too low-level Alex Shinn (25 May 2020 00:47 UTC)
Re: too low-level Amirouche Boubekki (28 May 2020 07:05 UTC)
Re: too low-level Alex Shinn (28 May 2020 08:27 UTC)
Re: too low-level Alex Shinn (28 May 2020 09:10 UTC)
Re: too low-level Amirouche Boubekki (03 Jun 2020 06:01 UTC)
Re: too low-level Alex Shinn (04 Jun 2020 08:17 UTC)
Re: too low-level Marc Nieper-Wißkirchen (04 Jun 2020 08:53 UTC)
Re: too low-level Alex Shinn (04 Jun 2020 10:12 UTC)
Re: too low-level Marc Nieper-Wißkirchen (04 Jun 2020 10:28 UTC)
Re: too low-level John Cowan (04 Jun 2020 15:59 UTC)
Re: too low-level Marc Nieper-Wißkirchen (04 Jun 2020 16:29 UTC)
Re: too low-level John Cowan (04 Jun 2020 19:54 UTC)
Re: too low-level Marc Nieper-Wißkirchen (04 Jun 2020 20:55 UTC)
Re: too low-level Alex Shinn (04 Jun 2020 22:50 UTC)
Re: too low-level Arthur A. Gleckler (04 Jun 2020 23:32 UTC)
Re: too low-level John Cowan (05 Jun 2020 00:17 UTC)
Re: too low-level Alex Shinn (05 Jun 2020 04:07 UTC)
Re: too low-level John Cowan (05 Jun 2020 17:46 UTC)

Re: too low-level Marc Nieper-Wißkirchen 04 Jun 2020 20:54 UTC

Am Do., 4. Juni 2020 um 21:53 Uhr schrieb John Cowan <xxxxxx@ccil.org>:

> I don't know what would count as a truly universal solution, except maybe an arena-based allocator.  But I do have the same facilities for my data interchange pre-SRFIs, which provide for both S-expression and ASN.1 binary formats.  See <https://bitbucket.org/cowan/r7rs-wg1-infra/src/default/CoreSexps.md>.  (The term "Core" has no particular meaning any more; if anyone can think of a better term, let me know.)

When it comes to binary formats for data exchange, I would also like
to see canonical s-expressions ([1]), especially for their simplicity.

As for a truly universal solution: There is a technical side (maybe
complicated), but there is also the API to this side. We could now at
least think of an API that would work for JSON as for the READ
procedure, but can also deal with gzip-versions of these formats, etc.
(to protect against gzip-bombs). One way to handle this could be to
start a new process for data whose total binary size is higher than a
certain limit (of course, for this we need some knowledge about the
dependency of the complexity of the problem on the binary size).

--

[1] https://en.wikipedia.org/wiki/Canonical_S-expressions