So happy that the Meltano Slack is back online. Delayed Christmas present 🎄 🙂 I am migrating from Gitlab CI to Github actions. Horrible, will write an article about it. One issue with GHA is how env variables can(not) be propagated to reusable workflows. As workaround I set variables in the UI and access them using

${{ vars.X }}

. But it seems that I forgot some variable (they are many). meltano run fails with the following error:

Cannot start plugin tap-github-org: expected string or bytes-like object, got 'NoneType'

It fails the same way for all taps(github, csv-s3, ...). The same target is used - snowflake. Anyone has a clue what could be the root cause? Could it be caused by not propagated env variable?

Oh, that error message is very vague and doesn't help narrow down where it's coming from 😥. None of the usual suspects produce that message:

>>> open(None)
TypeError: expected str, bytes or os.PathLike object, not NoneType

>>> import json; json.loads(None)
TypeError: the JSON object must be str, bytes or bytearray, not NoneType

>>> from pathlib import Path; Path(None)
TypeError: expected str, bytes or os.PathLike object, not NoneType

Thanks to `meltano --log-level debug`:

Cannot start plugin tap-github-org: Executable 'tap-github' could not be found. Extractor 'tap-github-org' may not have been installed yet using `meltano install extractor tap-github-org`, or the executable name may be incorrect

I build the docker image in the same way like on Gitlab (same Dockefile). It works on Gitlab. When I switch to related venv on my localhost and run

which tap-github

, it does not find anything. I was always wandering how this works. What could be the reason that the tap/target executable cannot be found?