I don’t know… It’s still an allocatable at the beginning and a pointer at the end. Looks really doubtful to me, and I wouldn’t really want to rely on such code.
I’ve tested, and len(f_string) after the c_f_pointer in FREE_STRING() returns a length of 0. Which is not really surprising, as the compiler has no way to know what was the original length of the character(:) variable. In these conditions, how can we be sure that deallocate does what is expected?