https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127478
Bug ID: 127478
Summary: [reflection] missing "consteval-only value" error for
init of auto var
Product: gcc
Version: 17.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: mpolacek at gcc dot gnu.org
Target Milestone: ---
In
```
using info = decltype(^^int);
consteval auto f () { return ^^int; }
consteval auto g () { return info{}; }
void fn1 () { bool b = f () != info{}; } // { dg-error
"consteval-only value" }
void fn2 () { bool b = g () != info{}; }
void fn3 () { bool b = ^^int != info{}; } // { dg-error
"consteval-only value" }
void fn4 () { bool b = info{} != info{}; }
constexpr void fn5 () { bool b = f () != info{}; } // { dg-error
"consteval-only value" }
constexpr void fn6 () { bool b = g () != info{}; }
constexpr void fn7 () { bool b = ^^int != info{}; } // { dg-error
"consteval-only value" }
constexpr void fn8 () { bool b = info{} != info{}; }
```
we don't give the error for f1, because of:
/* Don't fold initializers of automatic variables in constexpr functions,
that might fold away something that needs to be diagnosed at constexpr
evaluation time. */
if (!current_function_decl
|| !maybe_constexpr_fn (current_function_decl)
|| TREE_STATIC (decl))
value = cp_fully_fold_init (value);
folds the initializer to 1. We also check the unfolded version but
check_out_of_consteval_use doesn't evaluate a CALL_EXPR so it won't see what it
evaluates to.
