mirror of https://github.com/ArduPilot/ardupilot
4c9156eab9
We were casting the location in integer 32 bits to float, and making fabsf in float as well, so we were losing precision translated in about 40 cm minimum variation in position for the NMEA output. Also, even if using double and fabsF, we were still rounding up last 2 decimals, so now the logic is done in degree * 10e7 to not loose precision and then converted properly before building the string |
||
|---|---|---|
| .. | ||
| examples/NMEA_Output | ||
| AP_NMEA_Output.cpp | ||
| AP_NMEA_Output.h | ||