Revert "[analyzer] Check that a member expr is valid even when the result is an lvalue."

The original intent of this commit was to catch potential null dereferences
early, but it breaks the common "home-grown offsetof" idiom (PR13927):

 (((struct Foo *)0)->member - ((struct foo *)0))

As it turns out, this appears to be legal in C, per a footnote in
C11 6.5.3.2: "Thus, &*E is equivalent to E (even if E is a null pointer)".
In C++ this issue is still open:
  http://www.open-std.org/jtc1/sc22/wg21/docs/cwg_active.html#232

We'll just have to make sure we have good path notes in the future.

This reverts r164441 / 9be016dcd1ca3986873a7b66bd4bc027309ceb59.

llvm-svn: 164958

1 files changed, 4 insertions, 5 deletions

diff --git a/clang/test/Analysis/nullptr.cpp b/clang/test/Analysis/nullptr.cpp
index 80ef5fbf6d0..050c3f8dc56 100644
--- a/clang/test/Analysis/nullptr.cpp
+++ b/clang/test/Analysis/nullptr.cpp
@@ -23,11 +23,10 @@ void foo3(void) {
   };
   char *np = nullptr;
   // casting a nullptr to anything should be caught eventually
-  int *ip = &(((struct foo *)np)->f);  // expected-warning{{Access to field 'f' results in a dereference of a null pointer (loaded from variable 'np')}}
-
-  // Analysis stops at the first problem case, so we won't actually warn here.
-  *ip = 0;
-  *np = 0;
+  int *ip = &(((struct foo *)np)->f);
+  *ip = 0;  // expected-warning{{Dereference of null pointer}}
+  // should be error here too, but analysis gets stopped
+//  *np = 0;
 }
 
 // nullptr is implemented as a zero integer value, so should be able to compare