13
15
# You should have received a copy of the GNU General Public License
14
16
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
19
"""A collection of function for handling URL operations."""
66
68
relpath = relpath.encode('utf-8')
67
69
# After quoting and encoding, the path should be perfectly
68
70
# safe as a plain ASCII string, str() just enforces this
69
return str(urllib.quote(relpath, safe='/~'))
71
return str(urllib.quote(relpath))
72
74
def file_relpath(base, path):
73
75
"""Compute just the relative sub-portion of a url
75
77
This assumes that both paths are already fully specified file:// URLs.
77
79
if len(base) < MIN_ABS_FILEURL_LENGTH:
78
raise ValueError('Length of base (%r) must equal or'
80
raise ValueError('Length of base must be equal or'
79
81
' exceed the platform minimum url length (which is %d)' %
80
(base, MIN_ABS_FILEURL_LENGTH))
82
MIN_ABS_FILEURL_LENGTH)
81
83
base = local_path_from_url(base)
82
84
path = local_path_from_url(path)
83
85
return escape(osutils.relpath(base, path))
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:]
123
match = _url_scheme_re.match(base)
126
scheme = match.group('scheme')
127
path = match.group('path').split('/')
128
if path[-1:] == ['']:
129
# Strip off a trailing slash
130
# This helps both when we are at the root, and when
131
# 'base' has an extra slash at the end
134
path = base.split('/')
136
if scheme is not None and len(path) >= 1:
138
# the path should be represented as an abs path.
139
# we know this must be absolute because of the presence of a URL scheme.
141
path = [''] + path[1:]
143
# create an empty host, but dont alter the path - this might be a
144
# 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
149
match = _url_scheme_re.match(arg)
152
scheme = match.group('scheme')
153
# this skips .. normalisation, making http://host/../../..
155
path = match.group('path').split('/')
156
# set the host and path according to new absolute URL, discarding
157
# any previous values.
158
# XXX: duplicates mess from earlier in this function. This URL
159
# manipulation code needs some cleaning up.
160
if scheme is not None and len(path) >= 1:
163
# url scheme implies absolute path.
166
# no url scheme we take the path as is.
169
path = '/'.join(path)
144
170
path = joinpath(path, arg)
145
return base[:path_start] + path
171
path = path.split('/')
172
if remove_root and path[0:1] == ['']:
175
# Remove the leading slash from the path, so long as it isn't also the
176
# trailing slash, which we want to keep if present.
177
if path and path[0] == '' and len(path) > 1:
182
return '/'.join(path)
183
return scheme + '://' + '/'.join(path)
148
186
def joinpath(base, *args):
149
187
"""Join URL path segments to a URL path segment.
151
189
This is somewhat like osutils.joinpath, but intended for URLs.
153
191
XXX: this duplicates some normalisation logic, and also duplicates a lot of
181
219
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
182
220
def _posix_local_path_from_url(url):
183
221
"""Convert a url like file:///path/to/foo into /path/to/foo"""
184
file_localhost_prefix = 'file://localhost/'
185
if url.startswith(file_localhost_prefix):
186
path = url[len(file_localhost_prefix) - 1:]
187
elif not url.startswith('file:///'):
188
raise errors.InvalidURL(
189
url, 'local urls must start with file:/// or file://localhost/')
191
path = url[len('file://'):]
222
if not url.startswith('file:///'):
223
raise errors.InvalidURL(url, 'local urls must start with file:///')
192
224
# We only strip off 2 slashes
193
return unescape(path)
225
return unescape(url[len('file://'):])
196
228
def _posix_local_path_to_url(path):
239
271
This also handles transforming escaping unicode characters, etc.
241
# importing directly from ntpath allows us to test this
273
# importing directly from ntpath allows us to test this
242
274
# on non-win32 platform
243
275
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
244
276
# which actually strips trailing space characters.
245
# The worst part is that on linux ntpath.abspath has different
277
# The worst part is that under linux ntpath.abspath has different
246
278
# semantics, since 'nt' is not an available module.
248
280
return 'file:///'
267
299
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH
270
_url_scheme_re = re.compile(r'^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$')
302
_url_scheme_re = re.compile(r'^(?P<scheme>[^:/]{2,})://(?P<path>.*)$')
271
303
_url_hex_escapes_re = re.compile(r'(%[0-9a-fA-F]{2})')
274
306
def _unescape_safe_chars(matchobj):
275
307
"""re.sub callback to convert hex-escapes to plain characters (if safe).
277
309
e.g. '%7E' will be converted to '~'.
279
311
hex_digits = matchobj.group(0)[1:]
303
335
:param url: Either a hybrid URL or a local path
304
336
:return: A normalized URL which only includes 7-bit ASCII characters.
306
scheme_end, path_start = _find_scheme_and_separator(url)
307
if scheme_end is None:
338
m = _url_scheme_re.match(url)
308
340
return local_path_to_url(url)
309
prefix = url[:path_start]
310
path = url[path_start:]
341
scheme = m.group('scheme')
342
path = m.group('path')
311
343
if not isinstance(url, unicode):
313
345
if c not in _url_safe_characters:
314
346
raise errors.InvalidURL(url, 'URLs can only contain specific'
315
347
' safe characters (not %r)' % c)
316
348
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
317
return str(prefix + ''.join(path))
349
return str(scheme + '://' + ''.join(path))
319
351
# We have a unicode (hybrid) url
320
352
path_chars = list(path)
433
465
return url_base + head, tail
436
def split_segment_parameters_raw(url):
437
"""Split the subsegment of the last segment of a URL.
439
:param url: A relative or absolute URL
440
:return: (url, subsegments)
442
(parent_url, child_dir) = split(url)
443
subsegments = child_dir.split(",")
444
if len(subsegments) == 1:
446
return (join(parent_url, subsegments[0]), subsegments[1:])
449
def split_segment_parameters(url):
450
"""Split the segment parameters of the last segment of a URL.
452
:param url: A relative or absolute URL
453
:return: (url, segment_parameters)
455
(base_url, subsegments) = split_segment_parameters_raw(url)
457
for subsegment in subsegments:
458
(key, value) = subsegment.split("=", 1)
459
parameters[key] = value
460
return (base_url, parameters)
463
def join_segment_parameters_raw(base, *subsegments):
464
"""Create a new URL by adding subsegments to an existing one.
466
This adds the specified subsegments to the last path in the specified
467
base URL. The subsegments should be bytestrings.
469
:note: You probably want to use join_segment_parameters instead.
473
for subsegment in subsegments:
474
if type(subsegment) is not str:
475
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
476
if "," in subsegment:
477
raise errors.InvalidURLJoin(", exists in subsegments",
479
return ",".join((base,) + subsegments)
482
def join_segment_parameters(url, parameters):
483
"""Create a new URL by adding segment parameters to an existing one.
485
The parameters of the last segment in the URL will be updated; if a
486
parameter with the same key already exists it will be overwritten.
488
:param url: A URL, as string
489
:param parameters: Dictionary of parameters, keys and values as bytestrings
491
(base, existing_parameters) = split_segment_parameters(url)
493
new_parameters.update(existing_parameters)
494
for key, value in parameters.iteritems():
495
if type(key) is not str:
496
raise TypeError("parameter key %r is not a bytestring" % key)
497
if type(value) is not str:
498
raise TypeError("parameter value %r for %s is not a bytestring" %
501
raise errors.InvalidURLJoin("= exists in parameter key", url,
503
new_parameters[key] = value
504
return join_segment_parameters_raw(base,
505
*["%s=%s" % item for item in sorted(new_parameters.items())])
508
468
def _win32_strip_local_trailing_slash(url):
509
469
"""Strip slashes after the drive letter"""
510
470
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
582
542
# These are characters that if escaped, should stay that way
583
543
_no_decode_chars = ';/?:@&=+$,#'
584
544
_no_decode_ords = [ord(c) for c in _no_decode_chars]
585
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
545
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
586
546
+ ['%02X' % o for o in _no_decode_ords])
587
547
_hex_display_map = dict(([('%02x' % o, chr(o)) for o in range(256)]
588
548
+ [('%02X' % o, chr(o)) for o in range(256)]))
614
574
This will turn file:// urls into local paths, and try to decode
615
575
any portions of a http:// style url that it can.
617
Any sections of the URL which can't be represented in the encoding or
577
Any sections of the URL which can't be represented in the encoding or
618
578
need to stay as escapes are left alone.
620
580
:param url: A 7-bit ASCII URL
621
581
:param encoding: The final output encoding
623
:return: A unicode string which can be safely encoded into the
583
:return: A unicode string which can be safely encoded into the
624
584
specified encoding.
626
586
if encoding is None:
727
687
if len(segments) == 0:
729
689
return osutils.pathjoin(*segments)
734
"""Extract the server address, the credentials and the path from the url.
736
user, password, host and path should be quoted if they contain reserved
739
:param url: an quoted url
741
:return: (scheme, user, password, host, port, path) tuple, all fields
744
if isinstance(url, unicode):
745
raise errors.InvalidURL('should be ascii:\n%r' % url)
746
url = url.encode('utf-8')
747
(scheme, netloc, path, params,
748
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
749
user = password = host = port = None
751
user, host = netloc.rsplit('@', 1)
753
user, password = user.split(':', 1)
754
password = urllib.unquote(password)
755
user = urllib.unquote(user)
759
if ':' in host and not (host[0] == '[' and host[-1] == ']'): #there *is* port
760
host, port = host.rsplit(':',1)
764
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
766
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
769
host = urllib.unquote(host)
770
path = urllib.unquote(path)
772
return (scheme, user, password, host, port, path)
added
removed
bzrlib/urlutils.py
