Fix incremental decoder and stream reader in the "raw-unicode-escape" codec

[serhiy-storchaka]

BPO	45467
Nosy	@malemburg, @vstinner, @ezio-melotti, @serhiy-storchaka
PRs	bpo-45467: Fix IncrementalDecoder and StreamReader in the "raw-unicode-escape" codec #28944 [3.10] bpo-45467: Fix IncrementalDecoder and StreamReader in the "raw… #28952 [3.9] bpo-45467: Fix IncrementalDecoder and StreamReader in the "raw-unicod… #28953
Dependencies	bpo-45461: UnicodeDecodeError: 'unicodeescape' codec can't decode byte 0x5c in position 8191: \ at end of string

^{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 2021-10-14.19:17:23.992>
created_at = <Date 2021-10-14.10:29:46.781>
labels = ['3.10', 'type-bug', '3.9', 'expert-unicode', '3.11']
title = 'Fix incremental decoder and stream reader in the "raw-unicode-escape" codec'
updated_at = <Date 2021-10-14.19:17:23.992>
user = 'https://github.com/serhiy-storchaka'

bugs.python.org fields:

activity = <Date 2021-10-14.19:17:23.992>
actor = 'serhiy.storchaka'
assignee = 'none'
closed = True
closed_date = <Date 2021-10-14.19:17:23.992>
closer = 'serhiy.storchaka'
components = ['Unicode']
creation = <Date 2021-10-14.10:29:46.781>
creator = 'serhiy.storchaka'
dependencies = ['45461']
files = []
hgrepos = []
issue_num = 45467
keywords = ['patch']
message_count = 5.0
messages = ['403893', '403921', '403927', '403928', '403935']
nosy_count = 4.0
nosy_names = ['lemburg', 'vstinner', 'ezio.melotti', 'serhiy.storchaka']
pr_nums = ['28944', '28952', '28953']
priority = 'normal'
resolution = 'fixed'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue45467'
versions = ['Python 3.9', 'Python 3.10', 'Python 3.11']

[serhiy-storchaka]

Similar to 45461, but with "raw-unicode-escape".

When an incremental decoder gets a part of escape sequence (\uXXXX or \UXXXXXXXX) it raises an exception or return a bare "\" if it was the only part instead of keeping it until getting the rest. It is exposed in text files (io.TextIOWrapper) when reads from the underlying binary stream splits an escape sequence between blocks. There is similar issue with stream readers (codecs.StreamReader).

[serhiy-storchaka]

New changeset 39aa983 by Serhiy Storchaka in branch 'main':
bpo-45467: Fix IncrementalDecoder and StreamReader in the "raw-unicode-escape" codec (GH-28944)
39aa983

[serhiy-storchaka]

New changeset 4641afe by Serhiy Storchaka in branch '3.10':
[3.10] bpo-45467: Fix IncrementalDecoder and StreamReader in the "raw-unicode-escape" codec (GH-28944) (GH-28952)
4641afe

[serhiy-storchaka]

New changeset 6848602 by Serhiy Storchaka in branch '3.9':
bpo-45467: Fix IncrementalDecoder and StreamReader in the "raw-unicode-escape" codec (GH-28944) (GH-28953)
6848602

[vstinner]

Serhiy: I suppose that this issue can now be closed?