merge_cubes: invalid spatial_extent merging

[soxofaan]

came up when trying to compile a small poc for cross-backend processing (Open-EO/openeo-aggregator#115):
merge_cubes only seems to consider the spatial extent of the first cube, ignoring the second.

e.g.

cube1 = connection.load_collection(
    "TERRASCOPE_S2_TOC_V2",
    temporal_extent=["2023-08-20", "2023-09-01"],
    spatial_extent={"west": 3.00, "south": 51.00, "east": 3.10, "north": 51.10},
    bands=["B02"],
 )
cube2 = connection.load_collection(
    "TERRASCOPE_S2_TOC_V2",
    temporal_extent=["2023-08-20", "2023-09-01"],
    spatial_extent={"west": 3.05, "south": 51.05, "east": 3.15, "north": 51.15},
    bands=["B02"],
)

cube = cube1.merge_cubes(cube2, overlap_resolver="max")
res = cube.save_result(format="netCDF")
job = res.execute_batch()

Result only covers {"west": 3.00, "south": 51.00, "east": 3.10, "north": 51.10} and does not extend to {..., "east": 3.15, "north": 51.15}

When switching both cubes (cube2.merge_cubes(cube1, overlap_resolver="max"), the result only covers {"west": 3.05, "south": 51.05, "east": 3.15, "north": 51.15}

(not sure yet if this is an issue in openeo-geopyspark-driver or openeo-python-driver, just reporting here for now)

[soxofaan]

(FYI, just tried to verify: merging of cubes where only temporal extent is different seems to work properly)