Ooops, that's right - the dependence is there indirectly via acs-datetime.
Just before I wanted to post my true and humble apologies for overlooking this fact I stumbled across this in acs-datetime.info: "Note: this version of the package requires acs-event(s) for a helper function. This requirement may go away in the future." Phew! That saves me 😉
Althoug "in the future" generally translates to "never" when written in software documentation, it might still be a good idea to put the dependency directly into calendar.info because there are tables of acs-events that are directly referenced by calendar tables.
