Re: [Ethereal-dev] Unicode strings ...

[Richard Sharpe <sharpe@xxxxxxxxxx>]

Guy Harris wrote:
> On Sun, Aug 12, 2001 at 05:12:34PM -0700, Guy Harris wrote:
>
> > My inclination would be to go with UTF-8 internally,

UTF-8 has a number of advantages, and we can use iconv for conversion, 
but that requires another library :-(

> ...although that would require either that "proto_tree_add_string()" 
> and so on take an argument specifying the character set to be used or 
> that some indication of the character set be associated with the 
> field.
>
> The latter has the disadvantage that not all occurrences of the field
> would necessarily have the same character set, so I'd be inclined to 
> add a character-set argument.
>
> "proto_tree_add_item()" might also have to have that argument added 
> (the byte-order argument couldn't, I think, be used, as if the 
> character set in the packet is UCS-2, you'd have to specify whether 
> it's big-endian or little-endian UCS-2).

A problem I have just encountered is that proto_tree_add_string etc all 
expect the string you are going to add to be the "value" argument. What 
I want is a routine, like proto_tree_add_ucs2_xx, that will take a 
string of length X from the TVB at offset, and mark its extent in the 
byte view, but then convert that string into internal format (UTF-8) for 
display. I would also want to do formatting as well, which means more 
work ...

--
Richard Sharpe, rsharpe@xxxxxxxxxx, LPIC-1
www.samba.org, www.ethereal.com, SAMS Teach Yourself Samba
in 24 Hours, Special Edition, Using Samba