Repository navigation
Conversation
Check Cargo dependencies and GitHub Actions weekly, grouping minor and patch updates. Ignore major updates of datafusion* and arrow*, which change only when main moves to the next DataFusion major version, and leave out the PySpark runner's pinned pyspark version. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
How do we want DataFusion updates to arrive? Minor and patch updates of If we want Dependabot to propose DataFusion updates, a group puts them in one PR. Leaving out datafusion:
applies-to: version-updates
patterns:
- "datafusion*"If we'd rather update DataFusion by hand, as Comet does in practice, we could drop |
|
Besides the
|
DataFusion is updated by hand, so Dependabot no longer proposes datafusion* updates (one PR per crate otherwise). Use increase-if-necessary so a requirement in Cargo.toml only moves when the new version needs it, and drop the redundant target-branch.
|
Thanks both. Pushed 2593ab0 and 155adcc. @mbutrovich I went with skipping
|
Which issue does this PR close?
Part of #56: the "Add a Dependabot configuration" item.
Rationale for this change
Nothing keeps the Cargo dependencies or the GitHub Actions in our workflows up to date. apache/datafusion and apache/datafusion-iceberg both use Dependabot for this.
What changes are included in this PR?
A
.github/dependabot.yml, based on apache/datafusion-iceberg's:datafusion*andarrow*are ignored. Under the versioning policy proposed in docs: add a versioning and API compatibility policy #319,datafusion-spark56.x stays on DataFusion 56 and thearrowversion it uses, so those major versions change only whenmainmoves to the next DataFusion major version.branch-56for now, so Dependabot can only update its locked commit.crates/sqllogictest/scripts/requirements.txtpinspyspark==4.2.0, the Spark version the sqllogictest files are checked against. Changing it is a decision, not a routine update.There are no
labels, because this repository has noauto-dependencieslabel yet.Are these changes tested?
The file is valid YAML and follows apache/datafusion-iceberg's working configuration. GitHub validates
dependabot.ymlonce it is on the default branch, and reports any errors in the repository's Dependabot settings.Are there any user-facing changes?
No.
This PR was written by an LLM (Claude) and reviewed before opening.
🤖 Generated with Claude Code