v0.1.2 Posit library with sin and cos functions

Added sin and cos function and their tests in the sketch.
Still TODO : implement sin and cos for PI * x as Posit8 results would be more precise that way.

v0.1.1 : 2's complement, sqrt, next and prior

• Now using 2's complement to store negative numbers
• Cast from/to Posit16 to/from Posit8
• Next and prior functions
• Sqrt for Posit8 and Posit 16
  • Different strategies are used for both, and Posit8 algorithm does not always converge
• Operator overloading for +=, -=, *=, /=

Full Changelog: https://github.com/tochinet/Posit/commits/0.1.1

v0.1.0rc1

This is the first release of the Posit8 and Posit16 library for Arduino.
It supports 8-bit and 16-bit Posits, with a variable size of exponent for the Posit 8 (0, 1 or 2)
Creation of posit is possible from raw value (unsigned byte or int), float or double.
The four usual operations are supported ( + _ * / ), with operator overloading.
No support for Posit32, Quire, dotproducts, comparison functions, functions of one argument, conversion between formats etc.

There are a few deviations from the Posit standard for the sake of simplicity and limited usage of program and memory. In particular, not all functions are implemented.
Also, while posit arithmetic should never overflow nor underflow, it is expected in AI applications that some factors should tend to zero. Hence the library will round down to zero for results below a defined EPSILON value, set to 1ppm (1E-6) by default for Posit8 and EPSILON^2 (1E-12) for Posit16. When EPSILON is defined as zero, no underflow occurs, conform to the standard.

First release contributed to the Arduino library website

Corrected library.properties for contribution to Arduino platform.

CAUTION: This release does not encode the negative Posits as 2's complements !