From: John Cowan <>
Date: Saturday, July 20, 2019 3:55 PM

On Sat, Jul 20, 2019 at 11:11 AM <> wrote:
(signal-process proc sig)     →     undefined         (procedure)         (POSIX signal(2), kill(1))

Works for me.

Hmmm, in the interests of people with narrower screens, the POSIX should be dropped in favor of the necessary explanatory introduction to this at the beginning, but probably not the enclosing parens, to match the (procedure) and signal this is advisory information.  Probably ought to decrease the number of spaces before it as well.

I deeply appreciate all the editorial work you have done and are doing, by the way.

Thanks!  I'm glad to find areas where the fact that my C system programming experience has overtaken my days of LISP in the 1980s allows me to contribute right now.
[ I forgot about the issues of sharing record types. ]

For the specific Chibi problem, I'd just write a bit of hand-rolled C to call clock_gettime....

Beyond my current and planned anytime soon capabilities, the automagic C FFI is very nice, but I'll take a look after I get lots else done.  And I have to have something above because clock_gettime can error, for example if you give it a bogus clock.  The monotonic one ought to be in any system that supports the system call, the obvious and very easy implementation is uptime in seconds, but you never know....

I just noticed that spawn* use both plists and alists....

The only reason the environment is an alist is `get-environment-variables` in R7RS, in turn derived from SRFI 98.

That could be an argument for replacing the use of plists with alists.

- Harold