From b1ccb277dad54287d4e9bcde9113a1bf429eea81 Mon Sep 17 00:00:00 2001
From: Adrian Chaves <adrian@zyte.com>
Date: Fri, 19 Jun 2026 12:40:20 +0200
Subject: [PATCH] Make _schedule_coro() behavior the same in asyncio and
 non-asyncio

---
 scrapy/utils/defer.py | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/scrapy/utils/defer.py b/scrapy/utils/defer.py
index 29a34d4ef..699b51505 100644
--- a/scrapy/utils/defer.py
+++ b/scrapy/utils/defer.py
@@ -531,9 +531,14 @@ def _schedule_coro(coro: Coroutine[Any, Any, Any]) -> None:
     alternative is calling :func:`scrapy.utils.defer.deferred_from_coro`,
     keeping the result, and adding proper exception handling (e.g. errbacks) to
     it.
+
+    In both asyncio and non-asyncio modes the coroutine starts in the next
+    event-loop iteration, never in the current call stack.
     """
     if not is_asyncio_available():
-        Deferred.fromCoroutine(coro)
+        from twisted.internet import reactor
+
+        reactor.callLater(0, Deferred.fromCoroutine, coro)
         return
     loop = asyncio.get_event_loop()
     loop.create_task(coro)  # noqa: RUF006