Microsoft/calculator

Inaccurate results for trigonometric functions with exact results

Open

#309 opened on 2019年3月14日

GitHub で見る
 (3 comments) (0 reactions) (0 assignees)C++ (16,196 stars) (2,422 forks)batch import
Area: User ExperienceBugPri: 2help wantedtriage approved

説明

Describe the bug The below examples of the trigonometric functions show a correct result at first. However, subtracting that result from itself results in a non-zero value, which shows that the results are represented as incorrect approximations internally, when it should be an exact result.

In DEG mode: arcsin(1) - 90 != 0 sin(30) - 0.5 != 0 In RAD mode: sin(π/2) - 1 != 0 but interestingly: cos(±π) + 1 == 0

This is a new issue spawned from the similar issue #222.

Steps To Reproduce

  1. Try the above calculations in scientific mode

Expected behavior Exact internal representation and result 0 for above calculations.

Device and Application Information (please complete the following information):

  • OS Build: 10 0 17763 0
  • Architecture: X64
  • Application Version: 10.1812.10048.0

コントリビューターガイド

技術スタック
cpp
領域
desktop
Issue 種別
bug
難度新しい貢献者にとっての実装難度です。1 は非常に小さな変更、5 は専門的な作業です。
3
推定時間経験ある貢献者が調査、実装、テスト、pull request 準備にかけるおおよその時間範囲です。
1-3 hours
活動状況Issue が今どれだけ取り組みやすいかを示します。新しい、活発、古い、ブロック中、またはメンテナー入力待ちなどです。
stale
明確さ期待される変更、受け入れ条件、次の手順がどれだけ明確かを示します。
clear
前提条件
C++floating point arithmeticmath
初心者向け度初回貢献者にどれだけ取り組みやすいかを 1-100 で推定したスコアです。
65
調査方針
Look at the `src/CalcManager` directory for the implementation of trigonometric functions. The issue references #222, so read that issue for more context. Check how the calculator handles floating point precision and whether it uses double or a custom type. The fix might involve rounding results to exact values when the input is special (e.g., sin(pi/2)). Modify the relevant C++ code to handle these cases.