[clang] [analyzer] Fix security.VAList false positives with C23 va_start (PR #192024)

[Björn Svensson via cfe-commits]

bjosv wrote:

> The code change LGTM, but on the testing side I don't like that this PR introduces a new tests instead of using the same set tests for C23 and older standards.

Thanks! It got much better when guarding which mock to include in the existing test.
I just added one new C23 specific test to verify `va_start` with a single argument.

https://github.com/llvm/llvm-project/pull/192024