Add pauses and log repeats to avoid issues with detecting test completion on iOS/macOS.

tests/testbed.py

@@ -1,5 +1,7 @@
 import os
+import sys
 import tempfile
+import time
 from pathlib import Path
 
 import pytest
@@ -23,10 +25,30 @@ def run_tests():
         ]
     )
 
-    print(f">>>>>>>>>> EXIT {returncode} <<<<<<<<<<")
+    # On iOS and macOS, Briefcase needs to use a log streamer. However, this
+    # streamer can drop lines of test output due to system load. If the log
+    # result sentinel isn't seen in streamed output, the test will fail because
+    # the exit status of the test isn't known. As a safety measure on iOS and
+    # macOS, output the log sentinel multiple times to decrease the chance of
+    # this happening.
+    if sys.platform in {"darwin", "ios"}:
+        for i in range(0, 6):
+            time.sleep(0.5)
+            print(f">>>>>>>>>> EXIT {returncode} <<<<<<<<<<")
+    else:
+        print(f">>>>>>>>>> EXIT {returncode} <<<<<<<<<<")
 
 
 if __name__ == "__main__":
+    # On iOS and macOS, Briefcase needs to use a log streamer. However, this
+    # streamer takes time to start up. As a safety measure, wait a couple of seconds
+    # at app start, and output some lines to minimize the impact of the streamer
+    # startup.
+    if sys.platform in {"darwin", "ios"}:
+        print("Waiting for log streamer...")
+        time.sleep(2.0)
+        print("Log streamer should be ready.")
+
     # Run the app main to stimulate app creation and logging of test conditions.
     app_main()
     # Run the actual test suite.