[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Scheme-reports] Two issues regarding I/O

Vincent Manis scripsit:

> One additional editorial point about eof objects. This is, I think,
> the only type mentioned in the Report that is not listed in section
> 3.2. I would suggest adding at the bottom, wording such as `In addition,
> there is an implementation-defined set of end-of-file objects, which
> satisfy the predicate eof-object?. This type is not required to be
> disjoint from any other type, provided that the constraints specified
> by the input procedures that may return such objects are met.'

As noted in ticket #333, an eof-object already cannot be a boolean,
pair, symbol, number, character, string, vector, or bytevector, because
these objects have external representations and an eof-object MUST NOT.
It is theoretically possible for it to be a procedure or a port, but
none of the 40 Schemes actually exploits this possibility.  The ticket
therefore proposes that it be made formally disjoint.

Is not a patron, my Lord [Chesterfield],        John Cowan
one who looks with unconcern on a man           http://www.ccil.org/~cowan
struggling for life in the water, and when      cowan@x
he has reached ground encumbers him with help?
        --Samuel Johnson

Scheme-reports mailing list