The second reading was coming in wrong because we weren't turning off both
GPIO pins before we slept, so there was probably some remnant charge on the
first pin during the second read. Adding in a turn_pin_off call to ensure
that both pins are off and have time for the charge to dissipate.
Coverage increased (+0.03%) to 80.594% when pulling d06683a6b1d4fc3b7fb7e8af472f923eb46b8b4b on mtlynch:let-charge-dissipate into 1d3d41062a00b83050dcbdde525aa1fa154c292e on JeetShetty:master.
The second reading was coming in wrong because we weren't turning off both GPIO pins before we slept, so there was probably some remnant charge on the first pin during the second read. Adding in a turn_pin_off call to ensure that both pins are off and have time for the charge to dissipate.