Author martin.panter
Recipients John Jones, alex, benjamin.peterson, dhduvall, gennad, gregory.p.smith, martin.panter, neologix, pablogsal
Date 2018-01-29.10:03:27
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1517220207.17.0.467229070634.issue20104@psf.upfronthosting.co.za>
In-reply-to
Content
Does the PySequence_Fast result need releasing if the following “for” loop fails? There is a Py_DECREF only in the successful case, which seems inconsistent.

Does Python still support non-UTF-8 locales and bytes filenames? I haven’t been keeping up, but I assumed these things were still supported in many cases. It seems strange to only support UTF-8 in the “addopen” file action.

Pablo’s second PR currently <https://github.com/python/cpython/pull/5418/commits/e50bdb9> calls PyErr_SetString(PyExc_OSError, . . .) with a custom error message depending on which OS call failed. But I wonder if it is more important to set the “errno” attribute (which I think should choose an OSError subclass if appropriate). Perhaps you can do that by assigning the return values to “errno” and then calling PyErr_SetFromErrno. A disadvantage might be less context about which stage went wrong.
History
Date User Action Args
2018-01-29 10:03:27martin.pantersetrecipients: + martin.panter, gregory.p.smith, benjamin.peterson, alex, dhduvall, neologix, gennad, John Jones, pablogsal
2018-01-29 10:03:27martin.pantersetmessageid: <1517220207.17.0.467229070634.issue20104@psf.upfronthosting.co.za>
2018-01-29 10:03:27martin.panterlinkissue20104 messages
2018-01-29 10:03:27martin.pantercreate