Implement an indirect-goto optimization for goto *&&lbl and respect this
in the scope checker. With that done, turn an indirect goto into a
protected scope into a hard error; otherwise IR generation has to start
worrying about declarations not dominating their scopes, as exemplified
in PR8473.
If this really affects anyone, I can probably adjust this to only hard-error
on possible indirect gotos into VLA scopes rather than arbitrary scopes.
But we'll see how people cope with the aggressive change on the marginal
feature.
git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@117539 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/test/SemaCXX/scope-check.cpp b/test/SemaCXX/scope-check.cpp
index cdc3868..d462af0 100644
--- a/test/SemaCXX/scope-check.cpp
+++ b/test/SemaCXX/scope-check.cpp
@@ -66,7 +66,7 @@
C c0;
- goto *ip; // expected-warning {{indirect goto might cross protected scopes}}
+ goto *ip; // expected-error {{indirect goto might cross protected scopes}}
C c1; // expected-note {{jump bypasses variable initialization}}
lbl1: // expected-note {{possible target of indirect goto}}
return 0;
@@ -90,7 +90,7 @@
if (ip[1]) {
D d; // expected-note {{jump exits scope of variable with non-trivial destructor}}
ip += 2;
- goto *ip; // expected-warning {{indirect goto might cross protected scopes}}
+ goto *ip; // expected-error {{indirect goto might cross protected scopes}}
}
return 1;
}