[R6RS] Changing the transcoding mid-stream

[dyb at cs.indiana.edu]

>From Mike:

> From Kent:
> > Good points.  This raises a question: If a call to lookahead-u8 returns
> > eof, and the next operation on the port is a call to get-u8 or
> > lookahead-u8, can the second call return something other than eof if the
> > underlying file (or whatever) has been extended in the meantime?
>
> No.  The EOF is part of the data stream, or at least that was the
> intention.

Also from Mike:

> From Kent:
> > Here's a related question.  I presume that lookahead-char will raise an
> > exception if end-of-file is reached after some but before before enough
> > bytes have been read to make a whole character.  If this exception is
> > caught, the application continues, the underlying file is extended
> > independently with enough bytes to fill out the character, and get-u8 is
> > called again, must get-char also fail with the same exception?
>
> I presume the `get-u8' is really `get-char' and that its invocation is
> the same as the invocation of `get-char'.  In that case, yes, by same
> reasoning about the nature of EOFs as part of the data stream.

This assumes a different model from the one I had in mind, which is that
a port maintains a (probably double) buffer of bytes.  If the buffer is
empty, lookahead-whatever and get-whatever have to make a call into the
O/S to fill the buffer.  If the buffer has characters, but not enough to
read an entire char, lookahead-char and get-char again have to make a call
into the O/S to fill the buffer.  In the former case, the O/S may report
EOF, in which case the operation returns the eof object.  In the latter
case, the O/S may also report EOF, in which case the opertion raises an
exception.  In either case, if the underlying file (or other object) is
extended so that the O/S no longer reports EOF, the next call to either
procedure may complete normally.

I understand you have a model based on infinite streams, but I think my
model should be allowed as well, so I suggest that we not overspecify the
semantics on this edge case.  I can't imagine this will affect many
programs.  Most programs will simply quit trying to read from the port
when the lookahead procedure returns EOF or raise an exception.

Kent