A couple of options:
- MPA transition skips to the end, allowing the SPA transition to start.
- SPA transition skips to the end, allowing the MPA transition to continue (assuming the elements it needs are still there).
If an SPA transition starts while another SPA transition is in progress, the first transition skips to the end, allowing the second to start.
My opinion is option 1, given the consistency. I think we should avoid making MPA transitions different to SPA transitions, as much as possible. This thinking also seems in line with #8677