[ScriptedAlchemy]

Stack order (merge first → last): #4473 → #4478 → #4521 → #4526 → #4515 → #4510 → #4516 → #4524 → #4489 → #4488 → #4485

This PR is layer 3/11, based on #4478. Next: #4526.

Merge bottom to top using merge commits. After each parent merges, retarget its child to master before merging the child; this preserves the ancestry already integrated and avoids reintroducing conflicts.

Fixes #4518

Problem. lcm_effects::tests::refresh_begins_at_the_committed_frontier_after_background_summaries_publish fails intermittently on loaded runners with timed out waiting for background summaries to publish.

Root cause. Not a production ordering bug. The test pinned its fake cursor-agent script with timeout_secs = 5, the smallest admitted value. Under load the /bin/sh child is starved for more than 5 s, the provider kills it, and the convergence queue parks the session as unavailable / cursor_agent_unavailable. That state is only requeued when the summarizer binding changes, so the 30 s poll can never see a summary node. I confirmed this by dumping the queue row and the provider error on timeout (timed out waiting for .../cursor-agent) in every failing run.

Fix. The test no longer sets the 5 s provider budget, so the provider default applies and scheduling delay cannot kill the stand-in. The same pin in parked_sessions_converge_once_the_summarizer_becomes_available_without_restart had the same exposure and gets the same change. No poll timeout was raised and no assertion changed.

Verification. Test binary run directly, 150 concurrent copies at a time on an 18-core machine, --exact with the full path (1 passed confirmed):

	failures
before	187/300, then 113/150
after	0/150, 0/300

At 300 concurrent copies each run takes 80 to 100 s and any timed wait fails, with or without this change, so that load is not meaningful.

bazel test //crates/tracedecay-session-runtime:unit_test passes and bazel build --config=clippy //crates/tracedecay-session-runtime/... is clean.

🤖 Generated with Claude Code

[changeset-bot]

⚠️ No Changeset found

Latest commit: 37cd045

Merging this PR will not cause a version bump for any packages. If these changes should not result in a new version, you're good to go. If these changes should result in a version bump, you need to add a changeset.

Click here to learn what changesets are, and how to add one.

Click here if you're a maintainer who wants to add a changeset to this PR