From: Mike Bayer <mike_mp@zzzcomputing.com>
Date: Fri, 8 Apr 2022 13:48:26 +0000 (-0400)
Subject: update comment re: #2703
X-Git-Tag: rel_2_0_0b1~368
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=9b32d9040a7d5716ae3871acd76a598f6c402423;p=thirdparty%2Fsqlalchemy%2Fsqlalchemy.git

update comment re: #2703

Change-Id: Ia78db7601e98acd2562872e0fda639aac99650dd
---

diff --git a/lib/sqlalchemy/util/langhelpers.py b/lib/sqlalchemy/util/langhelpers.py
index 9e1194e231..00097fd9e7 100644
--- a/lib/sqlalchemy/util/langhelpers.py
+++ b/lib/sqlalchemy/util/langhelpers.py
@@ -106,10 +106,12 @@ class safe_reraise:
             with safe_reraise():
                 sess.rollback()
 
-    TODO: is this context manager getting us anything in Python 3?
-    Not sure of the coroutine issue stated above; we would assume this was
-    when using eventlet / gevent.  not sure if our own greenlet integration
-    is impacted.
+    TODO: we should at some point evaluate current behaviors in this regard
+    based on current greenlet, gevent/eventlet implementations in Python 3, and
+    also see the degree to which our own asyncio (based on greenlet also) is
+    impacted by this. .rollback() will cause IO / context switch to occur in
+    all these scenarios; what happens to the exception context from an
+    "except:" block if we don't explicitly store it? Original issue was #2703.
 
     """