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."""
23
from bzrlib.lazy_import import lazy_import
24
lazy_import(globals(), """
25
from posixpath import split as _posix_split
36
def basename(url, exclude_trailing_slash=True):
37
"""Return the last component of a URL.
39
:param url: The URL in question
40
:param exclude_trailing_slash: If the url looks like "path/to/foo/"
41
ignore the final slash and return 'foo' rather than ''
42
:return: Just the final component of the URL. This can return ''
43
if you don't exclude_trailing_slash, or if you are at the
46
return split(url, exclude_trailing_slash=exclude_trailing_slash)[1]
49
def dirname(url, exclude_trailing_slash=True):
50
"""Return the parent directory of the given path.
52
:param url: Relative or absolute URL
53
:param exclude_trailing_slash: Remove a final slash
54
(treat http://host/foo/ as http://host/foo, but
55
http://host/ stays http://host/)
56
:return: Everything in the URL except the last path chunk
58
# TODO: jam 20060502 This was named dirname to be consistent
59
# with the os functions, but maybe "parent" would be better
60
return split(url, exclude_trailing_slash=exclude_trailing_slash)[0]
64
"""Escape relpath to be a valid url."""
65
if isinstance(relpath, unicode):
66
relpath = relpath.encode('utf-8')
67
# After quoting and encoding, the path should be perfectly
68
# safe as a plain ASCII string, str() just enforces this
69
return str(urllib.quote(relpath, safe='/~'))
72
def file_relpath(base, path):
73
"""Compute just the relative sub-portion of a url
75
This assumes that both paths are already fully specified file:// URLs.
77
if len(base) < MIN_ABS_FILEURL_LENGTH:
78
raise ValueError('Length of base (%r) must equal or'
79
' exceed the platform minimum url length (which is %d)' %
80
(base, MIN_ABS_FILEURL_LENGTH))
81
base = osutils.normpath(local_path_from_url(base))
82
path = osutils.normpath(local_path_from_url(path))
83
return escape(osutils.relpath(base, path))
86
def _find_scheme_and_separator(url):
87
"""Find the scheme separator (://) and the first path separator
89
This is just a helper functions for other path utilities.
90
It could probably be replaced by urlparse
92
m = _url_scheme_re.match(url)
96
scheme = m.group('scheme')
97
path = m.group('path')
99
# Find the path separating slash
100
# (first slash after the ://)
101
first_path_slash = path.find('/')
102
if first_path_slash == -1:
103
return len(scheme), None
104
return len(scheme), first_path_slash+m.start('path')
108
"""Tests whether a URL is in actual fact a URL."""
109
return _url_scheme_re.match(url) is not None
112
def join(base, *args):
113
"""Create a URL by joining sections.
115
This will normalize '..', assuming that paths are absolute
116
(it assumes no symlinks in either path)
118
If any of *args is an absolute URL, it will be treated correctly.
120
join('http://foo', 'http://bar') => 'http://bar'
121
join('http://foo', 'bar') => 'http://foo/bar'
122
join('http://foo', 'bar', '../baz') => 'http://foo/baz'
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:]
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
144
path = joinpath(path, arg)
145
return base[:path_start] + path
148
def joinpath(base, *args):
149
"""Join URL path segments to a URL path segment.
151
This is somewhat like osutils.joinpath, but intended for URLs.
153
XXX: this duplicates some normalisation logic, and also duplicates a lot of
154
path handling logic that already exists in some Transport implementations.
155
We really should try to have exactly one place in the code base responsible
156
for combining paths of URLs.
158
path = base.split('/')
159
if len(path) > 1 and path[-1] == '':
160
#If the path ends in a trailing /, remove it.
163
if arg.startswith('/'):
165
for chunk in arg.split('/'):
170
raise errors.InvalidURLJoin('Cannot go above root',
178
return '/'.join(path)
181
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
182
def _posix_local_path_from_url(url):
183
"""Convert a url like file:///path/to/foo into /path/to/foo"""
184
url = split_segment_parameters_raw(url)[0]
185
file_localhost_prefix = 'file://localhost/'
186
if url.startswith(file_localhost_prefix):
187
path = url[len(file_localhost_prefix) - 1:]
188
elif not url.startswith('file:///'):
189
raise errors.InvalidURL(
190
url, 'local urls must start with file:/// or file://localhost/')
192
path = url[len('file://'):]
193
# We only strip off 2 slashes
194
return unescape(path)
197
def _posix_local_path_to_url(path):
198
"""Convert a local path like ./foo into a URL like file:///path/to/foo
200
This also handles transforming escaping unicode characters, etc.
202
# importing directly from posixpath allows us to test this
203
# on non-posix platforms
204
return 'file://' + escape(osutils._posix_abspath(path))
207
def _win32_local_path_from_url(url):
208
"""Convert a url like file:///C:/path/to/foo into C:/path/to/foo"""
209
if not url.startswith('file://'):
210
raise errors.InvalidURL(url, 'local urls must start with file:///, '
211
'UNC path urls must start with file://')
212
url = split_segment_parameters_raw(url)[0]
213
# We strip off all 3 slashes
214
win32_url = url[len('file:'):]
215
# check for UNC path: //HOST/path
216
if not win32_url.startswith('///'):
217
if (win32_url[2] == '/'
218
or win32_url[3] in '|:'):
219
raise errors.InvalidURL(url, 'Win32 UNC path urls'
220
' have form file://HOST/path')
221
return unescape(win32_url)
223
# allow empty paths so we can serve all roots
224
if win32_url == '///':
227
# usual local path with drive letter
228
if (len(win32_url) < 6
229
or win32_url[3] not in ('abcdefghijklmnopqrstuvwxyz'
230
'ABCDEFGHIJKLMNOPQRSTUVWXYZ')
231
or win32_url[4] not in '|:'
232
or win32_url[5] != '/'):
233
raise errors.InvalidURL(url, 'Win32 file urls start with'
234
' file:///x:/, where x is a valid drive letter')
235
return win32_url[3].upper() + u':' + unescape(win32_url[5:])
238
def _win32_local_path_to_url(path):
239
"""Convert a local path like ./foo into a URL like file:///C:/path/to/foo
241
This also handles transforming escaping unicode characters, etc.
243
# importing directly from ntpath allows us to test this
244
# on non-win32 platform
245
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
246
# which actually strips trailing space characters.
247
# The worst part is that on linux ntpath.abspath has different
248
# semantics, since 'nt' is not an available module.
252
win32_path = osutils._win32_abspath(path)
253
# check for UNC path \\HOST\path
254
if win32_path.startswith('//'):
255
return 'file:' + escape(win32_path)
256
return ('file:///' + str(win32_path[0].upper()) + ':' +
257
escape(win32_path[2:]))
260
local_path_to_url = _posix_local_path_to_url
261
local_path_from_url = _posix_local_path_from_url
262
MIN_ABS_FILEURL_LENGTH = len('file:///')
263
WIN32_MIN_ABS_FILEURL_LENGTH = len('file:///C:/')
265
if sys.platform == 'win32':
266
local_path_to_url = _win32_local_path_to_url
267
local_path_from_url = _win32_local_path_from_url
269
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH
272
_url_scheme_re = re.compile(r'^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$')
273
_url_hex_escapes_re = re.compile(r'(%[0-9a-fA-F]{2})')
276
def _unescape_safe_chars(matchobj):
277
"""re.sub callback to convert hex-escapes to plain characters (if safe).
279
e.g. '%7E' will be converted to '~'.
281
hex_digits = matchobj.group(0)[1:]
282
char = chr(int(hex_digits, 16))
283
if char in _url_dont_escape_characters:
286
return matchobj.group(0).upper()
289
def normalize_url(url):
290
"""Make sure that a path string is in fully normalized URL form.
292
This handles URLs which have unicode characters, spaces,
293
special characters, etc.
295
It has two basic modes of operation, depending on whether the
296
supplied string starts with a url specifier (scheme://) or not.
297
If it does not have a specifier it is considered a local path,
298
and will be converted into a file:/// url. Non-ascii characters
299
will be encoded using utf-8.
300
If it does have a url specifier, it will be treated as a "hybrid"
301
URL. Basically, a URL that should have URL special characters already
302
escaped (like +?&# etc), but may have unicode characters, etc
303
which would not be valid in a real URL.
305
:param url: Either a hybrid URL or a local path
306
:return: A normalized URL which only includes 7-bit ASCII characters.
308
scheme_end, path_start = _find_scheme_and_separator(url)
309
if scheme_end is None:
310
return local_path_to_url(url)
311
prefix = url[:path_start]
312
path = url[path_start:]
313
if not isinstance(url, unicode):
315
if c not in _url_safe_characters:
316
raise errors.InvalidURL(url, 'URLs can only contain specific'
317
' safe characters (not %r)' % c)
318
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
319
return str(prefix + ''.join(path))
321
# We have a unicode (hybrid) url
322
path_chars = list(path)
324
for i in xrange(len(path_chars)):
325
if path_chars[i] not in _url_safe_characters:
326
chars = path_chars[i].encode('utf-8')
327
path_chars[i] = ''.join(
328
['%%%02X' % ord(c) for c in path_chars[i].encode('utf-8')])
329
path = ''.join(path_chars)
330
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
331
return str(prefix + path)
334
def relative_url(base, other):
335
"""Return a path to other from base.
337
If other is unrelated to base, return other. Else return a relative path.
338
This assumes no symlinks as part of the url.
340
dummy, base_first_slash = _find_scheme_and_separator(base)
341
if base_first_slash is None:
344
dummy, other_first_slash = _find_scheme_and_separator(other)
345
if other_first_slash is None:
348
# this takes care of differing schemes or hosts
349
base_scheme = base[:base_first_slash]
350
other_scheme = other[:other_first_slash]
351
if base_scheme != other_scheme:
353
elif sys.platform == 'win32' and base_scheme == 'file://':
354
base_drive = base[base_first_slash+1:base_first_slash+3]
355
other_drive = other[other_first_slash+1:other_first_slash+3]
356
if base_drive != other_drive:
359
base_path = base[base_first_slash+1:]
360
other_path = other[other_first_slash+1:]
362
if base_path.endswith('/'):
363
base_path = base_path[:-1]
365
base_sections = base_path.split('/')
366
other_sections = other_path.split('/')
368
if base_sections == ['']:
370
if other_sections == ['']:
374
for b, o in zip(base_sections, other_sections):
377
output_sections.append(b)
379
match_len = len(output_sections)
380
output_sections = ['..' for x in base_sections[match_len:]]
381
output_sections.extend(other_sections[match_len:])
383
return "/".join(output_sections) or "."
386
def _win32_extract_drive_letter(url_base, path):
387
"""On win32 the drive letter needs to be added to the url base."""
388
# Strip off the drive letter
389
# path is currently /C:/foo
390
if len(path) < 4 or path[2] not in ':|' or path[3] != '/':
391
raise errors.InvalidURL(url_base + path,
392
'win32 file:/// paths need a drive letter')
393
url_base += path[0:3] # file:// + /C:
394
path = path[3:] # /foo
395
return url_base, path
398
def split(url, exclude_trailing_slash=True):
399
"""Split a URL into its parent directory and a child directory.
401
:param url: A relative or absolute URL
402
:param exclude_trailing_slash: Strip off a final '/' if it is part
403
of the path (but not if it is part of the protocol specification)
405
:return: (parent_url, child_dir). child_dir may be the empty string if we're at
408
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
410
if first_path_slash is None:
411
# We have either a relative path, or no separating slash
412
if scheme_loc is None:
414
if exclude_trailing_slash and url.endswith('/'):
416
return _posix_split(url)
418
# Scheme with no path
421
# We have a fully defined path
422
url_base = url[:first_path_slash] # http://host, file://
423
path = url[first_path_slash:] # /file/foo
425
if sys.platform == 'win32' and url.startswith('file:///'):
426
# Strip off the drive letter
427
# url_base is currently file://
428
# path is currently /C:/foo
429
url_base, path = _win32_extract_drive_letter(url_base, path)
430
# now it should be file:///C: and /foo
432
if exclude_trailing_slash and len(path) > 1 and path.endswith('/'):
434
head, tail = _posix_split(path)
435
return url_base + head, tail
438
def split_segment_parameters_raw(url):
439
"""Split the subsegment of the last segment of a URL.
441
:param url: A relative or absolute URL
442
:return: (url, subsegments)
444
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
445
# operates on urls not url+segments, and Transport classes
446
# should not be blindly adding slashes in the first place.
447
lurl = strip_trailing_slash(url)
448
# Segments begin at first comma after last forward slash, if one exists
449
segment_start = lurl.find(",", lurl.rfind("/")+1)
450
if segment_start == -1:
452
return (lurl[:segment_start], lurl[segment_start+1:].split(","))
455
def split_segment_parameters(url):
456
"""Split the segment parameters of the last segment of a URL.
458
:param url: A relative or absolute URL
459
:return: (url, segment_parameters)
461
(base_url, subsegments) = split_segment_parameters_raw(url)
463
for subsegment in subsegments:
464
(key, value) = subsegment.split("=", 1)
465
parameters[key] = value
466
return (base_url, parameters)
469
def join_segment_parameters_raw(base, *subsegments):
470
"""Create a new URL by adding subsegments to an existing one.
472
This adds the specified subsegments to the last path in the specified
473
base URL. The subsegments should be bytestrings.
475
:note: You probably want to use join_segment_parameters instead.
479
for subsegment in subsegments:
480
if type(subsegment) is not str:
481
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
482
if "," in subsegment:
483
raise errors.InvalidURLJoin(", exists in subsegments",
485
return ",".join((base,) + subsegments)
488
def join_segment_parameters(url, parameters):
489
"""Create a new URL by adding segment parameters to an existing one.
491
The parameters of the last segment in the URL will be updated; if a
492
parameter with the same key already exists it will be overwritten.
494
:param url: A URL, as string
495
:param parameters: Dictionary of parameters, keys and values as bytestrings
497
(base, existing_parameters) = split_segment_parameters(url)
499
new_parameters.update(existing_parameters)
500
for key, value in parameters.iteritems():
501
if type(key) is not str:
502
raise TypeError("parameter key %r is not a bytestring" % key)
503
if type(value) is not str:
504
raise TypeError("parameter value %r for %s is not a bytestring" %
507
raise errors.InvalidURLJoin("= exists in parameter key", url,
509
new_parameters[key] = value
510
return join_segment_parameters_raw(base,
511
*["%s=%s" % item for item in sorted(new_parameters.items())])
514
def _win32_strip_local_trailing_slash(url):
515
"""Strip slashes after the drive letter"""
516
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
522
def strip_trailing_slash(url):
523
"""Strip trailing slash, except for root paths.
525
The definition of 'root path' is platform-dependent.
526
This assumes that all URLs are valid netloc urls, such that they
529
It searches for ://, and then refuses to remove the next '/'.
530
It can also handle relative paths
532
path/to/foo => path/to/foo
533
path/to/foo/ => path/to/foo
534
http://host/path/ => http://host/path
535
http://host/path => http://host/path
536
http://host/ => http://host/
538
file:///foo/ => file:///foo
539
# This is unique on win32 platforms, and is the only URL
540
# format which does it differently.
541
file:///c|/ => file:///c:/
543
if not url.endswith('/'):
546
if sys.platform == 'win32' and url.startswith('file://'):
547
return _win32_strip_local_trailing_slash(url)
549
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
550
if scheme_loc is None:
551
# This is a relative path, as it has no scheme
552
# so just chop off the last character
555
if first_path_slash is None or first_path_slash == len(url)-1:
556
# Don't chop off anything if the only slash is the path
564
"""Unescape relpath from url format.
566
This returns a Unicode path from a URL
568
# jam 20060427 URLs are supposed to be ASCII only strings
569
# If they are passed in as unicode, urllib.unquote
570
# will return a UNICODE string, which actually contains
571
# utf-8 bytes. So we have to ensure that they are
572
# plain ASCII strings, or the final .decode will
573
# try to encode the UNICODE => ASCII, and then decode
577
except UnicodeError, e:
578
raise errors.InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,))
580
unquoted = urllib.unquote(url)
582
unicode_path = unquoted.decode('utf-8')
583
except UnicodeError, e:
584
raise errors.InvalidURL(url, 'Unable to encode the URL as utf-8: %s' % (e,))
588
# These are characters that if escaped, should stay that way
589
_no_decode_chars = ';/?:@&=+$,#'
590
_no_decode_ords = [ord(c) for c in _no_decode_chars]
591
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
592
+ ['%02X' % o for o in _no_decode_ords])
593
_hex_display_map = dict(([('%02x' % o, chr(o)) for o in range(256)]
594
+ [('%02X' % o, chr(o)) for o in range(256)]))
595
#These entries get mapped to themselves
596
_hex_display_map.update((hex,'%'+hex) for hex in _no_decode_hex)
598
# These characters shouldn't be percent-encoded, and it's always safe to
599
# unencode them if they are.
600
_url_dont_escape_characters = set(
601
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
602
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
603
"0123456789" # Numbers
604
"-._~" # Unreserved characters
607
# These characters should not be escaped
608
_url_safe_characters = set(
609
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
610
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
611
"0123456789" # Numbers
612
"_.-!~*'()" # Unreserved characters
613
"/;?:@&=+$," # Reserved characters
614
"%#" # Extra reserved characters
617
def unescape_for_display(url, encoding):
618
"""Decode what you can for a URL, so that we get a nice looking path.
620
This will turn file:// urls into local paths, and try to decode
621
any portions of a http:// style url that it can.
623
Any sections of the URL which can't be represented in the encoding or
624
need to stay as escapes are left alone.
626
:param url: A 7-bit ASCII URL
627
:param encoding: The final output encoding
629
:return: A unicode string which can be safely encoded into the
633
raise ValueError('you cannot specify None for the display encoding')
634
if url.startswith('file://'):
636
path = local_path_from_url(url)
637
path.encode(encoding)
642
# Split into sections to try to decode utf-8
644
for i in xrange(1, len(res)):
645
escaped_chunks = res[i].split('%')
646
for j in xrange(1, len(escaped_chunks)):
647
item = escaped_chunks[j]
649
escaped_chunks[j] = _hex_display_map[item[:2]] + item[2:]
651
# Put back the percent symbol
652
escaped_chunks[j] = '%' + item
653
except UnicodeDecodeError:
654
escaped_chunks[j] = unichr(int(item[:2], 16)) + item[2:]
655
unescaped = ''.join(escaped_chunks)
657
decoded = unescaped.decode('utf-8')
658
except UnicodeDecodeError:
659
# If this path segment cannot be properly utf-8 decoded
660
# after doing unescaping we will just leave it alone
664
decoded.encode(encoding)
665
except UnicodeEncodeError:
666
# If this chunk cannot be encoded in the local
667
# encoding, then we should leave it alone
670
# Otherwise take the url decoded one
672
return u'/'.join(res)
675
def derive_to_location(from_location):
676
"""Derive a TO_LOCATION given a FROM_LOCATION.
678
The normal case is a FROM_LOCATION of http://foo/bar => bar.
679
The Right Thing for some logical destinations may differ though
680
because no / may be present at all. In that case, the result is
681
the full name without the scheme indicator, e.g. lp:foo-bar => foo-bar.
682
This latter case also applies when a Windows drive
683
is used without a path, e.g. c:foo-bar => foo-bar.
684
If no /, path separator or : is found, the from_location is returned.
686
if from_location.find("/") >= 0 or from_location.find(os.sep) >= 0:
687
return os.path.basename(from_location.rstrip("/\\"))
689
sep = from_location.find(":")
691
return from_location[sep+1:]
696
def _is_absolute(url):
697
return (osutils.pathjoin('/foo', url) == url)
700
def rebase_url(url, old_base, new_base):
701
"""Convert a relative path from an old base URL to a new base URL.
703
The result will be a relative path.
704
Absolute paths and full URLs are returned unaltered.
706
scheme, separator = _find_scheme_and_separator(url)
707
if scheme is not None:
709
if _is_absolute(url):
711
old_parsed = urlparse.urlparse(old_base)
712
new_parsed = urlparse.urlparse(new_base)
713
if (old_parsed[:2]) != (new_parsed[:2]):
714
raise errors.InvalidRebaseURLs(old_base, new_base)
715
return determine_relative_path(new_parsed[2],
716
join(old_parsed[2], url))
719
def determine_relative_path(from_path, to_path):
720
"""Determine a relative path from from_path to to_path."""
721
from_segments = osutils.splitpath(from_path)
722
to_segments = osutils.splitpath(to_path)
724
for count, (from_element, to_element) in enumerate(zip(from_segments,
726
if from_element != to_element:
730
unique_from = from_segments[count:]
731
unique_to = to_segments[count:]
732
segments = (['..'] * len(unique_from) + unique_to)
733
if len(segments) == 0:
735
return osutils.pathjoin(*segments)
741
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
744
self.quoted_host = quoted_host
745
self.host = urllib.unquote(self.quoted_host)
746
self.quoted_user = quoted_user
747
if self.quoted_user is not None:
748
self.user = urllib.unquote(self.quoted_user)
751
self.quoted_password = quoted_password
752
if self.quoted_password is not None:
753
self.password = urllib.unquote(self.quoted_password)
757
self.quoted_path = _url_hex_escapes_re.sub(_unescape_safe_chars, quoted_path)
758
self.path = urllib.unquote(self.quoted_path)
760
def __eq__(self, other):
761
return (isinstance(other, self.__class__) and
762
self.scheme == other.scheme and
763
self.host == other.host and
764
self.user == other.user and
765
self.password == other.password and
766
self.path == other.path)
769
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
770
self.__class__.__name__,
771
self.scheme, self.quoted_user, self.quoted_password,
772
self.quoted_host, self.port, self.quoted_path)
775
def from_string(cls, url):
776
"""Create a URL object from a string.
778
:param url: URL as bytestring
780
if isinstance(url, unicode):
781
raise errors.InvalidURL('should be ascii:\n%r' % url)
782
url = url.encode('utf-8')
783
(scheme, netloc, path, params,
784
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
785
user = password = host = port = None
787
user, host = netloc.rsplit('@', 1)
789
user, password = user.split(':', 1)
793
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
795
host, port = host.rsplit(':',1)
799
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
801
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
804
return cls(scheme, user, password, host, port, path)
807
netloc = self.quoted_host
809
netloc = "[%s]" % netloc
810
if self.quoted_user is not None:
811
# Note that we don't put the password back even if we
812
# have one so that it doesn't get accidentally
814
netloc = '%s@%s' % (self.quoted_user, netloc)
815
if self.port is not None:
816
netloc = '%s:%d' % (netloc, self.port)
817
return urlparse.urlunparse(
818
(self.scheme, netloc, self.quoted_path, None, None, None))
821
def _combine_paths(base_path, relpath):
822
"""Transform a Transport-relative path to a remote absolute path.
824
This does not handle substitution of ~ but does handle '..' and '.'
829
t._combine_paths('/home/sarah', 'project/foo')
830
=> '/home/sarah/project/foo'
831
t._combine_paths('/home/sarah', '../../etc')
833
t._combine_paths('/home/sarah', '/etc')
836
:param base_path: base path
837
:param relpath: relative url string for relative part of remote path.
838
:return: urlencoded string for final path.
840
if not isinstance(relpath, str):
841
raise errors.InvalidURL(relpath)
842
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
843
if relpath.startswith('/'):
846
base_parts = base_path.split('/')
847
if len(base_parts) > 0 and base_parts[-1] == '':
848
base_parts = base_parts[:-1]
849
for p in relpath.split('/'):
851
if len(base_parts) == 0:
852
# In most filesystems, a request for the parent
853
# of root, just returns root.
860
path = '/'.join(base_parts)
861
if not path.startswith('/'):
865
def clone(self, offset=None):
866
"""Return a new URL for a path relative to this URL.
868
:param offset: A relative path, already urlencoded
869
:return: `URL` instance
871
if offset is not None:
872
relative = unescape(offset).encode('utf-8')
873
path = self._combine_paths(self.path, relative)
874
path = urllib.quote(path, safe="/~")
876
path = self.quoted_path
877
return self.__class__(self.scheme, self.quoted_user,
878
self.quoted_password, self.quoted_host, self.port,
883
"""Extract the server address, the credentials and the path from the url.
885
user, password, host and path should be quoted if they contain reserved
888
:param url: an quoted url
889
:return: (scheme, user, password, host, port, path) tuple, all fields
892
parsed_url = URL.from_string(url)
893
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
894
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
bzrlib/urlutils.py
