comparison planemo/lib/python3.7/site-packages/pip/_internal/utils/encoding.py @ 1:56ad4e20f292 draft

"planemo upload commit 6eee67778febed82ddd413c3ca40b3183a3898f1"
author guerler
date Fri, 31 Jul 2020 00:32:28 -0400
parents
children
comparison
equal deleted inserted replaced
0:d30785e31577 1:56ad4e20f292
1 import codecs
2 import locale
3 import re
4 import sys
5
6 from pip._internal.utils.typing import MYPY_CHECK_RUNNING
7
8 if MYPY_CHECK_RUNNING:
9 from typing import List, Tuple, Text
10
11 BOMS = [
12 (codecs.BOM_UTF8, 'utf-8'),
13 (codecs.BOM_UTF16, 'utf-16'),
14 (codecs.BOM_UTF16_BE, 'utf-16-be'),
15 (codecs.BOM_UTF16_LE, 'utf-16-le'),
16 (codecs.BOM_UTF32, 'utf-32'),
17 (codecs.BOM_UTF32_BE, 'utf-32-be'),
18 (codecs.BOM_UTF32_LE, 'utf-32-le'),
19 ] # type: List[Tuple[bytes, Text]]
20
21 ENCODING_RE = re.compile(br'coding[:=]\s*([-\w.]+)')
22
23
24 def auto_decode(data):
25 # type: (bytes) -> Text
26 """Check a bytes string for a BOM to correctly detect the encoding
27
28 Fallback to locale.getpreferredencoding(False) like open() on Python3"""
29 for bom, encoding in BOMS:
30 if data.startswith(bom):
31 return data[len(bom):].decode(encoding)
32 # Lets check the first two lines as in PEP263
33 for line in data.split(b'\n')[:2]:
34 if line[0:1] == b'#' and ENCODING_RE.search(line):
35 encoding = ENCODING_RE.search(line).groups()[0].decode('ascii')
36 return data.decode(encoding)
37 return data.decode(
38 locale.getpreferredencoding(False) or sys.getdefaultencoding(),
39 )