fix(e2e): Wait for native replay buffer before capturing exception

The captureReplay e2e test taps "Capture Exception" immediately after
the app reports ready. On slow CI simulators the native replay buffer
may not have captured its first frame yet, causing captureReplay() to
return false and the replay_id to be missing from the error event.

Add a short Maestro flow with UI interactions (~3s of swipes) between
app init and the exception capture to let the buffer populate.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: antonis

dev-packages/e2e-tests/maestro/captureReplay.yml

@@ -5,6 +5,10 @@ jsEngine: graaljs
     file: utils/launchTestAppClear.yml
     env:
       replaysOnErrorSampleRate: 1.0
+- runFlow:
+    file: utils/waitForReplayBuffer.yml
+    when:
+      platform: iOS
 - tapOn: "Capture Exception"
 - runFlow: utils/assertEventIdVisible.yml
 - runFlow:

dev-packages/e2e-tests/maestro/utils/waitForReplayBuffer.yml

@@ -0,0 +1,23 @@
+appId: ${APP_ID}
+jsEngine: graaljs
+---
+# The native replay integration captures screenshots at ~1 FPS.
+# After Sentry.init completes (signalled by "E2E Tests Ready"),
+# the replay buffer may still be empty. If captureReplay() is
+# called before any frames exist, it returns false and the
+# replay_id is never attached to the error event.
+#
+# Perform UI interactions to both create a natural delay (~3s)
+# and give the replay meaningful content to capture.
+- swipe:
+    direction: DOWN
+    duration: 800
+- swipe:
+    direction: UP
+    duration: 800
+- swipe:
+    direction: DOWN
+    duration: 800
+- swipe:
+    direction: UP
+    duration: 800