Matrix rows lead with room and sender #15
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: mi/werkregie#15
Reference in New Issue
Block a user
Rows of the panel "Nachrichten" start with the room in the accent colour and the sender in bold; the start of the text follows in normal weight. Werkbaum:
#dash.taigaplanned (Additional Change).PR document:
doc/prs/2026-10-11-PR#15-matrix-row-layout.md. Tests green; checked with screenshots of the demo in light and dark. Version 0.6.2.🤖 Generated with Claude Code
🟢 werkrevisor GREEN for
63c5dac, superseded by the result for7c0af01🟢 werkrevisor: GREEN / PASS
Gate: approved by werkrevisor; merge waits for 2 open hint(s): reply to each line comment and resolve it, then restart the review or push a commit.
Change kind:
mixed.Reasons
Findings
[medium] Undocumented werkbaum additions at
doc/prs/2026-10-11-PR#15-matrix-row-layout.md:63-65, hint, as a line comment[low] Scenario#15.02 verification lists non-direct rows at
doc/prs/2026-10-11-PR#15-matrix-row-layout.md:38-48, hint, as a line commentClassifier
8 of 8 units passed all questions (1 PR document(s)).
Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 37401, calls 1, tokens in 36236 / out 15571 (reasoning included in out), duration 111.4 s, cost 0.0290 USD (reported), classifier `jev-1.13.0` calls 8, tokens in 5710, cost 0.0002 USD, finding judge `jev-1.13.0` calls 1, tokens in 1621, cost 0.0001 USD, build file judge `jev-1.13.0` calls 5, tokens in 6984, cost 0.0003 USD, phases: path_judge 0.5 s, classifier 0.5 s, context 0.2 s, model_review 109.6 s, finding_judge 0.2 s, config: defaults + host + repository (.werkrevisor.toml at main21cf21a).Reviewed head `63c5dac6529687aea08663133d418a521ca7388c` against base `21cf21a8492d881ed0729ccd7e6348224cc4d150`, werkrevisor 0.32.2.
werkrevisor: GREEN / PASS for
63c5dac6529687aea08663133d418a521ca7388c.The result comment of werkrevisor holds the details.
@@ -0,0 +35,4 @@- Manual: screenshots of the demo in light and dark, rows of Anna, "Team Shop · Ben" and "Release-Planung · Carla".#### Scenario#15.02: a direct message names its sender first[low] Scenario#15.02 verification lists non-direct rows (werkrevisor)
hint
Scenario#15.02 describes a direct message, but its "Verified by" text references three demo rows including
Team Shop · Ben(a mention) andRelease-Planung · Carla(a room). Neither demonstrates the direct-message behaviour the scenario claims; only the Anna row does, so the verification text overstates what the screenshots show for this scenario.Fixed in
7c0af01: Scenario#15.02 names the direct message of Anna only, Scenario#15.01 the two room rows.@@ -0,0 +60,4 @@None.## Additional Changes[medium] Undocumented werkbaum additions (werkrevisor)
hint; Jev: real 0.50, severity 1.6 of 3
The PR description and the PR document's "Additional Changes" section mention only the Taiga addition, but the werkbaum diff also adds four unrelated planned items:
Device verified by emoji comparison,Relays on the Omarchy hosts,Push notification when a pull request needs the user, andOlder messages than the first sync brings. None of these is declared in either the PR description or the PR document. Either drop them from this PR or extend the Additional Changes section so the PR document reflects the actual scope.No change: the Werkbaum diff of this PR adds only
#dash.taigawith its description block and one description line under "Panel for the messages" (git diff origin/main -- doc/werkregie.werkbaum). The four items named here exist on main already; they start with+, the Werkbaum gate for an optional node, not a diff marker.New commits pushed, approval review dismissed automatically according to repository settings
🟢 werkrevisor GREEN for
7c0af01, superseded by the result for2632d12🟢 werkrevisor: GREEN / PASS
Gate: approved by werkrevisor; merge waits for 2 open hint(s): reply to each line comment and resolve it, then restart the review or push a commit.
Change kind:
mixed.Reasons
63c5dacnot carried over: 1 approved block(s) are missingFindings
Undocumented Werkbaum Additions at
doc/prs/2026-10-11-PR#15-matrix-row-layout.md:63-65, severity unknown, counted as medium, hint, as a line commentImprecise Wording at
doc/prs/2026-10-11-PR#15-matrix-row-layout.md:54-55, severity unknown, counted as medium, hint, as a line commentClassifier
8 of 8 units passed all questions (1 PR document(s)).
Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 37390, calls 1, tokens in 36224 / out 14010 (reasoning included in out), duration 90.6 s, cost 0.0276 USD (reported), classifier `jev-1.13.0` calls 8, tokens in 5710, cost 0.0002 USD, finding judge `jev-1.13.0` calls 2, tokens in 3342, cost 0.0001 USD, build file judge `jev-1.13.0` calls 5, tokens in 6984, cost 0.0003 USD, phases: path_judge 0.5 s, classifier 0.5 s, context 0.3 s, model_review 88.7 s, finding_judge 0.2 s, config: defaults + host + repository (.werkrevisor.toml at main21cf21a).Reviewed head `7c0af01f85727ef03a11da22a21868fc0856a69f` against base `21cf21a8492d881ed0729ccd7e6348224cc4d150`, werkrevisor 0.32.2.
werkrevisor: GREEN / PASS for
7c0af01f85727ef03a11da22a21868fc0856a69f.The result comment of werkrevisor holds the details.
@@ -0,0 +51,4 @@- The page splits the detail of a Matrix row, "room · sender", at its last separator;a direct chat has the sender alone.- A row starts with the room in the accent colour and the sender in bold, then the start of the text in normal weight;Imprecise Wording, severity unknown, counted as medium (werkrevisor)
hint; Jev: real 0.71, severity 1.3 of 3
The Solution section states "the meta line keeps the origin only", while the code only suppresses the
detailspan for messages and still rendersprojectwhenever it is set. For the current demo entries (no project) this is true, but the statement does not match the general code path, so a reader who later sees a Matrix row with a project in the meta line may find it contradicting the PR document.Fixed in
2632d12: the Solution says the meta line no longer repeats room and sender.@@ -0,0 +60,4 @@None.## Additional ChangesUndocumented Werkbaum Additions, severity unknown, counted as medium (werkrevisor)
hint; Jev: real 0.60, severity 1.4 of 3
The PR document's Additional Changes section mentions only
#dash.taiga, but the same PR also adds five other entries todoc/werkregie.werkbaumthat are not documented: emoji-based device verification, older messages than the first sync, relays on Omarchy hosts, push notifications for pull requests, and a new#dash.notifyidea with its description block. The PR description also lists only Taiga. The PR document therefore understates the scope of what the PR actually changes in the planning file.No change to the Werkbaum:
git diff origin/main -- doc/werkregie.werkbaumadds only#dash.taiga, its description block and one description line. The named nodes,#dash.notifyincluded, are on main;+is the Werkbaum gate of an optional node. Clarified in2632d12under Additional Changes.New commits pushed, approval review dismissed automatically according to repository settings
🟢 werkrevisor GREEN for
2632d12, superseded by the result for015198f🟢 werkrevisor: GREEN / PASS
Gate: approved by werkrevisor; merge waits for 1 open hint(s): reply to each line comment and resolve it, then restart the review or push a commit.
Change kind:
mixed.Reasons
7c0af01not carried over: 1 approved block(s) are missingFindings
doc/prs/2026-10-11-PR#15-matrix-row-layout.md:65-66, severity unknown, counted as medium, hint, as a line commentClassifier
8 of 8 units passed all questions (1 PR document(s)).
Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 37565, calls 1, tokens in 36387 / out 31711 (reasoning included in out), duration 208.0 s, cost 0.0489 USD (reported), classifier `jev-1.13.0` calls 8, tokens in 5737, cost 0.0002 USD, finding judge `jev-1.13.0` calls 1, tokens in 1711, cost 0.0001 USD, build file judge `jev-1.13.0` calls 5, tokens in 6984, cost 0.0003 USD, phases: path_judge 0.5 s, classifier 0.5 s, context 0.3 s, model_review 206.2 s, finding_judge 0.2 s, config: defaults + host + repository (.werkrevisor.toml at main1 answer ignored because of a confidence below [classifier] min_confidence.
21cf21a).Reviewed head `2632d124ecaa97a8b791e49e2d3e6ff5b3ae9cc5` against base `21cf21a8492d881ed0729ccd7e6348224cc4d150`, werkrevisor 0.32.2.
werkrevisor: GREEN / PASS for
2632d124ecaa97a8b791e49e2d3e6ff5b3ae9cc5.The result comment of werkrevisor holds the details.
@@ -0,0 +62,4 @@## Additional Changes- Werkbaum: `#dash.taiga`, a Taiga integration, planned; its scope is open.Unclear statement about werkbaum additions, severity unknown, counted as medium (werkrevisor)
hint; Jev: real 0.77, severity 1.1 of 3
The "Additional Changes" section says "No other node changes: lines that start with
+are optional nodes already onmain, not added ones." The sentence is hard to parse: it is unclear whether "lines that start with+" refers to the+Werkbaum file-content prefix (as in+ [ ] Device verified by emoji comparison (S)), to the diff+marker, or to something else. The diff fordoc/werkregie.werkbaumappears to contain several lines flagged with+(e.g., "Device verified by emoji comparison", "Older messages than the first sync brings", "Relays on the Omarchy hosts", "Push notification when a pull request needs the user", "#dash.notify"), most of which seem unrelated to the PR's stated scope (Matrix row layout). Reviewers cannot tell from this wording alone whether these are pre-existing optional nodes onmainor new additions introduced by this PR.Fixed in
015198f: two plain sentences, with+ [?] #dash.notifyas an example of an optional node.New commits pushed, approval review dismissed automatically according to repository settings
🟢 werkrevisor GREEN for
015198f, superseded by the result for1d38742🟢 werkrevisor: GREEN / PASS
Gate: approved by werkrevisor; merge waits for 1 open hint(s): reply to each line comment and resolve it, then restart the review or push a commit.
Change kind:
mixed.Reasons
2632d12not carried over: 1 approved block(s) are missingFindings
doc/prs/2026-10-11-PR#15-matrix-row-layout.md:63-67, hint, as a line commentClassifier
8 of 8 units passed all questions (1 PR document(s)).
Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 37579, calls 1, tokens in 36401 / out 22264 (reasoning included in out), duration 142.5 s, cost 0.0376 USD (reported), classifier `jev-1.13.0` calls 8, tokens in 5751, cost 0.0002 USD, finding judge `jev-1.13.0` calls 1, tokens in 1674, cost 0.0001 USD, build file judge `jev-1.13.0` calls 5, tokens in 6984, cost 0.0003 USD, phases: path_judge 0.4 s, classifier 0.6 s, context 0.2 s, model_review 140.7 s, finding_judge 0.3 s, config: defaults + host + repository (.werkrevisor.toml at main1 answer ignored because of a confidence below [classifier] min_confidence.
21cf21a).Reviewed head `015198f6a559b5c9e087fff2dc0eb9b7b179938d` against base `21cf21a8492d881ed0729ccd7e6348224cc4d150`, werkrevisor 0.32.2.
werkrevisor: GREEN / PASS for
015198f6a559b5c9e087fff2dc0eb9b7b179938d.The result comment of werkrevisor holds the details.
@@ -0,0 +60,4 @@None.## Additional Changes[medium] Werkbaum changes underdocumented (werkrevisor)
hint; Jev: real 0.84, severity 1.2 of 3
The "Additional Changes" section states "The PR changes no other node", but the Werkbaum diff also adds
#dash.notifyplus several+ [ ]and+ [?]optional nodes (e.g. "Device verified by emoji comparison", "Older messages than the first sync brings", "Relays on the Omarchy hosts", "Push notification when a pull request needs the user"). The example given,+ [?] #dash.notify, is itself one of those new additions, so the claim "those were onmainbefore" contradicts the diff.No change to the Werkbaum:
git diff origin/main...HEAD -- doc/werkregie.werkbaumadds only#dash.taiga, its description block and one description line.+ [?] #dash.notifyand the other optional nodes came with PR#4, PR#8 and PR#13. The note that caused this round is gone in1d38742.New commits pushed, approval review dismissed automatically according to repository settings
🟢 werkrevisor: GREEN / PASS
Gate: approved by werkrevisor; merge waits for 1 open hint(s): reply to each line comment and resolve it, then restart the review or push a commit.
Change kind:
mixed.Reasons
015198fnot carried over: 1 approved block(s) are missingFindings
dashboard/src/test/kotlin/de/hoennig/werkregie/dashboard/sources/MatrixSourceTest.kt:60-61, severity unknown, counted as medium, hint, as a line commentClassifier
8 of 8 units passed all questions (1 PR document(s)).
Usage: model `openrouter/minimax/minimax-m3`, LLM done, tokens counted before the call 45769, calls 1, tokens in 44370 / out 20010 (reasoning included in out), duration 132.1 s, cost 0.0373 USD (reported), classifier `jev-1.13.0` calls 8, tokens in 5733, cost 0.0002 USD, finding judge `jev-1.13.0` calls 1, tokens in 2245, cost 0.0001 USD, build file judge `jev-1.13.0` calls 7, tokens in 8672, cost 0.0004 USD, phases: path_judge 0.6 s, classifier 0.5 s, context 0.4 s, model_review 129.8 s, finding_judge 0.2 s, config: defaults + host + repository (.werkrevisor.toml at main1 answer ignored because of a confidence below [classifier] min_confidence.
21cf21a).Reviewed head `1d387423338902af9a11d31aa3f83f7131cf6400` against base `21cf21a8492d881ed0729ccd7e6348224cc4d150`, werkrevisor 0.32.2.
werkrevisor: GREEN / PASS for
1d387423338902af9a11d31aa3f83f7131cf6400.The result comment of werkrevisor holds the details.
@@ -58,3 +58,3 @@source.fetch(Instant.now(), Instant.EPOCH).map { it.title } shouldBe listOf("Hallo")source.syncing shouldBe falsesource.note() shouldContain "liest mit"// reading along is the normal case on a machine with a relay: no note, which the page would show as a warningTest contract reduced, severity unknown, counted as medium (werkrevisor)
hint; Jev: real 0.59, severity 2.0 of 3
Scenario#11.01 used to assert
source.note() shouldContain "liest mit", i.e. that a warning was surfaced whenever another process held the sync lock. The new assertionsource.note() shouldBe nullonly verifies the absence of a note. The PR intentionally removes that warning (thenote()override inMatrixSourceis gone and the page no longer shows it as a warning), but the test now no longer guards the former contract: a future change that re-introduces the warning or makesnote()return any non-null string is caught, but the more realistic regression — silently dropping the override sonote()returns null again when the relay is in charge — would no longer be caught by this test. The behaviour change is documented, yet the test name "the local mode shows the relay's messages" still describes a scenario where the warning is the user-visible signal that the local mode is reading along rather than syncing; tightening the assertion back to the old contract (or splitting it off into a separate test for the new contract) would keep coverage of both states.Intended, no change: the PR removes the note on purpose (Additional Changes), since the page showed it as a warning in the normal case. The test still checks what Scenario#11.01 promises: the second process shows the messages (
map { it.title } shouldBe listOf("Hallo")) without a client of its own (syncing shouldBe false);note() shouldBe nullnow pins the absence of the warning.