Re: tools to make non-conservative GC feasible.

[Greg Harvey <Greg.Harvey@thezone.net>]

Chris.Bitmead@misys.com.au writes:

> >I wouldn't trust it in cases where someone could end up dead if
> >something goes wrong (and I don't trust nuclear arsenals to anything),
> >but I do trust it for guile, which isn't really targetted towards life
> >and death situations.
> 
> Well, the thing is Guile's aim seems to be that it shall become ubiquitous.
> I mean, the aim of guile is that it shall become a part of nearly every
> GNU program as a standard extension language. And GNU's aim, you might
> argue might be said to be Linus's "World Domination". The upshot is that
> guile could one day be in places that you wouldn't dream of. No, I'm not
> really worried, but it does give pause for thought.

Pause to think about the stupidity of the person who'd use a
non-realtime, conservative gc using scheme extension language in a
place where it could cause someone to get killed, maybe; but if that's
the case, we're screwed anyway, conservative gc or no. The possibility
won't keep me awake at nights.

In any case, there's not much more to say about it. We know the
benefits and drawbacks; I don't think it's a problem, and I don't feel
any particular motivation to take on the task of changing guile to use
explicit root marking. If you want to give it a shot, by all means go
for it (jeez, you could probably get that done faster than the gengc
;); but we're really just wasting our time reiterating the maybe,
maybe nots of conservative scanning.

-- 
Greg