Add dummy hash functions for arb types

[lgoettgens]

Some progress towards oscar-system/Oscar.jl#2222, actually resolves all Nemo cases listed there.

These hash functions are valid, but not good. Once we have a way to safely access the mid and rad, these can be improved by delegating to hash functions for arf or mag (which then need someone to look at flint implementations).

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 8 lines in your changes missing coverage. Please review.

Project coverage is 87.88%. Comparing base (73e3de8) to head (1e7251a).
Report is 11 commits behind head on master.

Files with missing lines	Patch %	Lines
src/arb/Complex.jl	0.00%	2 Missing ⚠️
src/arb/Real.jl	0.00%	2 Missing ⚠️
src/arb/acb.jl	0.00%	2 Missing ⚠️
src/arb/arb.jl	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1954      +/-   ##
==========================================
- Coverage   87.93%   87.88%   -0.05%     
==========================================
  Files          99       99              
  Lines       36402    36424      +22     
==========================================
+ Hits        32010    32013       +3     
- Misses       4392     4411      +19     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[fingolfin]

Perhaps @albinahlback or @fredrik-johansson have suggestions for what might be "good" (or at least not completely horrible) hash functions for arb types?

[albinahlback]

They should fit inside a UInt? If so, I suppose concatenating trailing bits of mantissa and exponent would not be a terrible idea? Perhaps also fit the sign bit in some way.

[albinahlback]

I believe fmpz already have a straight forward hash function (it just takes the first limb) that you should be able to use, but I cannot seem to find it at the moment.

Edit: Found it. It is calcium_fmpz_hash, and is defined in calcium.h.

Edit2: You can utilize this to access the exponent, I meant to say.

[fingolfin]

Fine by me. But perhaps also open an issue to remind us we need to improve these??