No issue: Add new cron target task for legacy API UI tests

Setups up a new daily cron task to run select UI tests on API 28 (we’ll call it legacy). We don’t necessarily want to invoke Flank twice (serial) with a different configuration (FTL APIs dont support API splitting per test-target) on event CI but we do want them running elsewhere (cron).

.cron.yml

@@ -30,3 +30,9 @@ jobs:
           treeherder-symbol: screenshots-D
           target-tasks-method: screenshots
       when: [{weekday: 'Monday', hour: 10, minute: 0}]
+    - name: legacy-api-ui-tests
+      job:
+          type: decision-task
+          treeherder-symbol: legacy-api-ui
+          target-tasks-method: legacy_api_ui_tests
+      when: [{hour: 10, minute: 30}]

automation/taskcluster/androidTest/flank-x86-legacy-api-tests.yml

@@ -0,0 +1,77 @@
+# gcloud args match the official gcloud cli
+# https://cloud.google.com/sdk/gcloud/reference/firebase/test/android/run
+gcloud:
+  results-bucket: fenix_test_artifacts
+  record-video: true
+
+  # The maximum possible testing time is 30m on physical devices and 60m on virtual devices.
+  timeout: 30m
+  # will start test then close socket. no reports will be generated.
+  # to retrieve results later, use the "refresh" command
+  # reports will be generated from /results/matrix_ids.json
+  #async: true
+  # will start test then leave socket open. reports will be published
+  # to /results
+  # see: https://github.com/TestArmada/flank/issues/339
+  async: false
+
+  # results-history-name
+  # by default, set to app name
+  # declare results-history-name to create a separate dropdown menu in Firebase
+  # see: https://github.com/TestArmada/flank/issues/341
+  #results-history-name: tmp_parallel
+
+  # The number of times a test execution should be re-attempted if one or more failures occur.
+  # The maximum number of reruns allowed is 10. Default is 0, which implies no reruns.
+  num-flaky-test-attempts: 1
+
+  # test and app are the only required args
+  app: /app/path
+  test: /test/path
+
+  auto-google-login: false
+  use-orchestrator: true
+  environment-variables:
+    clearPackageData: true
+  directories-to-pull:
+    - /sdcard/screenshots
+  performance-metrics: true
+
+  test-targets:
+    - class org.mozilla.fenix.ui.DownloadTest#pauseResumeCancelDownloadTest
+    - class org.mozilla.fenix.ui.SearchTest#scanButtonAllowPermissionTest
+    - class org.mozilla.fenix.ui.SearchTest#scanButtonDenyPermissionTest
+    - class org.mozilla.fenix.ui.SitePermissionsTest#audioVideoPermissionChoiceOnEachRequestTest
+    - class org.mozilla.fenix.ui.SitePermissionsTest#cameraPermissionChoiceOnEachRequestTest
+    - class org.mozilla.fenix.ui.SitePermissionsTest#microphonePermissionChoiceOnEachRequestTest
+    - class org.mozilla.fenix.ui.SitePermissionsTest#rememberAllowAudioVideoPermissionChoiceTest
+    - class org.mozilla.fenix.ui.SitePermissionsTest#rememberAllowCameraPermissionChoiceTest
+    - class org.mozilla.fenix.ui.SitePermissionsTest#rememberAllowMicrophonePermissionChoiceTest
+    - class org.mozilla.fenix.ui.SitePermissionsTest#rememberBlockAudioVideoPermissionChoiceTest
+    - class org.mozilla.fenix.ui.SitePermissionsTest#rememberBlockCameraPermissionChoiceTest
+    - class org.mozilla.fenix.ui.SitePermissionsTest#rememberBlockMicrophonePermissionChoiceTest
+    - class org.mozilla.fenix.ui.SmokeTest#redirectToAppPermissionsSystemSettingsTest
+
+  device:
+    - model: Pixel2
+      version: 28
+      locale: en_US
+
+flank:
+  project: GOOGLE_PROJECT
+  # test shards - the amount of groups to split the test suite into
+  # set to -1 to use one shard per test. default: 1
+  max-test-shards: -1
+  # num-test-runs: the amount of times to run the tests.
+  # 1 runs the tests once. 10 runs all the tests 10x
+  num-test-runs: 1
+  ### Output Style flag
+  ## Output style of execution status. May be one of [verbose, multi, single, compact].
+  ## For runs with only one test execution the default value is 'verbose', in other cases
+  ## 'multi' is used as the default. The output style 'multi' is not displayed correctly on consoles
+  ## which don't support ansi codes, to avoid corrupted output use single or verbose.
+  ## The output style `compact` is used to produce less detailed output, it prints just Args, test and matrix count, weblinks, cost, and result reports.
+  output-style: compact
+  ### Full Junit Result flag
+  ## Enable create additional local junit result on local storage with failure nodes on passed flaky tests.
+  full-junit-result: true

automation/taskcluster/androidTest/ui-test.sh

@@ -84,6 +84,8 @@ elif [[ "${device_type}" == "x86-screenshots-tests" ]]; then
     flank_template="${PATH_TEST}/flank-x86-screenshots-tests.yml"
 elif [[ "${device_type}" == "x86-beta-tests" ]]; then
     flank_template="${PATH_TEST}/flank-x86-beta.yml"
+elif [[ "${device_type}" == "x86-legacy-api-tests" ]]; then
+    flank_template="${PATH_TEST}/flank-x86-legacy-api-tests.yml"
 else
     echo "FAILURE: flank config file not found!"
     exitcode=1

taskcluster/ci/ui-test/kind.yml

@@ -148,3 +148,14 @@ jobs:
                 - [automation/taskcluster/androidTest/ui-test.sh, x86-start-test, app.apk, android-test.apk, '-1']
         treeherder:
             symbol: nightly(ui-test-x86-nightly)
+    legacy-x86:
+        attributes:
+            legacy: true
+        description: Run select UI tests on older Android API on cron
+        run-on-tasks-for: []
+        run-on-git-branches: [main]
+        run:
+            commands:
+                - [automation/taskcluster/androidTest/ui-test.sh, x86-legacy-api-tests, app.apk, android-test.apk, '-1']
+        treeherder:
+            symbol: debug(legacy-x86)

taskcluster/fenix_taskgraph/target_tasks.py

@@ -110,3 +110,12 @@ def target_tasks_screnshots(full_task_graph, parameters, graph_config):
         return task.attributes.get("screenshots", False)
 
     return [l for l, t in full_task_graph.tasks.items() if filter(t, parameters)]
+
+@_target_task("legacy_api_ui_tests")
+def target_tasks_legacy_api_ui_tests(full_task_graph, parameters, graph_config):
+    """Select the set of tasks required to run select UI tests on other API."""
+
+    def filter(task, parameters):
+        return task.attributes.get("legacy", False)
+
+    return [l for l, t in full_task_graph.tasks.items() if filter(t, parameters)]