Regression in pathlib.path.read_text

[kusma]

BPO	44487
Nosy	@pfmoore, @tjguk, @methane, @zware, @zooba, @kusma

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2021-06-22.13:29:19.054>
created_at = <Date 2021-06-22.10:24:16.413>
labels = ['library', '3.10', 'OS-windows', '3.11']
title = 'Regression in pathlib.path.read_text'
updated_at = <Date 2021-06-22.13:40:22.840>
user = 'https://github.com/kusma'

bugs.python.org fields:

activity = <Date 2021-06-22.13:40:22.840>
actor = 'methane'
assignee = 'none'
closed = True
closed_date = <Date 2021-06-22.13:29:19.054>
closer = 'kusma'
components = ['Library (Lib)', 'Windows']
creation = <Date 2021-06-22.10:24:16.413>
creator = 'kusma'
dependencies = []
files = []
hgrepos = []
issue_num = 44487
keywords = []
message_count = 4.0
messages = ['396318', '396328', '396330', '396331']
nosy_count = 6.0
nosy_names = ['paul.moore', 'tim.golden', 'methane', 'zach.ware', 'steve.dower', 'kusma']
pr_nums = []
priority = 'normal'
resolution = None
stage = 'resolved'
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue44487'
versions = ['Python 3.10', 'Python 3.11']

[kusma]

This commit lead to a regression when using Meson on Windows to build the Mesa project:

4827483

The reason is that pathlib.read_text now uses the locale as the encoding when reading files when there's no encoding explicitly passed in. That means that C++ source files are attempted read as CP1252 on Windows, which throws exceptions when source files contain UTF-8 code-points.

[kusma]

After digging some more, I no longer suspect that this commit is to blame, but instead some logic in Meson that effectively disabled the problematic code under in our use-case before we upgraded our Visual Studio version.

The reason is that I was able to reproduce this using Python 3.9, and as far as I can tell that doesn't contain this change.

So, sorry for the noise. I'll close this, and rather reopen if I find out I'm wrong.

[methane]

Please write a link to the original issue next time.
mesonbuild/meson#8916