Monthly Recurring Rules Behavior?

[rlanvin]

Monthly Recurring Rules Behavior?

[rlanvin]

The lib is based on RFC 5545 so invalid dates generated while iterating the rule are ignored. However in your example you would easily get the last day of the month with BYMONTHDAY=-1 instead.

This behavior is not configurable, and SKIP has not been implemented - in fact to be honest I hadn't heard about RFC 7529 until now.

The example I used was misleading. I may want another specific date (when available) other than simply the last day.

Example - 30th of every month, fallback to last day of month:
2013-01-30, 2013-02-28, 2013-03-30, 2013-04-30, 2013-05-30

Example - 29th of every month, fallback to last day of month:
2013-01-29, 2013-02-28, 2013-03-29, 2013-04-29, 2013-05-29

so BYMONTHDAY=-1 won't work.

---

Is there any hope that RFC 7529 might be integrated(unsure of word choice) in the future?

[rlanvin]

Did you try RSet? Because I think this could be achieved with the RSet class that allows combining multiple rules, e.g. "29th of every month" + "every last day of February". While I agree it's not as elegant, it's nevertheless doable.

Is there any hope that RFC 7529 might be integrated(unsure of word choice) in the future?

As much as I would like to say yes, I honestly believe this is highly unlikely. The problem is that a date such as 30 February doesn't not exists. It's not "skipped" by the algorithm (and we could somehow modify it to, let's say, replace it by the previous valid date instead of "skipping" it), no, instead it isn't considered at all. So to implement SKIP would require (I think, based on a quick assessment) a major rewrite of the core algorithm. On top of that, most helpers functions (such as occursAt) would also need to be rewritten because a date that does not match the rule according to RFC 5545, actually could maybe match the rule according to RFC 7529 depending on the SKIP parameter. The cherry on top of the cake is that the RFC states that SKIP must only be used when RSCALE is present, and that comes with a gazillion of other issues...

The use case for this SKIP is extremely narrow (if I understand correctly is only about leap days and "last day of the month" type of rules) so I don't think it's worth such an extensive rewrite and all the problems associated with it.

I understand, thank you for the consideration.
I do believe I could make something work with RSet, and that might be what I end up doing ultimately.