#22650: test suite: load Unicode test data files from www.pythontest.net

This commit is contained in:
Georg Brandl 2014-11-06 14:37:49 +01:00
parent a1137fba89
commit 6246f2a103
6 changed files with 15 additions and 23 deletions

View File

@ -10,19 +10,17 @@ import unittest
class TestGB2312Map(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'gb2312'
mapfileurl = 'http://people.freebsd.org/~perky/i18n/EUC-CN.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/EUC-CN.TXT'
class TestGBKMap(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'gbk'
mapfileurl = 'http://www.unicode.org/Public/MAPPINGS/VENDORS/' \
'MICSFT/WINDOWS/CP936.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/CP936.TXT'
class TestGB18030Map(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'gb18030'
mapfileurl = 'http://source.icu-project.org/repos/icu/data/' \
'trunk/charset/data/xml/gb-18030-2000.xml'
mapfileurl = 'http://www.pythontest.net/unicode/gb-18030-2000.xml'
def test_main():

View File

@ -10,7 +10,7 @@ import unittest
class TestBig5HKSCSMap(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'big5hkscs'
mapfileurl = 'http://people.freebsd.org/~perky/i18n/BIG5HKSCS-2004.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/BIG5HKSCS-2004.TXT'
def test_main():
test_support.run_unittest(__name__)

View File

@ -10,8 +10,7 @@ import unittest
class TestCP932Map(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'cp932'
mapfileurl = 'http://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/' \
'WINDOWS/CP932.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/CP932.TXT'
supmaps = [
('\x80', u'\u0080'),
('\xa0', u'\uf8f0'),
@ -27,15 +26,14 @@ class TestEUCJPCOMPATMap(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'euc_jp'
mapfilename = 'EUC-JP.TXT'
mapfileurl = 'http://people.freebsd.org/~perky/i18n/EUC-JP.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/EUC-JP.TXT'
class TestSJISCOMPATMap(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'shift_jis'
mapfilename = 'SHIFTJIS.TXT'
mapfileurl = 'http://www.unicode.org/Public/MAPPINGS/OBSOLETE' \
'/EASTASIA/JIS/SHIFTJIS.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/SHIFTJIS.TXT'
pass_enctest = [
('\x81_', u'\\'),
]
@ -49,14 +47,14 @@ class TestEUCJISX0213Map(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'euc_jisx0213'
mapfilename = 'EUC-JISX0213.TXT'
mapfileurl = 'http://people.freebsd.org/~perky/i18n/EUC-JISX0213.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/EUC-JISX0213.TXT'
class TestSJISX0213Map(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'shift_jisx0213'
mapfilename = 'SHIFT_JISX0213.TXT'
mapfileurl = 'http://people.freebsd.org/~perky/i18n/SHIFT_JISX0213.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/SHIFT_JISX0213.TXT'
def test_main():

View File

@ -10,14 +10,13 @@ import unittest
class TestCP949Map(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'cp949'
mapfileurl = 'http://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT' \
'/WINDOWS/CP949.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/CP949.TXT'
class TestEUCKRMap(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'euc_kr'
mapfileurl = 'http://people.freebsd.org/~perky/i18n/EUC-KR.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/EUC-KR.TXT'
# A4D4 HANGUL FILLER indicates the begin of 8-bytes make-up sequence.
pass_enctest = [('\xa4\xd4', u'\u3164')]
@ -27,8 +26,7 @@ class TestEUCKRMap(test_multibytecodec_support.TestBase_Mapping,
class TestJOHABMap(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'johab'
mapfileurl = 'http://www.unicode.org/Public/MAPPINGS/OBSOLETE/EASTASIA/' \
'KSC/JOHAB.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/JOHAB.TXT'
# KS X 1001 standard assigned 0x5c as WON SIGN.
# but, in early 90s that is the only era used johab widely,
# the most softwares implements it as REVERSE SOLIDUS.

View File

@ -10,14 +10,12 @@ import unittest
class TestBIG5Map(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'big5'
mapfileurl = 'http://www.unicode.org/Public/MAPPINGS/OBSOLETE/' \
'EASTASIA/OTHER/BIG5.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/BIG5.TXT'
class TestCP950Map(test_multibytecodec_support.TestBase_Mapping,
unittest.TestCase):
encoding = 'cp950'
mapfileurl = 'http://www.unicode.org/Public/MAPPINGS/VENDORS/MICSFT/' \
'WINDOWS/CP950.TXT'
mapfileurl = 'http://www.pythontest.net/unicode/CP950.TXT'
pass_enctest = [
('\xa2\xcc', u'\u5341'),
('\xa2\xce', u'\u5345'),

View File

@ -7,7 +7,7 @@ import os
from unicodedata import normalize, unidata_version
TESTDATAFILE = "NormalizationTest.txt"
TESTDATAURL = "http://www.unicode.org/Public/" + unidata_version + "/ucd/" + TESTDATAFILE
TESTDATAURL = "http://www.pythontest.net/unicode/" + unidata_version + "/" + TESTDATAFILE
def check_version(testfile):
hdr = testfile.readline()