Add a new KvikIO compatibility mode "AUTO"

[kingcrimsontianyu]

This PR addresses issue #529. A new compatibility mode AUTO (previously proposed as ALLOW) is added that explicitly asks KvikIO to determine whether ON or OFF is eventually used for the I/O.

Closes #529

[madsbk]

Looks great @kingcrimsontianyu but I am wondering if we should use the CompatMode enum instead of a string throughout?

[kingcrimsontianyu]

I'm still debating on this. My thought is that no matter what value the user specifies (via the environment variable or compat_mode_reset), internally defaults::_compat_mode will fall into one of the two states (ON/OFF) at the end, and the getter method bool compat_mode() never has a chance to return CompatMode::AUTO. So I kept CompatMode as only an internal detail. But I'm happy to make changes. Let me know what you think @madsbk @vuule . Thank you!

PS: If we decide to expose CompatMode in the public interface, do we need another getter method CompatMode requested_compat_mode() that returns the value user initially passed?

[madsbk]

I have no strong opinion, but leaning towards exposing CompatMode::AUTO.

PS: If we decide to expose CompatMode in the public interface, do we need another getter method CompatMode requested_compat_mode() that returns the value user initially passed?

No, I don't think that is needed.

[madsbk]

Looks great @kingcrimsontianyu! Could you also update the C++ docs and the Python docs ?

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[kingcrimsontianyu]

/ok to test

[kingcrimsontianyu]

/ok to test

[kingcrimsontianyu]

/ok to test

[kingcrimsontianyu]

/ok to test

[kingcrimsontianyu]

Created #556 to improve our testing on GDS/POSIX I/O code in the future.