Environment

Version: v0.3.4
OS: ubuntu18.04.4 LTS

Description of this Issue

There are some cases where judge result of Test Run is different,
when the expected value of test data is a floating point number.
Below are examples.

• ABC117-A: https://atcoder.jp/contests/abc117/tasks/abc117_a

$ acd abc117/abc117_a test
Case 1: WA
Input:
8 3

Expected:
2.6666666667

Actually:
2.6666666666666665

Exit with: 0
Time: 3.3488ms

Case 2: WA
Input:
99 1

Expected:
99.0000000000

Actually:
99

Exit with: 0
Time: 3.1136ms

Case 3: WA
Input:
1 100

Expected:
0.0100000000

Actually:
0.01

Exit with: 0
Time: 2.8194ms

• ABC119-B: https://atcoder.jp/contests/abc119/tasks/abc119_b

$ acd abc119/abc119_b test
Case 1: WA
Input:
2
10000 JPY
0.10000000 BTC

Expected:
48000.0

Actually:
48000.000000

Exit with: 0
Time: 3.3426ms

Case 2: WA
Input:
3
100000000 JPY
100.00000000 BTC
0.00000001 BTC

Expected:
138000000.0038

Actually:
138000000.003800

Exit with: 0
Time: 3.1238ms

I think it is hard to absorb the difference of users environment (For example Mac, Linux, Windows?. where is the boost installed?).
So I think running test on container is better.

Ref https://qiita.com/knakajima3027/items/809be5abe5c3d9f29ee7

The following error occurs when executing the "acd prepare" command.

$ acd prepare abc100
Error: cannot detect lang

This error may be that the value of val specified in L96 of parse.go is becoming outdated. Probably should be changed to 5001.

https://atcoder.jp/contests/language-test-202001

Thanks for making very useful cli tool.
It would be great if browse command works by passing just problem name like abc100_a.