achieve a similar effect to stacker::maybe_grow

[loongs-zhang]

when #33 get merged, all CI of this PR should pass.

[Amanieu]

After thinking about this for a while, I think that this functionality doesn't belong in corosensei itself. Instead it should be provided in a separate crate, as a wrapper around a Coroutine. While corosensei can provide the base functionality for stack switching, the logic for querying the remaining amount of stack requires std (corosensei is #![no_std]). You can already query the limit of a stack using the limit() trait method.

[Amanieu]

In fact you should prefer using on_stack instead of Coroutine if you don't need support for yielding. It will be much faster.

[loongs-zhang]

After thinking about this for a while, I think that this functionality doesn't belong in corosensei itself. Instead it should be provided in a separate crate, as a wrapper around a Coroutine. While corosensei can provide the base functionality for stack switching, the logic for querying the remaining amount of stack requires std (corosensei is #![no_std]). You can already query the limit of a stack using the limit() trait method.

So does corosensei still need remaining_stack?

[loongs-zhang]

In fact you should prefer using on_stack instead of Coroutine if you don't need support for yielding. It will be much faster.

Haha, I need yielding.

[Amanieu]

No. In fact remaining_stack can return incorrect results with your implementation since the stack pointer may not necessarily be on the coroutine stack: it could have switched to another stack temporarily and then yielded.