author | Sverre Rabbelier <srabbelier@gmail.com> |
Wed, 21 Jan 2009 16:13:03 +0000 | |
changeset 877 | b2575da44c36 |
parent 323 | ff1a9aa48cfd |
permissions | -rw-r--r-- |
54
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
1 |
import types |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
2 |
import urllib |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
3 |
import datetime |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
4 |
from django.utils.functional import Promise |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
5 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
6 |
class DjangoUnicodeDecodeError(UnicodeDecodeError): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
7 |
def __init__(self, obj, *args): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
8 |
self.obj = obj |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
9 |
UnicodeDecodeError.__init__(self, *args) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
10 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
11 |
def __str__(self): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
12 |
original = UnicodeDecodeError.__str__(self) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
13 |
return '%s. You passed in %r (%s)' % (original, self.obj, |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
14 |
type(self.obj)) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
15 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
16 |
class StrAndUnicode(object): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
17 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
18 |
A class whose __str__ returns its __unicode__ as a UTF-8 bytestring. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
19 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
20 |
Useful as a mix-in. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
21 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
22 |
def __str__(self): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
23 |
return self.__unicode__().encode('utf-8') |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
24 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
25 |
def smart_unicode(s, encoding='utf-8', strings_only=False, errors='strict'): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
26 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
27 |
Returns a unicode object representing 's'. Treats bytestrings using the |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
28 |
'encoding' codec. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
29 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
30 |
If strings_only is True, don't convert (some) non-string-like objects. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
31 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
32 |
if isinstance(s, Promise): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
33 |
# The input is the result of a gettext_lazy() call. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
34 |
return s |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
35 |
return force_unicode(s, encoding, strings_only, errors) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
36 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
37 |
def force_unicode(s, encoding='utf-8', strings_only=False, errors='strict'): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
38 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
39 |
Similar to smart_unicode, except that lazy instances are resolved to |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
40 |
strings, rather than kept as lazy objects. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
41 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
42 |
If strings_only is True, don't convert (some) non-string-like objects. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
43 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
44 |
if strings_only and isinstance(s, (types.NoneType, int, long, datetime.datetime, datetime.date, datetime.time, float)): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
45 |
return s |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
46 |
try: |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
47 |
if not isinstance(s, basestring,): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
48 |
if hasattr(s, '__unicode__'): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
49 |
s = unicode(s) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
50 |
else: |
323
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
51 |
try: |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
52 |
s = unicode(str(s), encoding, errors) |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
53 |
except UnicodeEncodeError: |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
54 |
if not isinstance(s, Exception): |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
55 |
raise |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
56 |
# If we get to here, the caller has passed in an Exception |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
57 |
# subclass populated with non-ASCII data without special |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
58 |
# handling to display as a string. We need to handle this |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
59 |
# without raising a further exception. We do an |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
60 |
# approximation to what the Exception's standard str() |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
61 |
# output should be. |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
62 |
s = ' '.join([force_unicode(arg, encoding, strings_only, |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
63 |
errors) for arg in s]) |
54
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
64 |
elif not isinstance(s, unicode): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
65 |
# Note: We use .decode() here, instead of unicode(s, encoding, |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
66 |
# errors), so that if s is a SafeString, it ends up being a |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
67 |
# SafeUnicode at the end. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
68 |
s = s.decode(encoding, errors) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
69 |
except UnicodeDecodeError, e: |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
70 |
raise DjangoUnicodeDecodeError(s, *e.args) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
71 |
return s |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
72 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
73 |
def smart_str(s, encoding='utf-8', strings_only=False, errors='strict'): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
74 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
75 |
Returns a bytestring version of 's', encoded as specified in 'encoding'. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
76 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
77 |
If strings_only is True, don't convert (some) non-string-like objects. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
78 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
79 |
if strings_only and isinstance(s, (types.NoneType, int)): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
80 |
return s |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
81 |
if isinstance(s, Promise): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
82 |
return unicode(s).encode(encoding, errors) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
83 |
elif not isinstance(s, basestring): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
84 |
try: |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
85 |
return str(s) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
86 |
except UnicodeEncodeError: |
323
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
87 |
if isinstance(s, Exception): |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
88 |
# An Exception subclass containing non-ASCII data that doesn't |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
89 |
# know how to print itself properly. We shouldn't raise a |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
90 |
# further exception. |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
91 |
return ' '.join([smart_str(arg, encoding, strings_only, |
ff1a9aa48cfd
Load ../vendor/django into trunk/app/django.
Pawel Solyga <Pawel.Solyga@gmail.com>
parents:
54
diff
changeset
|
92 |
errors) for arg in s]) |
54
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
93 |
return unicode(s).encode(encoding, errors) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
94 |
elif isinstance(s, unicode): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
95 |
return s.encode(encoding, errors) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
96 |
elif s and encoding != 'utf-8': |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
97 |
return s.decode('utf-8', errors).encode(encoding, errors) |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
98 |
else: |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
99 |
return s |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
100 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
101 |
def iri_to_uri(iri): |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
102 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
103 |
Convert an Internationalized Resource Identifier (IRI) portion to a URI |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
104 |
portion that is suitable for inclusion in a URL. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
105 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
106 |
This is the algorithm from section 3.1 of RFC 3987. However, since we are |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
107 |
assuming input is either UTF-8 or unicode already, we can simplify things a |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
108 |
little from the full method. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
109 |
|
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
110 |
Returns an ASCII string containing the encoded result. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
111 |
""" |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
112 |
# The list of safe characters here is constructed from the printable ASCII |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
113 |
# characters that are not explicitly excluded by the list at the end of |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
114 |
# section 3.1 of RFC 3987. |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
115 |
if iri is None: |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
116 |
return iri |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
117 |
return urllib.quote(smart_str(iri), safe='/#%[]=:;$&()+,!?*') |
03e267d67478
Major reorganization of the soc svn repo, to merge into a single App Engine
Todd Larsen <tlarsen@google.com>
parents:
diff
changeset
|
118 |