Split off from 4970.
There are many spurious inconsistencies across and within locales with the widths of dates, so I suggest that we change these during the data resolution phase.
We agreed on changing yyyy to y for Gregorian. However, we still need to look at the impact of changing yy to y.
Note, this changed all y+ to y except in stock short patterns, for all calendars (thus in availableFormats and intervalFormats patterns). It also changed y+ in skeletons to y, which it should not have done. I filed : to fix DateTimeCanonicalizer.java to stop doing that. Meanwhile I repaired this problem for some locales as part of the generic calendar changes in : and will continue to do as part of more generic calendar changes in :
Milestone 23dres deleted