Align NowFunc::new() with canonical ConfigOptions timezone and enhance documentation

[kosiew]

Which issue does this PR close?

• Closes NowFunc::new() seeds the timezone with "+00" while the canonical UTC offset remains "+00:00" in ConfigOptions::default(). #18219.

---

Rationale for this change

The deprecated NowFunc::new() constructor previously initialized its timezone using the shorthand offset "+00", which was inconsistent with the canonical UTC offset format "+00:00" used by ConfigOptions::default(). This mismatch could cause subtle inconsistencies in ScalarValue comparisons or downstream timezone handling.

This PR ensures backward compatibility while aligning NowFunc::new() with the canonical default configuration, making behavior consistent across both constructors. It also improves documentation to clarify this relationship and provides a regression test to confirm parity between the two initialization paths.

---

What changes are included in this PR?

• Updated the deprecated NowFunc::new() to delegate to NowFunc::new_with_config(&ConfigOptions::default()).
• Added detailed doc comments explaining the rationale and proper usage of the constructors.
• Introduced a new test module verifying that NowFunc::new() and NowFunc::new_with_config() produce identical return fields and scalar values.
• Updated user documentation (scalar_functions.md) to note the constructor preference and clarify the canonical default timezone format (+00:00).

---

Are these changes tested?

✅ Yes. A new test now_func_default_matches_config was added to confirm functional equivalence between the legacy and configuration-based constructors, including matching field outputs and scalar timezones.

---

Are there any user-facing changes?

• Yes, but backward-compatible:

  • NowFunc::new() remains available but now mirrors the canonical timezone offset (+00:00).
  • Documentation has been updated to guide users toward the preferred NowFunc::new_with_config() method.

No breaking API or behavior changes are expected, as this update standardizes the default timezone while maintaining prior function signatures.

[alamb]

Thank you @kosiew -- this makes sense to me

@Omega359 can you take a look too if you have a moment?

[Omega359]

we'll want to align the docs and test and the order of merges with #18359 where the default tz was updated to be None (vs +00:00 previously). Otherwise I think this is a great PR.