Re: rethinking syscall tapset

["Frank Ch. Eigler" <fche at redhat dot com>]

Hi -

> > > Also, I am planning to eventually make returnp indicate something
> > > about the return value's type.
> > Why, considering our static typing?
> How would that work with return values?

Normal functions' return values are put through the type
check/inference engine the same way as other variables, so that's the
basic model.  Do you believe the users of .return probe values (say,
open() rc and/or errno) need some sort of type polymorphism?  Can you
provide a script scenario where a run-time type-code conditional is
needed?

- FChE