Fix e2e on 11.12

.github/actions/archive-test-results/action.yaml

@@ -12,20 +12,24 @@ inputs:
     required: false
     default: ${{ github.workspace }}
 
+# Every step below is `if: ${{ !cancelled() }}` rather than `if: always()`. Both still run
+# when the test step failed — which is the case we need these artifacts for — but always()
+# is exempt from cancellation, so cancelling a run left each shard uploading five artifact
+# sets before it could die, which is why "Cancel workflow" appeared to do nothing.
 runs:
   using: composite
   steps:
     - name: Archive runtime logs
       uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f #v7
-      if: always()
+      if: ${{ !cancelled() }}
       with:
         name: ${{ inputs.platform }}-runtime-logs-${{ inputs.test-type }}
         path: log/*.log
         if-no-files-found: ignore
 
     - name: Archive test screenshots
       uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f #v7
-      if: always()
+      if: ${{ !cancelled() }}
       with:
         name: ${{ inputs.platform }}-screenshots-${{ inputs.test-type }}
         path: ${{ inputs.workspace-path }}/maestro/images/actual/${{ inputs.platform }}/**/*.png
@@ -38,7 +42,7 @@ runs:
       # resolves the diff name against the baseline's own relative path. That nested location
       # was never matched by the screenshots glob above, so diffs were silently lost. Gather
       # any *_diff.png from wherever it landed into one flat dir for upload.
-      if: always()
+      if: ${{ !cancelled() }}
       shell: bash
       run: |
         dest="${{ inputs.workspace-path }}/maestro/images/diff/${{ inputs.platform }}"
@@ -49,15 +53,15 @@ runs:
 
     - name: Archive screenshot diffs
       uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f #v7
-      if: always()
+      if: ${{ !cancelled() }}
       with:
         name: ${{ inputs.platform }}-diffs-${{ inputs.test-type }}
         path: ${{ inputs.workspace-path }}/maestro/images/diff/${{ inputs.platform }}/*.png
         if-no-files-found: ignore
 
     - name: Archive artifacts
       uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f #v7
-      if: always()
+      if: ${{ !cancelled() }}
       with:
         name: ${{ inputs.platform }}-artifacts-${{ inputs.test-type }}
         path: packages/pluggableWidgets/**/artifacts/
@@ -68,7 +72,7 @@ runs:
     # empty and the upload is a no-op (if-no-files-found: ignore).
     - name: Archive failure videos
       uses: actions/upload-artifact@bbbca2ddaa5d8feaa63e36b76fdaad77386f024f #v7
-      if: always()
+      if: ${{ !cancelled() }}
       with:
         name: ${{ inputs.platform }}-videos-${{ inputs.test-type }}
         path: maestro/videos/*.mp4

.github/workflows/Build.yml

@@ -9,6 +9,15 @@ on:
     branches:
       - main
       - 'mx/**'
+
+# Supersede in-flight runs: a new push to the same PR/branch cancels the previous run
+# instead of queueing beside it (this repo shares GitHub Free's 20-concurrent-job cap with
+# NativePipeline, so stale runs directly delay the ones that matter). Only PR runs are
+# cancelled — a push run on main/mx-* is the recorded result for that commit, so let it finish.
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
+
 jobs:
   test:
     name: "Build (${{ matrix.os }})"

.github/workflows/NativePipeline.yml

@@ -105,7 +105,7 @@ env:
   # Pin the test project (mendix/Native-Mobile-Resources) to a SHA for reproducibility instead
   # of tracking a moving `main`. Single source of truth — bump deliberately. The repo publishes
   # no tags, so a commit SHA is the only stable ref. Override per run via the dispatch input.
-  NATIVE_MOBILE_RESOURCES_REF: ${{ github.event.inputs.test_project_ref || 'a915484ef2cfed403cfbdfa531638df3c46c9d00' }}
+  NATIVE_MOBILE_RESOURCES_REF: ${{ github.event.inputs.test_project_ref || '05255f5ad19de7b109f6a68088e4fb198447ed00' }}
 permissions:
   packages: write
 jobs:
@@ -418,7 +418,7 @@ jobs:
     runs-on: ubuntu-26.04
     # Skip this job if we're using artifacts from a specific run; the native bundles now come
     # from the project job (deploy + native-packager), so gate on project success.
-    if: ${{ github.event.inputs.LOCAL_TEST_ARTIFACT_RUN_ID == '' && always() && (needs.project.result == 'success') }}
+    if: ${{ github.event.inputs.LOCAL_TEST_ARTIFACT_RUN_ID == '' && !cancelled() && (needs.project.result == 'success') }}
     steps:
       - name: Debug branch value
         run: echo "Using branch ${{ needs.determine-nt-version.outputs.nt_branch }}"
@@ -535,7 +535,7 @@ jobs:
     runs-on: macos-26
     # Skip this job if we're using artifacts from a specific run; the native bundles now come
     # from the project job (deploy + native-packager), so gate on project success.
-    if: ${{ github.event.inputs.LOCAL_TEST_ARTIFACT_RUN_ID == '' && always() && (needs.project.result == 'success') }}
+    if: ${{ github.event.inputs.LOCAL_TEST_ARTIFACT_RUN_ID == '' && !cancelled() && (needs.project.result == 'success') }}
     steps:
       - name: "Check out Native Template for Native Components Test Project"
         uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
@@ -647,7 +647,7 @@ jobs:
   android-widget-tests:
     needs: [scope, mendix-version, project, android-app, android-avd-cache]
     # Run if widgets need testing (widgets_to_test is not empty) and project succeeds
-    if: ${{ needs.scope.outputs.widgets_to_test != '[]' && always() && needs.project.result == 'success' && (needs.android-app.result == 'success' || needs.android-app.result == 'skipped') }}
+    if: ${{ needs.scope.outputs.widgets_to_test != '[]' && !cancelled() && needs.project.result == 'success' && (needs.android-app.result == 'success' || needs.android-app.result == 'skipped') }}
     runs-on: ubuntu-26.04
     # 30 min hard cap per widget shard. A healthy widget's flows run in ~30s each; the cap is a
     # backstop for genuinely wedged shards. Bumped 20→30 after a run where shards were cancelled
@@ -731,10 +731,12 @@ jobs:
             bash maestro/run_maestro_widget_tests.sh android "${{ matrix.widget }}" || [ "$UPDATE_BASELINES" = "true" ]
 
       - name: Archive test results
-        # always(): the test step exits non-zero on a failing shard, which would otherwise
+        # !cancelled(): the test step exits non-zero on a failing shard, which would otherwise
         # SKIP this step — i.e. we'd capture screenshots/logs/videos for green shards but NOT
-        # for the failing ones we actually need to debug. Run regardless of test outcome.
-        if: ${{ always() }}
+        # for the failing ones we actually need to debug. This covers pass AND fail, but
+        # unlike always() it is NOT exempt from cancellation, so a cancelled run stops here
+        # instead of first uploading five artifact sets per shard.
+        if: ${{ !cancelled() }}
         uses: ./.github/actions/archive-test-results
         with:
           platform: android
@@ -743,15 +745,15 @@ jobs:
   ios-widget-tests:
     needs: [scope, mendix-version, project, ios-app]
     # Run if widgets need testing (widgets_to_test is not empty) and project succeeds
-    if: ${{ needs.scope.outputs.widgets_to_test != '[]' && always() && needs.project.result == 'success' && (needs.ios-app.result == 'success' || needs.ios-app.result == 'skipped') }}
+    if: ${{ needs.scope.outputs.widgets_to_test != '[]' && !cancelled() && needs.project.result == 'success' && (needs.ios-app.result == 'success' || needs.ios-app.result == 'skipped') }}
     runs-on: macos-26
     # 30 min per widget shard, matching android-widget-tests. iOS needs the budget more than
     # Android: sim cold-boot + Maestro driver startup + runtime-ready wait eat several minutes
     # before any flow runs, so on a free macOS runner slow-but-healthy shards were getting
     # truncated — at 20 min several shards (accordion, background-image, bottom-sheet,
     # color-picker, safe-area-view) were cancelled mid-flow. If a shard still hits 30 the cause
     # is real app instability, not the cap.
-    timeout-minutes: 35
+    timeout-minutes: 30
     strategy:
       max-parallel: 4
       matrix:
@@ -803,8 +805,10 @@ jobs:
           bash maestro/run_maestro_widget_tests.sh ios "${{ matrix.widget }}" || [ "$UPDATE_BASELINES" = "true" ]
 
       - name: Archive test results
-        # always(): capture failure artifacts (the failing shard's test step exits non-zero).
-        if: ${{ always() }}
+        # !cancelled(): capture failure artifacts (the failing shard's test step exits
+        # non-zero), but do NOT hold a cancelled run open uploading artifacts nobody wants
+        # -- always() is exempt from cancellation, !cancelled() is not.
+        if: ${{ !cancelled() }}
         uses: ./.github/actions/archive-test-results
         with:
           platform: ios
@@ -813,7 +817,7 @@ jobs:
   android-js-tests:
     needs: [scope, mendix-version, project, android-app, android-avd-cache]
     # Run if JS actions changed and project succeeds and either android-app succeeds OR we're using custom artifacts (android-app was skipped)
-    if: ${{ needs.scope.outputs.js_actions_changed == 'true' && always() && needs.project.result == 'success' && (needs.android-app.result == 'success' || needs.android-app.result == 'skipped') }}
+    if: ${{ needs.scope.outputs.js_actions_changed == 'true' && !cancelled() && needs.project.result == 'success' && (needs.android-app.result == 'success' || needs.android-app.result == 'skipped') }}
     runs-on: ubuntu-26.04
     timeout-minutes: 90
     steps:
@@ -883,8 +887,10 @@ jobs:
             bash maestro/run_maestro_jsactions_tests.sh android || [ "$UPDATE_BASELINES" = "true" ]
 
       - name: Archive test results
-        # always(): capture failure artifacts (the failing shard's test step exits non-zero).
-        if: ${{ always() }}
+        # !cancelled(): capture failure artifacts (the failing shard's test step exits
+        # non-zero), but do NOT hold a cancelled run open uploading artifacts nobody wants
+        # -- always() is exempt from cancellation, !cancelled() is not.
+        if: ${{ !cancelled() }}
         uses: ./.github/actions/archive-test-results
         with:
           platform: android
@@ -893,7 +899,7 @@ jobs:
   ios-js-tests:
     needs: [scope, mendix-version, project, ios-app]
     # Run if JS actions changed and project succeeds and either ios-app succeeds OR we're using custom artifacts (ios-app was skipped)
-    if: ${{ needs.scope.outputs.js_actions_changed == 'true' && always() && needs.project.result == 'success' && (needs.ios-app.result == 'success' || needs.ios-app.result == 'skipped') }}
+    if: ${{ needs.scope.outputs.js_actions_changed == 'true' && !cancelled() && needs.project.result == 'success' && (needs.ios-app.result == 'success' || needs.ios-app.result == 'skipped') }}
     runs-on: macos-26
     timeout-minutes: 90
     steps:
@@ -941,8 +947,10 @@ jobs:
           bash maestro/run_maestro_jsactions_tests.sh ios || [ "$UPDATE_BASELINES" = "true" ]
 
       - name: Archive test results
-        # always(): capture failure artifacts (the failing shard's test step exits non-zero).
-        if: ${{ always() }}
+        # !cancelled(): capture failure artifacts (the failing shard's test step exits
+        # non-zero), but do NOT hold a cancelled run open uploading artifacts nobody wants
+        # -- always() is exempt from cancellation, !cancelled() is not.
+        if: ${{ !cancelled() }}
         uses: ./.github/actions/archive-test-results
         with:
           platform: ios
@@ -954,9 +962,11 @@ jobs:
   # sources after a successful merge, leaving just the combined artifacts.
   aggregate-test-results:
     needs: [android-widget-tests, ios-widget-tests, android-js-tests, ios-js-tests]
-    # Run even when shards fail/cancel — the failing shards are exactly the ones whose
-    # screenshots/logs we want to inspect.
-    if: ${{ always() }}
+    # Run even when shards fail or time out — those are exactly the ones whose
+    # screenshots/logs we want to inspect. !cancelled() rather than always() so an
+    # explicitly cancelled run isn't held open merging artifacts nobody asked for
+    # (a shard that hits timeout-minutes does not make cancelled() true).
+    if: ${{ !cancelled() }}
     runs-on: ubuntu-26.04
     # The workflow-level block grants only packages:write, so every other scope defaults to
     # none. This job needs contents:read (to check out the triage script) and actions:read

.github/workflows/Release.yml

@@ -6,6 +6,14 @@ on:
   pull_request:
     branches: [main, 'mx/11.12.x']
 
+# Supersede in-flight runs: a new push to the same PR/branch cancels the previous run
+# instead of queueing beside it (this repo shares GitHub Free's 20-concurrent-job cap with
+# NativePipeline, so stale runs directly delay the ones that matter). Only PR runs are
+# cancelled — a push run on main/mx-* is the recorded result for that commit, so let it finish.
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
+
 jobs:
   test:
     name: "Release (${{ matrix.os }})"

.github/workflows/ShaCheck.yml

@@ -10,6 +10,14 @@ on:
     paths:
       - ".github/workflows/*.yml"
 
+# Supersede in-flight runs: a new push to the same PR/branch cancels the previous run
+# instead of queueing beside it (this repo shares GitHub Free's 20-concurrent-job cap with
+# NativePipeline, so stale runs directly delay the ones that matter). Only PR runs are
+# cancelled — a push run on main/mx-* is the recorded result for that commit, so let it finish.
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
+
 jobs:
   harden_security:
     name: Check SHA in GH Actions

.github/workflows/UnitTests.yml

@@ -6,6 +6,14 @@ on:
   pull_request:
     branches: [main, 'mx/11.12.x']
 
+# Supersede in-flight runs: a new push to the same PR/branch cancels the previous run
+# instead of queueing beside it (this repo shares GitHub Free's 20-concurrent-job cap with
+# NativePipeline, so stale runs directly delay the ones that matter). Only PR runs are
+# cancelled — a push run on main/mx-* is the recorded result for that commit, so let it finish.
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: ${{ github.event_name == 'pull_request' }}
+
 jobs:
   test:
     name: "Unit tests"

maestro/helpers/helpers.sh

@@ -78,13 +78,62 @@ stop_recording() {
   REC_FILE=""
 }
 
+# Function to clean up stale XCTest processes (iOS only)
+cleanup_xctest_processes() {
+    if [ "$PLATFORM" != "ios" ]; then
+        return 0
+    fi
+    echo "🧹 Cleaning up stale XCTest processes..."
+    # Kill any leftover XCTest runner processes that might be blocking driver initialization
+    pkill -9 -f "XCTRunner" 2>/dev/null || true
+    pkill -9 -f "xctest" 2>/dev/null || true
+    pkill -9 -f "maestro.*driver" 2>/dev/null || true
+    sleep 2
+}
+
 # Function to restart the iOS simulator
+#
+# This is the recovery path taken when the Maestro/XCTest driver wedges ("iOS driver not
+# ready in time"), so it has to ACTUALLY restart the device. It previously did not: the
+# shutdown result was discarded with `|| true`, and the following prepare_ios.sh reported
+#
+#     Monitoring boot status for iPhone 17 Pro (...).
+#     Device already booted, nothing to do.
+#
+# i.e. the device never went down and the shard only got an app reinstall on top of the
+# same wedged runtime. So we now verify the device reaches Shutdown and say so loudly if
+# it doesn't, rather than silently continuing with a half-reset simulator.
 restart_simulator() {
     echo "🔄 Restarting iOS Simulator..."
-    # Shut down whatever is booted; the device is auto-selected in prepare_ios.sh,
-    # so we don't depend on a hardcoded device name here.
-    xcrun simctl shutdown all || true
-    sleep 10
+    # Kill the XCTest runner FIRST: while it holds the device, shutdown can fail.
+    cleanup_xctest_processes
+
+    # Target the device prepare_ios.sh selected when we know it; `all` otherwise (first
+    # call in a shard, where SIMULATOR_DEVICE_ID isn't exported into this shell yet).
+    local target="${SIMULATOR_DEVICE_ID:-all}"
+    echo "Shutting down simulator ($target)..."
+    if ! xcrun simctl shutdown "$target" 2>&1; then
+        # Already-shutdown is a benign error, so don't fail here — the poll below is what
+        # decides whether the device is actually down.
+        echo "shutdown reported an error; verifying device state anyway"
+    fi
+
+    # `simctl shutdown <udid>` is synchronous (measured ~3s), but poll rather than trust
+    # it: this replaces a blind `sleep 10` that padded every retry whether or not it was
+    # needed. Bounded so a stuck device surfaces as a warning instead of hanging.
+    local deadline=$((SECONDS + 30))
+    while [ "$SECONDS" -lt "$deadline" ]; do
+        if [ -n "${SIMULATOR_DEVICE_ID:-}" ]; then
+            xcrun simctl list devices | grep -q "$SIMULATOR_DEVICE_ID.*Shutdown" && break
+        else
+            xcrun simctl list devices | grep -q "(Booted)" || break
+        fi
+        sleep 1
+    done
+    if [ "$SECONDS" -ge "$deadline" ]; then
+        echo "::warning::Simulator did not reach Shutdown within 30s; the driver may still be wedged after this restart"
+    fi
+
     bash ./maestro/helpers/prepare_ios.sh
 }
 

maestro/helpers/prepare_ios.sh

@@ -64,6 +64,34 @@ elif best:
     xcrun simctl bootstatus "$DEVICE_ID" -b || echo "Simulator booted"
 }
 
+# Cut UIKit animation time, mirroring `disable-animations: true` on the Android emulator
+# (NativePipeline.yml) — iOS previously had no equivalent, so every navigation push/modal
+# played at full length on a 3-vCPU runner while Android's ran instantly.
+#
+# Scope: this shortens SYSTEM transitions (nav pushes, modal presents, alerts). React
+# Native's `Animated` does NOT read this flag automatically, so widget-internal animations
+# (carousel, animation-native) are unaffected — this is not a substitute for the per-flow
+# waits those need.
+#
+# Requires a BOOTED device (`simctl spawn`), so call after start_simulator. The value is
+# written to the device's persistent domain and survives shutdown/boot, so a
+# restart_simulator in helpers.sh keeps it.
+reduce_motion() {
+    echo "Reducing motion on iOS Simulator..."
+    if [ -z "$SIMULATOR_DEVICE_ID" ]; then
+        echo "Error: SIMULATOR_DEVICE_ID not set"
+        return 1
+    fi
+    # Best-effort: a future runtime that drops/renames these keys should not fail the shard,
+    # since tests still pass with animations on (just slower).
+    xcrun simctl spawn "$SIMULATOR_DEVICE_ID" defaults write com.apple.Accessibility ReduceMotionEnabled -bool true \
+        || echo "::warning::Could not set ReduceMotionEnabled; animations stay at full length"
+    # Replaces the sliding push/pop transition with a cross-fade — the slide is what Maestro
+    # most often races against when asserting right after a navigation tap.
+    xcrun simctl spawn "$SIMULATOR_DEVICE_ID" defaults write com.apple.Accessibility ReduceMotionReduceSlideTransitionsEnabled -bool true \
+        || echo "::warning::Could not set ReduceMotionReduceSlideTransitionsEnabled"
+}
+
 set_status_bar() {
     echo "Setting status bar on iOS Simulator..."
     if [ -z "$SIMULATOR_DEVICE_ID" ]; then
@@ -92,6 +120,7 @@ verify_installed_app() {
 }
 
 start_simulator
+reduce_motion
 set_status_bar
 install_ios_app
 verify_installed_app

maestro/run_maestro_jsactions_tests.sh

@@ -97,6 +97,16 @@ run_jsactions_tests() {
   fi
 }
 
+# Clean up any stale XCTest processes before starting (iOS only)
+if [ "$PLATFORM" == "ios" ]; then
+  cleanup_xctest_processes
+fi
+
+# Fast-fail smoke check before running jsActions tests (same as widget tests)
+if ! smoke_check; then
+  exit 1
+fi
+
 # Run jsActions tests
 run_jsactions_tests
 if [ $? -ne 0 ]; then