build failure on clang 19: deleted operator '=' on atomic

[katrinafyi]

after fixing #2529, there is another build failure with clang 19. the same failure happens on macos and linux.

souffle> /tmp/nix-build-souffle-2.4.1.drv-0/source/src/interpreter/Index.h:417:14: error: overload resolution selected deleted operator '='
souffle>   417 |         data = src.data;
souffle>       |         ~~~~ ^ ~~~~~~~~
souffle> /nix/store/cvl21fnjrwqcw59vkvhfdfl5w4d0xfaw-libcxx-19.1.6-dev/include/c++/v1/__atomic/atomic.h:58:11: note: candidate function has been explicitly deleted
souffle>    58 |   atomic& operator=(const atomic&)          = delete;
souffle>       |           ^
souffle> /nix/store/cvl21fnjrwqcw59vkvhfdfl5w4d0xfaw-libcxx-19.1.6-dev/include/c++/v1/__atomic/atomic.h:59:11: note: candidate function has been explicitly deleted
souffle>    59 |   atomic& operator=(const atomic&) volatile = delete;
souffle>       |           ^
souffle> /nix/store/cvl21fnjrwqcw59vkvhfdfl5w4d0xfaw-libcxx-19.1.6-dev/include/c++/v1/__atomic/atomic.h:49:29: note: candidate function
souffle>    49 |   _LIBCPP_HIDE_FROM_ABI _Tp operator=(_Tp __d) volatile _NOEXCEPT {
souffle>       |                             ^
souffle> /nix/store/cvl21fnjrwqcw59vkvhfdfl5w4d0xfaw-libcxx-19.1.6-dev/include/c++/v1/__atomic/atomic.h:53:29: note: candidate function
souffle>    53 |   _LIBCPP_HIDE_FROM_ABI _Tp operator=(_Tp __d) _NOEXCEPT {
souffle>       |                             ^

souffle/src/interpreter/Index.h

Lines 422 to 424 in 040a962

	void insert(const Index& src) {
	data = src.data;
	}

it looks like the code as written wants to implicitly load/store the atomic by coercing it to a bool, but the compiler is falling back to copy assignment for atomics which is deleted. it looks like the implicit load was deprecated (removed?) in C++ 20: https://en.cppreference.com/w/cpp/atomic/atomic/operator_T

to fix this, it might be possible to add explicit .load() and .store() calls. or, a std::atomic_flag could be used instead. i haven't tried any patches since i'm not familiar with the code.

[quentin]

I think it is safe to simply replace with data = true here.