Use engine commit hash when vendoring sources #172
Conversation
This PR brings #duckdb/duckdb-odbc/84 changes to JDBC repo. Testing: no code changes, no new tests.
| @@ -58,6 +59,10 @@ jobs: | |||
| run: | | |||
| git push -u origin HEAD | |||
|
|
|||
| - name: Wait for push to complete | |||
| if: steps.vendor.outputs.vendor != '' && github.event_name != 'pull_request' | |||
| run: sleep 10 # Give GitHub a moment to process the push | |||
There was a problem hiding this comment.
Could you expand on why this sleep is necessary? Doesn't the push complete otherwise?
There was a problem hiding this comment.
M, I've just copied this bit from the latest ODBC workflow to keep JDBC in sync with it. I will re-check.
There was a problem hiding this comment.
As an update on this, I've got some input on GH workflows from Carlo, as a result the vendoring process in ODBC is being changed to import engine sources into a separate branch, create a PR on this branch and merge this PR automatically. In this variant the testing workflow is dispatched by triggering the pull_request event instead of a direct dispatch, so we do not need to wait manually until the push completes. I intend to bring the same workflow changes to JDBC once they are approved for ODBC, keeping this PR as a "draft" for now.
This PR brings duckdb/duckdb-odbc#84 changes to JDBC repo.
Testing: no code changes, no new tests.