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
305
:param url: Either a hybrid URL or a local path
340
306
:return: A normalized URL which only includes 7-bit ASCII characters.
342
m = _url_scheme_re.match(url)
308
scheme_end, path_start = _find_scheme_and_separator(url)
309
if scheme_end is None:
344
310
return local_path_to_url(url)
345
scheme = m.group('scheme')
346
path = m.group('path')
311
prefix = url[:path_start]
312
path = url[path_start:]
347
313
if not isinstance(url, unicode):
349
315
if c not in _url_safe_characters:
350
316
raise errors.InvalidURL(url, 'URLs can only contain specific'
351
317
' safe characters (not %r)' % c)
352
318
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
353
return str(scheme + '://' + ''.join(path))
319
return str(prefix + ''.join(path))
355
321
# We have a unicode (hybrid) url
356
322
path_chars = list(path)
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
(parent_url, child_dir) = split(url)
445
subsegments = child_dir.split(",")
446
if len(subsegments) == 1:
448
return (join(parent_url, subsegments[0]), subsegments[1:])
451
def split_segment_parameters(url):
452
"""Split the segment parameters of the last segment of a URL.
454
:param url: A relative or absolute URL
455
:return: (url, segment_parameters)
457
(base_url, subsegments) = split_segment_parameters_raw(url)
459
for subsegment in subsegments:
460
(key, value) = subsegment.split("=", 1)
461
parameters[key] = value
462
return (base_url, parameters)
465
def join_segment_parameters_raw(base, *subsegments):
466
"""Create a new URL by adding subsegments to an existing one.
468
This adds the specified subsegments to the last path in the specified
469
base URL. The subsegments should be bytestrings.
471
:note: You probably want to use join_segment_parameters instead.
475
for subsegment in subsegments:
476
if type(subsegment) is not str:
477
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
478
if "," in subsegment:
479
raise errors.InvalidURLJoin(", exists in subsegments",
481
return ",".join((base,) + subsegments)
484
def join_segment_parameters(url, parameters):
485
"""Create a new URL by adding segment parameters to an existing one.
487
The parameters of the last segment in the URL will be updated; if a
488
parameter with the same key already exists it will be overwritten.
490
:param url: A URL, as string
491
:param parameters: Dictionary of parameters, keys and values as bytestrings
493
(base, existing_parameters) = split_segment_parameters(url)
495
new_parameters.update(existing_parameters)
496
for key, value in parameters.iteritems():
497
if type(key) is not str:
498
raise TypeError("parameter key %r is not a bytestring" % key)
499
if type(value) is not str:
500
raise TypeError("parameter value %r for %s is not a bytestring" %
503
raise errors.InvalidURLJoin("= exists in parameter key", url,
505
new_parameters[key] = value
506
return join_segment_parameters_raw(base,
507
*["%s=%s" % item for item in sorted(new_parameters.items())])
472
510
def _win32_strip_local_trailing_slash(url):
473
511
"""Strip slashes after the drive letter"""
474
512
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
693
731
return osutils.pathjoin(*segments)
737
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
740
self.quoted_host = quoted_host
741
self.host = urllib.unquote(self.quoted_host)
742
self.quoted_user = quoted_user
743
if self.quoted_user is not None:
744
self.user = urllib.unquote(self.quoted_user)
747
self.quoted_password = quoted_password
748
if self.quoted_password is not None:
749
self.password = urllib.unquote(self.quoted_password)
753
self.quoted_path = quoted_path
754
self.path = urllib.unquote(self.quoted_path)
756
def __eq__(self, other):
757
return (isinstance(other, self.__class__) and
758
self.scheme == other.scheme and
759
self.host == other.host and
760
self.user == other.user and
761
self.password == other.password and
762
self.path == other.path)
765
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
766
self.__class__.__name__,
767
self.scheme, self.quoted_user, self.quoted_password,
768
self.quoted_host, self.port, self.quoted_path)
771
def from_string(cls, url):
772
"""Create a URL object from a string.
774
:param url: URL as bytestring
776
if isinstance(url, unicode):
777
raise errors.InvalidURL('should be ascii:\n%r' % url)
778
url = url.encode('utf-8')
779
(scheme, netloc, path, params,
780
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
781
user = password = host = port = None
783
user, host = netloc.rsplit('@', 1)
785
user, password = user.split(':', 1)
789
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
791
host, port = host.rsplit(':',1)
795
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
797
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
800
return cls(scheme, user, password, host, port, path)
803
netloc = self.quoted_host
805
netloc = "[%s]" % netloc
806
if self.quoted_user is not None:
807
# Note that we don't put the password back even if we
808
# have one so that it doesn't get accidentally
810
netloc = '%s@%s' % (self.quoted_user, netloc)
811
if self.port is not None:
812
netloc = '%s:%d' % (netloc, self.port)
813
return urlparse.urlunparse(
814
(self.scheme, netloc, self.quoted_path, None, None, None))
817
def _combine_paths(base_path, relpath):
818
"""Transform a Transport-relative path to a remote absolute path.
820
This does not handle substitution of ~ but does handle '..' and '.'
825
t._combine_paths('/home/sarah', 'project/foo')
826
=> '/home/sarah/project/foo'
827
t._combine_paths('/home/sarah', '../../etc')
829
t._combine_paths('/home/sarah', '/etc')
832
:param base_path: base path
833
:param relpath: relative url string for relative part of remote path.
834
:return: urlencoded string for final path.
836
if not isinstance(relpath, str):
837
raise errors.InvalidURL(relpath)
838
if relpath.startswith('/'):
841
base_parts = base_path.split('/')
842
if len(base_parts) > 0 and base_parts[-1] == '':
843
base_parts = base_parts[:-1]
844
for p in relpath.split('/'):
846
if len(base_parts) == 0:
847
# In most filesystems, a request for the parent
848
# of root, just returns root.
855
path = '/'.join(base_parts)
856
if not path.startswith('/'):
860
def clone(self, offset=None):
861
"""Return a new URL for a path relative to this URL.
863
:param offset: A relative path, already urlencoded
864
:return: `URL` instance
866
if offset is not None:
867
relative = unescape(offset).encode('utf-8')
868
path = self._combine_paths(self.path, relative)
869
path = urllib.quote(path)
871
path = self.quoted_path
872
return self.__class__(self.scheme, self.quoted_user,
873
self.quoted_password, self.quoted_host, self.port,
697
877
def parse_url(url):
698
878
"""Extract the server address, the credentials and the path from the url.
703
883
:param url: an quoted url
705
884
: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)
887
parsed_url = URL.from_string(url)
888
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
889
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
bzrlib/urlutils.py
