1
# Copyright (C) 2006-2010 Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# 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
"""A collection of function for handling URL operations."""
19
from __future__ import absolute_import
28
from urllib import parse as urlparse
30
from .lazy_import import lazy_import
31
lazy_import(globals(), """
32
from posixpath import split as _posix_split
46
def basename(url, exclude_trailing_slash=True):
47
"""Return the last component of a URL.
49
:param url: The URL in question
50
:param exclude_trailing_slash: If the url looks like "path/to/foo/"
51
ignore the final slash and return 'foo' rather than ''
52
:return: Just the final component of the URL. This can return ''
53
if you don't exclude_trailing_slash, or if you are at the
56
return split(url, exclude_trailing_slash=exclude_trailing_slash)[1]
59
def dirname(url, exclude_trailing_slash=True):
60
"""Return the parent directory of the given path.
62
:param url: Relative or absolute URL
63
:param exclude_trailing_slash: Remove a final slash
64
(treat http://host/foo/ as http://host/foo, but
65
http://host/ stays http://host/)
66
:return: Everything in the URL except the last path chunk
68
# TODO: jam 20060502 This was named dirname to be consistent
69
# with the os functions, but maybe "parent" would be better
70
return split(url, exclude_trailing_slash=exclude_trailing_slash)[0]
73
# Private copies of quote and unquote, copied from Python's
74
# urllib module because urllib unconditionally imports socket, which imports
77
always_safe = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
78
'abcdefghijklmnopqrstuvwxyz'
81
for i, c in zip(range(256), ''.join(map(chr, range(256)))):
82
_safe_map[c] = c if (i < 128 and c in always_safe) else '%{0:02X}'.format(i)
86
def quote(s, safe='/'):
87
"""quote('abc def') -> 'abc%20def'
89
Each part of a URL, e.g. the path info, the query, etc., has a
90
different set of reserved characters that must be quoted.
92
RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists
93
the following reserved characters.
95
reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
98
Each of these characters is reserved in some component of a URL,
99
but not necessarily in all of them.
101
By default, the quote function is intended for quoting the path
102
section of a URL. Thus, it will not encode '/'. This character
103
is reserved, but in typical usage the quote function is being
104
called on a path where the existing slash characters are used as
110
raise TypeError('None object cannot be quoted')
112
cachekey = (safe, always_safe)
114
(quoter, safe) = _safe_quoters[cachekey]
116
safe_map = _safe_map.copy()
117
safe_map.update([(c, c) for c in safe])
118
quoter = safe_map.__getitem__
119
safe = always_safe + safe
120
_safe_quoters[cachekey] = (quoter, safe)
121
if not s.rstrip(safe):
123
return ''.join(map(quoter, s))
126
unquote = urlparse.unquote
130
"""Escape relpath to be a valid url."""
131
if not isinstance(relpath, str):
132
relpath = relpath.encode('utf-8')
133
return quote(relpath, safe='/~')
136
def file_relpath(base, path):
137
"""Compute just the relative sub-portion of a url
139
This assumes that both paths are already fully specified file:// URLs.
141
if len(base) < MIN_ABS_FILEURL_LENGTH:
142
raise ValueError('Length of base (%r) must equal or'
143
' exceed the platform minimum url length (which is %d)' %
144
(base, MIN_ABS_FILEURL_LENGTH))
145
base = osutils.normpath(local_path_from_url(base))
146
path = osutils.normpath(local_path_from_url(path))
147
return escape(osutils.relpath(base, path))
150
def _find_scheme_and_separator(url):
151
"""Find the scheme separator (://) and the first path separator
153
This is just a helper functions for other path utilities.
154
It could probably be replaced by urlparse
156
m = _url_scheme_re.match(url)
160
scheme = m.group('scheme')
161
path = m.group('path')
163
# Find the path separating slash
164
# (first slash after the ://)
165
first_path_slash = path.find('/')
166
if first_path_slash == -1:
167
return len(scheme), None
168
return len(scheme), first_path_slash+m.start('path')
172
"""Tests whether a URL is in actual fact a URL."""
173
return _url_scheme_re.match(url) is not None
176
def join(base, *args):
177
"""Create a URL by joining sections.
179
This will normalize '..', assuming that paths are absolute
180
(it assumes no symlinks in either path)
182
If any of *args is an absolute URL, it will be treated correctly.
184
join('http://foo', 'http://bar') => 'http://bar'
185
join('http://foo', 'bar') => 'http://foo/bar'
186
join('http://foo', 'bar', '../baz') => 'http://foo/baz'
190
scheme_end, path_start = _find_scheme_and_separator(base)
191
if scheme_end is None and path_start is None:
193
elif path_start is None:
194
path_start = len(base)
195
path = base[path_start:]
197
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
198
if arg_scheme_end is None and arg_path_start is None:
200
elif arg_path_start is None:
201
arg_path_start = len(arg)
202
if arg_scheme_end is not None:
204
path = arg[arg_path_start:]
205
scheme_end = arg_scheme_end
206
path_start = arg_path_start
208
path = joinpath(path, arg)
209
return base[:path_start] + path
212
def joinpath(base, *args):
213
"""Join URL path segments to a URL path segment.
215
This is somewhat like osutils.joinpath, but intended for URLs.
217
XXX: this duplicates some normalisation logic, and also duplicates a lot of
218
path handling logic that already exists in some Transport implementations.
219
We really should try to have exactly one place in the code base responsible
220
for combining paths of URLs.
222
path = base.split('/')
223
if len(path) > 1 and path[-1] == '':
224
#If the path ends in a trailing /, remove it.
227
if arg.startswith('/'):
229
for chunk in arg.split('/'):
234
raise errors.InvalidURLJoin('Cannot go above root',
242
return '/'.join(path)
245
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
246
def _posix_local_path_from_url(url):
247
"""Convert a url like file:///path/to/foo into /path/to/foo"""
248
url = split_segment_parameters_raw(url)[0]
249
file_localhost_prefix = 'file://localhost/'
250
if url.startswith(file_localhost_prefix):
251
path = url[len(file_localhost_prefix) - 1:]
252
elif not url.startswith('file:///'):
253
raise errors.InvalidURL(
254
url, 'local urls must start with file:/// or file://localhost/')
256
path = url[len('file://'):]
257
# We only strip off 2 slashes
258
return unescape(path)
261
def _posix_local_path_to_url(path):
262
"""Convert a local path like ./foo into a URL like file:///path/to/foo
264
This also handles transforming escaping unicode characters, etc.
266
# importing directly from posixpath allows us to test this
267
# on non-posix platforms
268
return 'file://' + escape(osutils._posix_abspath(path))
271
def _win32_local_path_from_url(url):
272
"""Convert a url like file:///C:/path/to/foo into C:/path/to/foo"""
273
if not url.startswith('file://'):
274
raise errors.InvalidURL(url, 'local urls must start with file:///, '
275
'UNC path urls must start with file://')
276
url = split_segment_parameters_raw(url)[0]
277
# We strip off all 3 slashes
278
win32_url = url[len('file:'):]
279
# check for UNC path: //HOST/path
280
if not win32_url.startswith('///'):
281
if (win32_url[2] == '/'
282
or win32_url[3] in '|:'):
283
raise errors.InvalidURL(url, 'Win32 UNC path urls'
284
' have form file://HOST/path')
285
return unescape(win32_url)
287
# allow empty paths so we can serve all roots
288
if win32_url == '///':
291
# usual local path with drive letter
292
if (len(win32_url) < 6
293
or win32_url[3] not in ('abcdefghijklmnopqrstuvwxyz'
294
'ABCDEFGHIJKLMNOPQRSTUVWXYZ')
295
or win32_url[4] not in '|:'
296
or win32_url[5] != '/'):
297
raise errors.InvalidURL(url, 'Win32 file urls start with'
298
' file:///x:/, where x is a valid drive letter')
299
return win32_url[3].upper() + u':' + unescape(win32_url[5:])
302
def _win32_local_path_to_url(path):
303
"""Convert a local path like ./foo into a URL like file:///C:/path/to/foo
305
This also handles transforming escaping unicode characters, etc.
307
# importing directly from ntpath allows us to test this
308
# on non-win32 platform
309
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
310
# which actually strips trailing space characters.
311
# The worst part is that on linux ntpath.abspath has different
312
# semantics, since 'nt' is not an available module.
316
win32_path = osutils._win32_abspath(path)
317
# check for UNC path \\HOST\path
318
if win32_path.startswith('//'):
319
return 'file:' + escape(win32_path)
320
return ('file:///' + str(win32_path[0].upper()) + ':' +
321
escape(win32_path[2:]))
324
local_path_to_url = _posix_local_path_to_url
325
local_path_from_url = _posix_local_path_from_url
326
MIN_ABS_FILEURL_LENGTH = len('file:///')
327
WIN32_MIN_ABS_FILEURL_LENGTH = len('file:///C:/')
329
if sys.platform == 'win32':
330
local_path_to_url = _win32_local_path_to_url
331
local_path_from_url = _win32_local_path_from_url
333
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH
336
_url_scheme_re = re.compile('^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$')
337
_url_hex_escapes_re = re.compile('(%[0-9a-fA-F]{2})')
340
def _unescape_safe_chars(matchobj):
341
"""re.sub callback to convert hex-escapes to plain characters (if safe).
343
e.g. '%7E' will be converted to '~'.
345
hex_digits = matchobj.group(0)[1:]
346
char = chr(int(hex_digits, 16))
347
if char in _url_dont_escape_characters:
350
return matchobj.group(0).upper()
353
def normalize_url(url):
354
"""Make sure that a path string is in fully normalized URL form.
356
This handles URLs which have unicode characters, spaces,
357
special characters, etc.
359
It has two basic modes of operation, depending on whether the
360
supplied string starts with a url specifier (scheme://) or not.
361
If it does not have a specifier it is considered a local path,
362
and will be converted into a file:/// url. Non-ascii characters
363
will be encoded using utf-8.
364
If it does have a url specifier, it will be treated as a "hybrid"
365
URL. Basically, a URL that should have URL special characters already
366
escaped (like +?&# etc), but may have unicode characters, etc
367
which would not be valid in a real URL.
369
:param url: Either a hybrid URL or a local path
370
:return: A normalized URL which only includes 7-bit ASCII characters.
372
scheme_end, path_start = _find_scheme_and_separator(url)
373
if scheme_end is None:
374
return local_path_to_url(url)
375
prefix = url[:path_start]
376
path = url[path_start:]
377
if not isinstance(url, text_type):
379
if c not in _url_safe_characters:
380
raise errors.InvalidURL(url, 'URLs can only contain specific'
381
' safe characters (not %r)' % c)
382
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
383
return str(prefix + ''.join(path))
385
# We have a unicode (hybrid) url
386
path_chars = list(path)
388
for i in range(len(path_chars)):
389
if path_chars[i] not in _url_safe_characters:
390
chars = path_chars[i].encode('utf-8')
391
path_chars[i] = ''.join(
392
['%%%02X' % ord(c) for c in path_chars[i].encode('utf-8')])
393
path = ''.join(path_chars)
394
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
395
return str(prefix + path)
398
def relative_url(base, other):
399
"""Return a path to other from base.
401
If other is unrelated to base, return other. Else return a relative path.
402
This assumes no symlinks as part of the url.
404
dummy, base_first_slash = _find_scheme_and_separator(base)
405
if base_first_slash is None:
408
dummy, other_first_slash = _find_scheme_and_separator(other)
409
if other_first_slash is None:
412
# this takes care of differing schemes or hosts
413
base_scheme = base[:base_first_slash]
414
other_scheme = other[:other_first_slash]
415
if base_scheme != other_scheme:
417
elif sys.platform == 'win32' and base_scheme == 'file://':
418
base_drive = base[base_first_slash+1:base_first_slash+3]
419
other_drive = other[other_first_slash+1:other_first_slash+3]
420
if base_drive != other_drive:
423
base_path = base[base_first_slash+1:]
424
other_path = other[other_first_slash+1:]
426
if base_path.endswith('/'):
427
base_path = base_path[:-1]
429
base_sections = base_path.split('/')
430
other_sections = other_path.split('/')
432
if base_sections == ['']:
434
if other_sections == ['']:
438
for b, o in zip(base_sections, other_sections):
441
output_sections.append(b)
443
match_len = len(output_sections)
444
output_sections = ['..' for x in base_sections[match_len:]]
445
output_sections.extend(other_sections[match_len:])
447
return "/".join(output_sections) or "."
450
def _win32_extract_drive_letter(url_base, path):
451
"""On win32 the drive letter needs to be added to the url base."""
452
# Strip off the drive letter
453
# path is currently /C:/foo
454
if len(path) < 4 or path[2] not in ':|' or path[3] != '/':
455
raise errors.InvalidURL(url_base + path,
456
'win32 file:/// paths need a drive letter')
457
url_base += path[0:3] # file:// + /C:
458
path = path[3:] # /foo
459
return url_base, path
462
def split(url, exclude_trailing_slash=True):
463
"""Split a URL into its parent directory and a child directory.
465
:param url: A relative or absolute URL
466
:param exclude_trailing_slash: Strip off a final '/' if it is part
467
of the path (but not if it is part of the protocol specification)
469
:return: (parent_url, child_dir). child_dir may be the empty string if we're at
472
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
474
if first_path_slash is None:
475
# We have either a relative path, or no separating slash
476
if scheme_loc is None:
478
if exclude_trailing_slash and url.endswith('/'):
480
return _posix_split(url)
482
# Scheme with no path
485
# We have a fully defined path
486
url_base = url[:first_path_slash] # http://host, file://
487
path = url[first_path_slash:] # /file/foo
489
if sys.platform == 'win32' and url.startswith('file:///'):
490
# Strip off the drive letter
491
# url_base is currently file://
492
# path is currently /C:/foo
493
url_base, path = _win32_extract_drive_letter(url_base, path)
494
# now it should be file:///C: and /foo
496
if exclude_trailing_slash and len(path) > 1 and path.endswith('/'):
498
head, tail = _posix_split(path)
499
return url_base + head, tail
502
def split_segment_parameters_raw(url):
503
"""Split the subsegment of the last segment of a URL.
505
:param url: A relative or absolute URL
506
:return: (url, subsegments)
508
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
509
# operates on urls not url+segments, and Transport classes
510
# should not be blindly adding slashes in the first place.
511
lurl = strip_trailing_slash(url)
512
# Segments begin at first comma after last forward slash, if one exists
513
segment_start = lurl.find(",", lurl.rfind("/")+1)
514
if segment_start == -1:
516
return (lurl[:segment_start], lurl[segment_start+1:].split(","))
519
def split_segment_parameters(url):
520
"""Split the segment parameters of the last segment of a URL.
522
:param url: A relative or absolute URL
523
:return: (url, segment_parameters)
525
(base_url, subsegments) = split_segment_parameters_raw(url)
527
for subsegment in subsegments:
528
(key, value) = subsegment.split("=", 1)
529
parameters[key] = value
530
return (base_url, parameters)
533
def join_segment_parameters_raw(base, *subsegments):
534
"""Create a new URL by adding subsegments to an existing one.
536
This adds the specified subsegments to the last path in the specified
537
base URL. The subsegments should be bytestrings.
539
:note: You probably want to use join_segment_parameters instead.
543
for subsegment in subsegments:
544
if not isinstance(subsegment, str):
545
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
546
if "," in subsegment:
547
raise errors.InvalidURLJoin(", exists in subsegments",
549
return ",".join((base,) + subsegments)
552
def join_segment_parameters(url, parameters):
553
"""Create a new URL by adding segment parameters to an existing one.
555
The parameters of the last segment in the URL will be updated; if a
556
parameter with the same key already exists it will be overwritten.
558
:param url: A URL, as string
559
:param parameters: Dictionary of parameters, keys and values as bytestrings
561
(base, existing_parameters) = split_segment_parameters(url)
563
new_parameters.update(existing_parameters)
564
for key, value in parameters.items():
565
if not isinstance(key, str):
566
raise TypeError("parameter key %r is not a bytestring" % key)
567
if not isinstance(value, str):
568
raise TypeError("parameter value %r for %s is not a bytestring" %
571
raise errors.InvalidURLJoin("= exists in parameter key", url,
573
new_parameters[key] = value
574
return join_segment_parameters_raw(base,
575
*["%s=%s" % item for item in sorted(new_parameters.items())])
578
def _win32_strip_local_trailing_slash(url):
579
"""Strip slashes after the drive letter"""
580
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
586
def strip_trailing_slash(url):
587
"""Strip trailing slash, except for root paths.
589
The definition of 'root path' is platform-dependent.
590
This assumes that all URLs are valid netloc urls, such that they
593
It searches for ://, and then refuses to remove the next '/'.
594
It can also handle relative paths
596
path/to/foo => path/to/foo
597
path/to/foo/ => path/to/foo
598
http://host/path/ => http://host/path
599
http://host/path => http://host/path
600
http://host/ => http://host/
602
file:///foo/ => file:///foo
603
# This is unique on win32 platforms, and is the only URL
604
# format which does it differently.
605
file:///c|/ => file:///c:/
607
if not url.endswith('/'):
610
if sys.platform == 'win32' and url.startswith('file://'):
611
return _win32_strip_local_trailing_slash(url)
613
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
614
if scheme_loc is None:
615
# This is a relative path, as it has no scheme
616
# so just chop off the last character
619
if first_path_slash is None or first_path_slash == len(url)-1:
620
# Don't chop off anything if the only slash is the path
628
"""Unescape relpath from url format.
630
This returns a Unicode path from a URL
632
# jam 20060427 URLs are supposed to be ASCII only strings
633
# If they are passed in as unicode, unquote
634
# will return a UNICODE string, which actually contains
635
# utf-8 bytes. So we have to ensure that they are
636
# plain ASCII strings, or the final .decode will
637
# try to encode the UNICODE => ASCII, and then decode
639
if isinstance(url, text_type):
641
url = url.encode("ascii")
642
except UnicodeError as e:
643
raise errors.InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,))
645
unquoted = urlparse.unquote_to_bytes(url)
647
unquoted = unquote(url)
649
unicode_path = unquoted.decode('utf-8')
650
except UnicodeError as e:
651
raise errors.InvalidURL(url, 'Unable to encode the URL as utf-8: %s' % (e,))
655
# These are characters that if escaped, should stay that way
656
_no_decode_chars = ';/?:@&=+$,#'
657
_no_decode_ords = [ord(c) for c in _no_decode_chars]
658
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
659
+ ['%02X' % o for o in _no_decode_ords])
660
_hex_display_map = dict(([('%02x' % o, chr(o)) for o in range(256)]
661
+ [('%02X' % o, chr(o)) for o in range(256)]))
662
#These entries get mapped to themselves
663
_hex_display_map.update((hex,'%'+hex) for hex in _no_decode_hex)
665
# These characters shouldn't be percent-encoded, and it's always safe to
666
# unencode them if they are.
667
_url_dont_escape_characters = set(
668
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
669
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
670
"0123456789" # Numbers
671
"-._~" # Unreserved characters
674
# These characters should not be escaped
675
_url_safe_characters = set(
676
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
677
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
678
"0123456789" # Numbers
679
"_.-!~*'()" # Unreserved characters
680
"/;?:@&=+$," # Reserved characters
681
"%#" # Extra reserved characters
684
def unescape_for_display(url, encoding):
685
"""Decode what you can for a URL, so that we get a nice looking path.
687
This will turn file:// urls into local paths, and try to decode
688
any portions of a http:// style url that it can.
690
Any sections of the URL which can't be represented in the encoding or
691
need to stay as escapes are left alone.
693
:param url: A 7-bit ASCII URL
694
:param encoding: The final output encoding
696
:return: A unicode string which can be safely encoded into the
700
raise ValueError('you cannot specify None for the display encoding')
701
if url.startswith('file://'):
703
path = local_path_from_url(url)
704
path.encode(encoding)
709
# Split into sections to try to decode utf-8
711
for i in range(1, len(res)):
712
escaped_chunks = res[i].split('%')
713
for j in range(1, len(escaped_chunks)):
714
item = escaped_chunks[j]
716
escaped_chunks[j] = _hex_display_map[item[:2]] + item[2:]
718
# Put back the percent symbol
719
escaped_chunks[j] = '%' + item
720
except UnicodeDecodeError:
721
escaped_chunks[j] = unichr(int(item[:2], 16)) + item[2:]
722
unescaped = ''.join(escaped_chunks)
724
decoded = unescaped.decode('utf-8')
725
except UnicodeDecodeError:
726
# If this path segment cannot be properly utf-8 decoded
727
# after doing unescaping we will just leave it alone
731
decoded.encode(encoding)
732
except UnicodeEncodeError:
733
# If this chunk cannot be encoded in the local
734
# encoding, then we should leave it alone
737
# Otherwise take the url decoded one
739
return u'/'.join(res)
742
def derive_to_location(from_location):
743
"""Derive a TO_LOCATION given a FROM_LOCATION.
745
The normal case is a FROM_LOCATION of http://foo/bar => bar.
746
The Right Thing for some logical destinations may differ though
747
because no / may be present at all. In that case, the result is
748
the full name without the scheme indicator, e.g. lp:foo-bar => foo-bar.
749
This latter case also applies when a Windows drive
750
is used without a path, e.g. c:foo-bar => foo-bar.
751
If no /, path separator or : is found, the from_location is returned.
753
if from_location.find("/") >= 0 or from_location.find(os.sep) >= 0:
754
return os.path.basename(from_location.rstrip("/\\"))
756
sep = from_location.find(":")
758
return from_location[sep+1:]
763
def _is_absolute(url):
764
return (osutils.pathjoin('/foo', url) == url)
767
def rebase_url(url, old_base, new_base):
768
"""Convert a relative path from an old base URL to a new base URL.
770
The result will be a relative path.
771
Absolute paths and full URLs are returned unaltered.
773
scheme, separator = _find_scheme_and_separator(url)
774
if scheme is not None:
776
if _is_absolute(url):
778
old_parsed = urlparse.urlparse(old_base)
779
new_parsed = urlparse.urlparse(new_base)
780
if (old_parsed[:2]) != (new_parsed[:2]):
781
raise errors.InvalidRebaseURLs(old_base, new_base)
782
return determine_relative_path(new_parsed[2],
783
join(old_parsed[2], url))
786
def determine_relative_path(from_path, to_path):
787
"""Determine a relative path from from_path to to_path."""
788
from_segments = osutils.splitpath(from_path)
789
to_segments = osutils.splitpath(to_path)
791
for count, (from_element, to_element) in enumerate(zip(from_segments,
793
if from_element != to_element:
797
unique_from = from_segments[count:]
798
unique_to = to_segments[count:]
799
segments = (['..'] * len(unique_from) + unique_to)
800
if len(segments) == 0:
802
return osutils.pathjoin(*segments)
808
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
811
self.quoted_host = quoted_host
812
self.host = unquote(self.quoted_host)
813
self.quoted_user = quoted_user
814
if self.quoted_user is not None:
815
self.user = unquote(self.quoted_user)
818
self.quoted_password = quoted_password
819
if self.quoted_password is not None:
820
self.password = unquote(self.quoted_password)
824
self.quoted_path = _url_hex_escapes_re.sub(_unescape_safe_chars, quoted_path)
825
self.path = unquote(self.quoted_path)
827
def __eq__(self, other):
828
return (isinstance(other, self.__class__) and
829
self.scheme == other.scheme and
830
self.host == other.host and
831
self.user == other.user and
832
self.password == other.password and
833
self.path == other.path)
836
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
837
self.__class__.__name__,
838
self.scheme, self.quoted_user, self.quoted_password,
839
self.quoted_host, self.port, self.quoted_path)
842
def from_string(cls, url):
843
"""Create a URL object from a string.
845
:param url: URL as bytestring
847
# GZ 2017-06-09: Actually validate ascii-ness
848
if not isinstance(url, str):
849
raise errors.InvalidURL('should be ascii:\n%r' % url)
850
(scheme, netloc, path, params,
851
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
852
user = password = host = port = None
854
user, host = netloc.rsplit('@', 1)
856
user, password = user.split(':', 1)
860
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
862
host, port = host.rsplit(':',1)
866
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
868
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
871
return cls(scheme, user, password, host, port, path)
874
netloc = self.quoted_host
876
netloc = "[%s]" % netloc
877
if self.quoted_user is not None:
878
# Note that we don't put the password back even if we
879
# have one so that it doesn't get accidentally
881
netloc = '%s@%s' % (self.quoted_user, netloc)
882
if self.port is not None:
883
netloc = '%s:%d' % (netloc, self.port)
884
return urlparse.urlunparse(
885
(self.scheme, netloc, self.quoted_path, None, None, None))
888
def _combine_paths(base_path, relpath):
889
"""Transform a Transport-relative path to a remote absolute path.
891
This does not handle substitution of ~ but does handle '..' and '.'
896
t._combine_paths('/home/sarah', 'project/foo')
897
=> '/home/sarah/project/foo'
898
t._combine_paths('/home/sarah', '../../etc')
900
t._combine_paths('/home/sarah', '/etc')
903
:param base_path: base path
904
:param relpath: relative url string for relative part of remote path.
905
:return: urlencoded string for final path.
907
if not isinstance(relpath, str):
908
raise errors.InvalidURL(relpath)
909
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
910
if relpath.startswith('/'):
913
base_parts = base_path.split('/')
914
if len(base_parts) > 0 and base_parts[-1] == '':
915
base_parts = base_parts[:-1]
916
for p in relpath.split('/'):
918
if len(base_parts) == 0:
919
# In most filesystems, a request for the parent
920
# of root, just returns root.
927
path = '/'.join(base_parts)
928
if not path.startswith('/'):
932
def clone(self, offset=None):
933
"""Return a new URL for a path relative to this URL.
935
:param offset: A relative path, already urlencoded
936
:return: `URL` instance
938
if offset is not None:
939
relative = unescape(offset).encode('utf-8')
940
path = self._combine_paths(self.path, relative)
941
path = quote(path, safe="/~")
943
path = self.quoted_path
944
return self.__class__(self.scheme, self.quoted_user,
945
self.quoted_password, self.quoted_host, self.port,
950
"""Extract the server address, the credentials and the path from the url.
952
user, password, host and path should be quoted if they contain reserved
955
:param url: an quoted url
956
:return: (scheme, user, password, host, port, path) tuple, all fields
959
parsed_url = URL.from_string(url)
960
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
961
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
breezy/urlutils.py
