[patch] urlize breaks when string.letters is changed by the locale

[Andrew Stoneman <astoneman@…>]

utils.html.urlize() depends on string.letters being automatically convertible into unicode. If it is changed by the locale module, however, it sometimes contains non-ascii characters, which causes very mysterious unicode decode errors on sites that use the function. I've included a patch to have it use string.ascii_letters instead, which does not change.

[Andrew Stoneman <astoneman@…>]

patch to use ascii_letters instead of letters

[Andrew Stoneman <astoneman@…>]

A sample session to show the problem:

>>> import locale
>>> import string
>>> from django.utils.html import urlize
>>> string.letters
'abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ'
>>> locale.setlocale(locale.LC_ALL, 'de_DE')
'de_DE'
>>> string.letters
'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz\xaa\xb5\xba\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9
\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4
\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf8\xf9\xfa\xfb\xfc\xfd\xfe\xff'
>>> urlize('abc')
Traceback (most recent call last):
  File "<console>", line 1, in ?
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.4/lib/python2.4/site-packages/django/utils/functional.py",
line 129, in wrapper
    return func(*args, **kwargs)
  File "/opt/local/Library/Frameworks/Python.framework/Versions/2.4/lib/python2.4/site-packages/django/utils/html.py",
line 82, in urlize
    if middle.startswith('www.') or ('@' not in middle and not middle.startswith('http://') and \
UnicodeDecodeError: 'ascii' codec can't decode byte 0xaa in position 52: ordinal not in range(128)

[Simon G <dev@…>]

Andrew - this looks good, but can we get a regression test (like if your example above)?

[shaleh]

updated patch, adds unittests

[shaleh]

unittests added

[Malcolm Tredinnick]

Unfortunately, the test case isn't sufficiently portable (for example, on my Ubuntu laptop, it fails with an "invalid locale" error). Rather than worrying too much about lots of different installation situations, I'm just going to commit the core patch. It's correct and I can live without tests for this small change.

[Malcolm Tredinnick]

(In [6856]) Fixed #5657 -- Use string.ascii_letters instead of ascii.letters in the urlize
filter to ensure consistent (and correct) results no matter what the server's
locale setting might be. Thanks, Andrew Stoneman.