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.
128
scheme_end, path_start = _find_scheme_and_separator(base)
129
if scheme_end is None and path_start is None:
131
elif path_start is None:
132
path_start = len(base)
133
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.
135
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
136
if arg_scheme_end is None and arg_path_start is None:
138
elif arg_path_start is None:
139
arg_path_start = len(arg)
140
if arg_scheme_end is not None:
142
path = arg[arg_path_start:]
143
scheme_end = arg_scheme_end
144
path_start = arg_path_start
167
path = '/'.join(path)
168
146
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)
147
return base[:path_start] + path
184
150
def joinpath(base, *args):
469
437
return url_base + head, tail
440
def split_segment_parameters_raw(url):
441
"""Split the subsegment of the last segment of a URL.
443
:param url: A relative or absolute URL
444
:return: (url, subsegments)
446
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
447
# operates on urls not url+segments, and Transport classes
448
# should not be blindly adding slashes in the first place.
449
lurl = strip_trailing_slash(url)
450
# Segments begin at first comma after last forward slash, if one exists
451
segment_start = lurl.find(",", lurl.rfind("/")+1)
452
if segment_start == -1:
454
return (lurl[:segment_start], lurl[segment_start+1:].split(","))
457
def split_segment_parameters(url):
458
"""Split the segment parameters of the last segment of a URL.
460
:param url: A relative or absolute URL
461
:return: (url, segment_parameters)
463
(base_url, subsegments) = split_segment_parameters_raw(url)
465
for subsegment in subsegments:
466
(key, value) = subsegment.split("=", 1)
467
parameters[key] = value
468
return (base_url, parameters)
471
def join_segment_parameters_raw(base, *subsegments):
472
"""Create a new URL by adding subsegments to an existing one.
474
This adds the specified subsegments to the last path in the specified
475
base URL. The subsegments should be bytestrings.
477
:note: You probably want to use join_segment_parameters instead.
481
for subsegment in subsegments:
482
if type(subsegment) is not str:
483
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
484
if "," in subsegment:
485
raise errors.InvalidURLJoin(", exists in subsegments",
487
return ",".join((base,) + subsegments)
490
def join_segment_parameters(url, parameters):
491
"""Create a new URL by adding segment parameters to an existing one.
493
The parameters of the last segment in the URL will be updated; if a
494
parameter with the same key already exists it will be overwritten.
496
:param url: A URL, as string
497
:param parameters: Dictionary of parameters, keys and values as bytestrings
499
(base, existing_parameters) = split_segment_parameters(url)
501
new_parameters.update(existing_parameters)
502
for key, value in parameters.iteritems():
503
if type(key) is not str:
504
raise TypeError("parameter key %r is not a bytestring" % key)
505
if type(value) is not str:
506
raise TypeError("parameter value %r for %s is not a bytestring" %
509
raise errors.InvalidURLJoin("= exists in parameter key", url,
511
new_parameters[key] = value
512
return join_segment_parameters_raw(base,
513
*["%s=%s" % item for item in sorted(new_parameters.items())])
472
516
def _win32_strip_local_trailing_slash(url):
473
517
"""Strip slashes after the drive letter"""
474
518
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
693
737
return osutils.pathjoin(*segments)
743
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
746
self.quoted_host = quoted_host
747
self.host = urllib.unquote(self.quoted_host)
748
self.quoted_user = quoted_user
749
if self.quoted_user is not None:
750
self.user = urllib.unquote(self.quoted_user)
753
self.quoted_password = quoted_password
754
if self.quoted_password is not None:
755
self.password = urllib.unquote(self.quoted_password)
759
self.quoted_path = _url_hex_escapes_re.sub(_unescape_safe_chars, quoted_path)
760
self.path = urllib.unquote(self.quoted_path)
762
def __eq__(self, other):
763
return (isinstance(other, self.__class__) and
764
self.scheme == other.scheme and
765
self.host == other.host and
766
self.user == other.user and
767
self.password == other.password and
768
self.path == other.path)
771
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
772
self.__class__.__name__,
773
self.scheme, self.quoted_user, self.quoted_password,
774
self.quoted_host, self.port, self.quoted_path)
777
def from_string(cls, url):
778
"""Create a URL object from a string.
780
:param url: URL as bytestring
782
if isinstance(url, unicode):
783
raise errors.InvalidURL('should be ascii:\n%r' % url)
784
url = url.encode('utf-8')
785
(scheme, netloc, path, params,
786
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
787
user = password = host = port = None
789
user, host = netloc.rsplit('@', 1)
791
user, password = user.split(':', 1)
795
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
797
host, port = host.rsplit(':',1)
801
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
803
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
806
return cls(scheme, user, password, host, port, path)
809
netloc = self.quoted_host
811
netloc = "[%s]" % netloc
812
if self.quoted_user is not None:
813
# Note that we don't put the password back even if we
814
# have one so that it doesn't get accidentally
816
netloc = '%s@%s' % (self.quoted_user, netloc)
817
if self.port is not None:
818
netloc = '%s:%d' % (netloc, self.port)
819
return urlparse.urlunparse(
820
(self.scheme, netloc, self.quoted_path, None, None, None))
823
def _combine_paths(base_path, relpath):
824
"""Transform a Transport-relative path to a remote absolute path.
826
This does not handle substitution of ~ but does handle '..' and '.'
831
t._combine_paths('/home/sarah', 'project/foo')
832
=> '/home/sarah/project/foo'
833
t._combine_paths('/home/sarah', '../../etc')
835
t._combine_paths('/home/sarah', '/etc')
838
:param base_path: base path
839
:param relpath: relative url string for relative part of remote path.
840
:return: urlencoded string for final path.
842
if not isinstance(relpath, str):
843
raise errors.InvalidURL(relpath)
844
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
845
if relpath.startswith('/'):
848
base_parts = base_path.split('/')
849
if len(base_parts) > 0 and base_parts[-1] == '':
850
base_parts = base_parts[:-1]
851
for p in relpath.split('/'):
853
if len(base_parts) == 0:
854
# In most filesystems, a request for the parent
855
# of root, just returns root.
862
path = '/'.join(base_parts)
863
if not path.startswith('/'):
867
def clone(self, offset=None):
868
"""Return a new URL for a path relative to this URL.
870
:param offset: A relative path, already urlencoded
871
:return: `URL` instance
873
if offset is not None:
874
relative = unescape(offset).encode('utf-8')
875
path = self._combine_paths(self.path, relative)
876
path = urllib.quote(path, safe="/~")
878
path = self.quoted_path
879
return self.__class__(self.scheme, self.quoted_user,
880
self.quoted_password, self.quoted_host, self.port,
697
884
def parse_url(url):
698
885
"""Extract the server address, the credentials and the path from the url.
703
890
:param url: an quoted url
705
891
: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)
894
parsed_url = URL.from_string(url)
895
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
896
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
bzrlib/urlutils.py
