It was introduced a (very) long time ago (c7bf03e), was barely changed since then and seems unused by now.
Apparently it was used mainly to get installation's Python to PATH (`set PATH=%PYTHONHOME%;%PATH%`). It was also setting and env variables based on the cache as a side effect.
If needed, we'll be able reintroduce it later, delegating most of the logic to Python.
`TARGET_ARCH` has exactly the same values as `VS_PLATFORM` - "x64", "ARM", "ARM64") except it has `x86` instead of `Win32`, but this value is never actually used, so both variables end up using interchangeably, making code harder to reason about - you needed to remember that they're exactly the same except the case when the variable is passed directly to VS (either through `cmake -A` or some `msbuild` command), then user must use `VS_PLATFORM`. So we might just use one variable always.
`START` is set to `0` for "amd64" / "x64" and `1` in all other cases, effectively just skipping `GENERATORS[0]` in case when "amd64" or "x64" is present.
Which is now useless, since `GENERATORS[0]` doesn't exist anymore.
They were added awhile ago in 67666ed for overcoming some win32 issues, which is not around anymore.
And now `VS_HOST` is never set (and therefore `VS_TOOLSET_HOST` is always empty too) and likely to be never used.
CI was failing:
```
The system library `dbus-1` required by crate `libdbus-sys` was not found.
The file `dbus-1.pc` needs to be installed and the PKG_CONFIG_PATH environment variable must contain its parent directory.
The PKG_CONFIG_PATH environment variable is not set.
```
There was an error building a standalone ifcwrap when previously built IfcOpenShell had rocksdb - `_ifcopenshell_wrapper.cpython-314-x86_64-linux-gnu.so: undefined symbol: _ZTV17RocksDbSerializer`.
`IFOPSH_WITH_ROCKSDB` propagated to the build, requiring rocksdb symbols, but `WITH_ROCKSDB` wasn't set and `document_serializer_rdb` target wasn't provided.
Otherwise it will never be able to reach standalone ifcwrap build as in this case ifcopenshell is built first as a dependency without `BUILD_IFCPYTHON`.
When fixing warnings, noticed that was working a bit inproperly - `bnode` always end up being an empty list (because there's no `brick, A, REF.IFCReference` triple), so then passing empty list to `triples` resulted in selecting all nodes isntead of just the expected `bnode` (that's the behaviour `sqlachemy` was sending the warnings about - when empty lists unexpectedly selected everything).
`None, None` assumed by default.
Though this was introduced in Python 3.13, before 3.13 it only breaks if we'd do `typing.Generator[T]` (which is deprecated) -`collections.abc.Generator[T]` works fine, it seems it never had an arity check.
Andrej730