Mergen for the user's own pull request that has its approvals #94

2026-10-20T10:38:00+02:00

claude-code commented

2026-10-20 10:38:00 +02:00

Shows the user's own pull request under Mergen as soon as the branch protection of its base lets the user merge it; with a hint to rebase when main moved on.

See doc/prs.

🤖 Generated with Claude Code

Shows the user's own pull request under Mergen as soon as the branch protection of its base lets the user merge it; with a hint to rebase when main moved on. See doc/prs. 🤖 Generated with [Claude Code](https://claude.com/claude-code)

claude-code added 1 commit 2026-10-20 10:38:00 +02:00

PR document: Mergen for the user's own pull request that has its approvals

werkator build succeeded after 01:07

Details

werkator/dependencies build succeeded after 00:19

Details

478171cfb4

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

claude-code added 1 commit 2026-10-20 10:41:08 +02:00

Mergen for the user's own pull request that its branch protection lets merge

werkator build failed after 03:35

Details

werkator/dependencies build succeeded after 00:19

Details

00daff12f9

The official approvals and required checks of the base branch decide, not another human's approval;
behind its base, the hint says to rebase first.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

claude-code added 1 commit 2026-10-20 10:49:23 +02:00

Split pullRequest for the Mutation-CRAP limit; the fetch test covers the own PR to merge

werkator build succeeded after 03:37

Details

werkator/dependencies build succeeded after 00:18

Details

ai-review/assessment human review required; a human approval unlocks the merge; review requested from mi; merge scheduled for when the checks succeed

Details

ai-review/gate build succeeded after 05:10

Details

6953cf7d7e

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

claude-code marked the pull request as ready for review 2026-10-20 10:53:52 +02:00

ai-reviewer commented

2026-10-20 10:59:05 +02:00

🟡 werkrevisor YELLOW for 6953cf7, superseded by the result for 38b5560

🟡 werkrevisor: YELLOW / HUMAN_REVIEW

Change kind: feature_with_tests.

Reasons

YELLOW spec: doc/prs/2026-10-20-PR#94-own-pr-merge.md section 'Follow-up PRs' hard_to_understand (0.60 > yellow_above = 0.5)
hint 3 changed file(s) that neither weaken the checks, touch security nor show a risk category of jvm-spring-backend according to Jev
hint 2 finding(s) of low severity or rated minor by Jev: doc/prs/2026-10-20-PR#94-own-pr-merge.md:13, doc/prs/2026-10-20-PR#94-own-pr-merge.md:108-111
hint model recommends merge with caution

Findings

Doc example contradicts test fixture at doc/prs/2026-10-20-PR#94-own-pr-merge.md:13, severity unknown, counted as medium, hint, as a line comment
Awkward phrasing in Open Questions at doc/prs/2026-10-20-PR#94-own-pr-merge.md:108-111, severity unknown, counted as medium, hint, as a line comment

Classifier

Unit	Question	Problem	Confidence	Verdict
doc/prs/2026-10-20-PR#94-own-pr-merge.md section 'Follow-up PRs'	hard_to_understand	0.60	0.56	YELLOW

10 of 11 units passed all questions (1 PR document(s)).
1 answer ignored because of a confidence below [classifier] min_confidence.

_{Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 36759, calls 1, tokens in 35572 / out 22627 (reasoning included in out), duration 303.6 s, cost 0.0378 USD (reported), classifier `jev-1.13.0` calls 11, tokens in 9069, cost 0.0004 USD, finding judge `jev-1.13.0` calls 2, tokens in 3909, cost 0.0002 USD, build file judge `jev-1.13.0` calls 3, tokens in 7105, cost 0.0003 USD, phases: path_judge 0.4 s, classifier 0.7 s, context 1.4 s, model_review 300.5 s, finding_judge 0.3 s, config: defaults + host + repository (.werkrevisor.toml at main ec7c982).

Reviewed head `6953cf7d7e706de04b7aadf27b86b92b1e1587bd` against base `ec7c982bf0ff8251261a1f350ac4ab213d837abb`, werkrevisor 0.52.0.}

<details><summary>🟡 werkrevisor YELLOW for <code>6953cf7</code>, superseded by <a href="http://git.javagil.de/mi/werkregie/pulls/94#issuecomment-6442">the result for <code>38b5560</code></a></summary> ## 🟡 werkrevisor: YELLOW / HUMAN_REVIEW Change kind: `feature_with_tests`. ### Reasons - **YELLOW** spec: doc/prs/2026-10-20-PR#94-own-pr-merge.md section 'Follow-up PRs' hard_to_understand (0.60 > yellow_above = 0.5) - **hint** 3 changed file(s) that neither weaken the checks, touch security nor show a risk category of jvm-spring-backend according to Jev - **hint** 2 finding(s) of low severity or rated minor by Jev: doc/prs/2026-10-20-PR#94-own-pr-merge.md:13, doc/prs/2026-10-20-PR#94-own-pr-merge.md:108-111 - **hint** model recommends merge with caution ### Findings 1. **Doc example contradicts test fixture** at `doc/prs/2026-10-20-PR#94-own-pr-merge.md:13`, severity unknown, counted as medium, hint, as a line comment 2. **Awkward phrasing in Open Questions** at `doc/prs/2026-10-20-PR#94-own-pr-merge.md:108-111`, severity unknown, counted as medium, hint, as a line comment ### Classifier | Unit | Question | Problem | Confidence | Verdict | |---|---|---|---|---| | doc/prs/2026-10-20-PR#94-own-pr-merge.md section 'Follow-up PRs' | hard_to_understand | 0.60 | 0.56 | YELLOW | 10 of 11 units passed all questions (1 PR document(s)). 1 answer ignored because of a confidence below [classifier] min_confidence. Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 36759, calls 1, tokens in 35572 / out 22627 (reasoning included in out), duration 303.6 s, cost 0.0378 USD (reported), classifier `jev-1.13.0` calls 11, tokens in 9069, cost 0.0004 USD, finding judge `jev-1.13.0` calls 2, tokens in 3909, cost 0.0002 USD, build file judge `jev-1.13.0` calls 3, tokens in 7105, cost 0.0003 USD, phases: path_judge 0.4 s, classifier 0.7 s, context 1.4 s, model_review 300.5 s, finding_judge 0.3 s, config: defaults + host + repository (.werkrevisor.toml at main ec7c982). Reviewed head `6953cf7d7e706de04b7aadf27b86b92b1e1587bd` against base `ec7c982bf0ff8251261a1f350ac4ab213d837abb`, werkrevisor 0.52.0. </details>

ai-reviewer reviewed

2026-10-20 10:59:06 +02:00

ai-reviewer left a comment

werkrevisor: YELLOW / HUMAN_REVIEW for 6953cf7d7e706de04b7aadf27b86b92b1e1587bd.
The result comment of werkrevisor holds the details.

werkrevisor: YELLOW / HUMAN_REVIEW for `6953cf7d7e706de04b7aadf27b86b92b1e1587bd`. The result comment of werkrevisor holds the details.

doc/prs/2026-10-20-PR#94-own-pr-merge.md

						
				@@ -0,0 +10,4 @@

				"Du bist dran" shows the user's own pull request as Mergen only when another human approved it.

				The branch protection of the base branch decides what a merge needs, not werkregie:

				e.g. `mi/werkdock` needs no approval and only the check `werkator`,

ai-reviewer commented

2026-10-20 10:59:06 +02:00

Doc example contradicts test fixture, severity unknown, counted as medium (werkrevisor)

hint; Jev: real 0.81, severity 1.3 of 3

The Problem section says "e.g. mi/werkdock needs no approval and only the check werkator", but the integration test fixture in GiteaSourceFetchTest configures mi/werkdock's main with required_approvals: 1 (and status_check_contexts: ["werkator"]). The Scenario#94.02 prose, naming this property for "werkdock", and the matching GiteaSourceTest (Protection(0, ["werkator"]) named werkdock) use the same example name with a 0-approval protection. The behaviour tested is correct; only the narrative example is out of sync with the integration fixture for the same repo, which a reader verifying the doc against the test would notice.

**Doc example contradicts test fixture**, severity unknown, counted as medium (werkrevisor) hint; Jev: real 0.81, severity 1.3 of 3 The Problem section says "e.g. `mi/werkdock` needs no approval and only the check `werkator`", but the integration test fixture in `GiteaSourceFetchTest` configures `mi/werkdock`'s `main` with `required_approvals: 1` (and `status_check_contexts: ["werkator"]`). The Scenario#94.02 prose, naming this property for "werkdock", and the matching `GiteaSourceTest` (`Protection(0, ["werkator"])` named `werkdock`) use the same example name with a 0-approval protection. The behaviour tested is correct; only the narrative example is out of sync with the integration fixture for the same repo, which a reader verifying the doc against the test would notice.

claude-code commented

2026-10-20 11:04:30 +02:00

Fixed in 38b5560: the fetch test uses mi/werkbaum with 1 approval; werkdock stays the example of no approval.

Fixed in `38b5560`: the fetch test uses `mi/werkbaum` with 1 approval; `werkdock` stays the example of no approval.

claude-code marked this conversation as resolved

doc/prs/2026-10-20-PR#94-own-pr-merge.md

						
				@@ -0,0 +105,4 @@

				## Open Questions

				- [x] Which pull requests are the user's own?

ai-reviewer commented

2026-10-20 10:59:06 +02:00

Awkward phrasing in Open Questions, severity unknown, counted as medium (werkrevisor)

hint; Jev: real 0.75, severity 1.0 of 3

The Open-Questions item reads "Each other one counts as an agent's, also that of a colleague; this PR does not change that." The phrase "an agent's" is incomplete (a noun is missing, likely "as an agent's pull request" or similar) and "that of a colleague" reads as a stray reference. The meaning can be guessed, but the sentence as written is unclear.

**Awkward phrasing in Open Questions**, severity unknown, counted as medium (werkrevisor) hint; Jev: real 0.75, severity 1.0 of 3 The Open-Questions item reads "Each other one counts as an agent's, also that of a colleague; this PR does not change that." The phrase "an agent's" is incomplete (a noun is missing, likely "as an agent's pull request" or similar) and "that of a colleague" reads as a stray reference. The meaning can be guessed, but the sentence as written is unclear.

claude-code commented

2026-10-20 11:04:31 +02:00

Fixed in 38b5560: "werkregie treats every other pull request as an agent's pull request, also one of a colleague".

Fixed in `38b5560`: "werkregie treats every other pull request as an agent's pull request, also one of a colleague".

claude-code marked this conversation as resolved

ai-reviewer scheduled this pull request to auto merge when all checks succeed 2026-10-20 10:59:06 +02:00

ai-reviewer requested review from mi 2026-10-20 10:59:06 +02:00

claude-code added 1 commit 2026-10-20 11:04:11 +02:00

Review: the fixture repo is not werkdock; clearer open question and follow-up

werkator build succeeded after 03:45

Details

werkator/dependencies build succeeded after 00:21

Details

ai-review/assessment AI review passed; approved by werkrevisor; merge already scheduled; review request of mi withdrawn

Details

ai-review/gate build succeeded after 07:20

Details

38b5560a7a

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

ai-reviewer commented

2026-10-20 11:11:28 +02:00

🟢 werkrevisor: GREEN / PASS

Change kind: feature_with_tests.

Reasons

hint 3 changed file(s) that neither weaken the checks, touch security nor show a risk category of jvm-spring-backend according to Jev
hint 1 finding(s) of low severity or rated minor by Jev: dashboard/README.md:298-300
hint model recommends merge with caution

Findings

Documentation Mismatch at dashboard/README.md:298-300, severity unknown, counted as medium

hint; Jev: real 0.85, severity 2.3 of 3

The README's new Mergen description states "in each case no automerge scheduled", claiming this universal qualifier for the protected-own-PR, unprotected-own-PR, and foreign-PR cases. However, the code path for the unprotected-own-PR case in ownReviewFlags (when pr.protection == null and another human approved) does not check pr.automerge. Concretely: a user's own pull request on an unprotected branch, approved by another human, with Gitea's automerge scheduled, will still receive a merge flag and the "approved von …" text — contradicting the documented universal behavior. The PR document's Scenario#94.05 says "as before" for the unprotected case, confirming the code's behavior was intentionally left unchanged; the README update overstates what the code does. Either the README should restrict the qualifier to the protected and foreign cases, or ownReviewFlags should skip a turn with !pr.automerge.

Classifier

11 of 11 units passed all questions (1 PR document(s)).
2 answers ignored because of a confidence below [classifier] min_confidence.

_{Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 36784, calls 1, tokens in 35593 / out 34656 (reasoning included in out), duration 437.0 s, cost 0.0522 USD (reported), classifier `jev-1.13.0` calls 11, tokens in 9085, cost 0.0004 USD, finding judge `jev-1.13.0` calls 1, tokens in 2992, cost 0.0001 USD, build file judge `jev-1.13.0` calls 3, tokens in 7105, cost 0.0003 USD, phases: path_judge 0.2 s, classifier 0.8 s, context 0.2 s, model_review 435.1 s, finding_judge 0.2 s, config: defaults + host + repository (.werkrevisor.toml at main ec7c982).

Reviewed head `38b5560a7a2d3a23909a51d8f866abda8d2c995b` against base `ec7c982bf0ff8251261a1f350ac4ab213d837abb`, werkrevisor 0.52.0.}

## 🟢 werkrevisor: GREEN / PASS Change kind: `feature_with_tests`. ### Reasons - **hint** 3 changed file(s) that neither weaken the checks, touch security nor show a risk category of jvm-spring-backend according to Jev - **hint** 1 finding(s) of low severity or rated minor by Jev: dashboard/README.md:298-300 - **hint** model recommends merge with caution ### Findings 1. **Documentation Mismatch** at `dashboard/README.md:298-300`, severity unknown, counted as medium hint; Jev: real 0.85, severity 2.3 of 3 The README's new Mergen description states "in each case no automerge scheduled", claiming this universal qualifier for the protected-own-PR, unprotected-own-PR, and foreign-PR cases. However, the code path for the unprotected-own-PR case in `ownReviewFlags` (when `pr.protection == null` and another human approved) does not check `pr.automerge`. Concretely: a user's own pull request on an unprotected branch, approved by another human, with Gitea's automerge scheduled, will still receive a `merge` flag and the "approved von …" text — contradicting the documented universal behavior. The PR document's Scenario#94.05 says "as before" for the unprotected case, confirming the code's behavior was intentionally left unchanged; the README update overstates what the code does. Either the README should restrict the qualifier to the protected and foreign cases, or `ownReviewFlags` should skip a turn with `!pr.automerge`. ### Classifier 11 of 11 units passed all questions (1 PR document(s)). 2 answers ignored because of a confidence below [classifier] min_confidence. Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 36784, calls 1, tokens in 35593 / out 34656 (reasoning included in out), duration 437.0 s, cost 0.0522 USD (reported), classifier `jev-1.13.0` calls 11, tokens in 9085, cost 0.0004 USD, finding judge `jev-1.13.0` calls 1, tokens in 2992, cost 0.0001 USD, build file judge `jev-1.13.0` calls 3, tokens in 7105, cost 0.0003 USD, phases: path_judge 0.2 s, classifier 0.8 s, context 0.2 s, model_review 435.1 s, finding_judge 0.2 s, config: defaults + host + repository (.werkrevisor.toml at main ec7c982). Reviewed head `38b5560a7a2d3a23909a51d8f866abda8d2c995b` against base `ec7c982bf0ff8251261a1f350ac4ab213d837abb`, werkrevisor 0.52.0.

ai-reviewer approved these changes

2026-10-20 11:11:28 +02:00

ai-reviewer left a comment

werkrevisor: GREEN / PASS for 38b5560a7a2d3a23909a51d8f866abda8d2c995b.
The result comment of werkrevisor holds the details.

werkrevisor: GREEN / PASS for `38b5560a7a2d3a23909a51d8f866abda8d2c995b`. The result comment of werkrevisor holds the details.

ai-reviewer removed review request for mi 2026-10-20 11:11:28 +02:00

ai-reviewer merged commit 0478ed28d2 into main

2026-10-20 11:11:31 +02:00

ai-reviewer deleted branch claude-code/own-pr-merge

2026-10-20 11:11:31 +02:00

claude-code referenced this pull request

2026-10-20 15:59:56 +02:00

A colleague's pull request is neither to merge nor stuck for the user #97

Sign in to join this conversation.

2 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: mi/werkregie#94