What about a DateTime.from_iso8601 where you *have* to pass in a timezone or whatever else missing information in to it as well? That way the onus falls on the user to figure out what it should be?
On Monday, November 28, 2016 at 9:50:53 AM UTC-7, Joe McIlvain wrote: > > We have `Date.to_iso8601`, `Date.from_iso8601`, `Time.to_iso8601`, > `Time.from_iso8601`, and `DateTime.to_iso8601` functions, but we do not > have a `DateTime.from_iso8601` function. > > Congruence led me to believe that this function *had to* exist somewhere, > so I spent a significant amount of time searching and confirming that it > was indeed missing. > > I assumed that there must be a good reason for it to be missing, so I > searched the issue tracker, this mailing list, and eventually asked in IRC. > Here's a link to the transcript of that discussion, for context: > https://botbot.me/freenode/elixir-lang/2016-11-28/?msg=77118328&page=6 > > In short, it was suggested that conversion from ISO8601 to DateTime format > was ambiguous, since there is no concrete time zone encoded in the ISO8601 > format, and the DateTime struct, expects one. > > However, we do have "generic" time zone codes that could be used to convey > the situation of "I know the UTC offset, but I don't know anything else > about the time zone". According to wikipedia > <https://en.wikipedia.org/wiki/List_of_tz_database_time_zones>, the IANA > tz database defines the "Etc" area, and codes like "Etc/GMT+3" for this > kind of situation. This seems to be the approach that the `timex` library > uses when parsing formats that only know the UTC offset. > > For whatever it's worth, Ruby's `Time.parse` method uses `nil` for the > time zone code when it is not known, and relies on only having `utc_offset` > defined, so we could also do something similar to that. > > In general, I'd argue that having bidirectional conversions for `DateTime` > (or no conversion at all) is the best way to violate the "principle of > least surprise", especially when bidirectional transformations are > implemented for `Date` and `Time` to/from ISO8601. > > Is there another reason to exclude this function? If so, we should at the > very least have the reasoning on record so that users who run into the same > question can answer it by searching instead of asking again. > > Thanks, > -- 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/f71a81b3-9ebd-4ab7-a7f4-f514afb80ea9%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
