Essential Type Category of Complex Floating Types?

[gs]

C99 provides for complex floating point types. What is the essential type category of each?

[misra-c]

The essential types of complex floating point numbers should be considered as the essential types of their real and imaginary parts.

Code:

float _Complex f1 = 3.0f + 2.0f*_Complex_I;

In this both the real and imaginary parts will have an "essentially float" type.

The MISRA C:2012 rules for conversions/arithmetic operations etc: will then apply to these types.
e.g.

Code:

double _Complex d1 =  3.0 + 2.0*_Complex_I;
float _Complex f2  = d1;

This will violate rule 10.3 since both the real and imaginary parts are assigned to a narrower essential type.