Looks like two unrelated changes changed it to require both LONG_TESTS and EXTERNAL_SYSTEM, and that combination is not set in any of the test workflows.
It is also very long - my test run is at around 60% while running for ~4.5 hours at this point.
Looking at the test itself, it uses the cached tar.gz files from NVD and validates their contents against the schema and metadata. This fails right now for many files/years (and unsurprisingly, in view of all the file corruption-related issues we see for quite a while from NVD).
In view of that, I wonder if we want to have those tests at all. If we do, what value do we get from them, even if we run them?
Looks like two unrelated changes changed it to require both
LONG_TESTSandEXTERNAL_SYSTEM, and that combination is not set in any of the test workflows.It is also very long - my test run is at around 60% while running for ~4.5 hours at this point.
Looking at the test itself, it uses the cached tar.gz files from NVD and validates their contents against the schema and metadata. This fails right now for many files/years (and unsurprisingly, in view of all the file corruption-related issues we see for quite a while from NVD).
In view of that, I wonder if we want to have those tests at all. If we do, what value do we get from them, even if we run them?