3 ms·

I'm don't understand why he needs '==' to specify namespace and type aliases, also mentioned here[1]. Why not just '='? Show your working, Herb! Also starting

I'm don't understand why he needs '==' to specify namespace and type aliases, also mentioned here[1]. Why not just '='? Show your working, Herb! Also starting enums at 1 instead of 0 is an odd choice, if you'll pardon the pun. [1] https://herbsutter.com/2023/04/30/cppfront-spring-update/ https://herbsutter.com/2023/04/30/cppfront-spring-update/

dwattttt 3y ago
I've always respected choosing 0 as an invalid state. It makes it much harder to accidentally have a function succeed if the default representation is "you messed up"
- naruhodo 3y ago
  Wouldn't that mean that you should just define ERROR as the first enumeration constant, with value 0, rather than break out of the enum type? It's really common though to use enums as array indices, which start at 0.
  - dwattttt 3y ago
    In practice I always choose the first enumeration to be a specific error of "you failed to initialise this". I don't know that I've ever directly used an enum as an index, but I guess if I did I'd have to bias it.
Someone 3y ago
FTA: In the April blog post, I mentioned I needed a way to write type and namespace aliases, and that because all Cpp2 declarations are of the form thing : type = value, I decided to try using the same syntax but with == to denote “always equal to.” So, I think we have // new type foo = std::vector<int>; // alias of existing type bar == std::vector<int>; I guess that could evolve to (or maybe already has evolved to?) // variable with initial value foo: int = 42; // constant bar: int == 42;
- naruhodo 3y ago
  Oh... so the new type, foo, is distinct (not assignment-compatible) from std::vector<int>?
  - Someone 3y ago
    Yes, but functionally identical. See for example Haskell type vs new type: https://typeclasses.com/coerce https://typeclasses.com/coerce Python NewType: https://docs.python.org/3/library/typing.html#newtype https://docs.python.org/3/library/typing.html#newtype
dylukes 3y ago
If you start enums at 1, then the first item in the enum isn't coercible to false, all items coerce to true.

dwattttt 3y ago

I've always respected choosing 0 as an invalid state. It makes it much harder to accidentally have a function succeed if the default representation is "you messed up"

naruhodo 3y ago
Wouldn't that mean that you should just define ERROR as the first enumeration constant, with value 0, rather than break out of the enum type? It's really common though to use enums as array indices, which start at 0.
- dwattttt 3y ago
  In practice I always choose the first enumeration to be a specific error of "you failed to initialise this". I don't know that I've ever directly used an enum as an index, but I guess if I did I'd have to bias it.

Someone 3y ago

FTA: In the April blog post, I mentioned I needed a way to write type and namespace aliases, and that because all Cpp2 declarations are of the form thing : type = value, I decided to try using the same syntax but with == to denote “always equal to.” So, I think we have // new type foo = std::vector<int>; // alias of existing type bar == std::vector<int>; I guess that could evolve to (or maybe already has evolved to?) // variable with initial value foo: int = 42; // constant bar: int == 42;

naruhodo 3y ago
Oh... so the new type, foo, is distinct (not assignment-compatible) from std::vector<int>?
- Someone 3y ago
  Yes, but functionally identical. See for example Haskell type vs new type: https://typeclasses.com/coerce https://typeclasses.com/coerce Python NewType: https://docs.python.org/3/library/typing.html#newtype https://docs.python.org/3/library/typing.html#newtype

dylukes 3y ago

If you start enums at 1, then the first item in the enum isn't coercible to false, all items coerce to true.