[Regression] Gerrit platform no longer assigns reviewers on change creation

[avm99963]

How are you running Renovate?

Self-hosted Renovate

Which platform you running Renovate on?

Gerrit

Which version of Renovate are you using?

All versions after commit fc3f7a8 (41.167.1 (2025-10-31))

Please tell us more about your question or problem

When setting the reviewers field in renovate.json, Renovate no longer assigns those reviewers when creating new changes in Gerrit.

After a Git bisection, it seems like the commit that introduced this regression is fc3f7a8 (CCing @felipecrs). But as I say below, I don't understand why it is causing this issue.

Git bisection details

Result:

fc3f7a8ee7eb74977153d8f03d401784f5041c00 is the first bad commit
commit fc3f7a8ee7eb74977153d8f03d401784f5041c00
Author: Felipe Santos <felipecassiors@gmail.com>
Date:   Fri Oct 31 04:47:19 2025 -0300

    fix(gerrit): remove memCache to fix prNo coming as null when pr-created (#38957)

 lib/modules/platform/gerrit/client.spec.ts |  5 +----
 lib/modules/platform/gerrit/client.ts      | 28 +++-------------------------
 lib/modules/platform/gerrit/index.spec.ts  | 14 --------------
 lib/modules/platform/gerrit/index.ts       | 19 ++-----------------
 lib/modules/platform/gerrit/scm.spec.ts    |  7 -------
 lib/modules/platform/gerrit/scm.ts         |  6 ------
 6 files changed, 6 insertions(+), 73 deletions(-)

Bisection log:

git bisect start
# status: waiting for both good and bad commits
# good: [6cedaa23058222d24705dac8f667b61a4c1d5cc1] chore(deps): update dependency esbuild to v0.25.11 (main) (#38792)
git bisect good 6cedaa23058222d24705dac8f667b61a4c1d5cc1
# status: waiting for bad commit, 1 good commit known
# bad: [b5ef5ed804cec5584f9a4ddddab0ee23f06919f0] chore(deps): update dependency renovatebot/github-action to v44.2.6 (main) (#40671)
git bisect bad b5ef5ed804cec5584f9a4ddddab0ee23f06919f0
# bad: [60f7ffc620047f1e63dcf672f7c9b4aedfa7a13c] feat(bitbucket-server): get username from header (#39554)
git bisect bad 60f7ffc620047f1e63dcf672f7c9b4aedfa7a13c
# bad: [d1062f1a2adb134c7341c1e0431a05035b920a45] chore(deps): update dependency @vitest/eslint-plugin to v1.4.1 (main) (#39285)
git bisect bad d1062f1a2adb134c7341c1e0431a05035b920a45
# bad: [19c396a44a4e02dbbe4f78a94fe788015772342a] chore(minimumReleaseAgeBehaviour): reduce number of `WARN` log lines (#39001)
git bisect bad 19c396a44a4e02dbbe4f78a94fe788015772342a
# good: [6963af4190f31fc110b32f4343693010d0b71ec6] build(deps): update node.js to v22.21.1 (main) (#38915)
git bisect good 6963af4190f31fc110b32f4343693010d0b71ec6
# bad: [f33076854f017ee2d247fc29921fd56da6ac8aaa] fix(presets): update repository URL for wasm-bindgen (#38976)
git bisect bad f33076854f017ee2d247fc29921fd56da6ac8aaa
# good: [b35e27d6bce26dd603a1e566a1c9f6492fa7883e] chore(renovate): mark major bumps with `breaking` label (#38942)
git bisect good b35e27d6bce26dd603a1e566a1c9f6492fa7883e
# good: [a6f52c8dc2099f5e8755a583fdcb54eef5e83b70] chore(deps): update dependency tar to v7.5.2 [security] (main) (#38964)
git bisect good a6f52c8dc2099f5e8755a583fdcb54eef5e83b70
# good: [e565f938f9ffe4e9f7b1c81997b90993c937efac] feat(deps): update ghcr.io/renovatebot/base-image docker tag to v10.69.1 (main) (#38971)
git bisect good e565f938f9ffe4e9f7b1c81997b90993c937efac
# good: [e565f938f9ffe4e9f7b1c81997b90993c937efac] feat(deps): update ghcr.io/renovatebot/base-image docker tag to v10.69.1 (main) (#38971)
git bisect good e565f938f9ffe4e9f7b1c81997b90993c937efac
# bad: [fc3f7a8ee7eb74977153d8f03d401784f5041c00] fix(gerrit): remove memCache to fix prNo coming as null when pr-created (#38957)
git bisect bad fc3f7a8ee7eb74977153d8f03d401784f5041c00
# bad: [fc3f7a8ee7eb74977153d8f03d401784f5041c00] fix(gerrit): remove memCache to fix prNo coming as null when pr-created (#38957)
git bisect bad fc3f7a8ee7eb74977153d8f03d401784f5041c00
# good: [cc86fce6593528facc4f1f363c8de36b330f1b8f] feat(deps): update ghcr.io/renovatebot/base-image docker tag to v10.70.0 (main) (#38972)
git bisect good cc86fce6593528facc4f1f363c8de36b330f1b8f
# first bad commit: [fc3f7a8ee7eb74977153d8f03d401784f5041c00] fix(gerrit): remove memCache to fix prNo coming as null when pr-created (#38957)

Technical details

The following line adds reviewers to newly created changes (or PRs using Renovate's terminology):

renovate/lib/workers/repository/update/pr/index.ts

Line 593 in dfa3798

await addParticipants(config, pr);

It only runs if existingPr is not truthy (there is a guard clause in lines 378-504).

existingPr is retrieved by calling platform.getBranchPr:

renovate/lib/workers/repository/update/pr/index.ts

Lines 173 to 176 in dfa3798

	// Check if PR already exists
	const existingPr =
	(await platform.getBranchPr(branchName, config.baseBranch)) ??
	(await tryReuseAutoclosedPr(branchName, prTitle));

Now, here is what happened with commit fc3f7a8:

• Before fc3f7a8, after Renovate created the branch (using Renovate's terminology) by pushing to refs/for/..., getBranchPr returned null and thus existingPr = null. This made Renovate add reviewers. And, in subsequent runs, getBranchPr returned the already existing PR correctly.
• Since fc3f7a8, getBranchPr always correctly returns the branch (using Renovate's terminology), so reviewers aren't added.

Now, I don't understand how that commit changed this behavior. My current hypothesis is that before that commit, for some reason the cache was still being used in some requests and that's why getBranchPr returned null the first time (but I can't see where). Maybe this was made in purpose so that Renovate behaved correctly?

However, I think the root cause is that Renovate treats branches and PRs as separate entities, but for Gerrit both are the same thing.

So I imagine we could do 2 things to fix this:

1. Like before, the first time that a branch/change is created in Gerrit, existingPr should be false. (this seems a little bit hacky to me!)
2. Change Renovate's business logic so that the nature of Gerrit can be integrated better, to make sure that relevant business logic will always run on PR/change creation.

@felipecrs: if you have any thoughts, I'm all ears! And thank you for the amazing work you're doing in the Gerrit platform!

Logs (if relevant)

No response

[felipecrs]

Hey, thanks for the report and the kind words!

This has been reported already, and I'm fixing it in #39250. I have been on vacation for the past 4 weeks, and I plan to resume work on it next week.

[felipecrs]

By the way, please feel free to chime in and let me know your opinion about my PR's approach.

[avm99963]

Thanks! I didn't find that report before, I see you're referring to #39250 (comment).

Will take a look at your PR a little bit later! I spent the entire morning troubleshooting the issue and need to do other things now :P

I hope you enjoyed/are enjoying your vacations! :)