perf: Implement faster implementation of `st_has(z/m)` using `WKBBytesExecutor`

[petern48]

This PR introduced a new WkbBytesExecutor that was used to optimize the implementation of st_geometrytype. It allowed us to extract what we needed without parsing the entire buffer, saving a lot of time. I had the bright idea of leveraging this for the dimension calculation, so we can implement functions like st_hasz and st_hasm without parsing the entire geometry. The logic turns out to be more complicated than I originally expected (due to edge cases relating to inferring the dimensionality), but I still think this is doable.

To properly get the dimensionality, we need to OR all of the following (short circuiting permitted of course):

• the dimensionality of the geometry type (the obvious one)
  • e.g POINT Z EMPTY
• the dimensionality of the first nested geometry (if it's some sort of collection)
  • e.g GEOMETRYCOLLECTION (POINT Z (0 0 0)) -> xyz
• the dimensionality of the first coordinate (in WKB this is actually represented as the dimensionality of the nested geometry object, not the first coordinate)
  • e.g POINT (0 0 0) -> xyz, POINT (0 0 0 0) -> xyzm

All of those examples are (apparently) valid.

[petern48]

Testing some of those cases properly is currently blocked by #162

[petern48]

Putting on a pause on that PR, but I figured I'd push it for now and share what I found.