Re: SRFI for ports from arbitrary sources or to arbitrary sinks? Lassi Kortela (19 Sep 2019 12:07 UTC)
Re: SRFI for ports from arbitrary sources or to arbitrary sinks? Alaric Snell-Pym (24 Sep 2019 08:38 UTC)

Re: SRFI for ports from arbitrary sources or to arbitrary sinks? Lassi Kortela 19 Sep 2019 12:06 UTC

> At the beginning of John Cowan's Simple SQL Blobs section, he notes:
>
>> ISSUE: Should this be gotten rid of? It's not strictly necessary, but handling really large blobs without it will be messy. Ideally it should give us ports, but adding ports to a Scheme implementation is hard.
>
> And that's far from the first time it's been noted it would be very nice if we could specify the creation of a port in an API.
>
> While I can't speak to the difficultly of creating ports with arbitrary sources or sinks in the various Scheme implementations out there, not even all that much for Chibi Scheme where I did "monkey see, monkey do" generic port creation using a supplied file descriptor for my SRFI 170 sample implementation, it at all practical, if likely to be adopted by enough Schemes, it would be very nice to have this capability.
>
> "Arbitrary" in that I suppose you'd hand the port creation procedures a procedure to supply or consume text or binary data.

I also need this for the archive SRFIs.

R6RS has custom ports in the standard. Common Lisp has Gray streams as a
popular extension to the standard. Bytevector generators and
accumulators (SRFI 158) are a poor man's approximation.