[utils] Revise `isinstance()` tests (especially for str/unicode/bytes) to complete Linter fix

This commit is contained in:
dirkf 2023-07-30 21:47:48 +01:00
parent 7d965e6b65
commit 2d2a4bc832
2 changed files with 73 additions and 82 deletions

View File

@ -36,7 +36,7 @@ try:
)
except NameError:
compat_str, compat_basestring, compat_chr = (
str, str, chr
str, (str, bytes), chr
)
# casefold

View File

@ -1826,11 +1826,11 @@ def write_json_file(obj, fn):
if sys.version_info < (3, 0) and sys.platform != 'win32':
encoding = get_filesystem_encoding()
# os.path.basename returns a bytes object, but NamedTemporaryFile
# will fail if the filename contains non ascii characters unless we
# will fail if the filename contains non-ascii characters unless we
# use a unicode object
path_basename = lambda f: os.path.basename(fn).decode(encoding)
path_basename = lambda f: os.path.basename(f).decode(encoding)
# the same for os.path.dirname
path_dirname = lambda f: os.path.dirname(fn).decode(encoding)
path_dirname = lambda f: os.path.dirname(f).decode(encoding)
else:
path_basename = os.path.basename
path_dirname = os.path.dirname
@ -1894,10 +1894,10 @@ else:
return f
return None
# On python2.6 the xml.etree.ElementTree.Element methods don't support
# the namespace parameter
def xpath_with_ns(path, ns_map):
components = [c.split(':') for c in path.split('/')]
replaced = []
@ -1914,7 +1914,7 @@ def xpath_element(node, xpath, name=None, fatal=False, default=NO_DEFAULT):
def _find_xpath(xpath):
return node.find(compat_xpath(xpath))
if isinstance(xpath, (str, compat_str)):
if isinstance(xpath, compat_basestring):
n = _find_xpath(xpath)
else:
for xp in xpath:
@ -2262,39 +2262,32 @@ def get_subprocess_encoding():
return encoding
def encodeFilename(s, for_subprocess=False):
"""
@param s The name of the file
"""
# Jython assumes filenames are Unicode strings though reported as Python 2.x compatible
if sys.version_info < (3, 0) and not sys.platform.startswith('java'):
assert type(s) == compat_str
def encodeFilename(s, for_subprocess=False):
"""
@param s The name of the file
"""
# Pass '' directly to use Unicode APIs on Windows 2000 and up
# (Detecting Windows NT 4 is tricky because 'major >= 4' would
# match Windows 9x series as well. Besides, NT 4 is obsolete.)
if (not for_subprocess
and sys.platform == 'win32'
and sys.getwindowsversion()[0] >= 5
and isinstance(s, compat_str)):
return s
return _encode_compat_str(s, get_subprocess_encoding(), 'ignore')
def decodeFilename(b, for_subprocess=False):
return _decode_compat_str(b, get_subprocess_encoding(), 'ignore')
else:
# Python 3 has a Unicode API
if sys.version_info >= (3, 0):
return s
# Pass '' directly to use Unicode APIs on Windows 2000 and up
# (Detecting Windows NT 4 is tricky because 'major >= 4' would
# match Windows 9x series as well. Besides, NT 4 is obsolete.)
if not for_subprocess and sys.platform == 'win32' and sys.getwindowsversion()[0] >= 5:
return s
# Jython assumes filenames are Unicode strings though reported as Python 2.x compatible
if sys.platform.startswith('java'):
return s
return s.encode(get_subprocess_encoding(), 'ignore')
def decodeFilename(b, for_subprocess=False):
if sys.version_info >= (3, 0):
return b
if not isinstance(b, bytes):
return b
return b.decode(get_subprocess_encoding(), 'ignore')
encodeFilename = decodeFilename = lambda *s, **k: s[0]
def encodeArgument(s):
@ -2313,11 +2306,7 @@ def decodeArgument(b):
def decodeOption(optval):
if optval is None:
return optval
if isinstance(optval, bytes):
optval = optval.decode(preferredencoding())
assert isinstance(optval, compat_str)
return optval
return _decode_compat_str(optval)
def formatSeconds(secs):
@ -2363,7 +2352,7 @@ def make_HTTPS_handler(params, **kwargs):
if sys.version_info < (3, 2):
return YoutubeDLHTTPSHandler(params, **kwargs)
else: # Python < 3.4
else: # Python3 < 3.4
context = ssl.SSLContext(ssl.PROTOCOL_TLSv1)
context.verify_mode = (ssl.CERT_NONE
if opts_no_check_certificate
@ -2818,8 +2807,7 @@ class YoutubeDLHandler(compat_urllib_request.HTTPHandler):
location_escaped = escape_url(location_fixed)
if location != location_escaped:
del resp.headers['Location']
# if sys.version_info < (3, 0):
if not isinstance(location_escaped, str):
if not isinstance(location_escaped, str): # Py 2 case
location_escaped = location_escaped.encode('utf-8')
resp.headers['Location'] = location_escaped
return resp
@ -3086,8 +3074,7 @@ class YoutubeDLRedirectHandler(compat_urllib_request.HTTPRedirectHandler):
# On python 2 urlh.geturl() may sometimes return redirect URL
# as a byte string instead of unicode. This workaround forces
# it to return unicode.
if sys.version_info[0] < 3:
newurl = compat_str(newurl)
newurl = _decode_compat_str(newurl)
# Be conciliant with URIs containing a space. This is mainly
# redundant with the more complete encoding done in http_error_302(),
@ -3333,11 +3320,7 @@ class DateRange(object):
def platform_name():
""" Returns the platform name as a compat_str """
res = platform.platform()
if isinstance(res, bytes):
res = res.decode(preferredencoding())
assert isinstance(res, compat_str)
return res
return _decode_compat_str(res)
def _windows_write_string(s, out):
@ -3567,9 +3550,8 @@ def shell_quote(args):
quoted_args = []
encoding = get_filesystem_encoding()
for a in args:
if isinstance(a, bytes):
# We may get a filename encoded with 'encodeFilename'
a = a.decode(encoding)
# We may get a filename encoded with 'encodeFilename'
a = _decode_compat_str(a, encoding)
quoted_args.append(compat_shlex_quote(a))
return ' '.join(quoted_args)
@ -3733,8 +3715,9 @@ def parse_resolution(s):
def parse_bitrate(s):
if not isinstance(s, compat_str):
return
s = txt_or_none(s)
if not s:
return None
mobj = re.search(r'\b(\d+)\s*kbps', s)
if mobj:
return int(mobj.group(1))
@ -3822,18 +3805,17 @@ def base_url(url):
def urljoin(base, path):
if isinstance(path, bytes):
path = path.decode('utf-8')
if not isinstance(path, compat_str) or not path:
path = _decode_compat_str(path, encoding='utf-8', or_none=True)
if not path:
return None
if re.match(r'^(?:[a-zA-Z][a-zA-Z0-9+-.]*:)?//', path):
return path
if isinstance(base, bytes):
base = base.decode('utf-8')
if not isinstance(base, compat_str) or not re.match(
r'^(?:https?:)?//', base):
base = _decode_compat_str(base, encoding='utf-8', or_none=True)
if not base:
return None
return compat_urllib_parse.urljoin(base, path)
return (
re.match(r'^(?:https?:)?//', base)
and compat_urllib_parse.urljoin(base, path))
class HEADRequest(compat_urllib_request.Request):
@ -3998,8 +3980,7 @@ def get_exe_version(exe, args=['--version'],
stdout=subprocess.PIPE, stderr=subprocess.STDOUT))
except OSError:
return False
if isinstance(out, bytes): # Python 2.x
out = out.decode('ascii', 'ignore')
out = _decode_compat_str(out, 'ascii', 'ignore')
return detect_exe_version(out, version_re, unrecognized)
@ -4218,8 +4199,8 @@ def lowercase_escape(s):
def escape_rfc3986(s):
"""Escape non-ASCII characters as suggested by RFC 3986"""
if sys.version_info < (3, 0) and isinstance(s, compat_str):
s = s.encode('utf-8')
if sys.version_info < (3, 0):
s = _encode_compat_str(s, 'utf-8')
# ensure unicode: after quoting, it can always be converted
return compat_str(compat_urllib_parse.quote(s, b"%/;:@&=+$,!~*'()?#[]"))
@ -4242,8 +4223,7 @@ def parse_qs(url, **kwargs):
def read_batch_urls(batch_fd):
def fixup(url):
if not isinstance(url, compat_str):
url = url.decode('utf-8', 'replace')
url = _decode_compat_str(url, 'utf-8', 'replace')
BOM_UTF8 = '\xef\xbb\xbf'
if url.startswith(BOM_UTF8):
url = url[len(BOM_UTF8):]
@ -4305,10 +4285,8 @@ def _multipart_encode_impl(data, boundary):
out = b''
for k, v in data.items():
out += b'--' + boundary.encode('ascii') + b'\r\n'
if isinstance(k, compat_str):
k = k.encode('utf-8')
if isinstance(v, compat_str):
v = v.encode('utf-8')
k = _encode_compat_str(k, 'utf-8')
v = _encode_compat_str(v, 'utf-8')
# RFC 2047 requires non-ASCII field names to be encoded, while RFC 7578
# suggests sending UTF-8 directly. Firefox sends UTF-8, too
content = b'Content-Disposition: form-data; name="' + k + b'"\r\n\r\n' + v + b'\r\n'
@ -4435,8 +4413,26 @@ def merge_dicts(*dicts, **kwargs):
return merged
def encode_compat_str(string, encoding=preferredencoding(), errors='strict'):
return string if isinstance(string, compat_str) else compat_str(string, encoding, errors)
# very poor choice of name, as if Python string encodings weren't confusing enough
def encode_compat_str(s, encoding=preferredencoding(), errors='strict'):
assert isinstance(s, compat_basestring)
return s if isinstance(s, compat_str) else compat_str(s, encoding, errors)
# what it could have been
def _decode_compat_str(s, encoding=preferredencoding(), errors='strict', or_none=False):
if not or_none:
assert isinstance(s, compat_basestring)
return (
s if isinstance(s, compat_str)
else compat_str(s, encoding, errors) if isinstance(s, compat_basestring)
else None)
# the real encode_compat_str, but only for internal use
def _encode_compat_str(s, encoding=preferredencoding(), errors='strict'):
assert isinstance(s, compat_basestring)
return s.encode(encoding, errors) if isinstance(s, compat_str) else s
US_RATINGS = {
@ -4639,12 +4635,7 @@ def args_to_str(args):
def error_to_compat_str(err):
err_str = str(err)
# On python 2 error byte string must be decoded with proper
# encoding rather than ascii
if sys.version_info[0] < 3:
err_str = err_str.decode(preferredencoding())
return err_str
return _decode_compat_str(str(err))
def mimetype2ext(mt):