Format retained coordinate components with integer long division so AVR logging can retain submillimeter detail without a floating-point type override. Add bounded-output and precision regressions.
Co-authored-by: Limor Fried <ladyada@users.noreply.github.com>
Cover basic and extended GPS builds with float and double, and include the previously tested Nano and HILBERT reception fixtures.
Co-authored-by: Limor Fried <ladyada@users.noreply.github.com>
Run sketch regressions with consistent float and double definitions across library and test translation units. Verify that exact GNSS components and double GPS results retain position changes smaller than the legacy E7 resolution, and record the precision requirement for future receiver APIs and logging.
Co-authored-by: Limor Fried <ladyada@users.noreply.github.com>
Discover the C++ tests and Arduino regression sketches automatically. Run the sketches against the actual library with host serial output and timing, including marine extensions, and propagate failures and timeouts to CI.
Co-authored-by: Limor Fried <ladyada@users.noreply.github.com>
ladyadaandLimor Fried