Sat Nov 24 13:20:11 2018 reps 3 secs 15
Sat Nov 24 13:20:26 2018 reps 2 secs 15
Sat Nov 24 13:20:41 2018 reps 1 secs 15
Sat Nov 24 13:20:56 2018 reps 0 secs 15
I would've expected the interval between the 1st and 2nd execution be the same 15 seconds as the later intervals. If x3 had resumed the timer at 29 instead of 22, the intervals between the 1st and 2nd execution wouldn've been only 1 second. And if x3 had resumed the timer at 30, the 1st and 2nd rep would execute during the same $tick.
But I also had expected the -p behavior for .rep #1 to be what the requested -P behavior was hoping to be. As for whether it would break -p scripts if the delay following the resume of that timer were the same 15 used by the later intervals, I can't see a likely use for the current -p behavior. I would've guessed that the original requestor was repeating a timer at 1 sec intervals and wanted to pause that timer temporarily. In that scenario the -p behavior is not very noticeable. The timer executing immediately after resuming and having a 0-1 sec delay prior to the .rep after that is very similar to that timer's normal behavior.
I can't speak for the others asking for the -P behavior, but I assume they would want -P to have the interval before all .reps be the same as the .delay parameter, except when increased by the duration between when it's paused then later resumed.
