================
@@ -9242,3 +9242,15 @@ Declares that a function potentially allocates heap
memory, and prevents any pot
of ``nonallocating`` by the compiler.
}];
}
+
+def ValueTypeDocs : Documentation {
+ let Category = DocCatDecl;
+ let Content = [{
+The ``value_type`` attribute can be used to mark user-defined types as 'value
+types'. When objects of value types are passed value to functions, the objects
+are always considered to be formally copied into a new object. This means the
+argument itself must be the only value referencing the passed object. This
----------------
jyknight wrote:
NRVO copy-elision can enable a function which returns by value to stash the
address of the returned value. Or, C++ constructors can do the same.
Then, copy-elision on the call can result in the callee receiving a object
whose address is escaped, despite it receiving it "passed by value". E.g. in
the following example, `&a == global` in the body of `foo`.
```
struct NonTrivial { NonTrivial(); };
NonTrivial *global;
NonTrivial frob() {
NonTrivial result;
global = &result;
return result;
}
void foo(NonTrivial a);
int main() {
foo(frob());
}
```
But, I don't think this new attribute is viable; the intent expressed by the PR
message is to use it in libc++ to mark stdlib types. But, this cannot be a type
property, because any code could do the above with any type.
https://github.com/llvm/llvm-project/pull/95004
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
