Pure fixed-point math functions · SFTtech/openage#1543

2023-09-17T16:18:17.000Z

**Required Skills:** C++ **Difficulty**: Easy We should add additional math operations to our fixed-point implementation, so that all calculations can be done purely deterministic. The [`fpm`](https://github.com/MikeLankamp/fpm) library already has these operations and could be used as an inspiration. Important here: These operations **must not** use any floating point arithmetic for the calculation. They must receive fixed-point values as input and must also return a fixed-point value. This means `std` library calculations should not be used for solving this issue. Interesting for us: - [x] sqrt - [x] min - [x] max - [x] abs - [x] hypot - [ ] atan2 Nice to have: - [ ] sin - [ ] cos - [ ] tan - [ ] pow - [ ] log **Further Reading** - [openage fixed-point implementation](https://github.com/SFTtech/openage/blob/master/libopenage/util/fixed_point.h) - [fpm library documentation](https://github.com/MikeLankamp/fpm/blob/master/docs/index.md)

コントリビューターガイド

技術スタック: cpp
領域: backend
Issue 種別: feature
難度: 2
推定時間: 1-2 days
活動状況: stale
明確さ: clear
前提条件: C++fixed point arithmeticfpm library understandingopenage fixed point header
初心者向け度: 70
調査方針: Study the existing fixed point.h in libopenage/util. Understand how sqrt, min, max, abs, hypot are implemented (they are already done). Then examine the fpm library's implementation of atan2 and the nice to have functions. Implement the functions ensuring no floating point arithmetic, and add tests and documentation. Pay attention to potential overflow and precision issues.

説明

コントリビューターガイド