Mercurial > cpython-lru_cache-weakref
annotate Lib/test/test_normalization.py @ 58925:eecf1bef052d 2.6
Merged revisions 78207 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk
........
r78207 | ezio.melotti | 2010-02-17 01:26:09 +0200 (Wed, 17 Feb 2010) | 1 line
#7930: fix stripid
........
| author | Ezio Melotti <ezio.melotti@gmail.com> |
|---|---|
| date | Tue, 16 Feb 2010 23:29:44 +0000 |
| parents | 57542b502d32 |
| children | 52a861746129 |
| rev | line source |
|---|---|
|
57050
57542b502d32
Merged revisions 76033 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents:
49131
diff
changeset
|
1 from test.test_support import run_unittest, open_urlresource, TestSkipped |
|
41393
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
2 import unittest |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
3 |
|
26379
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
4 import sys |
|
26391
d207d1d78c5d
Fiddled things so that test_normalization is expected to be skipped if
Tim Peters <tim.peters@gmail.com>
parents:
26388
diff
changeset
|
5 import os |
|
49131
2ab85c392a8d
Read unidata_version from unicodedata module.
Martin v. Löwis <martin@v.loewis.de>
parents:
49129
diff
changeset
|
6 from unicodedata import normalize, unidata_version |
|
26391
d207d1d78c5d
Fiddled things so that test_normalization is expected to be skipped if
Tim Peters <tim.peters@gmail.com>
parents:
26388
diff
changeset
|
7 |
|
36929
fab44c5c8df4
Update test data to 4.1; disable PRI #29 for now.
Martin v. Löwis <martin@v.loewis.de>
parents:
35998
diff
changeset
|
8 TESTDATAFILE = "NormalizationTest" + os.extsep + "txt" |
|
49131
2ab85c392a8d
Read unidata_version from unicodedata module.
Martin v. Löwis <martin@v.loewis.de>
parents:
49129
diff
changeset
|
9 TESTDATAURL = "http://www.unicode.org/Public/" + unidata_version + "/ucd/" + TESTDATAFILE |
|
2ab85c392a8d
Read unidata_version from unicodedata module.
Martin v. Löwis <martin@v.loewis.de>
parents:
49129
diff
changeset
|
10 |
|
2ab85c392a8d
Read unidata_version from unicodedata module.
Martin v. Löwis <martin@v.loewis.de>
parents:
49129
diff
changeset
|
11 if os.path.exists(TESTDATAFILE): |
|
2ab85c392a8d
Read unidata_version from unicodedata module.
Martin v. Löwis <martin@v.loewis.de>
parents:
49129
diff
changeset
|
12 f = open(TESTDATAFILE) |
|
2ab85c392a8d
Read unidata_version from unicodedata module.
Martin v. Löwis <martin@v.loewis.de>
parents:
49129
diff
changeset
|
13 l = f.readline() |
|
2ab85c392a8d
Read unidata_version from unicodedata module.
Martin v. Löwis <martin@v.loewis.de>
parents:
49129
diff
changeset
|
14 f.close() |
|
2ab85c392a8d
Read unidata_version from unicodedata module.
Martin v. Löwis <martin@v.loewis.de>
parents:
49129
diff
changeset
|
15 if not unidata_version in l: |
|
2ab85c392a8d
Read unidata_version from unicodedata module.
Martin v. Löwis <martin@v.loewis.de>
parents:
49129
diff
changeset
|
16 os.unlink(TESTDATAFILE) |
|
26379
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
17 |
|
37188
760ffb0fba37
Exceptions should inherit from Exception now.
Neal Norwitz <nnorwitz@gmail.com>
parents:
36937
diff
changeset
|
18 class RangeError(Exception): |
|
26379
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
19 pass |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
20 |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
21 def NFC(str): |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
22 return normalize("NFC", str) |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
23 |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
24 def NFKC(str): |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
25 return normalize("NFKC", str) |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
26 |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
27 def NFD(str): |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
28 return normalize("NFD", str) |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
29 |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
30 def NFKD(str): |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
31 return normalize("NFKD", str) |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
32 |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
33 def unistr(data): |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
34 data = [int(x, 16) for x in data.split(" ")] |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
35 for x in data: |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
36 if x > sys.maxunicode: |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
37 raise RangeError |
|
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
38 return u"".join([unichr(x) for x in data]) |
|
41532
4b9fa8092e82
Whitespace normalization. Ugh, we really need to do this more often.
Neal Norwitz <nnorwitz@gmail.com>
parents:
41472
diff
changeset
|
39 |
|
41393
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
40 class NormalizationTest(unittest.TestCase): |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
41 def test_main(self): |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
42 part1_data = {} |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
43 for line in open_urlresource(TESTDATAURL): |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
44 if '#' in line: |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
45 line = line.split('#')[0] |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
46 line = line.strip() |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
47 if not line: |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
48 continue |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
49 if line.startswith("@Part"): |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
50 part = line.split()[0] |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
51 continue |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
52 if part == "@Part3": |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
53 # XXX we don't support PRI #29 yet, so skip these tests for now |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
54 continue |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
55 try: |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
56 c1,c2,c3,c4,c5 = [unistr(x) for x in line.split(';')[:-1]] |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
57 except RangeError: |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
58 # Skip unsupported characters; |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
59 # try atleast adding c1 if we are in part1 |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
60 if part == "@Part1": |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
61 try: |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
62 c1 = unistr(line.split(';')[0]) |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
63 except RangeError: |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
64 pass |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
65 else: |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
66 part1_data[c1] = 1 |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
67 continue |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
68 |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
69 # Perform tests |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
70 self.failUnless(c2 == NFC(c1) == NFC(c2) == NFC(c3), line) |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
71 self.failUnless(c4 == NFC(c4) == NFC(c5), line) |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
72 self.failUnless(c3 == NFD(c1) == NFD(c2) == NFD(c3), line) |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
73 self.failUnless(c5 == NFD(c4) == NFD(c5), line) |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
74 self.failUnless(c4 == NFKC(c1) == NFKC(c2) == \ |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
75 NFKC(c3) == NFKC(c4) == NFKC(c5), |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
76 line) |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
77 self.failUnless(c5 == NFKD(c1) == NFKD(c2) == \ |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
78 NFKD(c3) == NFKD(c4) == NFKD(c5), |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
79 line) |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
80 |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
81 # Record part 1 data |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
82 if part == "@Part1": |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
83 part1_data[c1] = 1 |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
84 |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
85 # Perform tests for all other data |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
86 for c in range(sys.maxunicode+1): |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
87 X = unichr(c) |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
88 if X in part1_data: |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
89 continue |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
90 self.failUnless(X == NFC(X) == NFD(X) == NFKC(X) == NFKD(X), c) |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
91 |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
92 def test_bug_834676(self): |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
93 # Check for bug 834676 |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
94 normalize('NFC', u'\ud55c\uae00') |
|
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
95 |
|
26379
1885b2d6393a
Patch #626485: Support Unicode normalization.
Martin v. Löwis <martin@v.loewis.de>
parents:
diff
changeset
|
96 |
|
26391
d207d1d78c5d
Fiddled things so that test_normalization is expected to be skipped if
Tim Peters <tim.peters@gmail.com>
parents:
26388
diff
changeset
|
97 def test_main(): |
|
41472
cc707f1ae0b6
Check the availability of the urlfetch resource earlier than before.
Collin Winter <collinw@gmail.com>
parents:
41393
diff
changeset
|
98 # Hit the exception early |
|
57050
57542b502d32
Merged revisions 76033 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents:
49131
diff
changeset
|
99 try: |
|
57542b502d32
Merged revisions 76033 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents:
49131
diff
changeset
|
100 open_urlresource(TESTDATAURL) |
|
57542b502d32
Merged revisions 76033 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents:
49131
diff
changeset
|
101 except IOError: |
|
57542b502d32
Merged revisions 76033 via svnmerge from
Antoine Pitrou <solipsis@pitrou.net>
parents:
49131
diff
changeset
|
102 raise TestSkipped("could not retrieve " + TESTDATAURL) |
|
41393
ea6fbf404c34
Convert test_normalization to use unittest.
Collin Winter <collinw@gmail.com>
parents:
37188
diff
changeset
|
103 run_unittest(NormalizationTest) |
|
30867
04ca4d1f751f
Overallocate target buffer for normalization more early. Fixes #834676.
Martin v. Löwis <martin@v.loewis.de>
parents:
30056
diff
changeset
|
104 |
|
26391
d207d1d78c5d
Fiddled things so that test_normalization is expected to be skipped if
Tim Peters <tim.peters@gmail.com>
parents:
26388
diff
changeset
|
105 if __name__ == "__main__": |
|
d207d1d78c5d
Fiddled things so that test_normalization is expected to be skipped if
Tim Peters <tim.peters@gmail.com>
parents:
26388
diff
changeset
|
106 test_main() |
