Some explanation how the seemingly different concepts "enumerated
value" and "manifest constant" are actually related:
Let's start with the "classical" enums as they're known from
C/C++. They are used to create lists of symbolic names:
enum Color { YELLOW, PINK, BLUE };
Internally, each of its members has an integer value by which it
is represented. In the above example, YELLOW is 0, PINK is 1 and
BLUE is 2. They are also implicitly convertible to their values.
Whenever YELLOW, PINK or BLUE is used in the program, the
compiler will treat it as if it were a literal 0, 1 or 2 (mostly;
in some cases it is still possible to know that you're dealing
with YELLOW instead of 0 by using introspection).
Normally, the compiler will assign values to the members starting
from 0, but it's also possible to specify them explicitly:
enum SIPrefixes { KILO = 1_000, MEGA = 1_000_000 };
An extension (and generalization) of this concept is to allow
base types other than integers:
enum WeekDays : string { MON = "Monday", TUE = "Tuesday" };
They, too, are implicitly convertible to their base types. And
analogously to the Color enum, in this case the members are
treated as if they were the string literals "Monday", or
"Tuesday".
From here, it's only a short distance to manifest constants. We
can get there by making the braces optional when there is only
one member and the enum type has no name, and allowing the base
type to be automatically deduced:
enum ManifestArrayLiteral = [1, 2, 3];
(I believe, internally DMD does treat the two concepts
differently, but IMO this is a nice way to understanding how they
work.)
