fix nanosec_face correction_profile bounds from going negative

[analogist]

The new nanosec_face has a nanosec_state.correction_profile of type int8_t, and can turn negative when ALARM is pressed from profile P0. From there the setting screen will display a cryptic P-, until the modulo arithmetic brings the profile back around to P0 again.

This makes two minor bugfixes to enforce bounds on correction_profile:

• roll around the correction_profile back to the highest profile when ALARM is pressed from P0 (based on #define nanosec_profile_count)
• prevent long press of ALARM from decrementing by 50 while on the profile screen - instead the behavior should still be to decrement by 1.

[joeycastillo]

This makes sense, thank you for the PR!