Skip to content

Commit 2ab8716

Browse files
coolreader18coolreader18
authored
Merge pull request #5623 from coolreader18/refactor-codecs
Refactor codecs
2 parents 549cce2 + e3d96aa commit 2ab8716

15 files changed

+1445
-666
lines changed

‎Cargo.lock

Lines changed: 1 addition & 0 deletions
Some generated files are not rendered by default. Learn more about customizing how changed files appear on GitHub.

‎Lib/_pycodecs.py

Lines changed: 7 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1086,11 +1086,13 @@ def charmapencode_output(c, mapping):
10861086
rep = mapping[c]
10871087
if isinstance(rep, int) or isinstance(rep, int):
10881088
if rep < 256:
1089-
return rep
1089+
return [rep]
10901090
else:
10911091
raise TypeError("character mapping must be in range(256)")
10921092
elif isinstance(rep, str):
1093-
return ord(rep)
1093+
return [ord(rep)]
1094+
elif isinstance(rep, bytes):
1095+
return rep
10941096
elif rep == None:
10951097
raise KeyError("character maps to <undefined>")
10961098
else:
@@ -1113,12 +1115,13 @@ def PyUnicode_EncodeCharmap(p, size, mapping='latin-1', errors='strict'):
11131115
#/* try to encode it */
11141116
try:
11151117
x = charmapencode_output(ord(p[inpos]), mapping)
1116-
res += [x]
1118+
res += x
11171119
except KeyError:
11181120
x = unicode_call_errorhandler(errors, "charmap",
11191121
"character maps to <undefined>", p, inpos, inpos+1, False)
11201122
try:
1121-
res += [charmapencode_output(ord(y), mapping) for y in x[0]]
1123+
for y in x[0]:
1124+
res += charmapencode_output(ord(y), mapping)
11221125
except KeyError:
11231126
raise UnicodeEncodeError("charmap", p, inpos, inpos+1,
11241127
"character maps to <undefined>")

‎Lib/test/test_charmapcodec.py

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -33,8 +33,6 @@ def test_constructorx(self):
3333
self.assertEqual(str(b'dxf', codecname), 'dabcf')
3434
self.assertEqual(str(b'dxfx', codecname), 'dabcfabc')
3535

36-
# TODO: RUSTPYTHON
37-
@unittest.expectedFailure
3836
def test_encodex(self):
3937
self.assertEqual('abc'.encode(codecname), b'abc')
4038
self.assertEqual('xdef'.encode(codecname), b'abcdef')

‎Lib/test/test_codeccallbacks.py

Lines changed: 0 additions & 14 deletions
Original file line numberDiff line numberDiff line change
@@ -203,8 +203,6 @@ def relaxedutf8(exc):
203203
self.assertRaises(UnicodeDecodeError, sin.decode,
204204
"utf-8", "test.relaxedutf8")
205205

206-
# TODO: RUSTPYTHON
207-
@unittest.expectedFailure
208206
def test_charmapencode(self):
209207
# For charmap encodings the replacement string will be
210208
# mapped through the encoding again. This means, that
@@ -329,8 +327,6 @@ def check_exceptionobjectargs(self, exctype, args, msg):
329327
exc = exctype(*args)
330328
self.assertEqual(str(exc), msg)
331329

332-
# TODO: RUSTPYTHON
333-
@unittest.expectedFailure
334330
def test_unicodeencodeerror(self):
335331
self.check_exceptionobjectargs(
336332
UnicodeEncodeError,
@@ -363,8 +359,6 @@ def test_unicodeencodeerror(self):
363359
"'ascii' codec can't encode character '\\U00010000' in position 0: ouch"
364360
)
365361

366-
# TODO: RUSTPYTHON
367-
@unittest.expectedFailure
368362
def test_unicodedecodeerror(self):
369363
self.check_exceptionobjectargs(
370364
UnicodeDecodeError,
@@ -377,8 +371,6 @@ def test_unicodedecodeerror(self):
377371
"'ascii' codec can't decode bytes in position 1-2: ouch"
378372
)
379373

380-
# TODO: RUSTPYTHON
381-
@unittest.expectedFailure
382374
def test_unicodetranslateerror(self):
383375
self.check_exceptionobjectargs(
384376
UnicodeTranslateError,
@@ -467,8 +459,6 @@ def test_badandgoodignoreexceptions(self):
467459
("", 2)
468460
)
469461

470-
# TODO: RUSTPYTHON
471-
@unittest.expectedFailure
472462
def test_badandgoodreplaceexceptions(self):
473463
# "replace" complains about a non-exception passed in
474464
self.assertRaises(
@@ -509,8 +499,6 @@ def test_badandgoodreplaceexceptions(self):
509499
("\ufffd", 2)
510500
)
511501

512-
# TODO: RUSTPYTHON
513-
@unittest.expectedFailure
514502
def test_badandgoodxmlcharrefreplaceexceptions(self):
515503
# "xmlcharrefreplace" complains about a non-exception passed in
516504
self.assertRaises(
@@ -1017,8 +1005,6 @@ def __getitem__(self, key):
10171005
self.assertRaises(ValueError, codecs.charmap_decode, b"\xff", "strict", D())
10181006
self.assertRaises(TypeError, codecs.charmap_decode, b"\xff", "strict", {0xff: sys.maxunicode+1})
10191007

1020-
# TODO: RUSTPYTHON
1021-
@unittest.expectedFailure
10221008
def test_encodehelper(self):
10231009
# enhance coverage of:
10241010
# Objects/unicodeobject.c::unicode_encode_call_errorhandler()

‎Lib/test/test_logging.py

Lines changed: 0 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -5525,8 +5525,6 @@ def test_encoding_errors_default(self):
55255525
self.assertEqual(data, r'\U0001f602: \u2603\ufe0f: The \xd8resund '
55265526
r'Bridge joins Copenhagen to Malm\xf6')
55275527

5528-
# TODO: RustPython
5529-
@unittest.expectedFailure
55305528
def test_encoding_errors_none(self):
55315529
# Specifying None should behave as 'strict'
55325530
try:

‎common/Cargo.toml

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -29,6 +29,7 @@ num-traits = { workspace = true }
2929
once_cell = { workspace = true }
3030
parking_lot = { workspace = true, optional = true }
3131
rand = { workspace = true }
32+
unicode_names2 = { workspace = true }
3233

3334
lock_api = "0.4"
3435
radium = "0.7"

0 commit comments

Comments
 (0)