This is the mail archive of the
guile@cygnus.com
mailing list for the Guile project.
Re: Upcoming 1.3.2 release
Greg Harvey <Greg.Harvey@thezone.net> writes:
> Greg Badros <gjb@cs.washington.edu> writes:
>
>
> > Right-- I'm aware that it'd be a bad idea to call Scheme code from a
> > GC-start hook. We want Scwm to be able to have a hard-coded (in C)
> > behaviour (or perhaps several possible behaviours) to convey to the user
> > that garbage collection is taking place. E.g.,
>
> The problem isn't so much that it could be problematic to call scheme
> code, but pretty much anything in guile is suspect (except scm_cons &
> newcell, because we know that'll screw things up). A set of predefined
> behaviours could work, but you'd want to watch out for something like
> the second, since you can't reliably call things like display & stuff
> like that... you could do writes on the underlying fd, if you don't
> have to worry about buffering (or don't care about slightly odd
> output), though. I'll add a bit to allow registering c functions to
> my gc code, but I do think it'd be a really bad idea to add a general
> hook for this.
Agreed, just a single C function.
(Good point re: writing to a port).
Thanks,
Greg B