.wandisco.com/cvs



Arthur Barrett wrote:
> Maninder,
>
> CVS does not support Unicode files, instead they must be committed as binary, 
> in which case you cannot do diff/merge etc.
>
> CVSNT on Mac/Linux/Unix/Windows (free, open source, GPL, just like CVS) does 
> support unicode files (-ku).
>
> Please supply more information (file type in repository, cvs version etc).
>
> To determine the "status" of the file use:
> cvs status myfile
>
> To determine your version of CVS use:
> cvs version
>
> If you are using CVSNT then the CVSNT newsgroup is the place that the 
> developers of CVSNT answer questions:
>
> news://news.cvsnt.org/support.cvsnt <news://news.cvsnt.org/support.cvsnt>
>
> or
>
> http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt 
> <http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt>
>
>
> Regards,
>
>
> Arthur Barrett
>
>
>       -----Original Message-----
>       From: address@hidden [mailto:address@hidden On Behalf Of Maninder 
> Singh(SDG)
>       Sent: Thursday, 16 June 2005 1:51 AM
>       To: address@hidden
>       Subject: Regarding check-in of unicode characters.
>
>
>
>       Hi All,
>
>                   We are trying to check-in the following characters into CVS.
>
>       But while doing so, the highlighted characters get corrupt and upon 
> checking out, these are replaced by some other characters.
>
>       Sí (means yes in Spanish)
>
>       Sì (means yes in Italian)
>
>
>
>       Can anybody provide a solution?
> 
>        
> 
>       Thanks and Regards,
> 
>       Maninder Singh