[caldeveloper-l] SKIP in RRULEs

[Neil Jenkins]

OK, based on what Marten's expander seems to do and what I think makes sense, I've decide to go with: 
 * Deduplicate invalid dates that map (via skip) to the same valid date *before* applying bySetPosition. 
 * Ignore any dates that were produced by an earlier iteration *after* applying bySetPosition. 
 
There's another issue I've realised is also undefined in RFC7529: how you apply byWeekNo and byYearDay to an invalid date (which can happen because you don't apply the skip for an invalid day-of-month until after these are applied according to the spec). Since this is clearly nonsense, I would say that they simply never match an invalid date. 
 
I've updated the JSCalendar recurrence rule algorithm in this pull request <https://github.com/CalConnect/PUBLIC_DRAFTS/pull/16/files>; I believe this matches the semantics of RFC7529 <https://tools.ietf.org/html/rfc7529> as far as they are defined there. If anyone with knowledge or implementation experience of this could check and confirm/disagree, that would be much appreciated. 
 
JSCalendar is now in a two-week last call at the IETF and this is the only remaining issue, so I'd like to merge soon. 
 
Neil.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.calconnect.org/pipermail/caldeveloper-l-calconnect.org/attachments/20181107/1c816ead/attachment.html>