Merge pull request #22 from danabr/ldap_strings_are_utf8

Make Net::LDAP use UTF-8 strings
This commit is contained in:
Rory O'Connell 2011-09-24 11:33:10 -07:00
commit dac73f46d2
3 changed files with 31 additions and 2 deletions

View file

@ -295,6 +295,8 @@ class Net::BER::BerIdentifiedString < String
attr_accessor :ber_identifier attr_accessor :ber_identifier
def initialize args def initialize args
super args super args
# LDAP uses UTF-8 encoded strings
force_encoding('UTF-8') if respond_to?(:encoding)
end end
end end

View file

@ -12,9 +12,21 @@ module Net::BER::Extensions::String
# User code should call either #to_ber_application_string or # User code should call either #to_ber_application_string or
# #to_ber_contextspecific. # #to_ber_contextspecific.
def to_ber(code = 0x04) def to_ber(code = 0x04)
[code].pack('C') + length.to_ber_length_encoding + self raw_string = raw_utf8_encoded
[code].pack('C') + raw_string.length.to_ber_length_encoding + raw_string
end end
def raw_utf8_encoded
if self.respond_to?(:encode)
# Strings should be UTF-8 encoded according to LDAP.
# However, the BER code is not necessarily valid UTF-8
self.encode('UTF-8').force_encoding('ASCII-8BIT')
else
self
end
end
private :raw_utf8_encoded
## ##
# Creates an application-specific BER string encoded value with the # Creates an application-specific BER string encoded value with the
# provided syntax code value. # provided syntax code value.

View file

@ -75,6 +75,21 @@ describe "BER encoding of" do
end end
end end
end end
if "Ruby 1.9".respond_to?(:encoding)
context "strings" do
it "should properly encode UTF-8 strings" do
"\u00e5".force_encoding("UTF-8").to_ber.should ==
"\x04\x02\xC3\xA5"
end
it "should properly encode strings encodable as UTF-8" do
"teststring".encode("US-ASCII").to_ber.should == "\x04\nteststring"
end
it "should fail on strings that can not be converted to UTF-8" do
error = Encoding::UndefinedConversionError
lambda {"\x81".to_ber }.should raise_exception(error)
end
end
end
end end
describe "BER decoding of" do describe "BER decoding of" do