In the context of https://github.com/Open-EO/openeo-geotrellis-extensions/issues/109, I created a new branch for openeo-geotrellis-extensions, in which I changed the versions of its submodules to 2.3.0_2.12-spark331, as to not interfere with other (-SNAPSHOT) versions in Artifactory.
To run the tests in openeo-geopyspark-driver against these jars, I created a new branch there, in which I updated pre_test.sh. It seems, however, that there are a couple of tests in tests/deploy/test_deploy.py that run against the actual jar files that pre_test.sh downloads; they fail because the jars have versions that no longer match the expected (-SNAPSHOT) versions in the tests so I had to make the tests more lenient as well:
Maybe we should run those tests against dedicated jar files that just contain a META-INF/MANIFEST.MF instead of the real ones. That way we're no longer tied to particular versions and we can tighten the tests again.
In the context of https://github.com/Open-EO/openeo-geotrellis-extensions/issues/109, I created a new branch for openeo-geotrellis-extensions, in which I changed the versions of its submodules to
2.3.0_2.12-spark331
, as to not interfere with other (-SNAPSHOT
) versions in Artifactory.To run the tests in openeo-geopyspark-driver against these jars, I created a new branch there, in which I updated
pre_test.sh
. It seems, however, that there are a couple of tests intests/deploy/test_deploy.py
that run against the actual jar files thatpre_test.sh
downloads; they fail because the jars have versions that no longer match the expected (-SNAPSHOT
) versions in the tests so I had to make the tests more lenient as well:Maybe we should run those tests against dedicated jar files that just contain a
META-INF/MANIFEST.MF
instead of the real ones. That way we're no longer tied to particular versions and we can tighten the tests again.