[cvsnt] Question about CVSNT and Unicode files

[Tony Hoyle]

Arthur Barrett wrote:
> Bogdan,
>  
> Upgrade to CVSNT 2.5.03 (stable) or 2.5.04 (testing) documentation on keywords is here:
> http://www.cvsnt.org/manual/html/Substitution-modes.html
>  
> To change file.ext from -ku to to -kucs2le you would do this:
> cvs up -kucs2le file.ext
> cvs ci -f -m "change to ucs2le" file.ext
>  
> To change 'new' files with the extension .txt to be ucs2le by default you need to alter the cvswrappers.
>  
For codepages you need to surround them with {} so:

cvs up -k{ucs2le} file.ext

etc.


You probably don't want to do this though - visual studio understands 
(and in fact requires) a BOM at the start of ucs2 files.. as do nearly 
all unicode compliant editors (there are some odd ones about that don't, 
which is why the option is there).

CVSNT doesn't do anything with the characters - it just encodes the 
binary representation as UTF8 for storage then expands them again on the 
client side when checked out... this conversion is completely lossless.


Tony