bpo-28080: Add support for the fallback encoding in ZIP files (GH-32007) · python/cpython@a25a985 (original) (raw)

`@@ -21,8 +21,10 @@

21

`from random import randint, random, randbytes

22

23

`from test.support import script_helper

24

from test.support import (findfile, requires_zlib, requires_bz2,

25

requires_lzma, captured_stdout, requires_subprocess)

24

from test.support import (

25

findfile, requires_zlib, requires_bz2, requires_lzma,

26

captured_stdout, captured_stderr, requires_subprocess

27

)

26

28

`from test.support.os_helper import (

27

29

`TESTFN, unlink, rmtree, temp_dir, temp_cwd, fd_count

28

30

`@@ -3210,5 +3212,139 @@ def test_inheritance(self, alpharep):

3210

3212

`assert isinstance(file, cls)

3211

3213

3212

3214

3215

class EncodedMetadataTests(unittest.TestCase):

3216

file_names = ['\u4e00', '\u4e8c', '\u4e09'] # Han 'one', 'two', 'three'

3217

file_content = [

3218

"This is pure ASCII.\n".encode('ascii'),

3219

This is modern Japanese. (UTF-8)

3220

"\u3053\u308c\u306f\u73fe\u4ee3\u7684\u65e5\u672c\u8a9e\u3067\u3059\u3002\n".encode('utf-8'),

3221

This is obsolete Japanese. (Shift JIS)

3222

"\u3053\u308c\u306f\u53e4\u3044\u65e5\u672c\u8a9e\u3067\u3059\u3002\n".encode('shift_jis'),

3223

]

3224

+

3225

def setUp(self):

3226

self.addCleanup(unlink, TESTFN)

3227

Create .zip of 3 members with Han names encoded in Shift JIS.

3228

Each name is 1 Han character encoding to 2 bytes in Shift JIS.

3229

The ASCII names are arbitrary as long as they are length 2 and

3230

not otherwise contained in the zip file.

3231

Data elements are encoded bytes (ascii, utf-8, shift_jis).

3232

placeholders = ["n1", "n2"] + self.file_names[2:]

3233

with zipfile.ZipFile(TESTFN, mode="w") as tf:

3234

for temp, content in zip(placeholders, self.file_content):

3235

tf.writestr(temp, content, zipfile.ZIP_STORED)

3236

Hack in the Shift JIS names with flag bit 11 (UTF-8) unset.

3237

with open(TESTFN, "rb") as tf:

3238

data = tf.read()

3239

for name, temp in zip(self.file_names, placeholders[:2]):

3240

data = data.replace(temp.encode('ascii'),

3241

name.encode('shift_jis'))

3242

with open(TESTFN, "wb") as tf:

3243

tf.write(data)

3244

+

3245

def _test_read(self, zipfp, expected_names, expected_content):

3246

Check the namelist

3247

names = zipfp.namelist()

3248

self.assertEqual(sorted(names), sorted(expected_names))

3249

+

3250

Check infolist

3251

infos = zipfp.infolist()

3252

names = [zi.filename for zi in infos]

3253

self.assertEqual(sorted(names), sorted(expected_names))

3254

+

3255

check getinfo

3256

for name, content in zip(expected_names, expected_content):

3257

info = zipfp.getinfo(name)

3258

self.assertEqual(info.filename, name)

3259

self.assertEqual(info.file_size, len(content))

3260

self.assertEqual(zipfp.read(name), content)

3261

+

3262

def test_read_with_metadata_encoding(self):

3263

Read the ZIP archive with correct metadata_encoding

3264

with zipfile.ZipFile(TESTFN, "r", metadata_encoding='shift_jis') as zipfp:

3265

self._test_read(zipfp, self.file_names, self.file_content)

3266

+

3267

def test_read_without_metadata_encoding(self):

3268

Read the ZIP archive without metadata_encoding

3269

expected_names = [name.encode('shift_jis').decode('cp437')

3270

for name in self.file_names[:2]] + self.file_names[2:]

3271

with zipfile.ZipFile(TESTFN, "r") as zipfp:

3272

self._test_read(zipfp, expected_names, self.file_content)

3273

+

3274

def test_read_with_incorrect_metadata_encoding(self):

3275

Read the ZIP archive with incorrect metadata_encoding

3276

expected_names = [name.encode('shift_jis').decode('koi8-u')

3277

for name in self.file_names[:2]] + self.file_names[2:]

3278

with zipfile.ZipFile(TESTFN, "r", metadata_encoding='koi8-u') as zipfp:

3279

self._test_read(zipfp, expected_names, self.file_content)

3280

+

3281

def test_read_with_unsuitable_metadata_encoding(self):

3282

Read the ZIP archive with metadata_encoding unsuitable for

3283

decoding metadata

3284

with self.assertRaises(UnicodeDecodeError):

3285

zipfile.ZipFile(TESTFN, "r", metadata_encoding='ascii')

3286

with self.assertRaises(UnicodeDecodeError):

3287

zipfile.ZipFile(TESTFN, "r", metadata_encoding='utf-8')

3288

+

3289

def test_read_after_append(self):

3290

newname = '\u56db' # Han 'four'

3291

expected_names = [name.encode('shift_jis').decode('cp437')

3292

for name in self.file_names[:2]] + self.file_names[2:]

3293

expected_names.append(newname)

3294

expected_content = (*self.file_content, b"newcontent")

3295

+

3296

with zipfile.ZipFile(TESTFN, "a") as zipfp:

3297

zipfp.writestr(newname, "newcontent")

3298

self.assertEqual(sorted(zipfp.namelist()), sorted(expected_names))

3299

+

3300

with zipfile.ZipFile(TESTFN, "r") as zipfp:

3301

self._test_read(zipfp, expected_names, expected_content)

3302

+

3303

with zipfile.ZipFile(TESTFN, "r", metadata_encoding='shift_jis') as zipfp:

3304

self.assertEqual(sorted(zipfp.namelist()), sorted(expected_names))

3305

for i, (name, content) in enumerate(zip(expected_names, expected_content)):

3306

info = zipfp.getinfo(name)

3307

self.assertEqual(info.filename, name)

3308

self.assertEqual(info.file_size, len(content))

3309

if i < 2:

3310

with self.assertRaises(zipfile.BadZipFile):

3311

zipfp.read(name)

3312

else:

3313

self.assertEqual(zipfp.read(name), content)

3314

+

3315

def test_write_with_metadata_encoding(self):

3316

ZF = zipfile.ZipFile

3317

for mode in ("w", "x", "a"):

3318

with self.assertRaisesRegex(ValueError,

3319

"^metadata_encoding is only"):

3320

ZF("nonesuch.zip", mode, metadata_encoding="shift_jis")

3321

+

3322

def test_cli_with_metadata_encoding(self):

3323

errmsg = "Non-conforming encodings not supported with -c."

3324

args = ["--metadata-encoding=shift_jis", "-c", "nonesuch", "nonesuch"]

3325

with captured_stdout() as stdout:

3326

with captured_stderr() as stderr:

3327

self.assertRaises(SystemExit, zipfile.main, args)

3328

self.assertEqual(stdout.getvalue(), "")

3329

self.assertIn(errmsg, stderr.getvalue())

3330

+

3331

with captured_stdout() as stdout:

3332

zipfile.main(["--metadata-encoding=shift_jis", "-t", TESTFN])

3333

listing = stdout.getvalue()

3334

+

3335

with captured_stdout() as stdout:

3336

zipfile.main(["--metadata-encoding=shift_jis", "-l", TESTFN])

3337

listing = stdout.getvalue()

3338

for name in self.file_names:

3339

self.assertIn(name, listing)

3340

+

3341

os.mkdir(TESTFN2)

3342

self.addCleanup(rmtree, TESTFN2)

3343

zipfile.main(["--metadata-encoding=shift_jis", "-e", TESTFN, TESTFN2])

3344

listing = os.listdir(TESTFN2)

3345

for name in self.file_names:

3346

self.assertIn(name, listing)

3347

+

3348

+

3213

3349

`if name == "main":

3214

3350

`unittest.main()