fleet/fleetd
3
0
Fork 0
Code Issues 80 Pull Requests Actions Packages Projects Releases 1 Wiki Activity

fleetd #710: Remove CallerResolver members accessor #713

Closed
agent wants to merge 0 commits from worker/task-13-823ce2-10 into main
pull from: worker/task-13-823ce2-10
merge into: fleet:main
Conversation 1 Commits - Files Changed -
agent commented 2026-10-04 06:54:00 +02:00
Member
Copy Link
Copy Source

Removes the public CallerResolver.members() accessor and its only assertion. The accessor returns architectTerminals, so its name conflicts with fleet_list members, which lists spawned members. There is no production caller, and live architects already appear in fleet_list with role architect. Tests: mvn clean install passed: Tests run: 2005, Failures: 0, Errors: 0, Skipped: 0.

Removes the public CallerResolver.members() accessor and its only assertion. The accessor returns architectTerminals, so its name conflicts with fleet_list members, which lists spawned members. There is no production caller, and live architects already appear in fleet_list with role architect. Tests: mvn clean install passed: Tests run: 2005, Failures: 0, Errors: 0, Skipped: 0.
agent added 1 commit 2026-10-04 06:54:01 +02:00
Remove CallerResolver members accessor
CI / shell-tests (pull_request) Failing after 8s
Details
CI / contract (pull_request) Successful in 51s
Details
CI / build (pull_request) Failing after 1m47s
Details
b205bcc2aa
ltms commented 2026-10-04 07:01:42 +02:00
Owner
Copy Link
Copy Source

Merged locally in 25d53e6. Closing by hand, because a local merge never closes a PR here.

Verified

Merged with PR #712 onto origin/main in one throwaway worktree and ran mvn clean install
myself, output to a file and not piped: exit 0, BUILD SUCCESS,
Tests run: 2008, Failures: 0, Errors: 0, CallerResolverTest: 49.

49 both before and after confirms what you reported: an assertion was removed, not a test method.
The pushed tree is byte-identical to the tree I built (6939bf0).

Criterion 1 I re-ran myself: grep -rn '\.members()' fleetd/src --include='*.java' returns no
hit with a CallerResolver receiver.

Criterion 3 — I have to correct the answer

You reported "the sweep list is empty". On my own measurement it is not, and the method it misses
is the interesting one.

$ grep -rn "permits(" fleetd/src/main --include='*.java' | grep -v 'static boolean permits'
fleetd/src/main/.../auth/Authz.java:80:        return permits(caller, action, targetSession, NO_KNOWN_LEAD_OR_COLLABORATOR);
fleetd/src/main/.../mcp/FleetMcp.java:701:   if (Authz.permits(caller, action, target, callers.knownLeadOrCollaborator())) {
fleetd/src/main/.../rest/FleetApp.java:281:  return Authz.permits(caller, action, target, knownLeadOrCollaborator);

Both production gates call the four-argument form. The three-argument
Authz.permits(caller, action, targetSession) is called only by itself at Authz.java:80 and
from tests. It has no production caller.

Why I think the slip happened, and it is partly my wording: I wrote the criterion as "no caller
outside its own test file". Your grep found the three-argument form called from
CallerResolverTest.java, which is outside AuthzTest.java, so you applied my words correctly.
The property I actually wanted was "no caller in src/main". That is on me.

It matters more than a tidy-up, because that overload supplies
NO_KNOWN_LEAD_OR_COLLABORATOR — a deny-all classifier — as a silent default. A future call
site that reaches for the shorter signature compiles, passes, and refuses every collaborator send,
with nothing to indicate why. I am keeping it for now (that was a deliberate call during #669) but
it needs a comment saying it is a test convenience, and that is a separate unit.

The honest reading of criterion 3 is: one entry, found by measurement, not zero. The rest of
your sweep stands — I checked the other nine declarations and each has a production caller.

The change itself is correct and the evidence for criteria 1 and 2 was good, including flagging
your own zsh: read-only variable: status wrapper slip and re-running clean rather than reporting
the first result.

Merged locally in `25d53e6`. Closing by hand, because a local merge never closes a PR here. ## Verified Merged with PR #712 onto `origin/main` in one throwaway worktree and ran `mvn clean install` myself, output to a file and not piped: exit 0, `BUILD SUCCESS`, `Tests run: 2008, Failures: 0, Errors: 0`, `CallerResolverTest: 49`. 49 both before and after confirms what you reported: an assertion was removed, not a test method. The pushed tree is byte-identical to the tree I built (`6939bf0`). Criterion 1 I re-ran myself: `grep -rn '\.members()' fleetd/src --include='*.java'` returns no hit with a `CallerResolver` receiver. ## Criterion 3 — I have to correct the answer You reported "the sweep list is empty". On my own measurement it is not, and the method it misses is the interesting one. ``` $ grep -rn "permits(" fleetd/src/main --include='*.java' | grep -v 'static boolean permits' fleetd/src/main/.../auth/Authz.java:80: return permits(caller, action, targetSession, NO_KNOWN_LEAD_OR_COLLABORATOR); fleetd/src/main/.../mcp/FleetMcp.java:701: if (Authz.permits(caller, action, target, callers.knownLeadOrCollaborator())) { fleetd/src/main/.../rest/FleetApp.java:281: return Authz.permits(caller, action, target, knownLeadOrCollaborator); ``` Both production gates call the **four**-argument form. The three-argument `Authz.permits(caller, action, targetSession)` is called only by itself at `Authz.java:80` and from tests. It has no production caller. Why I think the slip happened, and it is partly my wording: I wrote the criterion as "no caller outside its *own test file*". Your grep found the three-argument form called from `CallerResolverTest.java`, which is outside `AuthzTest.java`, so you applied my words correctly. The property I actually wanted was "no caller in `src/main`". That is on me. It matters more than a tidy-up, because that overload supplies `NO_KNOWN_LEAD_OR_COLLABORATOR` — a **deny-all** classifier — as a silent default. A future call site that reaches for the shorter signature compiles, passes, and refuses every collaborator send, with nothing to indicate why. I am keeping it for now (that was a deliberate call during #669) but it needs a comment saying it is a test convenience, and that is a separate unit. The honest reading of criterion 3 is: **one entry, found by measurement, not zero.** The rest of your sweep stands — I checked the other nine declarations and each has a production caller. The change itself is correct and the evidence for criteria 1 and 2 was good, including flagging your own `zsh: read-only variable: status` wrapper slip and re-running clean rather than reporting the first result.
ltms closed this pull request 2026-10-04 07:01:45 +02:00
Some checks are pending
CI / shell-tests (pull_request) Failing after 8s
Details
CI / contract (pull_request) Successful in 51s
Details
CI / build (pull_request) Failing after 1m47s
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#713
Delete Branch "worker/task-13-823ce2-10"

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?