Coverage for /pythoncovmergedfiles/medio/medio/usr/local/lib/python3.11/site-packages/IPython/utils/encoding.py: 67%

1# coding: utf-8 

2""" 

3Utilities for dealing with text encodings 

4""" 

5 

6from __future__ import annotations 

7 

8# ----------------------------------------------------------------------------- 

9# Copyright (C) 2008-2012 The IPython Development Team 

10# 

11# Distributed under the terms of the BSD License. The full license is in 

12# the file COPYING, distributed as part of this software. 

13# ----------------------------------------------------------------------------- 

14 

15# ----------------------------------------------------------------------------- 

16# Imports 

17# ----------------------------------------------------------------------------- 

18import sys 

19import locale 

20import warnings 

21from typing import Any, Literal 

22 

23 

24# to deal with the possibility of sys.std* not being a stream at all 

25def get_stream_enc(stream: Any, default: str | None = None) -> str | None: 

26 """Return the given stream's encoding or a default. 

27 

28 There are cases where ``sys.std*`` might not actually be a stream, so 

29 check for the encoding attribute prior to returning it, and return 

30 a default if it doesn't exist or evaluates as False. ``default`` 

31 is None if not provided. 

32 """ 

33 if not hasattr(stream, "encoding") or not stream.encoding: 

34 return default 

35 else: 

36 return stream.encoding 

37 

38 

39_sentinel: object = object() 

40 

41 

42# Less conservative replacement for sys.getdefaultencoding, that will try 

43# to match the environment. 

44# Defined here as central function, so if we find better choices, we 

45# won't need to make changes all over IPython. 

46def getdefaultencoding(prefer_stream: object | bool = _sentinel) -> str: 

47 """Return IPython's guess for the default encoding for bytes as text. 

48 

49 If prefer_stream is True (default), asks for stdin.encoding first, 

50 to match the calling Terminal, but that is often None for subprocesses. 

51 

52 Then fall back on locale.getpreferredencoding(), 

53 which should be a sensible platform default (that respects LANG environment), 

54 and finally to sys.getdefaultencoding() which is the most conservative option, 

55 and usually UTF8 as of Python 3. 

56 """ 

57 if prefer_stream is not _sentinel: 

58 warnings.warn( 

59 "getpreferredencoding(prefer_stream=) argument is deprecated since " 

60 "IPython 9.0, getdefaultencoding() will take no argument in the " 

61 "future. If you rely on `prefer_stream`, please open an issue on " 

62 "the IPython repo.", 

63 DeprecationWarning, 

64 stacklevel=2, 

65 ) 

66 prefer_stream = True 

67 enc: str | None = None 

68 if prefer_stream: 

69 enc = get_stream_enc(sys.stdin) 

70 if not enc or enc == "ascii": 

71 try: 

72 # There are reports of getpreferredencoding raising errors 

73 # in some cases, which may well be fixed, but let's be conservative here. 

74 enc = locale.getpreferredencoding() 

75 except Exception: 

76 pass 

77 enc = enc or sys.getdefaultencoding() 

78 # On windows `cp0` can be returned to indicate that there is no code page. 

79 # Since cp0 is an invalid encoding return instead cp1252 which is the 

80 # Western European default. 

81 if enc == "cp0": 

82 warnings.warn( 

83 "Invalid code page cp0 detected - using cp1252 instead." 

84 "If cp1252 is incorrect please ensure a valid code page " 

85 "is defined for the process.", 

86 RuntimeWarning, 

87 ) 

88 return "cp1252" 

89 return enc 

90 

91 

92DEFAULT_ENCODING = getdefaultencoding()