Does SDP r= field stop time take precedence over the r= offset+duration?

Hi,

When the SDP spec. says the 't=' field defines the stop time for the
repeat schedule, what does that actually mean?  For example: assume the
't=' field says the stop time is 10:30 while the 'r=' field says the
session becomes active for one hour starting at 10:00.

Is this a protocol violation because the 't=' and 'r=' fields are
inconsistent?

or

Does this indicate that the final repetition is, in fact, only 30
minutes long?  That is, the stop time defined by the 't=' field is a
hard limit on the duration of any active periods defined by the 'r='
field.

or

Is the session actually active until 11:00 regardless of what the 't='
field says?  That is, the 't=' field actually defines the window during
which an active period may start but says nothing about when the
activity ceases.

Thanks,
Bryan.