It sounds like this can be made safe. It sounds a lot ilke the connection that gracket makes to the underlying OS to do GUI things and that is shared (and safe).
Robby On Fri, Sep 10, 2010 at 10:17 PM, Eli Barzilay <e...@barzilay.org> wrote: > On Sep 10, Robby Findler wrote: >> FWIW, this is one path that tools can use to circumvent DrRacket's >> property (well, the property we work towards anyways) that no program >> can cause DrRacket itself to crash or freeze. So if you do provide >> things to the user's program in this manner, please try to keep that >> invariant in mind. > > Actually, for John's case I think that what he really needs is to be > able to break more than just drracket... There was a plan at some > point to have a single process-global hash table which he'd be able to > use for cases where some external resource is concerned -- and this > sounds like what is really needed here. For example, he'll need it if > he wants to make sounds at syntax time. [This is also related to some > huge and bloody flamewar on c.l.s a few years ago, re the ability to > implement IO buffering in racket.] > > This should be available as an unsafe feature, of course. > > -- > ((lambda (x) (x x)) (lambda (x) (x x))) Eli Barzilay: > http://barzilay.org/ Maze is Life! > _________________________________________________ > For list-related administrative tasks: > http://lists.racket-lang.org/listinfo/dev > _________________________________________________ For list-related administrative tasks: http://lists.racket-lang.org/listinfo/dev