[issue9797] wrong assumption in pystate.c

Antoine Pitrou added the comment: Committed in r84623 (3.x), r84627 (2.7) and r84628 (3.1). -- resolution: -> fixed stage: patch review -> committed/rejected status: open -> closed ___ Python tracker _

[issue9797] wrong assumption in pystate.c

Antoine Pitrou added the comment: Ok, it is simpler indeed. Here is an updated patch. -- Added file: http://bugs.python.org/file18795/autotlskey2.patch ___ Python tracker ___ ___

[issue9797] wrong assumption in pystate.c

Amaury Forgeot d'Arc added the comment: Couldn't the patch use autoInterpreterState instead? It is already modified at the same times as autoTLSkey. -- ___ Python tracker ___ ___

[issue9797] wrong assumption in pystate.c

New submission from Antoine Pitrou : pystate.c assumes that when autoTLSkey is 0, it hasn't been created yet. However, some TLS implementations can return 0 as a valid key value. Lots of interesting things then happen. Here is a patch. -- files: autotlskey.patch keywords: patch messag