Fix ST_Affine()

[yutannihilation]

Fix #591

I believe this is a simple mistake around the indices. I still haven't succeeded to build duckdb-spatial locally, so all I can do is to rely on CI... (edit: I confirmed the test passes on my local)

[yutannihilation]

Should I have created pull requests against v1.3.0 branch instead of main? It seems all checks are failing...

[Maxxen]

Hello! Thanks a lot for the PR! Looks great!

The branch configuration is a bit confusing... generally we do most development on the branch corresponding to the latest stable duckdb version (v1.3.0) which has the DuckDB version in the CI and submodule "pinned". Whatever gets built by CI on this branch also becomes available for users by installing from the 'core_nightly' repository, e.g. FORCE INSTALL spatial FROM core_nightly. Although, sometimes new features in spatial are dependent on yet-to-be-released DuckDB changes, in which case we merge them into main. However, when we are getting close to a new release, we merge the latest stable branch into main and start to shift development to the main branch, to consolidate the changes thats happened both in DuckDB core and spatial so far and make sure spatial is in a good state for the next DuckDB release.

[yutannihilation]

Thanks for the detailed explanation! I got it. Since my branches are still based on main, the CI probably will keep failing. I will merge the latest v1.3.0 branch.

[Maxxen]

No problem! The windows mingw CI still fails randomly sometime due to a vcpkg openssl issue, but dont worry about that!