fleet/fleetd
3
0
Fork 0
Code Issues 73 Pull Requests Actions Packages Projects Releases 1 Wiki Activity

CB-586: prune refs/wip/* whose tree is reachable from main and older than 24h #108

Closed
agent wants to merge 0 commits from worker/cb586-10500c-1 into main
pull from: worker/cb586-10500c-1
merge into: fleet:main
Conversation 1 Commits - Files Changed -
agent commented 2026-08-16 19:06:54 +02:00
Member
Copy Link
Copy Source

CB-586: nothing ever pruned refs/wip/*, so CB-578 stage C snapshots pin their whole trees forever.

This adds the decided retention rule (reachability + 24h age floor) to GitWorktrees and drives it from the existing SessionReaper loop:

  • A snapshot ref is deleted only when BOTH hold: its commit's tree content is already reachable from main, and the ref is older than 24h.
  • Reachability is the safety floor: a snapshot not reachable from main is the last copy of a worker's work and is never deleted automatically (proved by test).
  • Every deletion logs the ref name and commit sha so it is recoverable from the reflog.
  • The /members response now carries wipRefs{count,costBytes} — operator census without shelling into the repo (chosen surface for criterion 4).

Scope: session package + tests; the /members field and reaper wiring are the only touches outside.

Build: mvn -f bridged/pom.xml clean install -> Tests run: 813, Failures: 0, Errors: 0, Skipped: 0; BUILD SUCCESS (unpiped).

CB-586: nothing ever pruned refs/wip/*, so CB-578 stage C snapshots pin their whole trees forever. This adds the decided retention rule (reachability + 24h age floor) to GitWorktrees and drives it from the existing SessionReaper loop: - A snapshot ref is deleted only when BOTH hold: its commit's tree content is already reachable from main, and the ref is older than 24h. - Reachability is the safety floor: a snapshot not reachable from main is the last copy of a worker's work and is never deleted automatically (proved by test). - Every deletion logs the ref name and commit sha so it is recoverable from the reflog. - The /members response now carries wipRefs{count,costBytes} — operator census without shelling into the repo (chosen surface for criterion 4). Scope: session package + tests; the /members field and reaper wiring are the only touches outside. Build: mvn -f bridged/pom.xml clean install -> Tests run: 813, Failures: 0, Errors: 0, Skipped: 0; BUILD SUCCESS (unpiped).
agent added 1 commit 2026-08-16 19:06:54 +02:00
CB-586: prune refs/wip/* older than 24h whose tree is reachable from main
CI / build (pull_request) Successful in 1m13s
Details
CI / contract (pull_request) Successful in 1m21s
Details
0efe1567c0 
Add the CB-586 retention rule to GitWorktrees and drive it from the reaper:
a snapshot is deleted only when its tree content is already reachable from
main AND the ref is older than 24h. Reachability keeps the last copy of a
worker's work; the age floor stops a fresh snapshot being swept while a
lead is still looking at it. Every deletion logs the ref name and commit
sha so it is recoverable from the reflog. The /members response gains a
wipRefs{count,costBytes} census the operator can read without shelling
into the repo.
ltms commented 2026-08-16 20:14:17 +02:00
Owner
Copy Link
Copy Source

Merged locally into main and pushed — merge commit 15ff6bc, plus a follow-up fix at b14b66a. Gitea cannot mark a locally merged PR as merged, so I am closing it by hand — merged, not rejected.

The review detail is on #67. The short version: the seam was right and I checked it against a real git repo, but the caller in SessionReaper never reached it. lastWipSweepNanos started at Long.MIN_VALUE, now - Long.MIN_VALUE overflowed, and the interval gate returned early on every iteration forever. Your tests all passed because they call SessionManager.sweepWipRefs directly, which walks around the gate. Fixed with a sweptOnce flag and a test that asserts through the loop.

860 tests, mvn clean install, BUILD SUCCESS. Good work — the safety rule itself held up under every case I threw at it.

Merged locally into `main` and pushed — merge commit `15ff6bc`, plus a follow-up fix at `b14b66a`. Gitea cannot mark a locally merged PR as merged, so I am closing it by hand — **merged, not rejected**. The review detail is on #67. The short version: the seam was right and I checked it against a real git repo, but the caller in `SessionReaper` never reached it. `lastWipSweepNanos` started at `Long.MIN_VALUE`, `now - Long.MIN_VALUE` overflowed, and the interval gate returned early on every iteration forever. Your tests all passed because they call `SessionManager.sweepWipRefs` directly, which walks around the gate. Fixed with a `sweptOnce` flag and a test that asserts through the loop. 860 tests, `mvn clean install`, BUILD SUCCESS. Good work — the safety rule itself held up under every case I threw at it.
ltms closed this pull request 2026-08-16 20:14:21 +02:00
Some checks are pending
CI / build (pull_request) Successful in 1m13s
Details
CI / contract (pull_request) Successful in 1m21s
Details

Pull request closed

Please reopen this pull request to perform a merge.
Sign in to join this conversation.
Reviewers
No Reviewers
Labels
Clear labels
blocked
Cannot start until something else lands. The body says what.
 needs-live-proof
Merged and green, but never shown working on the running daemon. Not the same as done.
 ready-to-delegate
Scope, files and acceptance criteria are written. A worker can be briefed from the body alone.
 silent-default
A feature that compiles, passes tests, and ships turned off. Nine recurrences and counting.
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
agent kevin ltms
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: fleet/fleetd#108
Delete Branch "worker/cb586-10500c-1"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?