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.
126
scheme_end, path_start = _find_scheme_and_separator(base)
127
if scheme_end is None and path_start is None:
129
elif path_start is None:
130
path_start = len(base)
131
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.
133
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
134
if arg_scheme_end is None and arg_path_start is None:
136
elif arg_path_start is None:
137
arg_path_start = len(arg)
138
if arg_scheme_end is not None:
140
path = arg[arg_path_start:]
141
scheme_end = arg_scheme_end
142
path_start = arg_path_start
167
path = '/'.join(path)
168
144
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)
145
return base[:path_start] + path
184
148
def joinpath(base, *args):
339
304
:param url: Either a hybrid URL or a local path
340
305
:return: A normalized URL which only includes 7-bit ASCII characters.
342
m = _url_scheme_re.match(url)
307
scheme_end, path_start = _find_scheme_and_separator(url)
308
if scheme_end is None:
344
309
return local_path_to_url(url)
345
scheme = m.group('scheme')
346
path = m.group('path')
310
prefix = url[:path_start]
311
path = url[path_start:]
347
312
if not isinstance(url, unicode):
349
314
if c not in _url_safe_characters:
350
315
raise errors.InvalidURL(url, 'URLs can only contain specific'
351
316
' safe characters (not %r)' % c)
352
317
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
353
return str(scheme + '://' + ''.join(path))
318
return str(prefix + ''.join(path))
355
320
# We have a unicode (hybrid) url
356
321
path_chars = list(path)
469
434
return url_base + head, tail
437
def split_segment_parameters_raw(url):
438
"""Split the subsegment of the last segment of a URL.
440
:param url: A relative or absolute URL
441
:return: (url, subsegments)
443
(parent_url, child_dir) = split(url)
444
subsegments = child_dir.split(",")
445
if len(subsegments) == 1:
447
return (join(parent_url, subsegments[0]), subsegments[1:])
450
def split_segment_parameters(url):
451
"""Split the segment parameters of the last segment of a URL.
453
:param url: A relative or absolute URL
454
:return: (url, segment_parameters)
456
(base_url, subsegments) = split_segment_parameters_raw(url)
458
for subsegment in subsegments:
459
(key, value) = subsegment.split("=", 1)
460
parameters[key] = value
461
return (base_url, parameters)
464
def join_segment_parameters_raw(base, *subsegments):
465
"""Create a new URL by adding subsegments to an existing one.
467
This adds the specified subsegments to the last path in the specified
468
base URL. The subsegments should be bytestrings.
470
:note: You probably want to use join_segment_parameters instead.
474
for subsegment in subsegments:
475
if type(subsegment) is not str:
476
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
477
if "," in subsegment:
478
raise errors.InvalidURLJoin(", exists in subsegments",
480
return ",".join((base,) + subsegments)
483
def join_segment_parameters(url, parameters):
484
"""Create a new URL by adding segment parameters to an existing one.
486
The parameters of the last segment in the URL will be updated; if a
487
parameter with the same key already exists it will be overwritten.
489
:param url: A URL, as string
490
:param parameters: Dictionary of parameters, keys and values as bytestrings
492
(base, existing_parameters) = split_segment_parameters(url)
494
new_parameters.update(existing_parameters)
495
for key, value in parameters.iteritems():
496
if type(key) is not str:
497
raise TypeError("parameter key %r is not a bytestring" % key)
498
if type(value) is not str:
499
raise TypeError("parameter value %r for %s is not a bytestring" %
502
raise errors.InvalidURLJoin("= exists in parameter key", url,
504
new_parameters[key] = value
505
return join_segment_parameters_raw(base,
506
*["%s=%s" % item for item in sorted(new_parameters.items())])
472
509
def _win32_strip_local_trailing_slash(url):
473
510
"""Strip slashes after the drive letter"""
474
511
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
693
730
return osutils.pathjoin(*segments)
736
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
739
self.quoted_host = quoted_host
740
self.host = urllib.unquote(self.quoted_host)
741
self.quoted_user = quoted_user
742
if self.quoted_user is not None:
743
self.user = urllib.unquote(self.quoted_user)
746
self.quoted_password = quoted_password
747
if self.quoted_password is not None:
748
self.password = urllib.unquote(self.quoted_password)
752
self.quoted_path = quoted_path
753
self.path = urllib.unquote(self.quoted_path)
755
def __eq__(self, other):
756
return (isinstance(other, self.__class__) and
757
self.scheme == other.scheme and
758
self.host == other.host and
759
self.user == other.user and
760
self.password == other.password and
761
self.path == other.path)
764
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
765
self.__class__.__name__,
766
self.scheme, self.quoted_user, self.quoted_password,
767
self.quoted_host, self.port, self.quoted_path)
770
def from_string(cls, url):
771
"""Create a URL object from a string.
773
:param url: URL as bytestring
775
if isinstance(url, unicode):
776
raise errors.InvalidURL('should be ascii:\n%r' % url)
777
url = url.encode('utf-8')
778
(scheme, netloc, path, params,
779
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
780
user = password = host = port = None
782
user, host = netloc.rsplit('@', 1)
784
user, password = user.split(':', 1)
788
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
790
host, port = host.rsplit(':',1)
794
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
796
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
799
return cls(scheme, user, password, host, port, path)
802
netloc = self.quoted_host
804
netloc = "[%s]" % netloc
805
if self.quoted_user is not None:
806
# Note that we don't put the password back even if we
807
# have one so that it doesn't get accidentally
809
netloc = '%s@%s' % (self.quoted_user, netloc)
810
if self.port is not None:
811
netloc = '%s:%d' % (netloc, self.port)
812
return urlparse.urlunparse(
813
(self.scheme, netloc, self.quoted_path, None, None, None))
816
def _combine_paths(base_path, relpath):
817
"""Transform a Transport-relative path to a remote absolute path.
819
This does not handle substitution of ~ but does handle '..' and '.'
824
t._combine_paths('/home/sarah', 'project/foo')
825
=> '/home/sarah/project/foo'
826
t._combine_paths('/home/sarah', '../../etc')
828
t._combine_paths('/home/sarah', '/etc')
831
:param base_path: base path
832
:param relpath: relative url string for relative part of remote path.
833
:return: urlencoded string for final path.
835
if not isinstance(relpath, str):
836
raise errors.InvalidURL(relpath)
837
if relpath.startswith('/'):
840
base_parts = base_path.split('/')
841
if len(base_parts) > 0 and base_parts[-1] == '':
842
base_parts = base_parts[:-1]
843
for p in relpath.split('/'):
845
if len(base_parts) == 0:
846
# In most filesystems, a request for the parent
847
# of root, just returns root.
854
path = '/'.join(base_parts)
855
if not path.startswith('/'):
859
def clone(self, offset=None):
860
"""Return a new URL for a path relative to this URL.
862
:param offset: A relative path, already urlencoded
863
:return: `URL` instance
865
if offset is not None:
866
relative = unescape(offset).encode('utf-8')
867
path = self._combine_paths(self.path, relative)
868
path = urllib.quote(path)
870
path = self.quoted_path
871
return self.__class__(self.scheme, self.quoted_user,
872
self.quoted_password, self.quoted_host, self.port,
697
876
def parse_url(url):
698
877
"""Extract the server address, the credentials and the path from the url.
703
882
:param url: an quoted url
705
883
: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)
886
parsed_url = URL.from_string(url)
887
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
888
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
bzrlib/urlutils.py
