Should I pull the plug again? John Cowan (30 Jun 2020 22:39 UTC)
Re: Should I pull the plug again? Marc Nieper-Wißkirchen (01 Jul 2020 06:34 UTC)
Re: Should I pull the plug again? John Cowan (06 Jul 2020 03:56 UTC)
Re: Should I pull the plug again? Arthur A. Gleckler (06 Jul 2020 04:59 UTC)
Re: Should I pull the plug again? Marc Nieper-Wißkirchen (06 Jul 2020 20:34 UTC)
Re: Should I pull the plug again? John Cowan (06 Jul 2020 21:29 UTC)
Re: Should I pull the plug again? Shiro Kawai (07 Jul 2020 00:52 UTC)
Re: Should I pull the plug again? John Cowan (07 Jul 2020 04:12 UTC)
Re: Should I pull the plug again? Shiro Kawai (07 Jul 2020 04:16 UTC)
Re: Should I pull the plug again? Shiro Kawai (06 Jul 2020 04:37 UTC)
(missing)
(missing)
(missing)
Re: Should I pull the plug again? John Cowan (06 Jul 2020 20:10 UTC)
Re: Should I pull the plug again? Shiro Kawai (06 Jul 2020 20:18 UTC)

Re: Should I pull the plug again? Marc Nieper-Wißkirchen 06 Jul 2020 20:34 UTC

Am Mo., 6. Juli 2020 um 05:56 Uhr schrieb John Cowan <xxxxxx@ccil.org>:

> Well, if the system has weak hash tables, which are not yet standardized, though ephemerons, the main building block for them, are standardized.

By the argument given in SRFI 124 (GC is not observable), weak hash
tables can be portably implemented. :)

The nice thing about using weak hash tables to attack properties to
procedures (which are called closures in other languages) is that it
doesn't slow down the actual procedure call. The idea of using a
special argument has this problem because some dispatch has to happen.
And the weak hash table idea will work with every procedure.