Monthly recurrence rule is generating 2 appointments on same day due to daylight saving change

[otzy007]

Hi,

I've found an interesting case on both v 0.12.1 and master where for a recurrence rule there are 2 appointments created on the same day Sun, 05 Nov 2017 01:00:00 AKDT -08:00, Sun, 05 Nov 2017 01:00:00 AKST -09:00.

I suppose this has something to do with daylight saving time.

The code to reproduce this issue is

      Time.zone = "UTC"
      
      r = Rule.monthly.day_of_month(5)

      ocurrences_date_from = Time.zone.parse('Fri, 05 May 2017 19:11:17 UTC +00:00')
      schedule = IceCube::Schedule.new
      schedule.add_recurrence_rule r
      schedule.start_time = Time.zone.parse("Wed, 05 Oct 2016 09:00:00 UTC +00:00").in_time_zone('America/Nome')

      p schedule.occurrences_between(ocurrences_date_from, ocurrences_date_from +  16.months)

and the output is

[Mon, 05 Jun 2017 01:00:00 AKDT -08:00, Wed, 05 Jul 2017 01:00:00 AKDT -08:00, Sat, 05 Aug 2017 01:00:00 AKDT -08:00, Tue, 05 Sep 2017 01:00:00 AKDT -08:00, Thu, 05 Oct 2017 01:00:00 AKDT -08:00, Sun, 05 Nov 2017 01:00:00 AKDT -08:00, Sun, 05 Nov 2017 01:00:00 AKST -09:00, Tue, 05 Dec 2017 01:00:00 AKST -09:00, Fri, 05 Jan 2018 01:00:00 AKST -09:00, Mon, 05 Feb 2018 01:00:00 AKST -09:00, Mon, 05 Mar 2018 01:00:00 AKST -09:00, Thu, 05 Apr 2018 01:00:00 AKDT -08:00, Sat, 05 May 2018 01:00:00 AKDT -08:00, Tue, 05 Jun 2018 01:00:00 AKDT -08:00, Thu, 05 Jul 2018 01:00:00 AKDT -08:00, Sun, 05 Aug 2018 01:00:00 AKDT -08:00, Wed, 05 Sep 2018 01:00:00 AKDT -08:00]

As you can see there are 2 appointments on 5th of November: Sun, 05 Nov 2017 01:00:00 AKDT -08:00, Sun, 05 Nov 2017 01:00:00 AKST -09:00

Any hints on why this could happen? Or how we can fix it?

[toxaq]

Have you seen #335 ? Timezones not fully supported but no decision was made about using ActiveSupport solution offered...

[avit]

Thanks, closing as a duplicate.

If you load ActiveSupport and use ActiveSupport::TimeWithZone objects (Time.zone.now etc.), then it will work correctly.

[avit]

Actually, reading this again it looks like that's what you're doing. I'll have a closer look.