Fixed clash in p1_display's plot_gnss_skyplot.

[thoshizaki]

Fixes

• Conditioned for the cases idx is empty or not.
• Let max_cn0_dbhz be always exactly len(p1_time) long and aligned to sv_data's epochs by timestamp, so the later decimation indexing at analyzer.py:1281 (max_cn0_dbhz[idx]) can no longer go out of bounds.
• Forced satellite timestamps to be increasing (were not likely because this is a merged/"duo" log).

[thoshizaki]

Comparison by 3 Logs

Left: master 2a8a893
(It gives error.)

Right: fix-cn0 a20fba3
(It does run.)

Amazon Leo (1) HITL Log

• python p1_display.py aeb01 -p fix-cn0

Amazon Leo (2) HITL Log

• python p1_display.py 7f32e -p fix-cn0

Atlas Duo, Test Drive

• python p1_display.py dfa47f9 -p fix-cn0 --log-type user

[adamshapiro0]

Closing in favor of #416. The underlying issue here was that the GNSSSignals messages from the two antennas were conflicting and p1_display was not separating them, but instead trying to plot them together.