[CALCITE-6074] The size of REAL, DOUBLE, and FLOAT is not consistent

[mihaibudiu]

This is based on a code audit.
The assumption is that FLOAT=DOUBLE=64 bit and REAL=32 bit
I also deleted some deprecated functions instead of fixing them.

[mihaibudiu]

Does anyone want to review this PR?
It relies on the core assumption that in Calcite FLOAT=DOUBLE.

This PR clearly fixes multiple bugs caused by misunderstandings about the floating point types.
In our compiler backend we have disabled the FLOAT type altogether (using a shuttle), because I think it's too dangerous for users to use correctly. My original proposal was to remove the FLOAT type from the Calcite IR and generate DOUBLE in the parser. Then such mistakes would become impossible in the future. This PR is less radical.

[dssysolyatin]

@mihaibudiu PR looks good for me. I faced with the same problem when tried to convert sql from postgresql dialect to spark dialect. I was really surprise that spark float is 4 byte and postgres float is double precision by default.

[mihaibudiu]

Apparently I can't delete deprecated functions, so I will put these back.

[mihaibudiu]

I am planning to merge this PR soon. If you have comments now is the time.
It touches a lot of different parts.

[sonarcloud]

Quality Gate passed

The SonarCloud Quality Gate passed, but some issues were introduced.

6 New issues
0 Security Hotspots
57.1% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarCloud