Re: [Chicken-users] BOM in a Scheme source file

[Pierpaolo Bernardi]

On 9/10/07, John Cowan <address@hidden> wrote:
> Pierpaolo Bernardi scripsit:

> > which says that you can put a bom in a utf8 file (of course, you can
> > put whatever character you want in a file), but it is a character
> > like every other character, it has no particular meaning wrt the encoding.
>
> BOMs serve two purposes: in UTF-16 and UTF-32 it specifies the actual
> byte order, but in all UTFs it helps to provide a signature specifing
> the encoding.  As such, when a UTF-8 file begins with U+FEFF, the decoder
> MAY use it to assume UTF-8 input,

Ot it may not, and all will work better.

I hope that no programming language implementation will start *guessing*
things in its parser or in its standard file io library.

> So you
> are right to say that a BOM in a UTF-8 file does not affect the format
> of the encoding, but it can and does affect the overall decoding process.

Random decoding.  :)

P.