https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121876
Bug ID: 121876
Summary: [16 Regression] recursive ICE in ICE handling with
nested diagnostics, in set_diagnostic_buffer, at
buffering.cc:49
Product: gcc
Version: 16.0
Status: UNCONFIRMED
Keywords: diagnostic, error-recovery, ice-on-invalid-code,
needs-bisection, needs-reduction
Severity: normal
Priority: P3
Component: diagnostics
Assignee: dmalcolm at redhat dot com
Reporter: dmalcolm at gcc dot gnu.org
CC: dmalcolm at gcc dot gnu.org, hewillk at gmail dot com,
unassigned at gcc dot gnu.org, webrown.cpp at gmail dot com
Target Milestone: ---
+++ This bug was initially created as a clone of Bug #121859 +++
#include <functional>
struct S {
static int f();
};
int main() {
S* s = nullptr;
std::function_ref<int()> r(std::nontype<&S::f>, s);
}
https://godbolt.org/z/hrjE4e81E
There are two things going on here:
(1) an ICE due to a read-through null in the C++ frontend here:
VVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVVV
at ../../src/gcc/cp/constraint.cc:3363
3363 if (TREE_CODE (expr) == TRAIT_EXPR)
(gdb) list
3363 if (TREE_CODE (expr) == TRAIT_EXPR)
3364 {
3365 diagnose_trait_expr (loc, expr, args);
3366 return true;
3367 }
(gdb) p expr
$1 = <tree 0x0>
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(2) during handling of the above ICE, the diagnostics subsystem attempts to
emit the ICE, but it's already midway through handling a nested diagnostic, and
so during the emission of the internal error, we hit an assertion.
This bug is to track (2) above: the ICE handler needs to gracefully handle the
case of an ICE occurring within a nested diagnostic.
