[htoyoda18]

What this PR does

When an attached docker compose up stops the application itself, on Ctrl+C or on --abort-on-container-exit / --exit-code-from, it ignored -t/--timeout and waited for each service's stop_grace_period (or the engine default). The flag is documented as the shutdown timeout "when attached", so this PR passes the -t value to those stops. Without -t, nothing changes.

Related issue

Fixes #14394

Changes made

• pkg/compose/up.go: stopApplication (Ctrl+C and abort on container exit) and stopLateStarter (a service that starts after termination began) now set Timeout: u.options.Create.Timeout in their api.StopOptions. This is the value -t already fills. It is nil when -t is not given, so the current behavior is unchanged in that case.
• pkg/compose/up_test.go: add TestUpSessionStopsWithUpTimeout, which checks that both stops pass the -t value to ContainerStop. Before the fix, both subtests failed with an empty Timeout.
• pkg/e2e/cascade_test.go, pkg/e2e/testdata/TestCascadeStopHonorsTimeout/compose.yaml: add a regression scenario. up -t 1 --abort-on-container-exit --exit-code-from tests must end with exit code 3 within 60s while another service ignores SIGTERM and has stop_grace_period: 2m. Before the fix it hit the 60s bound; now it finishes in ~3.5s. The Ctrl+C path goes through the same stopApplication, but the Scenario DSL cannot send signals, so it is covered by the unit test only.

Testing done

• Unit tests pass: make test
• E2E tests pass: make e2e-compose or make e2e-compose-standalone
• Linting passes: make lint
• Code formatted: make fmt
• Manually tested (describe):

AI tool used (if applicable)

AI Tool:

Additional context

(not mandatory) A picture of a cute animal, if possible in relation to what you did

[ndeloof]

LGTM, thanks.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!