2 # -*- coding: utf-8 -*-
5 # :Copyright: © 2011 Günter Milde.
6 # :License: Released under the terms of the `2-Clause BSD license`_, in short:
8 # Copying and distribution of this file, with or without modification,
9 # are permitted in any medium without royalty provided the copyright
10 # notice and this notice are preserved.
11 # This file is offered as-is, without any warranty.
13 # .. _2-Clause BSD license: http://www.spdx.org/licenses/BSD-2-Clause
16 Error reporting should be safe from encoding/decoding errors.
17 However, implicit conversions of strings and exceptions like
19 >>> u'%s world: %s' % ('H\xe4llo', Exception(u'H\xe4llo')
21 fail in some Python versions:
23 * In Python <= 2.6, ``unicode(<exception instance>)`` uses
24 `__str__` and fails with non-ASCII chars in`unicode` arguments.
25 (work around http://bugs.python.org/issue2517):
27 * In Python 2, unicode(<exception instance>) fails, with non-ASCII
28 chars in arguments. (Use case: in some locales, the errstr
29 argument of IOError contains non-ASCII chars.)
31 * In Python 2, str(<exception instance>) fails, with non-ASCII chars
32 in `unicode` arguments.
34 The `SafeString`, `ErrorString` and `ErrorOutput` classes handle
40 # Guess the locale's encoding.
41 # If no valid guess can be made, locale_encoding is set to `None`:
43 import locale
# module missing in Jython
45 locale_encoding
= None
47 locale_encoding
= locale
.getlocale()[1] or locale
.getdefaultlocale()[1]
48 # locale.getpreferredencoding([do_setlocale=True|False])
49 # has side-effects | might return a wrong guess.
50 # (cf. Update 1 in http://stackoverflow.com/questions/4082645/using-python-2-xs-locale-module-to-format-numbers-and-currency)
52 codecs
.lookup(locale_encoding
or '') # None -> ''
54 locale_encoding
= None
58 class SafeString(object):
60 A wrapper providing robust conversion to `str` and `unicode`.
63 def __init__(self
, data
, encoding
=None, encoding_errors
='backslashreplace',
64 decoding_errors
='replace'):
66 self
.encoding
= (encoding
or getattr(data
, 'encoding', None) or
67 locale_encoding
or 'ascii')
68 self
.encoding_errors
= encoding_errors
69 self
.decoding_errors
= decoding_errors
75 except UnicodeEncodeError, err
:
76 if isinstance(self
.data
, Exception):
77 args
= [str(SafeString(arg
, self
.encoding
,
78 self
.encoding_errors
))
79 for arg
in self
.data
.args
]
80 return ', '.join(args
)
81 if isinstance(self
.data
, unicode):
82 if sys
.version_info
> (3,0):
85 return self
.data
.encode(self
.encoding
,
89 def __unicode__(self
):
91 Return unicode representation of `self.data`.
93 Try ``unicode(self.data)``, catch `UnicodeError` and
95 * if `self.data` is an Exception instance, work around
96 http://bugs.python.org/issue2517 with an emulation of
97 Exception.__unicode__,
99 * else decode with `self.encoding` and `self.decoding_errors`.
102 u
= unicode(self
.data
)
103 if isinstance(self
.data
, EnvironmentError):
104 u
= u
.replace(": u'", ": '") # normalize filename quoting
106 except UnicodeError, error
: # catch ..Encode.. and ..Decode.. errors
107 if isinstance(self
.data
, EnvironmentError):
108 return u
"[Errno %s] %s: '%s'" % (self
.data
.errno
,
109 SafeString(self
.data
.strerror
, self
.encoding
,
110 self
.decoding_errors
),
111 SafeString(self
.data
.filename
, self
.encoding
,
112 self
.decoding_errors
))
113 if isinstance(self
.data
, Exception):
114 args
= [unicode(SafeString(arg
, self
.encoding
,
115 decoding_errors
=self
.decoding_errors
))
116 for arg
in self
.data
.args
]
117 return u
', '.join(args
)
118 if isinstance(error
, UnicodeDecodeError):
119 return unicode(self
.data
, self
.encoding
, self
.decoding_errors
)
122 class ErrorString(SafeString
):
124 Safely report exception type and message.
127 return '%s: %s' % (self
.data
.__class
__.__name
__,
128 super(ErrorString
, self
).__str
__())
130 def __unicode__(self
):
131 return u
'%s: %s' % (self
.data
.__class
__.__name
__,
132 super(ErrorString
, self
).__unicode
__())
135 class ErrorOutput(object):
137 Wrapper class for file-like error streams with
138 failsave de- and encoding of `str`, `bytes`, `unicode` and
139 `Exception` instances.
142 def __init__(self
, stream
=None, encoding
=None,
143 encoding_errors
='backslashreplace',
144 decoding_errors
='replace'):
147 - `stream`: a file-like object,
148 a string (path to a file),
149 `None` (write to `sys.stderr`, default), or
150 evaluating to `False` (write() requests are ignored).
151 - `encoding`: `stream` text encoding. Guessed if None.
152 - `encoding_errors`: how to treat encoding errors.
158 # if `stream` is a file name, open it
159 elif isinstance(stream
, str):
160 stream
= open(stream
, 'w')
161 elif isinstance(stream
, unicode):
162 stream
= open(stream
.encode(sys
.getfilesystemencoding()), 'w')
165 """Where warning output is sent."""
167 self
.encoding
= (encoding
or getattr(stream
, 'encoding', None) or
168 locale_encoding
or 'ascii')
169 """The output character encoding."""
171 self
.encoding_errors
= encoding_errors
172 """Encoding error handler."""
174 self
.decoding_errors
= decoding_errors
175 """Decoding error handler."""
177 def write(self
, data
):
179 Write `data` to self.stream. Ignore, if self.stream is False.
181 `data` can be a `string`, `unicode`, or `Exception` instance.
183 if self
.stream
is False:
185 if isinstance(data
, Exception):
186 data
= unicode(SafeString(data
, self
.encoding
,
187 self
.encoding_errors
, self
.decoding_errors
))
189 self
.stream
.write(data
)
190 except UnicodeEncodeError:
191 self
.stream
.write(data
.encode(self
.encoding
, self
.encoding_errors
))
193 if isinstance(data
, unicode): # passed stream may expect bytes
194 self
.stream
.write(data
.encode(self
.encoding
,
195 self
.encoding_errors
))
197 if self
.stream
in (sys
.stderr
, sys
.stdout
):
198 self
.stream
.buffer.write(data
) # write bytes to raw stream
200 self
.stream
.write(unicode(data
, self
.encoding
,
201 self
.decoding_errors
))
205 Close the error-output stream.
207 Ignored if the stream is` sys.stderr` or `sys.stdout` or has no
210 if self
.stream
in (sys
.stdout
, sys
.stderr
):
214 except AttributeError: