You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
flows::launch has no unit pin on the liveness ask #434
flows::launch has zero unit coverage of whether run_dotfiles asks devpod
for liveness. Give it some.
The evidence is a measurement, not a suspicion. While building the liveness pins
(#417), two separate mutations
were applied to production code:
run_dotfiles's workspace_state call replaced by placement.is_running();
the full Placement::Creatable arm-split stand-in, with the new arm grouped
into Naming::Anonymous.
Under both, all 1253 core tests stayed green. Only the dl integration
pins caught either. So the behaviour is pinned exclusively at the binary
boundary, and a refactor inside devlaunch-core can break it with cargo test -p devlaunch-core entirely green.
That is a real gap and it is not the same argument #417 made, which is why it is
its own ticket rather than a rider: #417's three rows are launch-shape
observations at the integration seam, and its author deliberately declined to add
a unit-level pin on the grounds that the integration seam catches the defect
directly. It does. But it is the only thing that does, and the two seams fail
independently.
What is wanted: a run_dotfiles-level pin in flows/launch.rs's own unit tests
that fails when the liveness ask is removed or rerouted. Scope it to the seam flows/launch.rs already uses for recorded calls; do not reach for a new
harness.
Demonstrate the pin can fail, by the same method: apply one of the two
mutations above, watch it go red, revert, and report what you saw. Pin the launch shapes that had no liveness pin #417's whole
finding was a green test sitting on broken code.
Standing refusals from #393,
none of which this may reintroduce: no liveness cache, no
try-the-attach-and-fall-back, no always-up, no Remembered arm.
Interacts with the arm split
(#418) and with #419. If it is cheaper to
write this pin after the split lands, say so and get it blocked rather than
writing a pin you will immediately rewrite.
Cut from origin/main; use your own worktree. Guards in rust/: cargo test --workspace, cargo clippy --locked --all-targets -- -D warnings, cargo fmt --check.
Closing against actual usage. The seam gap is real and well evidenced, but the function it pins, run_dotfiles, has one caller and it is the dotfiles verb (flows/launch.rs:3020), which this install never runs. The defect class is already caught by #417's integration pins (PR #430), as the ticket concedes. A second test seam on a dead-for-this-user verb is maintenance surface with no risk retired, which is the wrong trade under "small and focused".
Activity
blooop commentedon Aug 29, 2026
Closing against actual usage. The seam gap is real and well evidenced, but the function it pins,
run_dotfiles, has one caller and it is thedotfilesverb (flows/launch.rs:3020), which this install never runs. The defect class is already caught by #417's integration pins (PR #430), as the ticket concedes. A second test seam on a dead-for-this-user verb is maintenance surface with no risk retired, which is the wrong trade under "small and focused".