Niels Lohmann
Niels Lohmann
> is this because of using concat amalgamation check seems to have passed this time around You need to include `#include `.
Maybe some issue with the includes. Please try: - remove the `string_concat.hpp` include - revert the code to using `std::string()`... like you did before Sorry for the confusion.
Because there is more to forward-declare. I'm sorry, but it seems to be a more complicated issue to use an exception in the macro header.
Yes, the CI is green on the base branch.
Sure! Sorry for not answering earlier.
```cpp #include #include int main() { std::vector input = {{0x5b, 0x5b, 0x24, 0x42, 0x23, 0x75, 0x00, 0x00, 0x5b, 0x24, 0x42, 0x23, 0x75, 0x19, 0x00, 0x64, 0x00, 0x5b, 0x5b, 0x5b, 0x5b,...
The documentation (https://json.nlohmann.me/api/macros/nlohmann_define_type_intrusive/#notes) states the type must be default constructible to use the macro.
I think a cleaner approach would be indeed to fix `is_usable_as_basic_json_key_type` to also work with string views.
- [ci_static_analysis_clang](https://github.com/nlohmann/json/actions/runs/18982378422/job/54218502063?pr=4958#logs) and [ci_test_standards_clang (17, libstdcxx)](https://github.com/nlohmann/json/actions/runs/18982378422/job/54218502161?pr=4958#logs) complain about a missing new line at the end of a file - The MinGW job looks like a glitch - I restarted it....
@RogerSanders I'm sorry that the bug wasted your time :-(. I will try to make a 3.12.1 release soon.