60
70
return split(url, exclude_trailing_slash=exclude_trailing_slash)[0]
73
# Private copies of quote and unquote, copied from Python's
74
# urllib module because urllib unconditionally imports socket, which imports
77
always_safe = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
78
'abcdefghijklmnopqrstuvwxyz'
81
for i, c in zip(range(256), ''.join(map(chr, range(256)))):
82
_safe_map[c] = c if (i < 128 and c in always_safe) else '%{0:02X}'.format(i)
86
def quote(s, safe='/'):
87
"""quote('abc def') -> 'abc%20def'
89
Each part of a URL, e.g. the path info, the query, etc., has a
90
different set of reserved characters that must be quoted.
92
RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists
93
the following reserved characters.
95
reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
98
Each of these characters is reserved in some component of a URL,
99
but not necessarily in all of them.
101
By default, the quote function is intended for quoting the path
102
section of a URL. Thus, it will not encode '/'. This character
103
is reserved, but in typical usage the quote function is being
104
called on a path where the existing slash characters are used as
110
raise TypeError('None object cannot be quoted')
112
cachekey = (safe, always_safe)
114
(quoter, safe) = _safe_quoters[cachekey]
116
safe_map = _safe_map.copy()
117
safe_map.update([(c, c) for c in safe])
118
quoter = safe_map.__getitem__
119
safe = always_safe + safe
120
_safe_quoters[cachekey] = (quoter, safe)
121
if not s.rstrip(safe):
123
return ''.join(map(quoter, s))
126
unquote = urlparse.unquote
63
129
def escape(relpath):
64
130
"""Escape relpath to be a valid url."""
65
if isinstance(relpath, unicode):
131
if not isinstance(relpath, str):
66
132
relpath = relpath.encode('utf-8')
67
# After quoting and encoding, the path should be perfectly
68
# safe as a plain ASCII string, str() just enforces this
69
return str(urllib.quote(relpath, safe='/~'))
133
return quote(relpath, safe='/~')
72
136
def file_relpath(base, path):
121
match = _url_scheme_re.match(base)
124
scheme = match.group('scheme')
125
path = match.group('path').split('/')
126
if path[-1:] == ['']:
127
# Strip off a trailing slash
128
# This helps both when we are at the root, and when
129
# 'base' has an extra slash at the end
132
path = base.split('/')
134
if scheme is not None and len(path) >= 1:
136
# the path should be represented as an abs path.
137
# we know this must be absolute because of the presence of a URL scheme.
139
path = [''] + path[1:]
141
# create an empty host, but dont alter the path - this might be a
142
# relative url fragment.
190
scheme_end, path_start = _find_scheme_and_separator(base)
191
if scheme_end is None and path_start is None:
193
elif path_start is None:
194
path_start = len(base)
195
path = base[path_start:]
147
match = _url_scheme_re.match(arg)
150
scheme = match.group('scheme')
151
# this skips .. normalisation, making http://host/../../..
153
path = match.group('path').split('/')
154
# set the host and path according to new absolute URL, discarding
155
# any previous values.
156
# XXX: duplicates mess from earlier in this function. This URL
157
# manipulation code needs some cleaning up.
158
if scheme is not None and len(path) >= 1:
161
# url scheme implies absolute path.
164
# no url scheme we take the path as is.
197
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
198
if arg_scheme_end is None and arg_path_start is None:
200
elif arg_path_start is None:
201
arg_path_start = len(arg)
202
if arg_scheme_end is not None:
204
path = arg[arg_path_start:]
205
scheme_end = arg_scheme_end
206
path_start = arg_path_start
167
path = '/'.join(path)
168
208
path = joinpath(path, arg)
169
path = path.split('/')
170
if remove_root and path[0:1] == ['']:
173
# Remove the leading slash from the path, so long as it isn't also the
174
# trailing slash, which we want to keep if present.
175
if path and path[0] == '' and len(path) > 1:
180
return '/'.join(path)
181
return scheme + '://' + '/'.join(path)
209
return base[:path_start] + path
184
212
def joinpath(base, *args):
339
369
:param url: Either a hybrid URL or a local path
340
370
:return: A normalized URL which only includes 7-bit ASCII characters.
342
m = _url_scheme_re.match(url)
372
scheme_end, path_start = _find_scheme_and_separator(url)
373
if scheme_end is None:
344
374
return local_path_to_url(url)
345
scheme = m.group('scheme')
346
path = m.group('path')
347
if not isinstance(url, unicode):
375
prefix = url[:path_start]
376
path = url[path_start:]
377
if not isinstance(url, text_type):
349
379
if c not in _url_safe_characters:
350
380
raise errors.InvalidURL(url, 'URLs can only contain specific'
351
381
' safe characters (not %r)' % c)
352
382
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
353
return str(scheme + '://' + ''.join(path))
383
return str(prefix + ''.join(path))
355
385
# We have a unicode (hybrid) url
356
386
path_chars = list(path)
358
for i in xrange(len(path_chars)):
388
for i in range(len(path_chars)):
359
389
if path_chars[i] not in _url_safe_characters:
360
390
chars = path_chars[i].encode('utf-8')
361
391
path_chars[i] = ''.join(
362
392
['%%%02X' % ord(c) for c in path_chars[i].encode('utf-8')])
363
393
path = ''.join(path_chars)
364
394
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
365
return str(scheme + '://' + path)
395
return str(prefix + path)
368
398
def relative_url(base, other):
469
499
return url_base + head, tail
502
def split_segment_parameters_raw(url):
503
"""Split the subsegment of the last segment of a URL.
505
:param url: A relative or absolute URL
506
:return: (url, subsegments)
508
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
509
# operates on urls not url+segments, and Transport classes
510
# should not be blindly adding slashes in the first place.
511
lurl = strip_trailing_slash(url)
512
# Segments begin at first comma after last forward slash, if one exists
513
segment_start = lurl.find(",", lurl.rfind("/")+1)
514
if segment_start == -1:
516
return (lurl[:segment_start], lurl[segment_start+1:].split(","))
519
def split_segment_parameters(url):
520
"""Split the segment parameters of the last segment of a URL.
522
:param url: A relative or absolute URL
523
:return: (url, segment_parameters)
525
(base_url, subsegments) = split_segment_parameters_raw(url)
527
for subsegment in subsegments:
528
(key, value) = subsegment.split("=", 1)
529
parameters[key] = value
530
return (base_url, parameters)
533
def join_segment_parameters_raw(base, *subsegments):
534
"""Create a new URL by adding subsegments to an existing one.
536
This adds the specified subsegments to the last path in the specified
537
base URL. The subsegments should be bytestrings.
539
:note: You probably want to use join_segment_parameters instead.
543
for subsegment in subsegments:
544
if not isinstance(subsegment, str):
545
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
546
if "," in subsegment:
547
raise errors.InvalidURLJoin(", exists in subsegments",
549
return ",".join((base,) + subsegments)
552
def join_segment_parameters(url, parameters):
553
"""Create a new URL by adding segment parameters to an existing one.
555
The parameters of the last segment in the URL will be updated; if a
556
parameter with the same key already exists it will be overwritten.
558
:param url: A URL, as string
559
:param parameters: Dictionary of parameters, keys and values as bytestrings
561
(base, existing_parameters) = split_segment_parameters(url)
563
new_parameters.update(existing_parameters)
564
for key, value in parameters.items():
565
if not isinstance(key, str):
566
raise TypeError("parameter key %r is not a bytestring" % key)
567
if not isinstance(value, str):
568
raise TypeError("parameter value %r for %s is not a bytestring" %
571
raise errors.InvalidURLJoin("= exists in parameter key", url,
573
new_parameters[key] = value
574
return join_segment_parameters_raw(base,
575
*["%s=%s" % item for item in sorted(new_parameters.items())])
472
578
def _win32_strip_local_trailing_slash(url):
473
579
"""Strip slashes after the drive letter"""
474
580
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
524
630
This returns a Unicode path from a URL
526
632
# jam 20060427 URLs are supposed to be ASCII only strings
527
# If they are passed in as unicode, urllib.unquote
633
# If they are passed in as unicode, unquote
528
634
# will return a UNICODE string, which actually contains
529
635
# utf-8 bytes. So we have to ensure that they are
530
636
# plain ASCII strings, or the final .decode will
531
637
# try to encode the UNICODE => ASCII, and then decode
535
except UnicodeError, e:
536
raise errors.InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,))
538
unquoted = urllib.unquote(url)
639
if isinstance(url, text_type):
641
url = url.encode("ascii")
642
except UnicodeError as e:
643
raise errors.InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,))
645
unquoted = urlparse.unquote_to_bytes(url)
647
unquoted = unquote(url)
540
649
unicode_path = unquoted.decode('utf-8')
541
except UnicodeError, e:
650
except UnicodeError as e:
542
651
raise errors.InvalidURL(url, 'Unable to encode the URL as utf-8: %s' % (e,))
543
652
return unicode_path
693
802
return osutils.pathjoin(*segments)
808
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
811
self.quoted_host = quoted_host
812
self.host = unquote(self.quoted_host)
813
self.quoted_user = quoted_user
814
if self.quoted_user is not None:
815
self.user = unquote(self.quoted_user)
818
self.quoted_password = quoted_password
819
if self.quoted_password is not None:
820
self.password = unquote(self.quoted_password)
824
self.quoted_path = _url_hex_escapes_re.sub(_unescape_safe_chars, quoted_path)
825
self.path = unquote(self.quoted_path)
827
def __eq__(self, other):
828
return (isinstance(other, self.__class__) and
829
self.scheme == other.scheme and
830
self.host == other.host and
831
self.user == other.user and
832
self.password == other.password and
833
self.path == other.path)
836
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
837
self.__class__.__name__,
838
self.scheme, self.quoted_user, self.quoted_password,
839
self.quoted_host, self.port, self.quoted_path)
842
def from_string(cls, url):
843
"""Create a URL object from a string.
845
:param url: URL as bytestring
847
# GZ 2017-06-09: Actually validate ascii-ness
848
if not isinstance(url, str):
849
raise errors.InvalidURL('should be ascii:\n%r' % url)
850
(scheme, netloc, path, params,
851
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
852
user = password = host = port = None
854
user, host = netloc.rsplit('@', 1)
856
user, password = user.split(':', 1)
860
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
862
host, port = host.rsplit(':',1)
866
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
868
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
871
return cls(scheme, user, password, host, port, path)
874
netloc = self.quoted_host
876
netloc = "[%s]" % netloc
877
if self.quoted_user is not None:
878
# Note that we don't put the password back even if we
879
# have one so that it doesn't get accidentally
881
netloc = '%s@%s' % (self.quoted_user, netloc)
882
if self.port is not None:
883
netloc = '%s:%d' % (netloc, self.port)
884
return urlparse.urlunparse(
885
(self.scheme, netloc, self.quoted_path, None, None, None))
888
def _combine_paths(base_path, relpath):
889
"""Transform a Transport-relative path to a remote absolute path.
891
This does not handle substitution of ~ but does handle '..' and '.'
896
t._combine_paths('/home/sarah', 'project/foo')
897
=> '/home/sarah/project/foo'
898
t._combine_paths('/home/sarah', '../../etc')
900
t._combine_paths('/home/sarah', '/etc')
903
:param base_path: base path
904
:param relpath: relative url string for relative part of remote path.
905
:return: urlencoded string for final path.
907
if not isinstance(relpath, str):
908
raise errors.InvalidURL(relpath)
909
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
910
if relpath.startswith('/'):
913
base_parts = base_path.split('/')
914
if len(base_parts) > 0 and base_parts[-1] == '':
915
base_parts = base_parts[:-1]
916
for p in relpath.split('/'):
918
if len(base_parts) == 0:
919
# In most filesystems, a request for the parent
920
# of root, just returns root.
927
path = '/'.join(base_parts)
928
if not path.startswith('/'):
932
def clone(self, offset=None):
933
"""Return a new URL for a path relative to this URL.
935
:param offset: A relative path, already urlencoded
936
:return: `URL` instance
938
if offset is not None:
939
relative = unescape(offset).encode('utf-8')
940
path = self._combine_paths(self.path, relative)
941
path = quote(path, safe="/~")
943
path = self.quoted_path
944
return self.__class__(self.scheme, self.quoted_user,
945
self.quoted_password, self.quoted_host, self.port,
697
949
def parse_url(url):
698
950
"""Extract the server address, the credentials and the path from the url.
703
955
:param url: an quoted url
705
956
:return: (scheme, user, password, host, port, path) tuple, all fields
708
if isinstance(url, unicode):
709
raise errors.InvalidURL('should be ascii:\n%r' % url)
710
url = url.encode('utf-8')
711
(scheme, netloc, path, params,
712
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
713
user = password = host = port = None
715
user, host = netloc.rsplit('@', 1)
717
user, password = user.split(':', 1)
718
password = urllib.unquote(password)
719
user = urllib.unquote(user)
723
if ':' in host and not (host[0] == '[' and host[-1] == ']'): #there *is* port
724
host, port = host.rsplit(':',1)
728
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
730
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
733
host = urllib.unquote(host)
734
path = urllib.unquote(path)
736
return (scheme, user, password, host, port, path)
959
parsed_url = URL.from_string(url)
960
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
961
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
breezy/urlutils.py
