This new regression is happening because indicator-datetime-service no longer knows when the menu is being popped down because it's been decoupled from those GTK+ events after 13.10 de-GTK+-ified the indicator stack.
So we need to re-introduce some kind of notification mechanism like we used to have in about-to-show s.t. indicator-datetime-service will know to snap the calendar to the current date when the indicator's menu gets pulled down.
Adding also-affects unity, larsu is looking into u-p-s to see how much work it will take to get about-to-show working properly there.
If that proves to be more effort than we have time for right now, we could probably do an interim fix in IDO. indicator-datetime is IdoCalendarMenuItem's only customer, so we could change the contract between them for an IDO-and-datetime-specific fix. Obviously this a "plan b" option; it would be better to just fix about-to-show.
This new regression is happening because indicator- datetime- service no longer knows when the menu is being popped down because it's been decoupled from those GTK+ events after 13.10 de-GTK+-ified the indicator stack.
So we need to re-introduce some kind of notification mechanism like we used to have in about-to-show s.t. indicator- datetime- service will know to snap the calendar to the current date when the indicator's menu gets pulled down.
Adding also-affects unity, larsu is looking into u-p-s to see how much work it will take to get about-to-show working properly there.
If that proves to be more effort than we have time for right now, we could probably do an interim fix in IDO. indicator-datetime is IdoCalendarMenu Item's only customer, so we could change the contract between them for an IDO-and- datetime- specific fix. Obviously this a "plan b" option; it would be better to just fix about-to-show.