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
35
from .lazy_import import lazy_import
36
lazy_import(globals(), """
37
from posixpath import split as _posix_split
46
class InvalidURL(errors.PathError):
48
_fmt = 'Invalid url supplied to transport: "%(path)s"%(extra)s'
51
class InvalidURLJoin(errors.PathError):
53
_fmt = "Invalid URL join request: %(reason)s: %(base)r + %(join_args)r"
55
def __init__(self, reason, base, join_args):
58
self.join_args = join_args
59
errors.PathError.__init__(self, base, reason)
62
class InvalidRebaseURLs(errors.PathError):
64
_fmt = "URLs differ by more than path: %(from_)r and %(to)r"
66
def __init__(self, from_, to):
69
errors.PathError.__init__(self, from_, 'URLs differ by more than path.')
72
def basename(url, exclude_trailing_slash=True):
73
"""Return the last component of a URL.
75
:param url: The URL in question
76
:param exclude_trailing_slash: If the url looks like "path/to/foo/"
77
ignore the final slash and return 'foo' rather than ''
78
:return: Just the final component of the URL. This can return ''
79
if you don't exclude_trailing_slash, or if you are at the
82
return split(url, exclude_trailing_slash=exclude_trailing_slash)[1]
85
def dirname(url, exclude_trailing_slash=True):
86
"""Return the parent directory of the given path.
88
:param url: Relative or absolute URL
89
:param exclude_trailing_slash: Remove a final slash
90
(treat http://host/foo/ as http://host/foo, but
91
http://host/ stays http://host/)
92
:return: Everything in the URL except the last path chunk
94
# TODO: jam 20060502 This was named dirname to be consistent
95
# with the os functions, but maybe "parent" would be better
96
return split(url, exclude_trailing_slash=exclude_trailing_slash)[0]
99
# Private copies of quote and unquote, copied from Python's
100
# urllib module because urllib unconditionally imports socket, which imports
103
always_safe = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
104
'abcdefghijklmnopqrstuvwxyz'
107
for i, c in zip(range(256), ''.join(map(chr, range(256)))):
108
_safe_map[c] = c if (i < 128 and c in always_safe) else '%{0:02X}'.format(i)
112
def quote(s, safe='/'):
113
"""quote('abc def') -> 'abc%20def'
115
Each part of a URL, e.g. the path info, the query, etc., has a
116
different set of reserved characters that must be quoted.
118
RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists
119
the following reserved characters.
121
reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
124
Each of these characters is reserved in some component of a URL,
125
but not necessarily in all of them.
127
By default, the quote function is intended for quoting the path
128
section of a URL. Thus, it will not encode '/'. This character
129
is reserved, but in typical usage the quote function is being
130
called on a path where the existing slash characters are used as
136
raise TypeError('None object cannot be quoted')
138
cachekey = (safe, always_safe)
140
(quoter, safe) = _safe_quoters[cachekey]
142
safe_map = _safe_map.copy()
143
safe_map.update([(c, c) for c in safe])
144
quoter = safe_map.__getitem__
145
safe = always_safe + safe
146
_safe_quoters[cachekey] = (quoter, safe)
147
if not s.rstrip(safe):
149
return ''.join(map(quoter, s))
152
unquote = urlparse.unquote
156
"""Escape relpath to be a valid url."""
157
if not isinstance(relpath, str):
158
relpath = relpath.encode('utf-8')
159
return quote(relpath, safe='/~')
162
def file_relpath(base, path):
163
"""Compute just the relative sub-portion of a url
165
This assumes that both paths are already fully specified file:// URLs.
167
if len(base) < MIN_ABS_FILEURL_LENGTH:
168
raise ValueError('Length of base (%r) must equal or'
169
' exceed the platform minimum url length (which is %d)' %
170
(base, MIN_ABS_FILEURL_LENGTH))
171
base = osutils.normpath(local_path_from_url(base))
172
path = osutils.normpath(local_path_from_url(path))
173
return escape(osutils.relpath(base, path))
176
def _find_scheme_and_separator(url):
177
"""Find the scheme separator (://) and the first path separator
179
This is just a helper functions for other path utilities.
180
It could probably be replaced by urlparse
182
m = _url_scheme_re.match(url)
186
scheme = m.group('scheme')
187
path = m.group('path')
189
# Find the path separating slash
190
# (first slash after the ://)
191
first_path_slash = path.find('/')
192
if first_path_slash == -1:
193
return len(scheme), None
194
return len(scheme), first_path_slash+m.start('path')
198
"""Tests whether a URL is in actual fact a URL."""
199
return _url_scheme_re.match(url) is not None
202
def join(base, *args):
203
"""Create a URL by joining sections.
205
This will normalize '..', assuming that paths are absolute
206
(it assumes no symlinks in either path)
208
If any of *args is an absolute URL, it will be treated correctly.
210
join('http://foo', 'http://bar') => 'http://bar'
211
join('http://foo', 'bar') => 'http://foo/bar'
212
join('http://foo', 'bar', '../baz') => 'http://foo/baz'
216
scheme_end, path_start = _find_scheme_and_separator(base)
217
if scheme_end is None and path_start is None:
219
elif path_start is None:
220
path_start = len(base)
221
path = base[path_start:]
223
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
224
if arg_scheme_end is None and arg_path_start is None:
226
elif arg_path_start is None:
227
arg_path_start = len(arg)
228
if arg_scheme_end is not None:
230
path = arg[arg_path_start:]
231
scheme_end = arg_scheme_end
232
path_start = arg_path_start
234
path = joinpath(path, arg)
235
return base[:path_start] + path
238
def joinpath(base, *args):
239
"""Join URL path segments to a URL path segment.
241
This is somewhat like osutils.joinpath, but intended for URLs.
243
XXX: this duplicates some normalisation logic, and also duplicates a lot of
244
path handling logic that already exists in some Transport implementations.
245
We really should try to have exactly one place in the code base responsible
246
for combining paths of URLs.
248
path = base.split('/')
249
if len(path) > 1 and path[-1] == '':
250
#If the path ends in a trailing /, remove it.
253
if arg.startswith('/'):
255
for chunk in arg.split('/'):
260
raise InvalidURLJoin('Cannot go above root',
268
return '/'.join(path)
271
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
272
def _posix_local_path_from_url(url):
273
"""Convert a url like file:///path/to/foo into /path/to/foo"""
274
url = split_segment_parameters_raw(url)[0]
275
file_localhost_prefix = 'file://localhost/'
276
if url.startswith(file_localhost_prefix):
277
path = url[len(file_localhost_prefix) - 1:]
278
elif not url.startswith('file:///'):
280
url, 'local urls must start with file:/// or file://localhost/')
282
path = url[len('file://'):]
283
# We only strip off 2 slashes
284
return unescape(path)
287
def _posix_local_path_to_url(path):
288
"""Convert a local path like ./foo into a URL like file:///path/to/foo
290
This also handles transforming escaping unicode characters, etc.
292
# importing directly from posixpath allows us to test this
293
# on non-posix platforms
294
return 'file://' + escape(osutils._posix_abspath(path))
297
def _win32_local_path_from_url(url):
298
"""Convert a url like file:///C:/path/to/foo into C:/path/to/foo"""
299
if not url.startswith('file://'):
300
raise InvalidURL(url, 'local urls must start with file:///, '
301
'UNC path urls must start with file://')
302
url = split_segment_parameters_raw(url)[0]
303
# We strip off all 3 slashes
304
win32_url = url[len('file:'):]
305
# check for UNC path: //HOST/path
306
if not win32_url.startswith('///'):
307
if (win32_url[2] == '/'
308
or win32_url[3] in '|:'):
309
raise InvalidURL(url, 'Win32 UNC path urls'
310
' have form file://HOST/path')
311
return unescape(win32_url)
313
# allow empty paths so we can serve all roots
314
if win32_url == '///':
317
# usual local path with drive letter
318
if (len(win32_url) < 6
319
or win32_url[3] not in ('abcdefghijklmnopqrstuvwxyz'
320
'ABCDEFGHIJKLMNOPQRSTUVWXYZ')
321
or win32_url[4] not in '|:'
322
or win32_url[5] != '/'):
323
raise InvalidURL(url, 'Win32 file urls start with'
324
' file:///x:/, where x is a valid drive letter')
325
return win32_url[3].upper() + u':' + unescape(win32_url[5:])
328
def _win32_local_path_to_url(path):
329
"""Convert a local path like ./foo into a URL like file:///C:/path/to/foo
331
This also handles transforming escaping unicode characters, etc.
333
# importing directly from ntpath allows us to test this
334
# on non-win32 platform
335
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
336
# which actually strips trailing space characters.
337
# The worst part is that on linux ntpath.abspath has different
338
# semantics, since 'nt' is not an available module.
342
win32_path = osutils._win32_abspath(path)
343
# check for UNC path \\HOST\path
344
if win32_path.startswith('//'):
345
return 'file:' + escape(win32_path)
346
return ('file:///' + str(win32_path[0].upper()) + ':' +
347
escape(win32_path[2:]))
350
local_path_to_url = _posix_local_path_to_url
351
local_path_from_url = _posix_local_path_from_url
352
MIN_ABS_FILEURL_LENGTH = len('file:///')
353
WIN32_MIN_ABS_FILEURL_LENGTH = len('file:///C:/')
355
if sys.platform == 'win32':
356
local_path_to_url = _win32_local_path_to_url
357
local_path_from_url = _win32_local_path_from_url
359
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH
362
_url_scheme_re = re.compile('^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$')
363
_url_hex_escapes_re = re.compile('(%[0-9a-fA-F]{2})')
366
def _unescape_safe_chars(matchobj):
367
"""re.sub callback to convert hex-escapes to plain characters (if safe).
369
e.g. '%7E' will be converted to '~'.
371
hex_digits = matchobj.group(0)[1:]
372
char = chr(int(hex_digits, 16))
373
if char in _url_dont_escape_characters:
376
return matchobj.group(0).upper()
379
def normalize_url(url):
380
"""Make sure that a path string is in fully normalized URL form.
382
This handles URLs which have unicode characters, spaces,
383
special characters, etc.
385
It has two basic modes of operation, depending on whether the
386
supplied string starts with a url specifier (scheme://) or not.
387
If it does not have a specifier it is considered a local path,
388
and will be converted into a file:/// url. Non-ascii characters
389
will be encoded using utf-8.
390
If it does have a url specifier, it will be treated as a "hybrid"
391
URL. Basically, a URL that should have URL special characters already
392
escaped (like +?&# etc), but may have unicode characters, etc
393
which would not be valid in a real URL.
395
:param url: Either a hybrid URL or a local path
396
:return: A normalized URL which only includes 7-bit ASCII characters.
398
scheme_end, path_start = _find_scheme_and_separator(url)
399
if scheme_end is None:
400
return local_path_to_url(url)
401
prefix = url[:path_start]
402
path = url[path_start:]
403
if not isinstance(url, text_type):
405
if c not in _url_safe_characters:
406
raise InvalidURL(url, 'URLs can only contain specific'
407
' safe characters (not %r)' % c)
408
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
409
return str(prefix + ''.join(path))
411
# We have a unicode (hybrid) url
412
path_chars = list(path)
414
for i in range(len(path_chars)):
415
if path_chars[i] not in _url_safe_characters:
416
chars = path_chars[i].encode('utf-8')
417
path_chars[i] = ''.join(
418
['%%%02X' % ord(c) for c in path_chars[i].encode('utf-8')])
419
path = ''.join(path_chars)
420
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
421
return str(prefix + path)
424
def relative_url(base, other):
425
"""Return a path to other from base.
427
If other is unrelated to base, return other. Else return a relative path.
428
This assumes no symlinks as part of the url.
430
dummy, base_first_slash = _find_scheme_and_separator(base)
431
if base_first_slash is None:
434
dummy, other_first_slash = _find_scheme_and_separator(other)
435
if other_first_slash is None:
438
# this takes care of differing schemes or hosts
439
base_scheme = base[:base_first_slash]
440
other_scheme = other[:other_first_slash]
441
if base_scheme != other_scheme:
443
elif sys.platform == 'win32' and base_scheme == 'file://':
444
base_drive = base[base_first_slash+1:base_first_slash+3]
445
other_drive = other[other_first_slash+1:other_first_slash+3]
446
if base_drive != other_drive:
449
base_path = base[base_first_slash+1:]
450
other_path = other[other_first_slash+1:]
452
if base_path.endswith('/'):
453
base_path = base_path[:-1]
455
base_sections = base_path.split('/')
456
other_sections = other_path.split('/')
458
if base_sections == ['']:
460
if other_sections == ['']:
464
for b, o in zip(base_sections, other_sections):
467
output_sections.append(b)
469
match_len = len(output_sections)
470
output_sections = ['..' for x in base_sections[match_len:]]
471
output_sections.extend(other_sections[match_len:])
473
return "/".join(output_sections) or "."
476
def _win32_extract_drive_letter(url_base, path):
477
"""On win32 the drive letter needs to be added to the url base."""
478
# Strip off the drive letter
479
# path is currently /C:/foo
480
if len(path) < 4 or path[2] not in ':|' or path[3] != '/':
481
raise InvalidURL(url_base + path,
482
'win32 file:/// paths need a drive letter')
483
url_base += path[0:3] # file:// + /C:
484
path = path[3:] # /foo
485
return url_base, path
488
def split(url, exclude_trailing_slash=True):
489
"""Split a URL into its parent directory and a child directory.
491
:param url: A relative or absolute URL
492
:param exclude_trailing_slash: Strip off a final '/' if it is part
493
of the path (but not if it is part of the protocol specification)
495
:return: (parent_url, child_dir). child_dir may be the empty string if we're at
498
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
500
if first_path_slash is None:
501
# We have either a relative path, or no separating slash
502
if scheme_loc is None:
504
if exclude_trailing_slash and url.endswith('/'):
506
return _posix_split(url)
508
# Scheme with no path
511
# We have a fully defined path
512
url_base = url[:first_path_slash] # http://host, file://
513
path = url[first_path_slash:] # /file/foo
515
if sys.platform == 'win32' and url.startswith('file:///'):
516
# Strip off the drive letter
517
# url_base is currently file://
518
# path is currently /C:/foo
519
url_base, path = _win32_extract_drive_letter(url_base, path)
520
# now it should be file:///C: and /foo
522
if exclude_trailing_slash and len(path) > 1 and path.endswith('/'):
524
head, tail = _posix_split(path)
525
return url_base + head, tail
528
def split_segment_parameters_raw(url):
529
"""Split the subsegment of the last segment of a URL.
531
:param url: A relative or absolute URL
532
:return: (url, subsegments)
534
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
535
# operates on urls not url+segments, and Transport classes
536
# should not be blindly adding slashes in the first place.
537
lurl = strip_trailing_slash(url)
538
# Segments begin at first comma after last forward slash, if one exists
539
segment_start = lurl.find(",", lurl.rfind("/")+1)
540
if segment_start == -1:
542
return (lurl[:segment_start], [str(s) for s in lurl[segment_start+1:].split(",")])
545
def split_segment_parameters(url):
546
"""Split the segment parameters of the last segment of a URL.
548
:param url: A relative or absolute URL
549
:return: (url, segment_parameters)
551
(base_url, subsegments) = split_segment_parameters_raw(url)
553
for subsegment in subsegments:
554
(key, value) = subsegment.split("=", 1)
555
if not isinstance(key, str):
557
if not isinstance(value, str):
558
raise TypeError(value)
559
parameters[key] = value
560
return (base_url, parameters)
563
def join_segment_parameters_raw(base, *subsegments):
564
"""Create a new URL by adding subsegments to an existing one.
566
This adds the specified subsegments to the last path in the specified
567
base URL. The subsegments should be bytestrings.
569
:note: You probably want to use join_segment_parameters instead.
573
for subsegment in subsegments:
574
if not isinstance(subsegment, str):
575
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
576
if "," in subsegment:
577
raise InvalidURLJoin(", exists in subsegments",
579
return ",".join((base,) + subsegments)
582
def join_segment_parameters(url, parameters):
583
"""Create a new URL by adding segment parameters to an existing one.
585
The parameters of the last segment in the URL will be updated; if a
586
parameter with the same key already exists it will be overwritten.
588
:param url: A URL, as string
589
:param parameters: Dictionary of parameters, keys and values as bytestrings
591
(base, existing_parameters) = split_segment_parameters(url)
593
new_parameters.update(existing_parameters)
594
for key, value in parameters.items():
595
if not isinstance(key, str):
596
raise TypeError("parameter key %r is not a bytestring" % key)
597
if not isinstance(value, str):
598
raise TypeError("parameter value %r for %s is not a bytestring" %
601
raise InvalidURLJoin("= exists in parameter key", url,
603
new_parameters[key] = value
604
return join_segment_parameters_raw(base,
605
*["%s=%s" % item for item in sorted(new_parameters.items())])
608
def _win32_strip_local_trailing_slash(url):
609
"""Strip slashes after the drive letter"""
610
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
616
def strip_trailing_slash(url):
617
"""Strip trailing slash, except for root paths.
619
The definition of 'root path' is platform-dependent.
620
This assumes that all URLs are valid netloc urls, such that they
623
It searches for ://, and then refuses to remove the next '/'.
624
It can also handle relative paths
626
path/to/foo => path/to/foo
627
path/to/foo/ => path/to/foo
628
http://host/path/ => http://host/path
629
http://host/path => http://host/path
630
http://host/ => http://host/
632
file:///foo/ => file:///foo
633
# This is unique on win32 platforms, and is the only URL
634
# format which does it differently.
635
file:///c|/ => file:///c:/
637
if not url.endswith('/'):
640
if sys.platform == 'win32' and url.startswith('file://'):
641
return _win32_strip_local_trailing_slash(url)
643
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
644
if scheme_loc is None:
645
# This is a relative path, as it has no scheme
646
# so just chop off the last character
649
if first_path_slash is None or first_path_slash == len(url)-1:
650
# Don't chop off anything if the only slash is the path
658
"""Unescape relpath from url format.
660
This returns a Unicode path from a URL
662
# jam 20060427 URLs are supposed to be ASCII only strings
663
# If they are passed in as unicode, unquote
664
# will return a UNICODE string, which actually contains
665
# utf-8 bytes. So we have to ensure that they are
666
# plain ASCII strings, or the final .decode will
667
# try to encode the UNICODE => ASCII, and then decode
669
if isinstance(url, text_type):
671
url = url.encode("ascii")
672
except UnicodeError as e:
673
raise InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,))
675
unquoted = urlparse.unquote_to_bytes(url)
677
unquoted = unquote(url)
679
unicode_path = unquoted.decode('utf-8')
680
except UnicodeError as e:
681
raise InvalidURL(url, 'Unable to encode the URL as utf-8: %s' % (e,))
685
# These are characters that if escaped, should stay that way
686
_no_decode_chars = ';/?:@&=+$,#'
687
_no_decode_ords = [ord(c) for c in _no_decode_chars]
688
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
689
+ ['%02X' % o for o in _no_decode_ords])
690
_hex_display_map = dict(([('%02x' % o, chr(o)) for o in range(256)]
691
+ [('%02X' % o, chr(o)) for o in range(256)]))
692
#These entries get mapped to themselves
693
_hex_display_map.update((hex, '%'+hex) for hex in _no_decode_hex)
695
# These characters shouldn't be percent-encoded, and it's always safe to
696
# unencode them if they are.
697
_url_dont_escape_characters = set(
698
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
699
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
700
"0123456789" # Numbers
701
"-._~" # Unreserved characters
704
# These characters should not be escaped
705
_url_safe_characters = set(
706
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
707
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
708
"0123456789" # Numbers
709
"_.-!~*'()" # Unreserved characters
710
"/;?:@&=+$," # Reserved characters
711
"%#" # Extra reserved characters
714
def unescape_for_display(url, encoding):
715
"""Decode what you can for a URL, so that we get a nice looking path.
717
This will turn file:// urls into local paths, and try to decode
718
any portions of a http:// style url that it can.
720
Any sections of the URL which can't be represented in the encoding or
721
need to stay as escapes are left alone.
723
:param url: A 7-bit ASCII URL
724
:param encoding: The final output encoding
726
:return: A unicode string which can be safely encoded into the
730
raise ValueError('you cannot specify None for the display encoding')
731
if url.startswith('file://'):
733
path = local_path_from_url(url)
734
path.encode(encoding)
739
# Split into sections to try to decode utf-8
741
for i in range(1, len(res)):
742
escaped_chunks = res[i].split('%')
743
for j in range(1, len(escaped_chunks)):
744
item = escaped_chunks[j]
746
escaped_chunks[j] = _hex_display_map[item[:2]] + item[2:]
748
# Put back the percent symbol
749
escaped_chunks[j] = '%' + item
750
except UnicodeDecodeError:
751
escaped_chunks[j] = unichr(int(item[:2], 16)) + item[2:]
752
unescaped = ''.join(escaped_chunks)
753
if sys.version_info[0] == 2:
755
decoded = unescaped.decode('utf-8')
756
except UnicodeDecodeError:
757
# If this path segment cannot be properly utf-8 decoded
758
# after doing unescaping we will just leave it alone
762
decoded.encode(encoding)
763
except UnicodeEncodeError:
764
# If this chunk cannot be encoded in the local
765
# encoding, then we should leave it alone
768
# Otherwise take the url decoded one
772
return u'/'.join(res)
775
def derive_to_location(from_location):
776
"""Derive a TO_LOCATION given a FROM_LOCATION.
778
The normal case is a FROM_LOCATION of http://foo/bar => bar.
779
The Right Thing for some logical destinations may differ though
780
because no / may be present at all. In that case, the result is
781
the full name without the scheme indicator, e.g. lp:foo-bar => foo-bar.
782
This latter case also applies when a Windows drive
783
is used without a path, e.g. c:foo-bar => foo-bar.
784
If no /, path separator or : is found, the from_location is returned.
786
if from_location.find("/") >= 0 or from_location.find(os.sep) >= 0:
787
return os.path.basename(from_location.rstrip("/\\"))
789
sep = from_location.find(":")
791
return from_location[sep+1:]
796
def _is_absolute(url):
797
return (osutils.pathjoin('/foo', url) == url)
800
def rebase_url(url, old_base, new_base):
801
"""Convert a relative path from an old base URL to a new base URL.
803
The result will be a relative path.
804
Absolute paths and full URLs are returned unaltered.
806
scheme, separator = _find_scheme_and_separator(url)
807
if scheme is not None:
809
if _is_absolute(url):
811
old_parsed = urlparse.urlparse(old_base)
812
new_parsed = urlparse.urlparse(new_base)
813
if (old_parsed[:2]) != (new_parsed[:2]):
814
raise InvalidRebaseURLs(old_base, new_base)
815
return determine_relative_path(new_parsed[2],
816
join(old_parsed[2], url))
819
def determine_relative_path(from_path, to_path):
820
"""Determine a relative path from from_path to to_path."""
821
from_segments = osutils.splitpath(from_path)
822
to_segments = osutils.splitpath(to_path)
824
for count, (from_element, to_element) in enumerate(zip(from_segments,
826
if from_element != to_element:
830
unique_from = from_segments[count:]
831
unique_to = to_segments[count:]
832
segments = (['..'] * len(unique_from) + unique_to)
833
if len(segments) == 0:
835
return osutils.pathjoin(*segments)
841
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
844
self.quoted_host = quoted_host
845
self.host = unquote(self.quoted_host)
846
self.quoted_user = quoted_user
847
if self.quoted_user is not None:
848
self.user = unquote(self.quoted_user)
851
self.quoted_password = quoted_password
852
if self.quoted_password is not None:
853
self.password = unquote(self.quoted_password)
857
self.quoted_path = _url_hex_escapes_re.sub(_unescape_safe_chars, quoted_path)
858
self.path = unquote(self.quoted_path)
860
def __eq__(self, other):
861
return (isinstance(other, self.__class__) and
862
self.scheme == other.scheme and
863
self.host == other.host and
864
self.user == other.user and
865
self.password == other.password and
866
self.path == other.path)
869
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
870
self.__class__.__name__,
871
self.scheme, self.quoted_user, self.quoted_password,
872
self.quoted_host, self.port, self.quoted_path)
875
def from_string(cls, url):
876
"""Create a URL object from a string.
878
:param url: URL as bytestring
880
# GZ 2017-06-09: Actually validate ascii-ness
881
if not isinstance(url, str):
882
raise InvalidURL('should be ascii:\n%r' % url)
883
(scheme, netloc, path, params,
884
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
885
user = password = host = port = None
887
user, host = netloc.rsplit('@', 1)
889
user, password = user.split(':', 1)
893
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
895
host, port = host.rsplit(':', 1)
899
raise InvalidURL('invalid port number %s in url:\n%s' %
901
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
904
return cls(scheme, user, password, host, port, path)
907
netloc = self.quoted_host
909
netloc = "[%s]" % netloc
910
if self.quoted_user is not None:
911
# Note that we don't put the password back even if we
912
# have one so that it doesn't get accidentally
914
netloc = '%s@%s' % (self.quoted_user, netloc)
915
if self.port is not None:
916
netloc = '%s:%d' % (netloc, self.port)
917
return urlparse.urlunparse(
918
(self.scheme, netloc, self.quoted_path, None, None, None))
921
def _combine_paths(base_path, relpath):
922
"""Transform a Transport-relative path to a remote absolute path.
924
This does not handle substitution of ~ but does handle '..' and '.'
929
t._combine_paths('/home/sarah', 'project/foo')
930
=> '/home/sarah/project/foo'
931
t._combine_paths('/home/sarah', '../../etc')
933
t._combine_paths('/home/sarah', '/etc')
936
:param base_path: base path
937
:param relpath: relative url string for relative part of remote path.
938
:return: urlencoded string for final path.
940
if not isinstance(relpath, str):
941
raise InvalidURL(relpath)
942
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
943
if relpath.startswith('/'):
946
base_parts = base_path.split('/')
947
if len(base_parts) > 0 and base_parts[-1] == '':
948
base_parts = base_parts[:-1]
949
for p in relpath.split('/'):
951
if len(base_parts) == 0:
952
# In most filesystems, a request for the parent
953
# of root, just returns root.
960
path = '/'.join(base_parts)
961
if not path.startswith('/'):
965
def clone(self, offset=None):
966
"""Return a new URL for a path relative to this URL.
968
:param offset: A relative path, already urlencoded
969
:return: `URL` instance
971
if offset is not None:
972
relative = unescape(offset)
973
if sys.version_info[0] == 2:
974
relative = relative.encode('utf-8')
975
path = self._combine_paths(self.path, relative)
976
path = quote(path, safe="/~")
978
path = self.quoted_path
979
return self.__class__(self.scheme, self.quoted_user,
980
self.quoted_password, self.quoted_host, self.port,
985
"""Extract the server address, the credentials and the path from the url.
987
user, password, host and path should be quoted if they contain reserved
990
:param url: an quoted url
991
:return: (scheme, user, password, host, port, path) tuple, all fields
994
parsed_url = URL.from_string(url)
995
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
996
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
breezy/urlutils.py
