Fix a bug codegening SwitchInts with only an otherwise branch

[bkirwi]

The first commit adds a failing test: it exposes a bug dealing with dropping single-variant enums.

thread 'rustc' panicked at kani-compiler/src/codegen_cprover_gotoc/codegen/statement.rs:499:9:
assertion failed: targets.len() > 1
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

Kani unexpectedly panicked during compilation.
Please file an issue here: https://github.com/model-checking/kani/issues/new?labels=bug&template=bug_report.md

[Kani] current codegen item: codegen_function: std::ptr::drop_in_place::<Reference>
_RINvNtCs4AkhfejoRTd_4core3ptr13drop_in_placeNtCs6S0fALEP9ee_19single_variant_enum9ReferenceEBI_
[Kani] current codegen location: Loc { file: "/Users/bkirwi/.rustup/toolchains/nightly-2025-05-20-aarch64-apple-darwin/lib/rustlib/src/rust/library/core/src/ptr/mod.rs", function: None, start_line: 524, start_col: Some(1), end_line: 524, end_col: Some(56), pragmas: [] }
error: /Users/bkirwi/Code/kani/target/kani/bin/kani-compiler exited with status exit status: 101

This seems to affect both main and the latest release.

Resolves #4103

[carolynzech]

Thanks! Your fix looks good. I moved your test to a different folder--we like to put any test that fixes a crash in the kani subdirectory instead of ui. (If you're curious why, the ui folder has expected files where we assert that some particular substring is in the printed output. Here, we just care that the exit code is 0, i.e. that the proof went through successfully, which all we check for tests in the kani folder).

I also filed #4103 to document the bug for posterity. Thanks a lot for finding and fixing this!