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):
469
435
return url_base + head, tail
438
def split_segment_parameters_raw(url):
439
"""Split the subsegment of the last segment of a URL.
441
:param url: A relative or absolute URL
442
:return: (url, subsegments)
444
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
445
# operates on urls not url+segments, and Transport classes
446
# should not be blindly adding slashes in the first place.
447
lurl = strip_trailing_slash(url)
448
# Segments begin at first comma after last forward slash, if one exists
449
segment_start = lurl.find(",", lurl.rfind("/")+1)
450
if segment_start == -1:
452
return (lurl[:segment_start], lurl[segment_start+1:].split(","))
455
def split_segment_parameters(url):
456
"""Split the segment parameters of the last segment of a URL.
458
:param url: A relative or absolute URL
459
:return: (url, segment_parameters)
461
(base_url, subsegments) = split_segment_parameters_raw(url)
463
for subsegment in subsegments:
464
(key, value) = subsegment.split("=", 1)
465
parameters[key] = value
466
return (base_url, parameters)
469
def join_segment_parameters_raw(base, *subsegments):
470
"""Create a new URL by adding subsegments to an existing one.
472
This adds the specified subsegments to the last path in the specified
473
base URL. The subsegments should be bytestrings.
475
:note: You probably want to use join_segment_parameters instead.
479
for subsegment in subsegments:
480
if type(subsegment) is not str:
481
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
482
if "," in subsegment:
483
raise errors.InvalidURLJoin(", exists in subsegments",
485
return ",".join((base,) + subsegments)
488
def join_segment_parameters(url, parameters):
489
"""Create a new URL by adding segment parameters to an existing one.
491
The parameters of the last segment in the URL will be updated; if a
492
parameter with the same key already exists it will be overwritten.
494
:param url: A URL, as string
495
:param parameters: Dictionary of parameters, keys and values as bytestrings
497
(base, existing_parameters) = split_segment_parameters(url)
499
new_parameters.update(existing_parameters)
500
for key, value in parameters.iteritems():
501
if type(key) is not str:
502
raise TypeError("parameter key %r is not a bytestring" % key)
503
if type(value) is not str:
504
raise TypeError("parameter value %r for %s is not a bytestring" %
507
raise errors.InvalidURLJoin("= exists in parameter key", url,
509
new_parameters[key] = value
510
return join_segment_parameters_raw(base,
511
*["%s=%s" % item for item in sorted(new_parameters.items())])
472
514
def _win32_strip_local_trailing_slash(url):
473
515
"""Strip slashes after the drive letter"""
474
516
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
693
735
return osutils.pathjoin(*segments)
741
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
744
self.quoted_host = quoted_host
745
self.host = urllib.unquote(self.quoted_host)
746
self.quoted_user = quoted_user
747
if self.quoted_user is not None:
748
self.user = urllib.unquote(self.quoted_user)
751
self.quoted_password = quoted_password
752
if self.quoted_password is not None:
753
self.password = urllib.unquote(self.quoted_password)
757
self.quoted_path = _url_hex_escapes_re.sub(_unescape_safe_chars, quoted_path)
758
self.path = urllib.unquote(self.quoted_path)
760
def __eq__(self, other):
761
return (isinstance(other, self.__class__) and
762
self.scheme == other.scheme and
763
self.host == other.host and
764
self.user == other.user and
765
self.password == other.password and
766
self.path == other.path)
769
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
770
self.__class__.__name__,
771
self.scheme, self.quoted_user, self.quoted_password,
772
self.quoted_host, self.port, self.quoted_path)
775
def from_string(cls, url):
776
"""Create a URL object from a string.
778
:param url: URL as bytestring
780
if isinstance(url, unicode):
781
raise errors.InvalidURL('should be ascii:\n%r' % url)
782
url = url.encode('utf-8')
783
(scheme, netloc, path, params,
784
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
785
user = password = host = port = None
787
user, host = netloc.rsplit('@', 1)
789
user, password = user.split(':', 1)
793
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
795
host, port = host.rsplit(':',1)
799
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
801
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
804
return cls(scheme, user, password, host, port, path)
807
netloc = self.quoted_host
809
netloc = "[%s]" % netloc
810
if self.quoted_user is not None:
811
# Note that we don't put the password back even if we
812
# have one so that it doesn't get accidentally
814
netloc = '%s@%s' % (self.quoted_user, netloc)
815
if self.port is not None:
816
netloc = '%s:%d' % (netloc, self.port)
817
return urlparse.urlunparse(
818
(self.scheme, netloc, self.quoted_path, None, None, None))
821
def _combine_paths(base_path, relpath):
822
"""Transform a Transport-relative path to a remote absolute path.
824
This does not handle substitution of ~ but does handle '..' and '.'
829
t._combine_paths('/home/sarah', 'project/foo')
830
=> '/home/sarah/project/foo'
831
t._combine_paths('/home/sarah', '../../etc')
833
t._combine_paths('/home/sarah', '/etc')
836
:param base_path: base path
837
:param relpath: relative url string for relative part of remote path.
838
:return: urlencoded string for final path.
840
if not isinstance(relpath, str):
841
raise errors.InvalidURL(relpath)
842
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
843
if relpath.startswith('/'):
846
base_parts = base_path.split('/')
847
if len(base_parts) > 0 and base_parts[-1] == '':
848
base_parts = base_parts[:-1]
849
for p in relpath.split('/'):
851
if len(base_parts) == 0:
852
# In most filesystems, a request for the parent
853
# of root, just returns root.
860
path = '/'.join(base_parts)
861
if not path.startswith('/'):
865
def clone(self, offset=None):
866
"""Return a new URL for a path relative to this URL.
868
:param offset: A relative path, already urlencoded
869
:return: `URL` instance
871
if offset is not None:
872
relative = unescape(offset).encode('utf-8')
873
path = self._combine_paths(self.path, relative)
874
path = urllib.quote(path, safe="/~")
876
path = self.quoted_path
877
return self.__class__(self.scheme, self.quoted_user,
878
self.quoted_password, self.quoted_host, self.port,
697
882
def parse_url(url):
698
883
"""Extract the server address, the credentials and the path from the url.
703
888
:param url: an quoted url
705
889
: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)
892
parsed_url = URL.from_string(url)
893
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
894
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
bzrlib/urlutils.py
