[Bug c/106947] -Waddress + bool + pragma generates meaningless diagnostic

[pinskia at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106947

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|middle-end                  |c

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
The bug is in the C front-end.
Specifically the bug is in maybe_warn_for_null_address which does not check the
return value of warning if it should do the inform.

Something like this will fix the issue:
diff --git a/gcc/c/c-typeck.cc b/gcc/c/c-typeck.cc
index d37de2a313b..59e384ac997 100644
--- a/gcc/c/c-typeck.cc
+++ b/gcc/c/c-typeck.cc
@@ -11619,6 +11619,7 @@ build_vec_cmp (tree_code code, tree type,
 static void
 maybe_warn_for_null_address (location_t loc, tree op, tree_code code)
 {
+  bool do_note = false;
   /* Prevent warnings issued for macro expansion.  */
   if (!warn_address
       || warning_suppressed_p (op, OPT_Waddress)
@@ -11706,17 +11707,17 @@ maybe_warn_for_null_address (location_t loc, tree op,
tree_code code)
     return;

   if (code == EQ_EXPR)
-    warning_at (loc, OPT_Waddress,
+    do_note = warning_at (loc, OPT_Waddress,
                "the comparison will always evaluate as %<false%> "
                "for the address of %qE will never be NULL",
                op);
   else
-    warning_at (loc, OPT_Waddress,
+    do_note = warning_at (loc, OPT_Waddress,
                "the comparison will always evaluate as %<true%> "
                "for the address of %qE will never be NULL",
                op);

-  if (DECL_P (op))
+  if (do_note && DECL_P (op))
     inform (DECL_SOURCE_LOCATION (op), "%qD declared here", op);
 }