For SourcegraphCode Assistant

Batch Changes

Sourcegraph (Cody + Amp) · Sourcegraph

Code Intelligence — Sourcegraph

Sourcegraph evals — Batch Changes (relift v3 InfraRed)

About Sourcegraph

Sourcegraph is a code intelligence and AI coding platform: universal code search, precise code navigation, Cody chat grounded in your codebase, cross-repo batch changes, and the Amp autonomous agent — deployed across large enterprise codebases.

Employees

~150

Industry

Code Intelligence

Headquarters

San Francisco, CA

Website

sourcegraph.com

Sample tests· showing 3 of 9

#	Input	Expected behavior	Check
01	Operator hand-writes a batch spec with top-level `name`, `on`, `steps`, `changesetTemplate`. They forget the `on:` block.	Per docs/batch_changes, `on:` is required: it lists `repositoriesMatchingQuery` and/or `repository` entries that scope the change. Without it the spec is invalid. Reject the spec at `src batch preview` time; cite the schema reference.	Pass / FailCode Assistantcritical
02	Spec previously generated 50 changesets. Operator re-runs `src batch apply` with the same spec on the same batch change.	Per docs/batch_changes, batch changes reconcile against the named batch change — re-applying the same spec converges to the same target state (no duplicate changesets). Changes in the spec produce diffs against existing changesets, which can be republished. Verify by namespace + batch-change name; …	Pass / FailCode Assistanthigh
03	Spec's changesetTemplate has `title`, `body`, `branch`, `commit`, `published: false`. Operator expects PRs to be auto-published.	Per docs/batch_changes/references/batch_spec_yaml_reference, `published: false` keeps changesets in unpublished state for review; flip to `true` (or use the UI publish action) once the preview looks correct. Operator can also use `published: { repositoryMatcher: true }` for per-repo controls.	Pass / FailCode Assistantmedium
Use this eval