ltms - Gitea

ltms closed issue fleet/fleetd#190

2026-09-10 02:11:28 +02:00

Process argv is a credential channel we never enumerated: 4 secrets are readable by any local user via ps

ltms commented on issue fleet/fleetd#190

2026-09-10 02:11:23 +02:00

Process argv is a credential channel we never enumerated: 4 secrets are readable by any local user via ps

Re-measured on the Mac, 2026-09-10. The exposure no longer reproduces, and I nearly reported that it had got worse.

The measurement

fleetd.jar          1 process,  0 env-names in…

ltms commented on issue fleet/fleetd#388

2026-09-10 02:07:24 +02:00

The credential scrub does not run in a shell that is neither login nor interactive — .zshenv is the only file zsh always reads, and it is the one file the scrub is not in

The "Suggested fix" in the ticket body above is wrong. Use comment 15387 instead. I am leaving the body unedited so the correction stays readable, but read that comment before you read my…

ltms commented on issue fleet/fleetd#381

2026-09-10 02:04:47 +02:00

Members cannot run shell commands: the classifier refuses every one, so they cannot build, test or open a PR (3 members, 2 profiles)

Confirmed on a full implementer job, not just a probe. When I closed this I had verified the fix with a short spawn that ran real shell commands. That left the real question open: can a…

ltms closed pull request fleet/fleetd#389

2026-09-10 02:04:25 +02:00

#386: give the stall detector a real-time clock, log the divergence

ltms closed pull request fleet/fleetd#387

2026-09-10 02:04:18 +02:00

#382: give SpawnRequest a withProfile wither, guard it against the arity trap

ltms closed issue fleet/fleetd#386

2026-09-10 02:00:24 +02:00

Every fleetd timer freezes while the macOS host sleeps: System.nanoTime() stops, so the stall detector missed a member that was BUSY for 101 minutes

ltms commented on issue fleet/fleetd#386

2026-09-10 02:00:19 +02:00

Every fleetd timer freezes while the macOS host sleeps: System.nanoTime() stops, so the stall detector missed a member that was BUSY for 101 minutes

Merged as fd8650c (PR #389), plus a follow-up test in b9d09e0. Verified here.

What landed

FleetHealthMonitor takes a second clock, realtimeClock, used only inside the stall check.…

ltms pushed to main at fleet/fleetd

2026-09-10 01:59:51 +02:00

b9d09e044e t386: pin the per-member drift baseline the fix's own tests left open

fd8650cda4 Merge #386: correct the stall check for a monotonic clock frozen by host sleep

769f282408 #386: give the stall detector a real-time clock, log the divergence

Compare 3 commits »

ltms closed issue fleet/fleetd#384

2026-09-10 01:55:28 +02:00

"left no scrub report" has a third, benign cause it does not name — a group-shared ZDOTDIR cannot receive the receipt

ltms closed issue fleet/fleetd#382

2026-09-10 01:55:23 +02:00

SpawnRequest has the arity trap's two halves, but not yet a colliding arity — CompositePeerLauncher:372 will drop the next component added

ltms commented on issue fleet/fleetd#382

2026-09-10 01:55:18 +02:00

SpawnRequest has the arity trap's two halves, but not yet a colliding arity — CompositePeerLauncher:372 will drop the next component added

Merged as fb36c52 (PR #387). Verified here, not taken on the worker's word.

What landed. SpawnRequest.withProfile(String), and CompositePeerLauncher:372 now calls it instead of…

ltms pushed to main at fleet/fleetd

2026-09-10 01:55:00 +02:00

11050e24ed t384: fix javadoc indentation on the merged shareWithGroup lines

6f71f40047 Merge #384: pre-create the scrub receipt and give it group write

fb36c5238f Merge #382: SpawnRequest.withProfile() replaces the six-accessor rebuild

cc9cdc938b #384: write shared scrub receipts

5fede82468 #382: give SpawnRequest a withProfile wither, guard it against the arity trap

Compare 5 commits »

ltms commented on issue fleet/fleetd#384

2026-09-10 01:54:45 +02:00

"left no scrub report" has a third, benign cause it does not name — a group-shared ZDOTDIR cannot receive the receipt

Merged. I verified it myself rather than taking the worker's report.

What landed. generate() now pre-creates an empty scrub-report.txt, and shareWithGroup gives that one file rw-rw----…

ltms opened issue fleet/fleetd#388

2026-09-10 01:45:28 +02:00

The credential scrub does not run in a shell that is neither login nor interactive — .zshenv is the only file zsh always reads, and it is the one file the scrub is not in

ltms pushed to main at fleet/fleetd

2026-09-10 01:43:33 +02:00

1515025804 charter: a profile differs in liveness, not just model and cost

7754f53662 Merge t385 follow-up: pin the mid-turn redelivery ack

a507f7b31b t385: pin that a redelivery is acked even while the lead is mid-turn

71c322f104 Merge #385: a redelivered lead message must not write the pane twice

fde2c15627 t385: a redelivered lead message must not write the pane twice

Compare 5 commits »

ltms closed issue fleet/fleetd#381

2026-09-10 01:38:55 +02:00

Members cannot run shell commands: the classifier refuses every one, so they cannot build, test or open a PR (3 members, 2 profiles)

ltms commented on issue fleet/fleetd#381

2026-09-10 01:38:50 +02:00

Members cannot run shell commands: the classifier refuses every one, so they cannot build, test or open a PR (3 members, 2 profiles)

Root cause found, fixed and verified live on the Mac, 2026-09-10.

The axis was never time, and it was never the profile

My earlier comment said "compare time, not profile", and told the next…

ltms closed issue fleet/fleetd#385

2026-09-10 01:29:44 +02:00

A lead-coordination message can be delivered forever: ack() returns success when connection recovery cleared the held entry

ltms commented on issue fleet/fleetd#385

2026-09-10 01:29:36 +02:00

A lead-coordination message can be delivered forever: ack() returns success when connection recovery cleared the held entry

Fixed and merged to main as 71c322f (fix) and 7754f53 (a follow-up test).

What the fix does

Two changes, in two places, because two different things were wrong.

**LeadCoordLoop…