Bug or local misconfiguration? On my local installation 3 Tests are red

[leider]

Transcript:

  1) RRule respects COUNT parts
   Message:
     Expected [ Date(Sat Jan 01 2011 00:00:00 GMT+0100 (CET)), Date(Tue Feb 01 2011 00:00:00 GMT+0100 (CET)) ] to equal [ Date(Sat Jan 01 2011 00:00:00 GMT+0100 (CET)), Date(Tue Feb 01 2011 00:00:00 GMT+0100 (CET)), Date(Tue Mar 01 2011 00:00:00 GMT+0100 (CET)) ].
   Stacktrace:
     Error: Expected [ Date(Sat Jan 01 2011 00:00:00 GMT+0100 (CET)), Date(Tue Feb 01 2011 00:00:00 GMT+0100 (CET)) ] to equal [ Date(Sat Jan 01 2011 00:00:00 GMT+0100 (CET)), Date(Tue Feb 01 2011 00:00:00 GMT+0100 (CET)), Date(Tue Mar 01 2011 00:00:00 GMT+0100 (CET)) ].
    at null.<anonymous> (/JavascriptDev/node-icalendar/spec/rrule-spec.js:43:18)

  2) RRule respects EXDATE full date parts
   Message:
     Expected [ Date(Sat Jan 01 2011 10:00:00 GMT+0100 (CET)), Date(Tue Feb 01 2011 10:00:00 GMT+0100 (CET)), Date(Fri Apr 01 2011 10:00:00 GMT+0200 (CEST)) ] to equal [ Date(Sat Jan 01 2011 10:00:00 GMT+0100 (CET)), Date(Fri Apr 01 2011 10:00:00 GMT+0200 (CEST)), Date(Sun May 01 2011 10:00:00 GMT+0200 (CEST)) ].
   Stacktrace:
     Error: Expected [ Date(Sat Jan 01 2011 10:00:00 GMT+0100 (CET)), Date(Tue Feb 01 2011 10:00:00 GMT+0100 (CET)), Date(Fri Apr 01 2011 10:00:00 GMT+0200 (CEST)) ] to equal [ Date(Sat Jan 01 2011 10:00:00 GMT+0100 (CET)), Date(Fri Apr 01 2011 10:00:00 GMT+0200 (CEST)), Date(Sun May 01 2011 10:00:00 GMT+0200 (CEST)) ].
    at null.<anonymous> (/JavascriptDev/node-icalendar/spec/rrule-spec.js:84:18)

  3) VEvent objects uses EXDATE properties when calculating recurrence
   Message:
     Expected [ Date(Sat Jan 01 2011 10:00:00 GMT+0100 (CET)), Date(Tue Feb 01 2011 10:00:00 GMT+0100 (CET)), Date(Fri Apr 01 2011 10:00:00 GMT+0200 (CEST)) ] to equal [ Date(Sat Jan 01 2011 10:00:00 GMT+0100 (CET)), Date(Fri Apr 01 2011 10:00:00 GMT+0200 (CEST)), Date(Sun May 01 2011 10:00:00 GMT+0200 (CEST)) ].
   Stacktrace:
     Error: Expected [ Date(Sat Jan 01 2011 10:00:00 GMT+0100 (CET)), Date(Tue Feb 01 2011 10:00:00 GMT+0100 (CET)), Date(Fri Apr 01 2011 10:00:00 GMT+0200 (CEST)) ] to equal [ Date(Sat Jan 01 2011 10:00:00 GMT+0100 (CET)), Date(Fri Apr 01 2011 10:00:00 GMT+0200 (CEST)), Date(Sun May 01 2011 10:00:00 GMT+0200 (CEST)) ].
    at null.<anonymous> (/JavascriptDev/node-icalendar/spec/vevent-spec.js:124:18)

[james-emerton]

I suspect this may have something to do with your timezone (although we use UTC for all recurrence calculations.) Is it possible that March 1st does not exist in your timezone? =)

[leider]

It definitely is a time zone problem. When I go west the errors disappear. So you could also experience failing tests by setting your machine's time zone east of Greenwich.

Am 21.11.2013 um 02:29 schrieb James Emerton notifications@github.com:

I suspect this may have something to do with your timezone (although we use UTC for all recurrence calculations.) Is it possible that March 1st does not exist in your timezone? =)

— Reply to this email directly or view it on GitHub.

[leider]

Diving throughout the code for some time. it is hard tp see where local dates are used or where utc comes in. There is a lot of stuff done manually that could be much easier by using http://momentjs.com/timezone/

Perhaps you should have a closer look at these.

Also, are you sure to always have UTC stuff when you think you do? UTC is a way to display a Date. Javascript Dates are UTC by default. The only thing that is timezone aware is the representation!

[RBLU]

I see the same behaviour when I try to use the RRule.getNextOccurances() with a COUNT rule

• when my env is CET (UTC + 1) then it misses the last occurance, because it thinks it is after the end by 1 hour.
• when my env is UTC I get the correct number of occurances

By debugging through the code I noticed several issues in RRule.js with timezones, it seems to me that the RRule.to_utc_date is actually correcting in the wrong direction whenever the system timezone is not UTC - on the other hand I don't really understand the code there, so I might be wrong with that analysis.

Our servers run on UTC so this not an issue there - but unittests run on developers laptops in all kinds of timezones...

I can also see 4 failing unittests whenever my System Timezone is not UTC, as soon as I switch the laptops timezone to UTC all tests pass.

I tried fixing this, but I keep breaking other tests