We wouldn't make it a macro per so, but a construct specific to Kernel.Typespec. It should be straight-forward. My biggest concern at this point is the name. :)
On Sat, May 16, 2020 at 10:21 AM Kip <[email protected]> wrote: > I had the `one_of` approach as my other alternative but figured adding > another function to Kernel (or elsewhere) may not be in favour. What do you > suggest I do next to progress this `one_of` idea as a poc? > > Thanks too for the reduction, my attempts were making that far too > complicated! > > On Saturday, May 16, 2020 at 3:29:27 PM UTC+8, José Valim wrote: >> >> Today you can do it like this: >> >> var = Enum.reduce(@list, &{:|, _. [&1, &2]} >> @type foo :: unquote(var) >> >> But it may be worth introducing something like you describe but it >> probably makes sense to do it via a construct, such as @type foo :: >> one_of(@var) that does the conversion for you. I think automatically >> converting the list can be confusing, because people may think that [:foo, >> :bar] implies a specific ordering, for example. >> >> I am just not sure of a good name. one_of may conflict with existing code. >> >> On Sat, May 16, 2020 at 6:25 AM Kip <[email protected]> wrote: >> >>> It's not uncommon to have domain overlap between lists of valid tokens >>> (used for validations) and type specs. For example: >>> >>> @standard_formats [ >>> :standard, >>> :accounting, >>> :currency, >>> :percent >>> ] >>> >>> @currency_formats [ >>> :currency, >>> :accounting, >>> :currency_long, >>> :currency_short >>> ] >>> >>> @currency_symbol [ >>> :standard, >>> :iso >>> ] >>> >>> @type standard_formats :: :standard | :currency | :accounting | :short >>> | :long >>> @type currency_formats :: :currency_short | :currency_long | >>> :decimal_short | :decimal_long >>> @type currency_symbol :: :standard | :iso >>> >>> It would go good to remove one source of error by being able to allow >>> compile time use of a list as the subject of @typespec. For example: >>> >>> @type currency_symbol :: @currency_symbol # Any compile-time resolvable >>> list >>> >>> Of course a macro can be introduced to do this but its quite difficult >>> to achieve since it requires manual manipulation of AST. >>> >>> Proposal worth considering? Or consign to the history of my >>> less-than-helpful ideas :-) >>> >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "elixir-lang-core" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to [email protected]. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/elixir-lang-core/36f821aa-d718-48aa-a9e8-2f6d5e440632%40googlegroups.com >>> <https://groups.google.com/d/msgid/elixir-lang-core/36f821aa-d718-48aa-a9e8-2f6d5e440632%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> >> -- > You received this message because you are subscribed to the Google Groups > "elixir-lang-core" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/elixir-lang-core/3577d145-b325-4c67-806c-d634df15ad5a%40googlegroups.com > <https://groups.google.com/d/msgid/elixir-lang-core/3577d145-b325-4c67-806c-d634df15ad5a%40googlegroups.com?utm_medium=email&utm_source=footer> > . > -- You received this message because you are subscribed to the Google Groups "elixir-lang-core" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/elixir-lang-core/CAGnRm4%2Bqvg4suaUa6CTDUavrjcZq-U3-2k08m7MMhKBEo7Fakw%40mail.gmail.com.
