output precision up to 1 decimal place

It is not bad if the data is output with 1 decimal place.
like in following example:

TinyDHT.h

uint8_t readHumidity(void); -> float readHumidity(void);
int16_t readTemperature(bool S = false); -> float readTemperature(bool S = false);

TinyDHT.cpp

int16_t DHT::readTemperature(bool S) { -> float DHT::readTemperature(bool S) {
int16_t f; -> float f;
f = (int16_t)data[2]; -> f = (float)data[2];
f = (int16_t)(data[2] & 0x7F); -> f = (float)(data[2] & 0x7F);
f += (int16_t)data[3]; -> f += (float)data[3];

int8_t DHT::readHumidity(void) { // 0-100 % -> float DHT::readHumidity(void) { // 0-100 %
uint8_t f; -> float f;
uint16_t f2; // bigger to allow for math operations -> float f2; // bigger to allow for math operations
f2 = (uint16_t)data[0]; -> f2 = (float)data[0];
f = (uint8_t)f2; -> f = (float)f2;

Start by comparing the declarations in TinyDHT.h with the implementations in TinyDHT.cpp, focusing on readTemperature and readHumidity and their current numeric types. The change is done when these readings support the requested fractional precision consistently in the library API and implementation.

Written by the indexing model from the issue text.