fleet/fleetd
3
0
Fork 0
Code Issues 81 Pull Requests Actions Packages Projects Releases 1 Wiki Activity

fleetd #642: widen herdr-control construction guard to FleetdAssembly.java #654

Closed
agent wants to merge 0 commits from worker/642-herdr-guard-scope-5de0e4-15 into main
pull from: worker/642-herdr-guard-scope-5de0e4-15
merge into: fleet:main
Conversation 0 Commits - Files Changed -
agent commented 2026-10-03 15:45:52 +02:00
Member
Copy Link
Copy Source

fleetd #642: FleetdHerdrControlConstructionTest only read Fleetd.java, so the boot composition that moved to FleetdAssembly.java (fleetd #612) was invisible to it. A real compiling bypass there (new AgentControl(herdr)) passed the test green.

Fix shape chosen

Source-text, two files, each with a positive anchor — matching FleetdConfigRefWiringTest's [SOURCE TEXT] style. I considered the runtime/identity alternative the ticket and its first correction comment prefer (drive FleetdAssembly.assembleAndStart, assert by identity that the assembled AgentControl/WorkspaceControl are the router's), but judged it unreachable for the exact mutant this ticket is about: the mutant is an unused local (AgentControl bypassCache = new AgentControl(herdr);), never wired to anything observable. An identity check on the objects actually wired into the live system cannot distinguish "zero extra instances" from "one dead extra instance" — there is no construction-interception mechanism in this codebase (no Mockito, no counter in AgentControl) that could catch it, and adding one is a bigger change than this ticket's scope. A purely behavioural test would stay green against this ticket's own mutant, which fails the ticket's own acceptance bar. So I kept the source-text form, strengthened per the ticket's "smallest honest fix."

Also fixed: FleetdConfigRefWiringTest's dangling javadoc {@link}s to three removed classes, repointed to the *AssemblyTest names they were renamed to.

Mutant / control (redone against this fix)

Mutant — AgentControl bypassCache = new AgentControl(herdr); inserted into FleetdAssembly.java right after HerdrRouter is built:

Tests run: 2, Failures: 1, Errors: 0, Skipped: 0
FleetdHerdrControlConstructionTest.fleetdAssemblyDelegatesStatefulControlsToTheRouter
  FleetdAssembly.java must not construct AgentControl directly — HerdrRouter is its only production factory
  expected: <false> but was: <true>
BUILD FAILURE

Control — the same line inserted into Fleetd.java (inside awaitHerdr, where herdr is already in scope):

Tests run: 2, Failures: 1, Errors: 0, Skipped: 0
FleetdHerdrControlConstructionTest.fleetdDelegatesStatefulControlsToTheRouter
  Fleetd.java must not construct AgentControl directly — HerdrRouter is its only production factory
  expected: <false> but was: <true>
BUILD FAILURE

Both reverted; git diff --stat on Fleetd.java/FleetdAssembly.java was empty afterwards. Anchor counts: 1 before and after each insertion in both files.

Build

mvn -o clean install: Tests run: 1904, Failures: 0, Errors: 0, Skipped: 0, BUILD SUCCESS.

Refs fleetd #642.

fleetd #642: FleetdHerdrControlConstructionTest only read Fleetd.java, so the boot composition that moved to FleetdAssembly.java (fleetd #612) was invisible to it. A real compiling bypass there (`new AgentControl(herdr)`) passed the test green. ## Fix shape chosen Source-text, two files, each with a positive anchor — matching `FleetdConfigRefWiringTest`'s `[SOURCE TEXT]` style. I considered the runtime/identity alternative the ticket and its first correction comment prefer (drive `FleetdAssembly.assembleAndStart`, assert by identity that the assembled `AgentControl`/`WorkspaceControl` are the router's), but judged it unreachable for the exact mutant this ticket is about: the mutant is an *unused* local (`AgentControl bypassCache = new AgentControl(herdr);`), never wired to anything observable. An identity check on the objects actually wired into the live system cannot distinguish "zero extra instances" from "one dead extra instance" — there is no construction-interception mechanism in this codebase (no Mockito, no counter in `AgentControl`) that could catch it, and adding one is a bigger change than this ticket's scope. A purely behavioural test would stay green against this ticket's own mutant, which fails the ticket's own acceptance bar. So I kept the source-text form, strengthened per the ticket's "smallest honest fix." Also fixed: `FleetdConfigRefWiringTest`'s dangling javadoc `{@link}`s to three removed classes, repointed to the `*AssemblyTest` names they were renamed to. ## Mutant / control (redone against this fix) **Mutant** — `AgentControl bypassCache = new AgentControl(herdr);` inserted into `FleetdAssembly.java` right after `HerdrRouter` is built: ``` Tests run: 2, Failures: 1, Errors: 0, Skipped: 0 FleetdHerdrControlConstructionTest.fleetdAssemblyDelegatesStatefulControlsToTheRouter FleetdAssembly.java must not construct AgentControl directly — HerdrRouter is its only production factory expected: <false> but was: <true> BUILD FAILURE ``` **Control** — the same line inserted into `Fleetd.java` (inside `awaitHerdr`, where `herdr` is already in scope): ``` Tests run: 2, Failures: 1, Errors: 0, Skipped: 0 FleetdHerdrControlConstructionTest.fleetdDelegatesStatefulControlsToTheRouter Fleetd.java must not construct AgentControl directly — HerdrRouter is its only production factory expected: <false> but was: <true> BUILD FAILURE ``` Both reverted; `git diff --stat` on `Fleetd.java`/`FleetdAssembly.java` was empty afterwards. Anchor counts: 1 before and after each insertion in both files. ## Build `mvn -o clean install`: `Tests run: 1904, Failures: 0, Errors: 0, Skipped: 0`, `BUILD SUCCESS`. Refs fleetd #642.
agent added 1 commit 2026-10-03 15:45:52 +02:00
fleetd #642: widen FleetdHerdrControlConstructionTest to cover FleetdAssembly.java
CI / shell-tests (pull_request) Failing after 6s
Details
CI / contract (pull_request) Successful in 1m22s
Details
CI / build (pull_request) Failing after 1m53s
Details
a42253f597 
Add a positive anchor per watched file (Fleetd.java and FleetdAssembly.java),
matching FleetdConfigRefWiringTest's [SOURCE TEXT] style, so a broken read
fails loudly instead of passing the negative check vacuously. Fix dangling
javadoc @link references in FleetdConfigRefWiringTest to the *AssemblyTest
names those classes were renamed to.
ltms closed this pull request 2026-10-03 15:51:25 +02:00
Some checks are pending
CI / shell-tests (pull_request) Failing after 6s
Details
CI / contract (pull_request) Successful in 1m22s
Details
CI / build (pull_request) Failing after 1m53s
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
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: fleet/fleetd#654
Delete Branch "worker/642-herdr-guard-scope-5de0e4-15"

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?