1
1
mirror of https://github.com/ytdl-org/youtube-dl synced 2024-11-23 15:56:49 +01:00

[utils] Process non-base 10 integers in js_to_json

This commit is contained in:
Sergey M․ 2016-05-14 20:39:58 +06:00
parent 640eea0a0c
commit 89ac4a19e6
No known key found for this signature in database
GPG Key ID: 2C393E0F18A9236D
2 changed files with 31 additions and 0 deletions

View File

@ -617,6 +617,15 @@ class TestUtil(unittest.TestCase):
json_code = js_to_json(inp) json_code = js_to_json(inp)
self.assertEqual(json.loads(json_code), json.loads(inp)) self.assertEqual(json.loads(json_code), json.loads(inp))
inp = '''{
0:{src:'skipped', type: 'application/dash+xml'},
1:{src:'skipped', type: 'application/vnd.apple.mpegURL'},
}'''
self.assertEqual(js_to_json(inp), '''{
"0":{"src":"skipped", "type": "application/dash+xml"},
"1":{"src":"skipped", "type": "application/vnd.apple.mpegURL"}
}''')
def test_js_to_json_edgecases(self): def test_js_to_json_edgecases(self):
on = js_to_json("{abc_def:'1\\'\\\\2\\\\\\'3\"4'}") on = js_to_json("{abc_def:'1\\'\\\\2\\\\\\'3\"4'}")
self.assertEqual(json.loads(on), {"abc_def": "1'\\2\\'3\"4"}) self.assertEqual(json.loads(on), {"abc_def": "1'\\2\\'3\"4"})
@ -652,6 +661,16 @@ class TestUtil(unittest.TestCase):
on = js_to_json("['a\\\nb']") on = js_to_json("['a\\\nb']")
self.assertEqual(json.loads(on), ['ab']) self.assertEqual(json.loads(on), ['ab'])
on = js_to_json('{0xff:0xff}')
self.assertEqual(json.loads(on), {'255': 255})
on = js_to_json('{077:077}')
self.assertEqual(json.loads(on), {'63': 63})
on = js_to_json('{42:42}')
self.assertEqual(json.loads(on), {'42': 42})
def test_extract_attributes(self): def test_extract_attributes(self):
self.assertEqual(extract_attributes('<e x="y">'), {'x': 'y'}) self.assertEqual(extract_attributes('<e x="y">'), {'x': 'y'})
self.assertEqual(extract_attributes("<e x='y'>"), {'x': 'y'}) self.assertEqual(extract_attributes("<e x='y'>"), {'x': 'y'})

View File

@ -1925,6 +1925,17 @@ def js_to_json(code):
'\\x': '\\u00', '\\x': '\\u00',
}.get(m.group(0), m.group(0)), v[1:-1]) }.get(m.group(0), m.group(0)), v[1:-1])
INTEGER_TABLE = (
(r'^(0[xX][0-9a-fA-F]+)', 16),
(r'^(0+[0-7]+)', 8),
)
for regex, base in INTEGER_TABLE:
im = re.match(regex, v)
if im:
i = int(im.group(1), base)
return '"%d":' % i if v.endswith(':') else '%d' % i
return '"%s"' % v return '"%s"' % v
return re.sub(r'''(?sx) return re.sub(r'''(?sx)
@ -1932,6 +1943,7 @@ def js_to_json(code):
'(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'| '(?:[^'\\]*(?:\\\\|\\['"nurtbfx/\n]))*[^'\\]*'|
/\*.*?\*/|,(?=\s*[\]}])| /\*.*?\*/|,(?=\s*[\]}])|
[a-zA-Z_][.a-zA-Z_0-9]*| [a-zA-Z_][.a-zA-Z_0-9]*|
(?:0[xX][0-9a-fA-F]+|0+[0-7]+)(?:\s*:)?|
[0-9]+(?=\s*:) [0-9]+(?=\s*:)
''', fix_kv, code) ''', fix_kv, code)