I'm sorry, this is probably false issue.
I think there should be Setmark(h) added after this line:
http://code.google.com/p/inferno-os/source/browse/libinterp/xec.c#1322
But I was unable to implement example to demonstrate real bug here. In my tests
h->color always was equal either to mutator or propagator, so looks like there
are really no need in Setmark(h) here. But I don't understand why - the pointer
to complex struct is copied to p->exval, the h->ref++ is done as usually after
such operations, but where is Setmark(h)? Or it doesn't needed because p->exval
is "special" in some way? My guess was it's handled by rootset(), but I don't
see this in code.
Original issue reported on code.google.com by powerman...@gmail.com on 16 Apr 2011 at 9:48
Original issue reported on code.google.com by
powerman...@gmail.com
on 16 Apr 2011 at 9:48