Re: Support for unexec:ing in libguile (Re: Code compiling (marshalling?))

[Mikael Djurfeldt <mdj at mdj dot nada dot kth dot se>]

Miroslav Silovic <silovic@zesoi.fer.hr> writes:

> Mikael Djurfeldt <mdj@mdj.nada.kth.se> writes:
> 
> > What about adding easy-to-use unexec support to libguile?
> > 
> > Can we use your old unexec code?
> 
> If I recall correctly, the general gripe about unexec-like boots was
> that unexec doesn't reinit system resources (like sockets).

But one should not create sockets before unexec:ing.

> Perhaps an implementation that uses GOOPS to call revive-after-unexec
> generic on all the live objects?

I think such a solution would work for some cases, but I also think
that you're in for trouble when you try to get such a think to work
for more than those special cases.

In any case, the fact that it's difficult to dump at an arbitrary time
is no argument against unexec:ing in general.

Normally, you dump at a controlled place just after initial loading.

BTW, is it possible/a good idea to port SCWM:s unexec code into
libguile proper?