https://gcc.gnu.org/bugzilla/show_bug.cgi?id=104642
Bug ID: 104642
Summary: Add __builtin_trap() for missing return at -O0
Product: gcc
Version: 12.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: redi at gcc dot gnu.org
Target Milestone: ---
Again and again and again users are surprised (and often angry) about the
effects of a missing return in C++. It's undefined behaviour but that doesn't
stop bug reports like PR 104635 and questions like
https://gcc.gnu.org/pipermail/gcc/2022-January/238200.html
Users expect the undefined behaviour for a missing return to be "semi-defined".
We should just make it trap at -O0, then at least we can tell them that for
unoptimized code, the result is guaranteed to crash. With optimisation, it's
still undefined and can invalidate all their assumptions.
As I said in https://gcc.gnu.org/pipermail/gcc/2022-January/238204.html
What if we inserted the trap for -O0?
1. Not everybody uses ubsan even when they should use it.
2. The code can use unreachable annotations if it really needs to leave
some paths unhandled, but really can't live with the branch and trap
instructions. (The C++ standard is getting std::unreachable and std::assume
to do that in a portable way, so there is less excuse for not doing it).
