C++: Refactor code to use predicate isGuardPhi/4

[andersfugmann]

Addressing review comment on #6568: This Pr uses the new predicate for RangeSSA::isGuardPhi/4.
I have not deleted the old predicate RangeSSA::isGuardPhi/3, even though its not used anymore to avoid braking of external queries which might rely on RangeSSA::isGuardPhi/3

[jbj]

I suggest marking RangeSSA::isGuardPhi/3 as deprecated. It seems there wasn't a single need for it in our own codebase, and we found out it can be a footgun. Then a change note is needed.

As these changes affect the join order, a DCA run would be appropriate.

[andersfugmann]

I suggest marking RangeSSA::isGuardPhi/3 as deprecated. It seems there wasn't a single need for it in our own codebase, and we found out it can be a footgun. Then a change note is needed.

I will deprecate the function.

As these changes affect the join order, a DCA run would be appropriate.

DCA timings looks ok.

[jbj]

There's an unrelated integration-test failure:

Running TRAP import for CodeQL database at /home/runner/work/semmle-code/semmle-code/integration-tests/posix-only/multi_extract/cluster/javascript...
foo.js.trap.gz, 3: com.semmle.util.exception.CatastrophicError: Tried to insert tuple with 2 columns into files which has 5 columns.

Maybe merging from main will fix it.

[jbj]

Whoops, this PR had fallen off my radar. LGTM.