unicode: Fixed iri_to_uri() to be not quite as broken (I think it's spec

compliant now). Removed use of iri_to_uri() in the urlencode filter and made a
new iriencode filter, because IRI->URI conversionis are not a superset of URL
quoting converstions.


git-svn-id: http://code.djangoproject.com/svn/django/branches/unicode@5283 bcc190cf-cafb-0310-a4f2-bffc1f526a37

3 files changed, 13 insertions, 4 deletions

diff --git a/django/template/defaultfilters.py b/django/template/defaultfilters.py
index e85780ab79..6414724c64 100644
--- a/django/template/defaultfilters.py
+++ b/django/template/defaultfilters.py
@@ -3,7 +3,7 @@
 from django.template import resolve_variable, Library
 from django.conf import settings
 from django.utils.translation import ugettext, ungettext
-from django.utils.encoding import smart_unicode, smart_str
+from django.utils.encoding import smart_unicode, smart_str, iri_to_uri
 import re
 import random as random_module
 
@@ -84,6 +84,11 @@ def floatformat(text, arg=-1):
         formatstr = u'%%.%df' % abs(d)
         return formatstr % f
 
+def iriencode(value):
+    "Escapes an IRI value for use in a URL"
+    return smart_unicode(iri_to_uri(value))
+iriencode = stringfilter(iriencode)
+
 def linenumbers(value):
     "Displays text with line numbers"
     from django.utils.html import escape
@@ -170,7 +175,7 @@ upper = stringfilter(upper)
 def urlencode(value):
     "Escapes a value for use in a URL"
     import urllib
-    return smart_unicode(urllib.quote(smart_str(value)))
+    return smart_unicode(urllib.quote(value))
 urlencode = stringfilter(urlencode)
 
 def urlize(value):
@@ -566,6 +571,7 @@ register.filter(first)
 register.filter(fix_ampersands)
 register.filter(floatformat)
 register.filter(get_digit)
+register.filter(iriencode)
 register.filter(join)
 register.filter(length)
 register.filter(length_is)
diff --git a/django/utils/encoding.py b/django/utils/encoding.py
index 7466e02300..5bb23a28bf 100644
--- a/django/utils/encoding.py
+++ b/django/utils/encoding.py
@@ -66,5 +66,8 @@ def iri_to_uri(iri):
 
     Returns an ASCII string containing the encoded result.
     """
-    return urllib.quote(smart_str(iri), safe='/#%[]')
+    # The list of safe characters here is constructed from the printable ASCII
+    # characters that are not explicitly excluded by the list at the end of
+    # section 3.1 of RFC 3987.
+    return urllib.quote(smart_str(iri), safe='/#%[]=:;$&()+,!?')
 
diff --git a/tests/regressiontests/defaultfilters/tests.py b/tests/regressiontests/defaultfilters/tests.py
index 2e6c565eb7..fb35448edb 100644
--- a/tests/regressiontests/defaultfilters/tests.py
+++ b/tests/regressiontests/defaultfilters/tests.py
@@ -115,7 +115,7 @@ u'\xcb'
 u'jack%20%26%20jill'
 >>> urlencode(1)
 u'1'
->>> urlencode(u'S\xf8r-Tr\xf8ndelag')
+>>> iriencode(u'S\xf8r-Tr\xf8ndelag')
 u'S%C3%B8r-Tr%C3%B8ndelag'
 
 >>> urlizetrunc(u'http://short.com/', 20)