Message 325008 - Python tracker

➜

This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author	ncoghlan
Recipients	ncoghlan, ned.deily, vstinner
Date	2018-09-11.14:20:56
SpamBayes Score	-1.0
Marked as misclassified	Yes
Message-id	<1536675656.27.0.0269046726804.issue34589@psf.upfronthosting.co.za>
In-reply-to

Content
The entire change affecting the PEP 538 implementation in https://github.com/python/cpython/commit/9454060e84a669dde63824d9e2fcaf295e34f687#diff-8c018c3ada66d06c8e101e47a313c2c7 needs to be reverted: the locale should be coerced before ANY calls are made to Py_DecodeLocale, as the whole point of the architecture of the PEP implementation was to ensure that nothing ever gets decoded incorrectly in the first place. I just initially missed that in Victor's enthusiasm for fixing the incorrect decodings that can arise in the absence of locale coercion he'd introduced incorrect decoding into the locale coercion case.

The entire change affecting the PEP 538 implementation in https://github.com/python/cpython/commit/9454060e84a669dde63824d9e2fcaf295e34f687#diff-8c018c3ada66d06c8e101e47a313c2c7 needs to be reverted: the locale should be coerced before *ANY* calls are made to Py_DecodeLocale, as the whole point of the architecture of the PEP implementation was to ensure that *nothing ever gets decoded incorrectly in the first place*.

I just initially missed that in Victor's enthusiasm for fixing the incorrect decodings that can arise in the absence of locale coercion he'd *introduced* incorrect decoding into the locale coercion case.

History
Date	User	Action	Args
2018-09-11 14:20:56	ncoghlan	set	recipients: + ncoghlan, vstinner, ned.deily
2018-09-11 14:20:56	ncoghlan	set	messageid: <1536675656.27.0.0269046726804.issue34589@psf.upfronthosting.co.za>
2018-09-11 14:20:56	ncoghlan	link	issue34589 messages
2018-09-11 14:20:56	ncoghlan	create