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:]
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.
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
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.
167
path = '/'.join(path)
144
168
path = joinpath(path, arg)
145
return base[:path_start] + path
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)
148
184
def joinpath(base, *args):
306
333
:param url: Either a hybrid URL or a local path
307
334
:return: A normalized URL which only includes 7-bit ASCII characters.
309
scheme_end, path_start = _find_scheme_and_separator(url)
310
if scheme_end is None:
336
m = _url_scheme_re.match(url)
311
338
return local_path_to_url(url)
312
prefix = url[:path_start]
313
path = url[path_start:]
339
scheme = m.group('scheme')
340
path = m.group('path')
314
341
if not isinstance(url, unicode):
316
343
if c not in _url_safe_characters:
317
344
raise errors.InvalidURL(url, 'URLs can only contain specific'
318
345
' safe characters (not %r)' % c)
319
346
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
320
return str(prefix + ''.join(path))
347
return str(scheme + '://' + ''.join(path))
322
349
# We have a unicode (hybrid) url
323
350
path_chars = list(path)
436
463
return url_base + head, tail
439
def split_segment_parameters_raw(url):
440
"""Split the subsegment of the last segment of a URL.
442
:param url: A relative or absolute URL
443
:return: (url, subsegments)
445
(parent_url, child_dir) = split(url)
446
subsegments = child_dir.split(",")
447
if len(subsegments) == 1:
449
return (join(parent_url, subsegments[0]), subsegments[1:])
452
def split_segment_parameters(url):
453
"""Split the segment parameters of the last segment of a URL.
455
:param url: A relative or absolute URL
456
:return: (url, segment_parameters)
458
(base_url, subsegments) = split_segment_parameters_raw(url)
460
for subsegment in subsegments:
461
(key, value) = subsegment.split("=", 1)
462
parameters[key] = value
463
return (base_url, parameters)
466
def join_segment_parameters_raw(base, *subsegments):
467
"""Create a new URL by adding subsegments to an existing one.
469
This adds the specified subsegments to the last path in the specified
470
base URL. The subsegments should be bytestrings.
472
:note: You probably want to use join_segment_parameters instead.
476
for subsegment in subsegments:
477
if type(subsegment) is not str:
478
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
479
if "," in subsegment:
480
raise errors.InvalidURLJoin(", exists in subsegments",
482
return ",".join((base,) + subsegments)
485
def join_segment_parameters(url, parameters):
486
"""Create a new URL by adding segment parameters to an existing one.
488
The parameters of the last segment in the URL will be updated; if a
489
parameter with the same key already exists it will be overwritten.
491
:param url: A URL, as string
492
:param parameters: Dictionary of parameters, keys and values as bytestrings
494
(base, existing_parameters) = split_segment_parameters(url)
496
new_parameters.update(existing_parameters)
497
for key, value in parameters.iteritems():
498
if type(key) is not str:
499
raise TypeError("parameter key %r is not a bytestring" % key)
500
if type(value) is not str:
501
raise TypeError("parameter value %r for %s is not a bytestring" %
504
raise errors.InvalidURLJoin("= exists in parameter key", url,
506
new_parameters[key] = value
507
return join_segment_parameters_raw(base,
508
*["%s=%s" % item for item in sorted(new_parameters.items())])
511
466
def _win32_strip_local_trailing_slash(url):
512
467
"""Strip slashes after the drive letter"""
513
468
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
732
687
return osutils.pathjoin(*segments)
738
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
741
self.quoted_host = quoted_host
742
self.host = urllib.unquote(self.quoted_host)
743
self.quoted_user = quoted_user
744
if self.quoted_user is not None:
745
self.user = urllib.unquote(self.quoted_user)
748
self.quoted_password = quoted_password
749
if self.quoted_password is not None:
750
self.password = urllib.unquote(self.quoted_password)
754
self.quoted_path = quoted_path
755
self.path = urllib.unquote(self.quoted_path)
757
def __eq__(self, other):
758
return (isinstance(other, self.__class__) and
759
self.scheme == other.scheme and
760
self.host == other.host and
761
self.user == other.user and
762
self.password == other.password and
763
self.path == other.path)
766
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
767
self.__class__.__name__,
768
self.scheme, self.quoted_user, self.quoted_password,
769
self.quoted_host, self.port, self.quoted_path)
772
def from_string(cls, url):
773
"""Create a URL object from a string.
775
:param url: URL as bytestring
777
if isinstance(url, unicode):
778
raise errors.InvalidURL('should be ascii:\n%r' % url)
779
url = url.encode('utf-8')
780
(scheme, netloc, path, params,
781
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
782
user = password = host = port = None
784
user, host = netloc.rsplit('@', 1)
786
user, password = user.split(':', 1)
790
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
792
host, port = host.rsplit(':',1)
796
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
798
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
801
return cls(scheme, user, password, host, port, path)
804
netloc = self.quoted_host
806
netloc = "[%s]" % netloc
807
if self.quoted_user is not None:
808
# Note that we don't put the password back even if we
809
# have one so that it doesn't get accidentally
811
netloc = '%s@%s' % (self.quoted_user, netloc)
812
if self.port is not None:
813
netloc = '%s:%d' % (netloc, self.port)
814
return urlparse.urlunparse(
815
(self.scheme, netloc, self.quoted_path, None, None, None))
818
def _combine_paths(base_path, relpath):
819
"""Transform a Transport-relative path to a remote absolute path.
821
This does not handle substitution of ~ but does handle '..' and '.'
826
t._combine_paths('/home/sarah', 'project/foo')
827
=> '/home/sarah/project/foo'
828
t._combine_paths('/home/sarah', '../../etc')
830
t._combine_paths('/home/sarah', '/etc')
833
:param base_path: base path
834
:param relpath: relative url string for relative part of remote path.
835
:return: urlencoded string for final path.
837
if not isinstance(relpath, str):
838
raise errors.InvalidURL(relpath)
839
if relpath.startswith('/'):
842
base_parts = base_path.split('/')
843
if len(base_parts) > 0 and base_parts[-1] == '':
844
base_parts = base_parts[:-1]
845
for p in relpath.split('/'):
847
if len(base_parts) == 0:
848
# In most filesystems, a request for the parent
849
# of root, just returns root.
856
path = '/'.join(base_parts)
857
if not path.startswith('/'):
861
def clone(self, offset=None):
862
"""Return a new URL for a path relative to this URL.
864
:param offset: A relative path, already urlencoded
865
:return: `URL` instance
867
if offset is not None:
868
relative = unescape(offset).encode('utf-8')
869
path = self._combine_paths(self.path, relative)
870
path = urllib.quote(path)
872
path = self.quoted_path
873
return self.__class__(self.scheme, self.quoted_user,
874
self.quoted_password, self.quoted_host, self.port,
878
691
def parse_url(url):
879
692
"""Extract the server address, the credentials and the path from the url.
884
697
:param url: an quoted url
885
699
:return: (scheme, user, password, host, port, path) tuple, all fields
888
parsed_url = URL.from_string(url)
889
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
890
parsed_url.host, parsed_url.port, parsed_url.path)
702
if isinstance(url, unicode):
703
raise errors.InvalidURL('should be ascii:\n%r' % url)
704
url = url.encode('utf-8')
705
(scheme, netloc, path, params,
706
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
707
user = password = host = port = None
709
user, host = netloc.rsplit('@', 1)
711
user, password = user.split(':', 1)
712
password = urllib.unquote(password)
713
user = urllib.unquote(user)
717
if ':' in host and not (host[0] == '[' and host[-1] == ']'): #there *is* port
718
host, port = host.rsplit(':',1)
722
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
724
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
727
host = urllib.unquote(host)
728
path = urllib.unquote(path)
730
return (scheme, user, password, host, port, path)
added
removed
bzrlib/urlutils.py
