Skip to content

[SPARK-59280][BUILD] Expand and group sparktestsupport ignore patterns - #58555

Open
nchammas wants to merge 4 commits into
apache:masterfrom
nchammas:more-test-ignores
Open

[SPARK-59280][BUILD] Expand and group sparktestsupport ignore patterns#58555
nchammas wants to merge 4 commits into
apache:masterfrom
nchammas:more-test-ignores

Conversation

@nchammas

@nchammas nchammas commented Sep 7, 2026

Copy link
Copy Markdown
Contributor

What changes were proposed in this pull request?

  1. Expand the ignore patterns in sparktestsupport to cover almost all cases where we do not want any test modules to fire.
  2. Group the ignore patterns by rough domain or topic.
  3. Sort the patterns within each group.

Why are the changes needed?

  1. I conducted an AI-assisted analysis of the most recent 2,000 PRs that were merged to master. I found that about 4% of them triggered a full CI run that they could have completely skipped.
    Here are 10 examples, including the CI time consumed: #55685 (30h), #58072 (28h), #55800 (28h), #55878 (27h), #57972 (27h), #56996 (27h), #57741 (27h), #58206 (27h), #58098 (27h), #55972 (27h)
  2. This is the first of multiple PRs I am planning to cut CI waste and improve the precision of what tests we run for each PR. Some of these groups will become new test modules down the line. Most will just stay as ignore patterns. The groups and group comments will make it easier for maintainers to track why certain files don't trigger tests.
  3. Sorting the patterns makes future inserts easier to place and makes it more obvious to the reader when there are unintentional gaps.

Does this PR introduce any user-facing change?

No.

How was this patch tested?

Existing doctests and CI.

Was this patch authored or co-authored using generative AI tooling?

I wrote this patch and the associated analysis scripts (not posted; happy to share if interested) with assistance from GitHub Copilot.

@nchammas
nchammas requested a review from uros-b September 10, 2026 06:56

@uros-b uros-b left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This looks good to me, thank you @nchammas! Let's please ping someone with a bit more context to review this too, e.g. @zhengruifeng and/or @HyukjinKwon

"/.github/workflows/build_codegen_jdk.yml",
"/.github/workflows/build_coverage.yml",
"/.github/workflows/build_java*.yml",
"/.github/workflows/build_main.yml",

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure if we should skip these (and the docker file). If our testing environment is changed, we should run a full test against the new environment right?

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

yes, basically a infra change triggers a full test

"/sbin/decommission-worker.sh",
"/sbin/spark-daemons.sh",
"/sbin/start-all.sh",
"/sbin/start-connect-server.sh",

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

If I remember correctly, at least some of these shell scripts are directly used by some tests (local_server_pool or something) and can't be ignored.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants