ChuanqiXu9 wrote:
> > Since the case a coroutine without a suspend point is relatively rare case
> > to me.
>
> For what it's worth, this issue isn't limited to coroutines without any
> suspend point. It can also occur:
>
> * If the coroutine conditionally doesn't await (e.g., `if (condition) {
> co_await foo(); }` and `condition` is false)
> * If the coroutine awaits, but the awaitor's `await_ready()` returns true so
> the coroutine doesn't suspend
> * If the coroutine suspends, but the awaitor's `await_suspend()` returns
> false so the coroutine is immediately resumed
> * If the coroutine suspends, but the awaitor's `await_suspend()` causes the
> coroutine to resume before the ramp function finishes (either by resuming on
> the same thread or losing a race)
> * If the coroutine suspends, but the awaitor's `await_suspend()` returns a
> coroutine handle whose resumption ultimately results in the current coroutine
> resuming before the ramp function finishes
Or we can model it as, if it is possible to reach the section after coro.end
from entry without reaching a real suspend point.
https://github.com/llvm/llvm-project/pull/127653
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
