Compiling Meta Hacker Cup Inputs

#	User	Rating
1	tourist	4009
2	jiangly	3823
3	Benq	3738
4	Radewoosh	3633
5	jqdai0815	3620
6	orzdevinwang	3529
7	ecnerwala	3446
8	Um_nik	3396
9	ksun48	3390
10	gamegame	3386

#	User	Contrib.
1	cry	167
2	maomao90	163
2	Um_nik	163
4	atcoder_official	161
5	adamant	159
6	-is-this-fft-	158
7	awoo	157
8	TheScrasse	154
9	nor	153
9	Dominater069	153

Because it's fun :)

Inspired by SecondThread's comment above, I thought I'd see how hard it would be to solve a Meta Hacker Cup problem at compile time. I chose to start with Round 2 A1, since the input is quite small. This is probably trivial since C++20 and above allow you to allocate memory in constexpr functions, so I gave it a whirl using C++17. This means no allocations are allowed, but the input/output parsing is luckily very straightforward for this problem.

The code is a bit long (118 lines), so I've included it in the link below. It appears to work on the three major compilers: https://godbolt.org/z/cYG1rTfWr (v1) https://godbolt.org/z/3GdPKxrcj (v2).

Note that the compiled binary returns a random value, To actually view the answer, you can view the compiled binary in a hex editor, or do what I did:

$ strings a.out | grep Case
Case #1: 1
Case #2: 4
Case #3: 10
Case #4: 1
Case #5: 1
Case #6: 0

This code is also sufficient to pass the full data set (136 cases) without running into the default constexpr-ops limit. On my 2,3GHz Intel i5 laptop CPU from 2018, clang compiles the full data set in .5 seconds.

Does anyone know a nicer way of ensuring that ANS is not discarded by the compiler for not being used? EDIT: Thanks to EnDeRBeaT for the volatile recommendation, this is much simpler.