gh-98608: Fix Failure-handling in new_interpreter() (gh-102658)

The error-handling code in new_interpreter() has been broken for a while.  We hadn't noticed because those code mostly doesn't fail.  (I noticed while working on gh-101660.)  The problem is that we try to clear/delete the newly-created thread/interpreter using itself, which just failed.  The solution is to switch back to the calling thread state first.

(cherry picked from commit d1b883b52a99427d234c20e4a92ddfa6a1da8880)

Co-authored-by: Eric Snow <ericsnowcurrently@gmail.com>
https: //github.com/python/cpython/issues/98608

diff --git a/Python/pylifecycle.c b/Python/pylifecycle.c
index 4060c23ace1c454b664bdc831a7a4d5cce9cecd5..9248e971d9c78fa10daf68652d558701184623b9 100644 (file)
--- a/Python/pylifecycle.c
+++ b/Python/pylifecycle.c
@@ -2028,10 +2028,10 @@ error:
 
     /* Oops, it didn't work.  Undo it all. */
     PyErr_PrintEx(0);
+    PyThreadState_Swap(save_tstate);
     PyThreadState_Clear(tstate);
     PyThreadState_Delete(tstate);
     PyInterpreterState_Delete(interp);
-    PyThreadState_Swap(save_tstate);
 
     return status;
 }