On Sun, May 07, 2000 at 05:27:55PM -0700, Guy Harris wrote:
> > This is indeed better. I think the most important ones are:
> > - FT_ASCII ( 7bit )
> > - FT_ISO_8859
> > - FT_UNICODE
> > - FT_UTF8
>
> Do we need to distinguish between FT_ASCII and FT_ISO_8859, given that
> all the 8859 character sets are supersets of ASCII? We might want to
> distinguish between different 8859 character sets, assuming that the
> dissector can figure out which character set it is (one of the 8859
> character sets, some PC code page, one of the Windows extensions of one
> of the 8859 character sets, the Mac character set, etc.).
The character set should not be a "field type" (FT_*). Rather, we should
use the "display" field in the header_field_info struct to differentiate
the encoding. (We might need a different field type for multi-byte
character strings, however).
The frame_data struct records a per-frame character encoding. This could
still be used as a "default" character encoding, to be overriden
for each field if needed.
--gilbert
