nfs-utils: remove automatic statd check and defer control to user

[slymattz]

Testing the changes

• I tested the changes in this PR: YES

Motivation for the change

Currently, the NFSV4_ONLY variable is only checked for a non-empty string meaning assigning any value including n/no/false etc. would make it pass the test.

The proposed change normalizes the check.

Please advise whether this should be revbumped or left as is.

Local build testing

• I built this PR locally for my native architecture: x86_64-glibc

[classabbyamp]

is it documented as accepting negative/false values = disable? "if variable is set = enable" is a pretty common pattern on void

[ahesford]

As I said six months ago, this was the wrong fix. It adds an extra source of truth that requires synchronization in two separate configurations or you can end up with a non-working configuration. Not only that, but sv check doesn't actually do what's intended. It is racy and incomplete, and what's worse, even in the cases where it might gate the remainder of this service, the failure will be silent and likely to go unnoticed.

The proposed change just doubles down on complexity without addressing the real issues. Just remove the sv check and let people worry for themselves whether they need to enable it. This lets you discard the extra source of truth and simplifies, rather than complicates, this service.

[slymattz]

This lets you discard the extra source of truth and simplifies, rather than complicates, this service.

I agree. This way it looks simpler and everything is documented on void docs. The hack wasn't documented at all except for the comment in the runit file.