Author martin.panter
Recipients Dhiraj_Mishra, martin.panter, mbussonn, python-dev, serhiy.storchaka, takluyver
Date 2016-02-25.12:12:00
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1456402321.15.0.658891183975.issue26039@psf.upfronthosting.co.za>
In-reply-to
Content
When you say concurrent writes should be impossible, I guess that only applies to a single-threaded program. There is no lock protecting the “self._fileRefCnt > 1” check and related manipulation (not that I am saying there should be).

For serializing concurrent writes to a single handle, if that is intended it should be documented. If it is not intended, maybe it should be removed (my preference)?

It would be good to wait if Serhiy can explain the purpose of the lock, seeing he was involved in adding it and probably knows a lot more about this module than I.
History
Date User Action Args
2016-02-25 12:12:01martin.pantersetrecipients: + martin.panter, python-dev, takluyver, serhiy.storchaka, mbussonn, Dhiraj_Mishra
2016-02-25 12:12:01martin.pantersetmessageid: <1456402321.15.0.658891183975.issue26039@psf.upfronthosting.co.za>
2016-02-25 12:12:01martin.panterlinkissue26039 messages
2016-02-25 12:12:00martin.pantercreate