Fallback to manually consume space if posix_fallocate isn't supported at runtime

[ironage]

Fixes realm/realm-js#6349 and realm/realm-dotnet#3482.

I think this has emerged as an issue since the change in #7067 where the compile time detection of POSIX_FALLOCATE changed.

According to the man pages, posix_fallocate can return not supported:

EOPNOTSUPP
              The filesystem containing the file referred to by fd does
              not support this operation.  This error code can be
              returned by C libraries that don't perform the emulation
              shown in NOTES, such as musl libc.

Fortunately for us, we do not rely on this and already have a fallback path so it is simply a matter of handling this error code in the same way. Testing this requires combination of a system that passes our compile time check but fails at runtime which we do not have yet.

☑️ ToDos

• 📝 Changelog update
• 🚦 Tests (or not relevant)
• C-API, if public C++ API changed
• bindgen/spec.yml, if public C++ API changed

[coveralls-official]

Pull Request Test Coverage Report for Build james.stone_470

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.03%) to 91.815%

---

Totals
Change from base Build 1965:	0.03%
Covered Lines:	232979
Relevant Lines:	253749

---

💛 - Coveralls

[kneth]

Should we also check for ENOTSUPP?

[ironage]

I think it is fine to only handle EOPNOTSUPP as that is what the man pages describes. From what I can find online, these codes are supposed to be different values, but due to historic reasons they are actually the same. Furthermore, the android sources define them as the same. So unless we encounter a device that breaks that rule, we should be fine.