RE: [csmith-project/csmith] a39f3c: fix a bug analyzing union field reads/writes

["Xuejun Yang" <jxyang@cs.utah.edu>]

yes.

-Xuejun
> 
> This is fixing the issue reported by Pascal today?
> 
> John
> 
> 
> On 08/05/2011 02:05 PM, Xuejun Yang wrote:
> >    Branch: refs/heads/master
> >    Home:   https://github.com/csmith-project/csmith
> >
> >    Commit: a39f3c2a62c59d11aecd914ad176b6e5b1ab412a
> >
> https://github.com/csmith-project/csmith/commit/a39f3c2a62c59d11aecd914
> ad176b6e5b1ab412a
> >    Author: Xuejun Yang<jxyang@cs.utah.edu>
> >    Date:   2011-08-05 (Fri, 05 Aug 2011)
> >
> >    Changed paths:
> >      M src/Effect.cpp
> >    M src/FactUnion.cpp
> >    M src/Type.cpp
> >    M src/Type.h
> >    M src/Variable.cpp
> >    M src/Variable.h
> >
> >    Log Message:
> >    -----------
> >    fix a bug analyzing union field reads/writes
> >
> > correct a wrong assumption: writing to a field in a struct that is in an union
> does not change the last written field property. It turns out if the field itself is a
> struct with padding bytes, all of the union field become non-readable!
> >