Michael137 wrote: > Looking up global variables is not completely ideal (can be slow and look up > unrelated things). Would index_obj.GetValueAs**Signed**() == -1 work by any > chance?
Actually after some more investigating it turns out the test was just flawed. A default-constructed variant has a valid index (being the first element of the variant). The only case where the index is `variant_npos` is when the variant is "valueless", which [according to cppreference](https://en.cppreference.com/w/cpp/utility/variant/valueless_by_exception.html) only happens when an exception is thrown during assignment to the variant. I adjusted the test to test that scenario, and the formatter seems to work. https://github.com/llvm/llvm-project/pull/147283 _______________________________________________ lldb-commits mailing list lldb-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
