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
25
from bzrlib.lazy_import import lazy_import
26
lazy_import(globals(), """
27
from posixpath import split as _posix_split
37
def basename(url, exclude_trailing_slash=True):
38
"""Return the last component of a URL.
40
:param url: The URL in question
41
:param exclude_trailing_slash: If the url looks like "path/to/foo/"
42
ignore the final slash and return 'foo' rather than ''
43
:return: Just the final component of the URL. This can return ''
44
if you don't exclude_trailing_slash, or if you are at the
47
return split(url, exclude_trailing_slash=exclude_trailing_slash)[1]
50
def dirname(url, exclude_trailing_slash=True):
51
"""Return the parent directory of the given path.
53
:param url: Relative or absolute URL
54
:param exclude_trailing_slash: Remove a final slash
55
(treat http://host/foo/ as http://host/foo, but
56
http://host/ stays http://host/)
57
:return: Everything in the URL except the last path chunk
59
# TODO: jam 20060502 This was named dirname to be consistent
60
# with the os functions, but maybe "parent" would be better
61
return split(url, exclude_trailing_slash=exclude_trailing_slash)[0]
64
# Private copies of quote and unquote, copied from Python's
65
# urllib module because urllib unconditionally imports socket, which imports
68
always_safe = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
69
'abcdefghijklmnopqrstuvwxyz'
72
for i, c in zip(xrange(256), str(bytearray(xrange(256)))):
73
_safe_map[c] = c if (i < 128 and c in always_safe) else '%{0:02X}'.format(i)
77
def quote(s, safe='/'):
78
"""quote('abc def') -> 'abc%20def'
80
Each part of a URL, e.g. the path info, the query, etc., has a
81
different set of reserved characters that must be quoted.
83
RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists
84
the following reserved characters.
86
reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
89
Each of these characters is reserved in some component of a URL,
90
but not necessarily in all of them.
92
By default, the quote function is intended for quoting the path
93
section of a URL. Thus, it will not encode '/'. This character
94
is reserved, but in typical usage the quote function is being
95
called on a path where the existing slash characters are used as
101
raise TypeError('None object cannot be quoted')
103
cachekey = (safe, always_safe)
105
(quoter, safe) = _safe_quoters[cachekey]
107
safe_map = _safe_map.copy()
108
safe_map.update([(c, c) for c in safe])
109
quoter = safe_map.__getitem__
110
safe = always_safe + safe
111
_safe_quoters[cachekey] = (quoter, safe)
112
if not s.rstrip(safe):
114
return ''.join(map(quoter, s))
117
_hexdig = '0123456789ABCDEFabcdef'
118
_hextochr = dict((a + b, chr(int(a + b, 16)))
119
for a in _hexdig for b in _hexdig)
122
"""unquote('abc%20def') -> 'abc def'."""
130
s += _hextochr[item[:2]] + item[2:]
133
except UnicodeDecodeError:
134
s += unichr(int(item[:2], 16)) + item[2:]
139
"""Escape relpath to be a valid url."""
140
if isinstance(relpath, unicode):
141
relpath = relpath.encode('utf-8')
142
# After quoting and encoding, the path should be perfectly
143
# safe as a plain ASCII string, str() just enforces this
144
return str(quote(relpath, safe='/~'))
147
def file_relpath(base, path):
148
"""Compute just the relative sub-portion of a url
150
This assumes that both paths are already fully specified file:// URLs.
152
if len(base) < MIN_ABS_FILEURL_LENGTH:
153
raise ValueError('Length of base (%r) must equal or'
154
' exceed the platform minimum url length (which is %d)' %
155
(base, MIN_ABS_FILEURL_LENGTH))
156
base = osutils.normpath(local_path_from_url(base))
157
path = osutils.normpath(local_path_from_url(path))
158
return escape(osutils.relpath(base, path))
161
def _find_scheme_and_separator(url):
162
"""Find the scheme separator (://) and the first path separator
164
This is just a helper functions for other path utilities.
165
It could probably be replaced by urlparse
167
m = _url_scheme_re.match(url)
171
scheme = m.group('scheme')
172
path = m.group('path')
174
# Find the path separating slash
175
# (first slash after the ://)
176
first_path_slash = path.find('/')
177
if first_path_slash == -1:
178
return len(scheme), None
179
return len(scheme), first_path_slash+m.start('path')
183
"""Tests whether a URL is in actual fact a URL."""
184
return _url_scheme_re.match(url) is not None
187
def join(base, *args):
188
"""Create a URL by joining sections.
190
This will normalize '..', assuming that paths are absolute
191
(it assumes no symlinks in either path)
193
If any of *args is an absolute URL, it will be treated correctly.
195
join('http://foo', 'http://bar') => 'http://bar'
196
join('http://foo', 'bar') => 'http://foo/bar'
197
join('http://foo', 'bar', '../baz') => 'http://foo/baz'
201
scheme_end, path_start = _find_scheme_and_separator(base)
202
if scheme_end is None and path_start is None:
204
elif path_start is None:
205
path_start = len(base)
206
path = base[path_start:]
208
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
209
if arg_scheme_end is None and arg_path_start is None:
211
elif arg_path_start is None:
212
arg_path_start = len(arg)
213
if arg_scheme_end is not None:
215
path = arg[arg_path_start:]
216
scheme_end = arg_scheme_end
217
path_start = arg_path_start
219
path = joinpath(path, arg)
220
return base[:path_start] + path
223
def joinpath(base, *args):
224
"""Join URL path segments to a URL path segment.
226
This is somewhat like osutils.joinpath, but intended for URLs.
228
XXX: this duplicates some normalisation logic, and also duplicates a lot of
229
path handling logic that already exists in some Transport implementations.
230
We really should try to have exactly one place in the code base responsible
231
for combining paths of URLs.
233
path = base.split('/')
234
if len(path) > 1 and path[-1] == '':
235
#If the path ends in a trailing /, remove it.
238
if arg.startswith('/'):
240
for chunk in arg.split('/'):
245
raise errors.InvalidURLJoin('Cannot go above root',
253
return '/'.join(path)
256
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
257
def _posix_local_path_from_url(url):
258
"""Convert a url like file:///path/to/foo into /path/to/foo"""
259
url = split_segment_parameters_raw(url)[0]
260
file_localhost_prefix = 'file://localhost/'
261
if url.startswith(file_localhost_prefix):
262
path = url[len(file_localhost_prefix) - 1:]
263
elif not url.startswith('file:///'):
264
raise errors.InvalidURL(
265
url, 'local urls must start with file:/// or file://localhost/')
267
path = url[len('file://'):]
268
# We only strip off 2 slashes
269
return unescape(path)
272
def _posix_local_path_to_url(path):
273
"""Convert a local path like ./foo into a URL like file:///path/to/foo
275
This also handles transforming escaping unicode characters, etc.
277
# importing directly from posixpath allows us to test this
278
# on non-posix platforms
279
return 'file://' + escape(osutils._posix_abspath(path))
282
def _win32_local_path_from_url(url):
283
"""Convert a url like file:///C:/path/to/foo into C:/path/to/foo"""
284
if not url.startswith('file://'):
285
raise errors.InvalidURL(url, 'local urls must start with file:///, '
286
'UNC path urls must start with file://')
287
url = split_segment_parameters_raw(url)[0]
288
# We strip off all 3 slashes
289
win32_url = url[len('file:'):]
290
# check for UNC path: //HOST/path
291
if not win32_url.startswith('///'):
292
if (win32_url[2] == '/'
293
or win32_url[3] in '|:'):
294
raise errors.InvalidURL(url, 'Win32 UNC path urls'
295
' have form file://HOST/path')
296
return unescape(win32_url)
298
# allow empty paths so we can serve all roots
299
if win32_url == '///':
302
# usual local path with drive letter
303
if (len(win32_url) < 6
304
or win32_url[3] not in ('abcdefghijklmnopqrstuvwxyz'
305
'ABCDEFGHIJKLMNOPQRSTUVWXYZ')
306
or win32_url[4] not in '|:'
307
or win32_url[5] != '/'):
308
raise errors.InvalidURL(url, 'Win32 file urls start with'
309
' file:///x:/, where x is a valid drive letter')
310
return win32_url[3].upper() + u':' + unescape(win32_url[5:])
313
def _win32_local_path_to_url(path):
314
"""Convert a local path like ./foo into a URL like file:///C:/path/to/foo
316
This also handles transforming escaping unicode characters, etc.
318
# importing directly from ntpath allows us to test this
319
# on non-win32 platform
320
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
321
# which actually strips trailing space characters.
322
# The worst part is that on linux ntpath.abspath has different
323
# semantics, since 'nt' is not an available module.
327
win32_path = osutils._win32_abspath(path)
328
# check for UNC path \\HOST\path
329
if win32_path.startswith('//'):
330
return 'file:' + escape(win32_path)
331
return ('file:///' + str(win32_path[0].upper()) + ':' +
332
escape(win32_path[2:]))
335
local_path_to_url = _posix_local_path_to_url
336
local_path_from_url = _posix_local_path_from_url
337
MIN_ABS_FILEURL_LENGTH = len('file:///')
338
WIN32_MIN_ABS_FILEURL_LENGTH = len('file:///C:/')
340
if sys.platform == 'win32':
341
local_path_to_url = _win32_local_path_to_url
342
local_path_from_url = _win32_local_path_from_url
344
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH
347
_url_scheme_re = re.compile(r'^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$')
348
_url_hex_escapes_re = re.compile(r'(%[0-9a-fA-F]{2})')
351
def _unescape_safe_chars(matchobj):
352
"""re.sub callback to convert hex-escapes to plain characters (if safe).
354
e.g. '%7E' will be converted to '~'.
356
hex_digits = matchobj.group(0)[1:]
357
char = chr(int(hex_digits, 16))
358
if char in _url_dont_escape_characters:
361
return matchobj.group(0).upper()
364
def normalize_url(url):
365
"""Make sure that a path string is in fully normalized URL form.
367
This handles URLs which have unicode characters, spaces,
368
special characters, etc.
370
It has two basic modes of operation, depending on whether the
371
supplied string starts with a url specifier (scheme://) or not.
372
If it does not have a specifier it is considered a local path,
373
and will be converted into a file:/// url. Non-ascii characters
374
will be encoded using utf-8.
375
If it does have a url specifier, it will be treated as a "hybrid"
376
URL. Basically, a URL that should have URL special characters already
377
escaped (like +?&# etc), but may have unicode characters, etc
378
which would not be valid in a real URL.
380
:param url: Either a hybrid URL or a local path
381
:return: A normalized URL which only includes 7-bit ASCII characters.
383
scheme_end, path_start = _find_scheme_and_separator(url)
384
if scheme_end is None:
385
return local_path_to_url(url)
386
prefix = url[:path_start]
387
path = url[path_start:]
388
if not isinstance(url, unicode):
390
if c not in _url_safe_characters:
391
raise errors.InvalidURL(url, 'URLs can only contain specific'
392
' safe characters (not %r)' % c)
393
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
394
return str(prefix + ''.join(path))
396
# We have a unicode (hybrid) url
397
path_chars = list(path)
399
for i in xrange(len(path_chars)):
400
if path_chars[i] not in _url_safe_characters:
401
chars = path_chars[i].encode('utf-8')
402
path_chars[i] = ''.join(
403
['%%%02X' % ord(c) for c in path_chars[i].encode('utf-8')])
404
path = ''.join(path_chars)
405
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
406
return str(prefix + path)
409
def relative_url(base, other):
410
"""Return a path to other from base.
412
If other is unrelated to base, return other. Else return a relative path.
413
This assumes no symlinks as part of the url.
415
dummy, base_first_slash = _find_scheme_and_separator(base)
416
if base_first_slash is None:
419
dummy, other_first_slash = _find_scheme_and_separator(other)
420
if other_first_slash is None:
423
# this takes care of differing schemes or hosts
424
base_scheme = base[:base_first_slash]
425
other_scheme = other[:other_first_slash]
426
if base_scheme != other_scheme:
428
elif sys.platform == 'win32' and base_scheme == 'file://':
429
base_drive = base[base_first_slash+1:base_first_slash+3]
430
other_drive = other[other_first_slash+1:other_first_slash+3]
431
if base_drive != other_drive:
434
base_path = base[base_first_slash+1:]
435
other_path = other[other_first_slash+1:]
437
if base_path.endswith('/'):
438
base_path = base_path[:-1]
440
base_sections = base_path.split('/')
441
other_sections = other_path.split('/')
443
if base_sections == ['']:
445
if other_sections == ['']:
449
for b, o in zip(base_sections, other_sections):
452
output_sections.append(b)
454
match_len = len(output_sections)
455
output_sections = ['..' for x in base_sections[match_len:]]
456
output_sections.extend(other_sections[match_len:])
458
return "/".join(output_sections) or "."
461
def _win32_extract_drive_letter(url_base, path):
462
"""On win32 the drive letter needs to be added to the url base."""
463
# Strip off the drive letter
464
# path is currently /C:/foo
465
if len(path) < 4 or path[2] not in ':|' or path[3] != '/':
466
raise errors.InvalidURL(url_base + path,
467
'win32 file:/// paths need a drive letter')
468
url_base += path[0:3] # file:// + /C:
469
path = path[3:] # /foo
470
return url_base, path
473
def split(url, exclude_trailing_slash=True):
474
"""Split a URL into its parent directory and a child directory.
476
:param url: A relative or absolute URL
477
:param exclude_trailing_slash: Strip off a final '/' if it is part
478
of the path (but not if it is part of the protocol specification)
480
:return: (parent_url, child_dir). child_dir may be the empty string if we're at
483
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
485
if first_path_slash is None:
486
# We have either a relative path, or no separating slash
487
if scheme_loc is None:
489
if exclude_trailing_slash and url.endswith('/'):
491
return _posix_split(url)
493
# Scheme with no path
496
# We have a fully defined path
497
url_base = url[:first_path_slash] # http://host, file://
498
path = url[first_path_slash:] # /file/foo
500
if sys.platform == 'win32' and url.startswith('file:///'):
501
# Strip off the drive letter
502
# url_base is currently file://
503
# path is currently /C:/foo
504
url_base, path = _win32_extract_drive_letter(url_base, path)
505
# now it should be file:///C: and /foo
507
if exclude_trailing_slash and len(path) > 1 and path.endswith('/'):
509
head, tail = _posix_split(path)
510
return url_base + head, tail
513
def split_segment_parameters_raw(url):
514
"""Split the subsegment of the last segment of a URL.
516
:param url: A relative or absolute URL
517
:return: (url, subsegments)
519
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
520
# operates on urls not url+segments, and Transport classes
521
# should not be blindly adding slashes in the first place.
522
lurl = strip_trailing_slash(url)
523
# Segments begin at first comma after last forward slash, if one exists
524
segment_start = lurl.find(",", lurl.rfind("/")+1)
525
if segment_start == -1:
527
return (lurl[:segment_start], lurl[segment_start+1:].split(","))
530
def split_segment_parameters(url):
531
"""Split the segment parameters of the last segment of a URL.
533
:param url: A relative or absolute URL
534
:return: (url, segment_parameters)
536
(base_url, subsegments) = split_segment_parameters_raw(url)
538
for subsegment in subsegments:
539
(key, value) = subsegment.split("=", 1)
540
parameters[key] = value
541
return (base_url, parameters)
544
def join_segment_parameters_raw(base, *subsegments):
545
"""Create a new URL by adding subsegments to an existing one.
547
This adds the specified subsegments to the last path in the specified
548
base URL. The subsegments should be bytestrings.
550
:note: You probably want to use join_segment_parameters instead.
554
for subsegment in subsegments:
555
if type(subsegment) is not str:
556
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
557
if "," in subsegment:
558
raise errors.InvalidURLJoin(", exists in subsegments",
560
return ",".join((base,) + subsegments)
563
def join_segment_parameters(url, parameters):
564
"""Create a new URL by adding segment parameters to an existing one.
566
The parameters of the last segment in the URL will be updated; if a
567
parameter with the same key already exists it will be overwritten.
569
:param url: A URL, as string
570
:param parameters: Dictionary of parameters, keys and values as bytestrings
572
(base, existing_parameters) = split_segment_parameters(url)
574
new_parameters.update(existing_parameters)
575
for key, value in parameters.iteritems():
576
if type(key) is not str:
577
raise TypeError("parameter key %r is not a bytestring" % key)
578
if type(value) is not str:
579
raise TypeError("parameter value %r for %s is not a bytestring" %
582
raise errors.InvalidURLJoin("= exists in parameter key", url,
584
new_parameters[key] = value
585
return join_segment_parameters_raw(base,
586
*["%s=%s" % item for item in sorted(new_parameters.items())])
589
def _win32_strip_local_trailing_slash(url):
590
"""Strip slashes after the drive letter"""
591
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
597
def strip_trailing_slash(url):
598
"""Strip trailing slash, except for root paths.
600
The definition of 'root path' is platform-dependent.
601
This assumes that all URLs are valid netloc urls, such that they
604
It searches for ://, and then refuses to remove the next '/'.
605
It can also handle relative paths
607
path/to/foo => path/to/foo
608
path/to/foo/ => path/to/foo
609
http://host/path/ => http://host/path
610
http://host/path => http://host/path
611
http://host/ => http://host/
613
file:///foo/ => file:///foo
614
# This is unique on win32 platforms, and is the only URL
615
# format which does it differently.
616
file:///c|/ => file:///c:/
618
if not url.endswith('/'):
621
if sys.platform == 'win32' and url.startswith('file://'):
622
return _win32_strip_local_trailing_slash(url)
624
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
625
if scheme_loc is None:
626
# This is a relative path, as it has no scheme
627
# so just chop off the last character
630
if first_path_slash is None or first_path_slash == len(url)-1:
631
# Don't chop off anything if the only slash is the path
639
"""Unescape relpath from url format.
641
This returns a Unicode path from a URL
643
# jam 20060427 URLs are supposed to be ASCII only strings
644
# If they are passed in as unicode, unquote
645
# will return a UNICODE string, which actually contains
646
# utf-8 bytes. So we have to ensure that they are
647
# plain ASCII strings, or the final .decode will
648
# try to encode the UNICODE => ASCII, and then decode
652
except UnicodeError, e:
653
raise errors.InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,))
655
unquoted = unquote(url)
657
unicode_path = unquoted.decode('utf-8')
658
except UnicodeError, e:
659
raise errors.InvalidURL(url, 'Unable to encode the URL as utf-8: %s' % (e,))
663
# These are characters that if escaped, should stay that way
664
_no_decode_chars = ';/?:@&=+$,#'
665
_no_decode_ords = [ord(c) for c in _no_decode_chars]
666
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
667
+ ['%02X' % o for o in _no_decode_ords])
668
_hex_display_map = dict(([('%02x' % o, chr(o)) for o in range(256)]
669
+ [('%02X' % o, chr(o)) for o in range(256)]))
670
#These entries get mapped to themselves
671
_hex_display_map.update((hex,'%'+hex) for hex in _no_decode_hex)
673
# These characters shouldn't be percent-encoded, and it's always safe to
674
# unencode them if they are.
675
_url_dont_escape_characters = set(
676
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
677
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
678
"0123456789" # Numbers
679
"-._~" # Unreserved characters
682
# These characters should not be escaped
683
_url_safe_characters = set(
684
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
685
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
686
"0123456789" # Numbers
687
"_.-!~*'()" # Unreserved characters
688
"/;?:@&=+$," # Reserved characters
689
"%#" # Extra reserved characters
692
def unescape_for_display(url, encoding):
693
"""Decode what you can for a URL, so that we get a nice looking path.
695
This will turn file:// urls into local paths, and try to decode
696
any portions of a http:// style url that it can.
698
Any sections of the URL which can't be represented in the encoding or
699
need to stay as escapes are left alone.
701
:param url: A 7-bit ASCII URL
702
:param encoding: The final output encoding
704
:return: A unicode string which can be safely encoded into the
708
raise ValueError('you cannot specify None for the display encoding')
709
if url.startswith('file://'):
711
path = local_path_from_url(url)
712
path.encode(encoding)
717
# Split into sections to try to decode utf-8
719
for i in xrange(1, len(res)):
720
escaped_chunks = res[i].split('%')
721
for j in xrange(1, len(escaped_chunks)):
722
item = escaped_chunks[j]
724
escaped_chunks[j] = _hex_display_map[item[:2]] + item[2:]
726
# Put back the percent symbol
727
escaped_chunks[j] = '%' + item
728
except UnicodeDecodeError:
729
escaped_chunks[j] = unichr(int(item[:2], 16)) + item[2:]
730
unescaped = ''.join(escaped_chunks)
732
decoded = unescaped.decode('utf-8')
733
except UnicodeDecodeError:
734
# If this path segment cannot be properly utf-8 decoded
735
# after doing unescaping we will just leave it alone
739
decoded.encode(encoding)
740
except UnicodeEncodeError:
741
# If this chunk cannot be encoded in the local
742
# encoding, then we should leave it alone
745
# Otherwise take the url decoded one
747
return u'/'.join(res)
750
def derive_to_location(from_location):
751
"""Derive a TO_LOCATION given a FROM_LOCATION.
753
The normal case is a FROM_LOCATION of http://foo/bar => bar.
754
The Right Thing for some logical destinations may differ though
755
because no / may be present at all. In that case, the result is
756
the full name without the scheme indicator, e.g. lp:foo-bar => foo-bar.
757
This latter case also applies when a Windows drive
758
is used without a path, e.g. c:foo-bar => foo-bar.
759
If no /, path separator or : is found, the from_location is returned.
761
if from_location.find("/") >= 0 or from_location.find(os.sep) >= 0:
762
return os.path.basename(from_location.rstrip("/\\"))
764
sep = from_location.find(":")
766
return from_location[sep+1:]
771
def _is_absolute(url):
772
return (osutils.pathjoin('/foo', url) == url)
775
def rebase_url(url, old_base, new_base):
776
"""Convert a relative path from an old base URL to a new base URL.
778
The result will be a relative path.
779
Absolute paths and full URLs are returned unaltered.
781
scheme, separator = _find_scheme_and_separator(url)
782
if scheme is not None:
784
if _is_absolute(url):
786
old_parsed = urlparse.urlparse(old_base)
787
new_parsed = urlparse.urlparse(new_base)
788
if (old_parsed[:2]) != (new_parsed[:2]):
789
raise errors.InvalidRebaseURLs(old_base, new_base)
790
return determine_relative_path(new_parsed[2],
791
join(old_parsed[2], url))
794
def determine_relative_path(from_path, to_path):
795
"""Determine a relative path from from_path to to_path."""
796
from_segments = osutils.splitpath(from_path)
797
to_segments = osutils.splitpath(to_path)
799
for count, (from_element, to_element) in enumerate(zip(from_segments,
801
if from_element != to_element:
805
unique_from = from_segments[count:]
806
unique_to = to_segments[count:]
807
segments = (['..'] * len(unique_from) + unique_to)
808
if len(segments) == 0:
810
return osutils.pathjoin(*segments)
816
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
819
self.quoted_host = quoted_host
820
self.host = unquote(self.quoted_host)
821
self.quoted_user = quoted_user
822
if self.quoted_user is not None:
823
self.user = unquote(self.quoted_user)
826
self.quoted_password = quoted_password
827
if self.quoted_password is not None:
828
self.password = unquote(self.quoted_password)
832
self.quoted_path = _url_hex_escapes_re.sub(_unescape_safe_chars, quoted_path)
833
self.path = unquote(self.quoted_path)
835
def __eq__(self, other):
836
return (isinstance(other, self.__class__) and
837
self.scheme == other.scheme and
838
self.host == other.host and
839
self.user == other.user and
840
self.password == other.password and
841
self.path == other.path)
844
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
845
self.__class__.__name__,
846
self.scheme, self.quoted_user, self.quoted_password,
847
self.quoted_host, self.port, self.quoted_path)
850
def from_string(cls, url):
851
"""Create a URL object from a string.
853
:param url: URL as bytestring
855
if isinstance(url, unicode):
856
raise errors.InvalidURL('should be ascii:\n%r' % url)
857
url = url.encode('utf-8')
858
(scheme, netloc, path, params,
859
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
860
user = password = host = port = None
862
user, host = netloc.rsplit('@', 1)
864
user, password = user.split(':', 1)
868
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
870
host, port = host.rsplit(':',1)
874
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
876
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
879
return cls(scheme, user, password, host, port, path)
882
netloc = self.quoted_host
884
netloc = "[%s]" % netloc
885
if self.quoted_user is not None:
886
# Note that we don't put the password back even if we
887
# have one so that it doesn't get accidentally
889
netloc = '%s@%s' % (self.quoted_user, netloc)
890
if self.port is not None:
891
netloc = '%s:%d' % (netloc, self.port)
892
return urlparse.urlunparse(
893
(self.scheme, netloc, self.quoted_path, None, None, None))
896
def _combine_paths(base_path, relpath):
897
"""Transform a Transport-relative path to a remote absolute path.
899
This does not handle substitution of ~ but does handle '..' and '.'
904
t._combine_paths('/home/sarah', 'project/foo')
905
=> '/home/sarah/project/foo'
906
t._combine_paths('/home/sarah', '../../etc')
908
t._combine_paths('/home/sarah', '/etc')
911
:param base_path: base path
912
:param relpath: relative url string for relative part of remote path.
913
:return: urlencoded string for final path.
915
if not isinstance(relpath, str):
916
raise errors.InvalidURL(relpath)
917
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
918
if relpath.startswith('/'):
921
base_parts = base_path.split('/')
922
if len(base_parts) > 0 and base_parts[-1] == '':
923
base_parts = base_parts[:-1]
924
for p in relpath.split('/'):
926
if len(base_parts) == 0:
927
# In most filesystems, a request for the parent
928
# of root, just returns root.
935
path = '/'.join(base_parts)
936
if not path.startswith('/'):
940
def clone(self, offset=None):
941
"""Return a new URL for a path relative to this URL.
943
:param offset: A relative path, already urlencoded
944
:return: `URL` instance
946
if offset is not None:
947
relative = unescape(offset).encode('utf-8')
948
path = self._combine_paths(self.path, relative)
949
path = quote(path, safe="/~")
951
path = self.quoted_path
952
return self.__class__(self.scheme, self.quoted_user,
953
self.quoted_password, self.quoted_host, self.port,
958
"""Extract the server address, the credentials and the path from the url.
960
user, password, host and path should be quoted if they contain reserved
963
:param url: an quoted url
964
:return: (scheme, user, password, host, port, path) tuple, all fields
967
parsed_url = URL.from_string(url)
968
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
969
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
bzrlib/urlutils.py
