Repository navigation
fix(test): add mimocode to the compaction table every selectable harness needs - #36290
Conversation
UX Review (Opus 5.5) — ⏭️ skippedRevision |
First Principles Review (Opus 5.5) — ⏭️ skippedRevision |
Design Review (Opus 5.5) — ✅ PASSDesign-level review of Design-Verdict: PASS A one-row test-table fix that restores the pin's intent; nothing blocks. [DESIGN-REVIEWED] 59902ba Description read[DESCRIPTION-READ] ebc47fb68e31905a0a9f0305cc064376fa666398f28aa22aee566d529aee377d That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with |
Opus 5.5 Review — ✅ no blocking findingsReviewed Verdict parsed from the review's SHA-scoped output markers for commit False positive or not applicable? A repository writer can comment: |
GPT 6.1 Review — ✅ no blocking findingsGPT 6.1 completed its review of This comment is updated in place on each push. Review detailsNo findings. False positive or not applicable? A repository writer can comment: |
buluoray
left a comment
There was a problem hiding this comment.
🤖 Kiro Crew Auto-Pipeline [operator: buluoray]:
Approving. This fixes the red on main in test_reply_format_once.py::TestEveryMessageWhereTheBlockCannotBeTakenOut::test_the_table_names_every_selectable_harness, on both Linux and Windows. #24732 made MiMo Code selectable, and this test requires every selectable harness to have a row in COMPACTION_SEEN.
False is the right value. AcpProvider.reports_compaction answers from ACP_BACKENDS_COMPACTION_REPORTED (providers/acp.py:1047), and MiMo Code is not a member, so the parametrized test_each_harness row would fail on True. It sits with OpenCode, its parent, which is False too.
What I ran. With this one line applied on current main, test/test_reply_format_once.py gives 88 passed.
CI. The one red check, Analyze (javascript-typescript), failed on the GitHub API rate limit, not on this change. It does not touch any JavaScript.
Problem / Motivation
Goal: Backend Tests pass on main again.
test/test_reply_format_once.py::TestEveryMessageWhereTheBlockCannotBeTakenOut::test_the_table_names_every_selectable_harnessfails on main and on every PR based after #24732. That PR made MiMo Code (mimocode) a selectable ACP backend, but the test's hand-writtenCOMPACTION_SEENtable has no row for it.Why it matters
Every open PR's
Backend Testslane is red for a reason that is not its own, so nothing can reachPR Readiness.Not a goal
What changed (motivation → approach → change)
Symptom: the set comparison reports
mimocodeas an extra item on the right.Cause:
BASELINE_SELECTABLE_BACKENDSgainedmimocode;COMPACTION_SEENdid not.Change:
COMPACTION_SEENgets"mimocode": False. False is the real answer:mimocodeis not inACP_BACKENDS_COMPACTION_REPORTED(src/kiro_crew/agent_sdk/backends_turns.py), so Kiro Crew does not see its compaction, and the parametrizedtest_each_harnessnow checks that for it too.Backwards compatibility
Compatible: test-only change.
Removes nothing: one table row is added.
Tests
test_the_table_names_every_selectable_harness: passes; red on unpatched main.test_each_harness[mimocode-*]: new parametrized cases, pass.test/test_flake_pattern_ratchet.py: 68 passed.Proof: not provable -- test-only change
Manual verification
N/A — unit coverage sufficient: the change is one row of a test table.
Related Issues
Closes #36279
Sibling sweep
Swept: git grep -n "BASELINE_SELECTABLE_BACKENDS" -- test
Sibling: test/test_reply_format_once.py -- changed here
Pattern harvest
Not generalizable: the table is hand-written on purpose so a new harness forces a decision here; the test did its job.
Checklist