Skip to content

Stabilise first app launch in mobile e2e on CI #62

Open

Description

@johnnyhuy

The mobile e2e job passed on #60, but the first active workout test failed its initial launch and passed on retry.

The failure occurred before the workout flow, in resetAndLaunch at e2e/tests/helpers.ts:60.

start workout, log a set, rest timer appears
Error: app did not come to foreground
at resetAndLaunch (e2e/tests/helpers.ts:60:13)
retry #1 passed
1 flaky
3 passed (6.2m)

Evidence is in the CI job log and e2e report artifact.

Investigate first-launch readiness on the fresh CI simulator and make the launch helper reliable without hiding real startup failures. Verify the active workout flow on a fresh simulator and retain launch failure diagnostics. The release configuration slice changed no app or e2e code.

Activity

  1. johnnyhuy commented on Nov 14, 2026

    @johnnyhuy
    ContributorAuthor

    The latest main build also failed before the workout flow in resetAndLaunch, at helpers.ts:60. The first attempt timed out waiting for the app to foreground, then retry #1 ended with WebSocket code 1006. Three other tests passed.

    Failed main e2e job.

    The separate daily record unit slice in #64 changes no app or e2e code. This issue remains the follow-up for simulator launch reliability.

  2. johnnyhuy commented on Nov 14, 2026

    @johnnyhuy
    ContributorAuthor

    The e2e job on #64 succeeded with the same first-launch flake: initial app foreground launch failed, retry #1 passed, 1 flaky and 3 passed (5.6m). No app or e2e files changed in that slice. Evidence: https://github.com/echohello-dev/betterfit/actions/runs/43551624472/job/132315150389

  3. johnnyhuy commented on Nov 14, 2026

    @johnnyhuy
    ContributorAuthor

    The e2e job on #67 passed with the first active-workout test flaky again. Its initial attempt exceeded the 120000ms test timeout; retry #1 passed in 57.5s. The suite reported 1 flaky and 3 passed (10.1m).

    The retained error context contains only the test timeout, so this run does not establish where the initial attempt stalled. No app or e2e files changed. XCUITest passed all 27 tests.

    CI job log and e2e report artifact.

  4. johnnyhuy commented on Nov 14, 2026

    @johnnyhuy
    ContributorAuthor

    The mobile e2e job on #70 passed with 1 flaky and 3 passed (5.4m). The initial active-workout attempt failed in resetAndLaunch at helpers.ts:60 because the app did not come to foreground. Retry #1 passed in 1.1m.

    No app or e2e files changed in this per-workout export slice.

    CI job log and retained e2e report.

  5. johnnyhuy commented on Nov 14, 2026

    @johnnyhuy
    ContributorAuthor

    The mobile e2e job on #73 passed with 1 flaky and 3 passed (4.5m). The initial active-workout attempt failed in resetAndLaunch at helpers.ts:60 because the app did not come to foreground. Retry #1 passed in 1.3m.

    This catalogue equipment lookup slice changes only the core package, its tests and API documentation. No app or e2e files changed.

    CI job log and retained e2e report.

  6. johnnyhuy commented on Nov 14, 2026

    @johnnyhuy
    ContributorAuthor

    Additional first-test CI evidence from #76 at unchanged commit 5524b4a.

    The mobile e2e job failed the first active-workout test, while the other three e2e tests and all 27 XCUITests passed.

    start workout, log a set, rest timer appears
    Test timeout of 120000ms exceeded.
    retry #1
    Error: WebSocket connection closed (code=1006, reason=no reason)
    1 failed
    3 passed (11.4m)
    

    The saved e2e report includes a screenshot showing the Plan screen with Start workout visible. The logs do not identify which awaited driver operation timed out. The retry error points to the mobilewright RPC connection.

    Only the failed e2e job was rerun on the same commit. No app or e2e files changed in this package API slice.

    The isolated rerun passed with one flaky test and three passed. The first active-workout test failed its initial foreground check and passed on its internal retry.

    Error: app did not come to foreground
    start workout, log a set, rest timer appears (retry #1) passed
    1 flaky
    3 passed (5.6m)
    

    The foreground readiness issue remains open. Both CI attempts are retained as evidence; no e2e behaviour was changed by #76.

  7. johnnyhuy commented on Nov 14, 2026

    @johnnyhuy
    ContributorAuthor

    The mobile e2e job on #79 passed with one flaky test and three passed (5.0m). The first active-workout test failed the initial foreground check in resetAndLaunch at tests/helpers.ts:60, then passed on its internal retry in 1.2m. XCUITest passed all 27 tests.

    This entry-specific equipment swap slice changes only the core package, its tests and API documentation. No app or e2e files changed.

    CI job log and retained e2e report.

  8. johnnyhuy commented on Nov 15, 2026

    @johnnyhuy
    ContributorAuthor

    The mobile e2e job on #82 passed with two flaky tests and two passed (12.3m). The first active-workout attempt exceeded the test timeout and failed the foreground check at tests/helpers.ts:60, then passed on retry #1. The skip-rest test lost its WebSocket connection with code 1006 during locator.tap(), then also passed on retry #1.

    This entry-aware suggestion slice changes only the core package, tests and API documentation. No app or e2e files changed.

    CI job log and retained e2e report.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions