I didn't mean that we would make a struct that people could pretend was a datetime and do things like call `.year` on it. I meant that if we are using a "magic value" we could make the "magic value" be something that can be expanded over time.
For example, if I want to make a function that takes "a thing that can be passed into `DateTime.shift`", it is much easier if I can say `DateTime.t() | DateTime.Lazy.t()` etc. Don't we have similar treatment for `Regex`? Like the internals aren't meant to be pattern matched on/used, but it can be accepted by internal functions. If we're sure there will only ever be one such atom then I guess it makes sense, but if I wanted go guard against "something that DateTime accepts" it feels like `when is_struct(datetime, DateTime) or is_struct(datetime, DateTime.Lazy)` would future proof that guard, where as `when is_struct(datetime, DateTime) or datetime == :utc_now` wouldn't. > On Jan 8, 2025, at 11:10 AM, José Valim <jose.va...@gmail.com> wrote: > > The problem is that the struct fields are public and we can't make them lazy. > So if the user does `DateTime.lazy(:utc_now).year`, it won't work, and if the > only use of said lazy types is to pass it to these functions, we might as > well make it an atom specific to these functions. :) > > > José Valim > https://dashbit.co/ > > > On Wed, Jan 8, 2025 at 4:46 PM Zachary Daniel <zachary.s.dan...@gmail.com > <mailto:zachary.s.dan...@gmail.com>> wrote: >> DateTime.shift(DateTime.lazy(:utc_now), duration) >> >> Doesn’t have the same feel of convenience but it seems to me to be a more >> resilient option. >> >>> On Jan 8, 2025, at 10:44 AM, Zachary Daniel <zachary.s.dan...@gmail.com >>> <mailto:zachary.s.dan...@gmail.com>> wrote: >>> >>> >>> I’m wondering if an opaque struct may be better, representing a >>> computed-on-demand datetime. Like `%DateTime.Lazy{}`. Then it could grow >>> over time without needing to expand a set of magic atoms. >>> >>>> On Jan 8, 2025, at 10:37 AM, 'Billy Lanchantin' via elixir-lang-core >>>> <elixir-lang-core@googlegroups.com >>>> <mailto:elixir-lang-core@googlegroups.com>> wrote: >>>> >>>> >>>> I like the functionality and neither option seems like the obvious choice. >>>> I agree with José here: >>>> >>>> > 2. If we go [the dedicated function] route, we may find ourselves adding >>>> > other functions, such as `add_to_utc_now` and `diff_to_utc_now`. >>>> >>>> So I somewhat favor `Date.shift(:utc_today, month: 1)`. It's naturally >>>> extensible to other bases (like `:now`, `:utc_now`) without polluting the >>>> temporal modules with helpers. >>>> >>>> The downside is that the types will be more awkward. `DateTime.shift/2` >>>> and friends will accept either a struct or one of a set of special atoms. >>>> So we're stuck increasing the either size of the types or the number >>>> functions. >>>> >>>> --- >>>> >>>> What does the `Date.range(date, duration)` return? >>>> On Wednesday, January 8, 2025 at 8:02:45 AM UTC-5 José Valim wrote: >>>>> Another scenario where :utc_now could be used is DateTime.after?(date, >>>>> :utc_now) >>>>> >>>>> >>>>> José Valim >>>>> https://dashbit.co/ >>>>> >>>>> >>>>> On Wed, Jan 8, 2025 at 1:57 PM Jon Rowe <ma...@jonrowe.co.uk <>> wrote: >>>>>> I'd love either of these proposals to be a reality, I often find myself >>>>>> building my own helpers in tests to do this sort of thing. >>>>>> >>>>>> Initially I thought the function varient was better but if you look at >>>>>> it as "then we might have to add all these other functions" I found >>>>>> myself leaning towards allowing `:utc_now` as a placeholder in the >>>>>> existing api, it might be slightly more verbose but it leans towards a >>>>>> more compact core api overall... so that gets my +1. >>>>>> >>>>>> Cheers >>>>>> Jon >>>>>> >>>>>> On Wed, 8 Jan 2025, at 11:06 AM, José Valim wrote: >>>>>>> I'd love to see something along those lines but I can't pick a favorite. >>>>>>> >>>>>>> 1. Supporting :utc_now in "shift" could be a welcome addition, as we >>>>>>> could also support it in "add" and "diff" functions. However, I'd say >>>>>>> it is more verbose than from_utc_today. >>>>>>> >>>>>>> 2. from_utc_now/from_utc_today is clearer but less applicable. If we go >>>>>>> this route, we may find ourselves adding other functions, such as >>>>>>> `add_to_utc_now` and `diff_to_utc_now`. >>>>>>> >>>>>>> So I would love to hear everyone's thoughts. >>>>>>> >>>>>>> "Date.range/2" with a duration is a no-brainer though and we could add >>>>>>> it today. >>>>>>> >>>>>>> José Valim >>>>>>> https://dashbit.co/ >>>>>>> >>>>>>> >>>>>>> On Tue, Jan 7, 2025 at 5:43 PM Wojtek Mach <woj...@wojtekmach.pl <>> >>>>>>> wrote: >>>>>>> Hello, >>>>>>> >>>>>>> I'd like to propose adding the following functions: >>>>>>> >>>>>>> - `Date.from_utc_today(duration)` >>>>>>> - `NaiveDateTime.from_utc_today(duration)` >>>>>>> - `DateTime.from_utc_today(duration)` >>>>>>> >>>>>>> For example: >>>>>>> >>>>>>> # Say, now is ~U[2025-01-07 16:22:40.003901Z] >>>>>>> >>>>>>> iex> Date.from_utc_now(month: 1, day: 1) >>>>>>> ~D[2025-02-08] >>>>>>> >>>>>>> iex> NaiveDateTime.from_utc_now(hour: -1) >>>>>>> ~N[2025-01-07 15:22:40.003901] >>>>>>> >>>>>>> iex> DateTime.from_utc_now(Duration.new!(hour: 1)) >>>>>>> ~U[2025-01-07 17:22:40.003901Z] >>>>>>> >>>>>>> I believe they are especially useful when writing tests and they might >>>>>>> give opportunity for some optimizations. >>>>>>> >>>>>>> Another idea is to instead allow passing `:utc_today` / `:utc_now` to >>>>>>> the existing shift/2 functions: >>>>>>> >>>>>>> iex> Date.shift(:utc_today, month: 1, day: 1) >>>>>>> ~D[2025-02-08] >>>>>>> >>>>>>> iex> NaiveDateTime.shift(:utc_now, hour: -1) >>>>>>> ~N[2025-01-07 15:22:40.003901] >>>>>>> >>>>>>> iex> DateTime.from_utc_now(:utc_now, hour: 1) >>>>>>> ~U[2025-01-07 17:22:40.003901Z] >>>>>>> >>>>>>> Btw and this is a related but separate conversion, I think a >>>>>>> `Date.range(date, duration)` would be a nice addition. And so, I >>>>>>> believe a `Date.range(:utc_today, month: 1)` would be a natural >>>>>>> extension of this. I'm not sure if supporting `Date.add(:utc_today, 1)` >>>>>>> and similar is worth it, perhaps just for consistency. >>>>>>> >>>>>>> >>>>>>> -- >>>>>>> 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 elixir-lang-co...@googlegroups.com <>. >>>>>>> To view this discussion visit >>>>>>> https://groups.google.com/d/msgid/elixir-lang-core/2851ea28-d20e-4e8d-b957-38a582f7fa39n%40googlegroups.com >>>>>>> >>>>>>> <https://groups.google.com/d/msgid/elixir-lang-core/2851ea28-d20e-4e8d-b957-38a582f7fa39n%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 elixir-lang-co...@googlegroups.com <>. >>>>>>> To view this discussion visit >>>>>>> https://groups.google.com/d/msgid/elixir-lang-core/CAGnRm4L9iSWc7zD8v_%2BqNhsOhx10atYksuPO5gbGCMvvz88sPA%40mail.gmail.com >>>>>>> >>>>>>> <https://groups.google.com/d/msgid/elixir-lang-core/CAGnRm4L9iSWc7zD8v_%2BqNhsOhx10atYksuPO5gbGCMvvz88sPA%40mail.gmail.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 elixir-lang-co...@googlegroups.com <>. >>>>> >>>>>> To view this discussion visit >>>>>> https://groups.google.com/d/msgid/elixir-lang-core/3952efd8-06a4-4855-9277-8208b5c8068f%40app.fastmail.com >>>>>> >>>>>> <https://groups.google.com/d/msgid/elixir-lang-core/3952efd8-06a4-4855-9277-8208b5c8068f%40app.fastmail.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 elixir-lang-core+unsubscr...@googlegroups.com >>>> <mailto:elixir-lang-core+unsubscr...@googlegroups.com>. >>>> To view this discussion visit >>>> https://groups.google.com/d/msgid/elixir-lang-core/d1c3407a-5c4c-4978-a813-43eaf9711554n%40googlegroups.com >>>> >>>> <https://groups.google.com/d/msgid/elixir-lang-core/d1c3407a-5c4c-4978-a813-43eaf9711554n%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 elixir-lang-core+unsubscr...@googlegroups.com >> <mailto:elixir-lang-core+unsubscr...@googlegroups.com>. >> To view this discussion visit >> https://groups.google.com/d/msgid/elixir-lang-core/CF7D089E-DBB9-432A-B15D-4EB5DFB046DB%40gmail.com >> >> <https://groups.google.com/d/msgid/elixir-lang-core/CF7D089E-DBB9-432A-B15D-4EB5DFB046DB%40gmail.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 elixir-lang-core+unsubscr...@googlegroups.com > <mailto:elixir-lang-core+unsubscr...@googlegroups.com>. > To view this discussion visit > https://groups.google.com/d/msgid/elixir-lang-core/CAGnRm4K0q1dOWpFxNH%2BxAsdJpt_G9DA-BDGJqeYxOTgD-AuDQg%40mail.gmail.com > > <https://groups.google.com/d/msgid/elixir-lang-core/CAGnRm4K0q1dOWpFxNH%2BxAsdJpt_G9DA-BDGJqeYxOTgD-AuDQg%40mail.gmail.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 elixir-lang-core+unsubscr...@googlegroups.com. To view this discussion visit https://groups.google.com/d/msgid/elixir-lang-core/7390B51F-96B5-44C3-8896-85B171EFB4C7%40gmail.com.
