ltms - Gitea

ltms commented on issue fleet/fleetd#247

2026-09-03 07:32:43 +02:00

Seeding ~/.claude.json can still lose an external writer's change

Measured live on 2026-09-03 after deploying #149. The external writer is worse than I described when I filed this, and it changes which fix is right.

I spawned a real claude-code member…

ltms closed issue fleet/fleetd#149

2026-09-03 07:32:29 +02:00

A claude-code member spawned into a fresh worktree blocks forever on the workspace trust dialog

ltms commented on issue fleet/fleetd#149

2026-09-03 07:32:26 +02:00

A claude-code member spawned into a fresh worktree blocks forever on the workspace trust dialog

Criterion 1 is now proven live. Closing.

Redeployed onto the merged jar (pid 36650, jar b8af736f44fc), then spawned a real claude-code member into a fresh worktree:

worktree `/Users/dai.…

ltms pushed to main at fleet/fleetd

2026-09-03 07:29:31 +02:00

9d37f3aa29 fleetd #201: the coverage line must name the key its caller actually means

ltms closed pull request fleet/fleetd#251

2026-09-03 07:25:17 +02:00

fleetd #248: prove main() wires CompletionResolver arguments, not just the class

ltms closed issue fleet/fleetd#248

2026-09-03 07:25:13 +02:00

Fleetd's composition root is untested: a feature can be silently unwired and every test stays green

ltms commented on issue fleet/fleetd#248

2026-09-03 07:25:10 +02:00

Fleetd's composition root is untested: a feature can be silently unwired and every test stays green

Fixed and merged to main (PR #251). Build: 1228 tests, 0 failures, 0 compile errors.

I re-ran both mutations myself against the merged code, since the whole ticket is about not trusting…

ltms pushed to main at fleet/fleetd

2026-09-03 07:24:56 +02:00

eaf89abaf6 fleetd #248: make Fleetd's CompletionResolver wiring provable

d895f02bc1 fleetd #248: prove main() wires CompletionResolver's arguments, not just the class

Compare 2 commits »

ltms commented on issue fleet/fleetd#115

2026-09-03 07:06:59 +02:00

CB-612: the daemon warns at every boot about a token no profile needs, teaching operators to skim startup warnings

Captured the before picture from the running daemon's log, so criterion 4 has something to compare against. It also turned up the strongest possible argument for this ticket, which I did not…

ltms closed pull request fleet/fleetd#250

2026-09-03 07:05:29 +02:00

fleetd #148 (point 2): drop .envrc from the default parityOverlay

ltms closed issue fleet/fleetd#148

2026-09-03 07:05:26 +02:00

CB-634: the parity overlay copies .env/.envrc into worker worktrees — the same shape as CB-633

ltms commented on issue fleet/fleetd#148

2026-09-03 07:05:24 +02:00

CB-634: the parity overlay copies .env/.envrc into worker worktrees — the same shape as CB-633

All three points are done. Closing.

ltms pushed to main at fleet/fleetd

2026-09-03 07:05:04 +02:00

43206cac2f fleetd #148 point 2: drop .envrc from the default parity overlay

8bba3a8184 fleetd #148 (point 2): drop .envrc from the default parityOverlay

Compare 2 commits »

ltms commented on issue fleet/fleetd#115

2026-09-03 07:04:20 +02:00

CB-612: the daemon warns at every boot about a token no profile needs, teaching operators to skim startup warnings

Already fixed and merged as 1c051c4, earlier today. Confirmed with git merge-base --is-ancestor 1c051c4 origin/main.

The current code on main:

FleetConfig.java:406 — `tokenEnv =…

ltms commented on issue fleet/fleetd#176

2026-09-03 07:02:15 +02:00

capacity counts panes, not subscription seats, so a subscription profile reports a free slot that cannot be filled

Hit the same shape again today, on a different profile and for a different underlying reason. Recording it because it shows the defect is wider than the subscription-seat case this ticket was…

ltms opened issue fleet/fleetd#249

2026-09-03 06:59:22 +02:00

fleet_list can report another session's agentSessionId, and fleet_spawn{resumeSessionId} will resume it

ltms closed issue fleet/fleetd#234

2026-09-03 06:58:26 +02:00

#175's model read-back fires a FALSE POSITIVE on any member spawned without a worktree — and the quarantine it announces never happens

ltms commented on issue fleet/fleetd#234

2026-09-03 06:58:23 +02:00

#175's model read-back fires a FALSE POSITIVE on any member spawned without a worktree — and the quarantine it announces never happens

Fixed and merged to main as 838a701.

ExhaustionSink's 3-argument method (carrying the profile) is now the single abstract method, so a caller must supply the profile. The false positive is…

ltms closed issue fleet/fleetd#227

2026-09-03 06:58:16 +02:00

A backend outage is invisible at fleet level: repeated backend errors are never correlated, never cool off a credential, and never reach the lead

ltms commented on issue fleet/fleetd#227

2026-09-03 06:58:14 +02:00

A backend outage is invisible at fleet level: repeated backend errors are never correlated, never cool off a credential, and never reach the lead

Done and merged, together with #201 — the two were refined into one five-unit series because a fleet-level correlation needs a per-target classification to correlate. See #201 for the unit-by-unit…