[kushuh]

Resolves #3711

---

Before the change?

• getFileCommit listed every commit that touched the file (GET /repos/{owner}/{repo}/commits?sha={ref}&path={file}, 30 per page, all pages) before checking any of them. Each read of the github_repository_file data source, and each read of the resource that does not know its commit_sha yet, cost one request per 30 commits of the file's history.

After the change?

• It checks each page as it arrives and returns at the first non-merge commit that changed the file without removing it, which is the same commit as before. A read now costs the contents request, one page of commits and the commit read, however long the history.
• The commit's file pagination moves, unchanged, into getCommitWithFiles.
• Measured on a real repository with a read-only token, reading 20 workflow files: v6.13.0 spent 109 requests and this build 54, with identical content, commit_sha, commit_author, commit_email and commit_message for every file.

Pull request checklist

• Schema migrations have been created if needed (none: the schema does not change)
• Tests for the changes have been added (for bug fixes / features): Test_getFileCommit asserts the commit returned and the exact requests made. On the current code it fails on the extra page request alone.
• Docs have been reviewed and added / updated if needed (for bug fixes / features): no documented behavior changes.

Does this introduce a breaking change?

Please see our docs on breaking changes to help!

• Yes
• No

---

This change is largely AI-generated, written with Claude Code. make test and golangci-lint (standard and the strict rules on new code) pass locally.

[github-actions]

👋 Hi, and thank you for this contribution!

This repo is maintained by GitHub and community members on a best-effort basis. We'll get to this as soon as we can.

You can help us prioritize by joining the discussion on open issues and PRs, sharing details on the changes you need, and reviewing other contributions.

---

_{🤖 This is an automated message.}