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
47
class InvalidURL(errors.PathError):
49
_fmt = 'Invalid url supplied to transport: "%(path)s"%(extra)s'
52
class InvalidURLJoin(errors.PathError):
54
_fmt = "Invalid URL join request: %(reason)s: %(base)r + %(join_args)r"
56
def __init__(self, reason, base, join_args):
59
self.join_args = join_args
60
errors.PathError.__init__(self, base, reason)
63
class InvalidRebaseURLs(errors.PathError):
65
_fmt = "URLs differ by more than path: %(from_)r and %(to)r"
67
def __init__(self, from_, to):
70
errors.PathError.__init__(self, from_, 'URLs differ by more than path.')
73
def basename(url, exclude_trailing_slash=True):
74
"""Return the last component of a URL.
76
:param url: The URL in question
77
:param exclude_trailing_slash: If the url looks like "path/to/foo/"
78
ignore the final slash and return 'foo' rather than ''
79
:return: Just the final component of the URL. This can return ''
80
if you don't exclude_trailing_slash, or if you are at the
83
return split(url, exclude_trailing_slash=exclude_trailing_slash)[1]
86
def dirname(url, exclude_trailing_slash=True):
87
"""Return the parent directory of the given path.
89
:param url: Relative or absolute URL
90
:param exclude_trailing_slash: Remove a final slash
91
(treat http://host/foo/ as http://host/foo, but
92
http://host/ stays http://host/)
93
:return: Everything in the URL except the last path chunk
95
# TODO: jam 20060502 This was named dirname to be consistent
96
# with the os functions, but maybe "parent" would be better
97
return split(url, exclude_trailing_slash=exclude_trailing_slash)[0]
101
quote_from_bytes = urlparse.quote_from_bytes
102
quote = urlparse.quote
103
unquote_to_bytes = urlparse.unquote_to_bytes
105
# Private copies of quote and unquote, copied from Python's
106
# urllib module because urllib unconditionally imports socket, which imports
109
always_safe = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
110
'abcdefghijklmnopqrstuvwxyz'
113
for i, c in zip(range(256), ''.join(map(chr, range(256)))):
114
_safe_map[c] = c if (i < 128 and c in always_safe) else '%{0:02X}'.format(i)
117
def quote_from_bytes(s, safe='/'):
118
"""quote('abc def') -> 'abc%20def'
120
Each part of a URL, e.g. the path info, the query, etc., has a
121
different set of reserved characters that must be quoted.
123
RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists
124
the following reserved characters.
126
reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
129
Each of these characters is reserved in some component of a URL,
130
but not necessarily in all of them.
132
By default, the quote function is intended for quoting the path
133
section of a URL. Thus, it will not encode '/'. This character
134
is reserved, but in typical usage the quote function is being
135
called on a path where the existing slash characters are used as
141
raise TypeError('None object cannot be quoted')
143
cachekey = (safe, always_safe)
145
(quoter, safe) = _safe_quoters[cachekey]
147
safe_map = _safe_map.copy()
148
safe_map.update([(c, c) for c in safe])
149
quoter = safe_map.__getitem__
150
safe = always_safe + safe
151
_safe_quoters[cachekey] = (quoter, safe)
152
if not s.rstrip(safe):
154
return ''.join(map(quoter, s))
156
quote = quote_from_bytes
157
unquote_to_bytes = urlparse.unquote
160
unquote = urlparse.unquote
164
"""Escape relpath to be a valid url."""
165
if not isinstance(relpath, str) and sys.version_info[0] == 2:
166
relpath = relpath.encode('utf-8')
167
return quote(relpath, safe='/~')
170
def file_relpath(base, path):
171
"""Compute just the relative sub-portion of a url
173
This assumes that both paths are already fully specified file:// URLs.
175
if len(base) < MIN_ABS_FILEURL_LENGTH:
176
raise ValueError('Length of base (%r) must equal or'
177
' exceed the platform minimum url length (which is %d)' %
178
(base, MIN_ABS_FILEURL_LENGTH))
179
base = osutils.normpath(local_path_from_url(base))
180
path = osutils.normpath(local_path_from_url(path))
181
return escape(osutils.relpath(base, path))
184
def _find_scheme_and_separator(url):
185
"""Find the scheme separator (://) and the first path separator
187
This is just a helper functions for other path utilities.
188
It could probably be replaced by urlparse
190
m = _url_scheme_re.match(url)
194
scheme = m.group('scheme')
195
path = m.group('path')
197
# Find the path separating slash
198
# (first slash after the ://)
199
first_path_slash = path.find('/')
200
if first_path_slash == -1:
201
return len(scheme), None
202
return len(scheme), first_path_slash+m.start('path')
206
"""Tests whether a URL is in actual fact a URL."""
207
return _url_scheme_re.match(url) is not None
210
def join(base, *args):
211
"""Create a URL by joining sections.
213
This will normalize '..', assuming that paths are absolute
214
(it assumes no symlinks in either path)
216
If any of *args is an absolute URL, it will be treated correctly.
218
join('http://foo', 'http://bar') => 'http://bar'
219
join('http://foo', 'bar') => 'http://foo/bar'
220
join('http://foo', 'bar', '../baz') => 'http://foo/baz'
224
scheme_end, path_start = _find_scheme_and_separator(base)
225
if scheme_end is None and path_start is None:
227
elif path_start is None:
228
path_start = len(base)
229
path = base[path_start:]
231
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
232
if arg_scheme_end is None and arg_path_start is None:
234
elif arg_path_start is None:
235
arg_path_start = len(arg)
236
if arg_scheme_end is not None:
238
path = arg[arg_path_start:]
239
scheme_end = arg_scheme_end
240
path_start = arg_path_start
242
path = joinpath(path, arg)
243
return base[:path_start] + path
246
def joinpath(base, *args):
247
"""Join URL path segments to a URL path segment.
249
This is somewhat like osutils.joinpath, but intended for URLs.
251
XXX: this duplicates some normalisation logic, and also duplicates a lot of
252
path handling logic that already exists in some Transport implementations.
253
We really should try to have exactly one place in the code base responsible
254
for combining paths of URLs.
256
path = base.split('/')
257
if len(path) > 1 and path[-1] == '':
258
#If the path ends in a trailing /, remove it.
261
if arg.startswith('/'):
263
for chunk in arg.split('/'):
268
raise InvalidURLJoin('Cannot go above root',
276
return '/'.join(path)
279
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
280
def _posix_local_path_from_url(url):
281
"""Convert a url like file:///path/to/foo into /path/to/foo"""
282
url = split_segment_parameters_raw(url)[0]
283
file_localhost_prefix = 'file://localhost/'
284
if url.startswith(file_localhost_prefix):
285
path = url[len(file_localhost_prefix) - 1:]
286
elif not url.startswith('file:///'):
288
url, 'local urls must start with file:/// or file://localhost/')
290
path = url[len('file://'):]
291
# We only strip off 2 slashes
292
return unescape(path)
295
def _posix_local_path_to_url(path):
296
"""Convert a local path like ./foo into a URL like file:///path/to/foo
298
This also handles transforming escaping unicode characters, etc.
300
# importing directly from posixpath allows us to test this
301
# on non-posix platforms
302
return 'file://' + escape(osutils._posix_abspath(path))
305
def _win32_local_path_from_url(url):
306
"""Convert a url like file:///C:/path/to/foo into C:/path/to/foo"""
307
if not url.startswith('file://'):
308
raise InvalidURL(url, 'local urls must start with file:///, '
309
'UNC path urls must start with file://')
310
url = split_segment_parameters_raw(url)[0]
311
# We strip off all 3 slashes
312
win32_url = url[len('file:'):]
313
# check for UNC path: //HOST/path
314
if not win32_url.startswith('///'):
315
if (win32_url[2] == '/'
316
or win32_url[3] in '|:'):
317
raise InvalidURL(url, 'Win32 UNC path urls'
318
' have form file://HOST/path')
319
return unescape(win32_url)
321
# allow empty paths so we can serve all roots
322
if win32_url == '///':
325
# usual local path with drive letter
326
if (len(win32_url) < 6
327
or win32_url[3] not in ('abcdefghijklmnopqrstuvwxyz'
328
'ABCDEFGHIJKLMNOPQRSTUVWXYZ')
329
or win32_url[4] not in '|:'
330
or win32_url[5] != '/'):
331
raise InvalidURL(url, 'Win32 file urls start with'
332
' file:///x:/, where x is a valid drive letter')
333
return win32_url[3].upper() + u':' + unescape(win32_url[5:])
336
def _win32_local_path_to_url(path):
337
"""Convert a local path like ./foo into a URL like file:///C:/path/to/foo
339
This also handles transforming escaping unicode characters, etc.
341
# importing directly from ntpath allows us to test this
342
# on non-win32 platform
343
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
344
# which actually strips trailing space characters.
345
# The worst part is that on linux ntpath.abspath has different
346
# semantics, since 'nt' is not an available module.
350
win32_path = osutils._win32_abspath(path)
351
# check for UNC path \\HOST\path
352
if win32_path.startswith('//'):
353
return 'file:' + escape(win32_path)
354
return ('file:///' + str(win32_path[0].upper()) + ':' +
355
escape(win32_path[2:]))
358
local_path_to_url = _posix_local_path_to_url
359
local_path_from_url = _posix_local_path_from_url
360
MIN_ABS_FILEURL_LENGTH = len('file:///')
361
WIN32_MIN_ABS_FILEURL_LENGTH = len('file:///C:/')
363
if sys.platform == 'win32':
364
local_path_to_url = _win32_local_path_to_url
365
local_path_from_url = _win32_local_path_from_url
367
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH
370
_url_scheme_re = re.compile('^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$')
371
_url_hex_escapes_re = re.compile('(%[0-9a-fA-F]{2})')
374
def _unescape_safe_chars(matchobj):
375
"""re.sub callback to convert hex-escapes to plain characters (if safe).
377
e.g. '%7E' will be converted to '~'.
379
hex_digits = matchobj.group(0)[1:]
380
char = chr(int(hex_digits, 16))
381
if char in _url_dont_escape_characters:
384
return matchobj.group(0).upper()
387
def normalize_url(url):
388
"""Make sure that a path string is in fully normalized URL form.
390
This handles URLs which have unicode characters, spaces,
391
special characters, etc.
393
It has two basic modes of operation, depending on whether the
394
supplied string starts with a url specifier (scheme://) or not.
395
If it does not have a specifier it is considered a local path,
396
and will be converted into a file:/// url. Non-ascii characters
397
will be encoded using utf-8.
398
If it does have a url specifier, it will be treated as a "hybrid"
399
URL. Basically, a URL that should have URL special characters already
400
escaped (like +?&# etc), but may have unicode characters, etc
401
which would not be valid in a real URL.
403
:param url: Either a hybrid URL or a local path
404
:return: A normalized URL which only includes 7-bit ASCII characters.
406
scheme_end, path_start = _find_scheme_and_separator(url)
407
if scheme_end is None:
408
return local_path_to_url(url)
409
prefix = url[:path_start]
410
path = url[path_start:]
411
if not isinstance(url, text_type):
413
if c not in _url_safe_characters:
414
raise InvalidURL(url, 'URLs can only contain specific'
415
' safe characters (not %r)' % c)
416
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
417
return str(prefix + ''.join(path))
419
# We have a unicode (hybrid) url
420
path_chars = list(path)
422
for i in range(len(path_chars)):
423
if path_chars[i] not in _url_safe_characters:
424
chars = path_chars[i].encode('utf-8')
425
path_chars[i] = ''.join(
426
['%%%02X' % c for c in bytearray(path_chars[i].encode('utf-8'))])
427
path = ''.join(path_chars)
428
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
429
return str(prefix + path)
432
def relative_url(base, other):
433
"""Return a path to other from base.
435
If other is unrelated to base, return other. Else return a relative path.
436
This assumes no symlinks as part of the url.
438
dummy, base_first_slash = _find_scheme_and_separator(base)
439
if base_first_slash is None:
442
dummy, other_first_slash = _find_scheme_and_separator(other)
443
if other_first_slash is None:
446
# this takes care of differing schemes or hosts
447
base_scheme = base[:base_first_slash]
448
other_scheme = other[:other_first_slash]
449
if base_scheme != other_scheme:
451
elif sys.platform == 'win32' and base_scheme == 'file://':
452
base_drive = base[base_first_slash+1:base_first_slash+3]
453
other_drive = other[other_first_slash+1:other_first_slash+3]
454
if base_drive != other_drive:
457
base_path = base[base_first_slash+1:]
458
other_path = other[other_first_slash+1:]
460
if base_path.endswith('/'):
461
base_path = base_path[:-1]
463
base_sections = base_path.split('/')
464
other_sections = other_path.split('/')
466
if base_sections == ['']:
468
if other_sections == ['']:
472
for b, o in zip(base_sections, other_sections):
475
output_sections.append(b)
477
match_len = len(output_sections)
478
output_sections = ['..' for x in base_sections[match_len:]]
479
output_sections.extend(other_sections[match_len:])
481
return "/".join(output_sections) or "."
484
def _win32_extract_drive_letter(url_base, path):
485
"""On win32 the drive letter needs to be added to the url base."""
486
# Strip off the drive letter
487
# path is currently /C:/foo
488
if len(path) < 4 or path[2] not in ':|' or path[3] != '/':
489
raise InvalidURL(url_base + path,
490
'win32 file:/// paths need a drive letter')
491
url_base += path[0:3] # file:// + /C:
492
path = path[3:] # /foo
493
return url_base, path
496
def split(url, exclude_trailing_slash=True):
497
"""Split a URL into its parent directory and a child directory.
499
:param url: A relative or absolute URL
500
:param exclude_trailing_slash: Strip off a final '/' if it is part
501
of the path (but not if it is part of the protocol specification)
503
:return: (parent_url, child_dir). child_dir may be the empty string if we're at
506
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
508
if first_path_slash is None:
509
# We have either a relative path, or no separating slash
510
if scheme_loc is None:
512
if exclude_trailing_slash and url.endswith('/'):
514
return _posix_split(url)
516
# Scheme with no path
519
# We have a fully defined path
520
url_base = url[:first_path_slash] # http://host, file://
521
path = url[first_path_slash:] # /file/foo
523
if sys.platform == 'win32' and url.startswith('file:///'):
524
# Strip off the drive letter
525
# url_base is currently file://
526
# path is currently /C:/foo
527
url_base, path = _win32_extract_drive_letter(url_base, path)
528
# now it should be file:///C: and /foo
530
if exclude_trailing_slash and len(path) > 1 and path.endswith('/'):
532
head, tail = _posix_split(path)
533
return url_base + head, tail
536
def split_segment_parameters_raw(url):
537
"""Split the subsegment of the last segment of a URL.
539
:param url: A relative or absolute URL
540
:return: (url, subsegments)
542
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
543
# operates on urls not url+segments, and Transport classes
544
# should not be blindly adding slashes in the first place.
545
lurl = strip_trailing_slash(url)
546
# Segments begin at first comma after last forward slash, if one exists
547
segment_start = lurl.find(",", lurl.rfind("/")+1)
548
if segment_start == -1:
550
return (lurl[:segment_start], [str(s) for s in lurl[segment_start+1:].split(",")])
553
def split_segment_parameters(url):
554
"""Split the segment parameters of the last segment of a URL.
556
:param url: A relative or absolute URL
557
:return: (url, segment_parameters)
559
(base_url, subsegments) = split_segment_parameters_raw(url)
561
for subsegment in subsegments:
562
(key, value) = subsegment.split("=", 1)
563
if not isinstance(key, str):
565
if not isinstance(value, str):
566
raise TypeError(value)
567
parameters[key] = value
568
return (base_url, parameters)
571
def join_segment_parameters_raw(base, *subsegments):
572
"""Create a new URL by adding subsegments to an existing one.
574
This adds the specified subsegments to the last path in the specified
575
base URL. The subsegments should be bytestrings.
577
:note: You probably want to use join_segment_parameters instead.
581
for subsegment in subsegments:
582
if not isinstance(subsegment, str):
583
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
584
if "," in subsegment:
585
raise InvalidURLJoin(", exists in subsegments",
587
return ",".join((base,) + subsegments)
590
def join_segment_parameters(url, parameters):
591
"""Create a new URL by adding segment parameters to an existing one.
593
The parameters of the last segment in the URL will be updated; if a
594
parameter with the same key already exists it will be overwritten.
596
:param url: A URL, as string
597
:param parameters: Dictionary of parameters, keys and values as bytestrings
599
(base, existing_parameters) = split_segment_parameters(url)
601
new_parameters.update(existing_parameters)
602
for key, value in parameters.items():
603
if not isinstance(key, str):
604
raise TypeError("parameter key %r is not a str" % key)
605
if not isinstance(value, str):
606
raise TypeError("parameter value %r for %r is not a str" %
609
raise InvalidURLJoin("= exists in parameter key", url,
611
new_parameters[key] = value
612
return join_segment_parameters_raw(base,
613
*["%s=%s" % item for item in sorted(new_parameters.items())])
616
def _win32_strip_local_trailing_slash(url):
617
"""Strip slashes after the drive letter"""
618
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
624
def strip_trailing_slash(url):
625
"""Strip trailing slash, except for root paths.
627
The definition of 'root path' is platform-dependent.
628
This assumes that all URLs are valid netloc urls, such that they
631
It searches for ://, and then refuses to remove the next '/'.
632
It can also handle relative paths
634
path/to/foo => path/to/foo
635
path/to/foo/ => path/to/foo
636
http://host/path/ => http://host/path
637
http://host/path => http://host/path
638
http://host/ => http://host/
640
file:///foo/ => file:///foo
641
# This is unique on win32 platforms, and is the only URL
642
# format which does it differently.
643
file:///c|/ => file:///c:/
645
if not url.endswith('/'):
648
if sys.platform == 'win32' and url.startswith('file://'):
649
return _win32_strip_local_trailing_slash(url)
651
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
652
if scheme_loc is None:
653
# This is a relative path, as it has no scheme
654
# so just chop off the last character
657
if first_path_slash is None or first_path_slash == len(url)-1:
658
# Don't chop off anything if the only slash is the path
666
"""Unescape relpath from url format.
668
This returns a Unicode path from a URL
670
# jam 20060427 URLs are supposed to be ASCII only strings
671
# If they are passed in as unicode, unquote
672
# will return a UNICODE string, which actually contains
673
# utf-8 bytes. So we have to ensure that they are
674
# plain ASCII strings, or the final .decode will
675
# try to encode the UNICODE => ASCII, and then decode
679
if isinstance(url, text_type):
682
except UnicodeError as e:
683
raise InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,))
684
return urlparse.unquote(url)
686
if isinstance(url, text_type):
688
url = url.encode("ascii")
689
except UnicodeError as e:
690
raise InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,))
691
unquoted = unquote(url)
693
unicode_path = unquoted.decode('utf-8')
694
except UnicodeError as e:
695
raise InvalidURL(url, 'Unable to encode the URL as utf-8: %s' % (e,))
699
# These are characters that if escaped, should stay that way
700
_no_decode_chars = ';/?:@&=+$,#'
701
_no_decode_ords = [ord(c) for c in _no_decode_chars]
702
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
703
+ ['%02X' % o for o in _no_decode_ords])
704
_hex_display_map = dict(([('%02x' % o, int2byte(o)) for o in range(256)]
705
+ [('%02X' % o, int2byte(o)) for o in range(256)]))
706
#These entries get mapped to themselves
707
_hex_display_map.update((hex, b'%'+hex.encode('ascii')) for hex in _no_decode_hex)
709
# These characters shouldn't be percent-encoded, and it's always safe to
710
# unencode them if they are.
711
_url_dont_escape_characters = set(
712
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
713
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
714
"0123456789" # Numbers
715
"-._~" # Unreserved characters
718
# These characters should not be escaped
719
_url_safe_characters = set(
720
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
721
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
722
"0123456789" # Numbers
723
"_.-!~*'()" # Unreserved characters
724
"/;?:@&=+$," # Reserved characters
725
"%#" # Extra reserved characters
729
def _unescape_segment_for_display(segment, encoding):
730
"""Unescape a segment for display.
732
Helper for unescape_for_display
734
:param url: A 7-bit ASCII URL
735
:param encoding: The final output encoding
737
:return: A unicode string which can be safely encoded into the
740
escaped_chunks = segment.split('%')
741
escaped_chunks[0] = escaped_chunks[0].encode('utf-8')
742
for j in range(1, len(escaped_chunks)):
743
item = escaped_chunks[j]
745
escaped_chunks[j] = _hex_display_map[item[:2]]
747
# Put back the percent symbol
748
escaped_chunks[j] = b'%' + (item[:2].encode('utf-8') if PY3 else item[:2])
749
except UnicodeDecodeError:
750
escaped_chunks[j] = unichr(int(item[:2], 16)).encode('utf-8')
751
escaped_chunks[j] += (item[2:].encode('utf-8') if PY3 else item[2:])
752
unescaped = b''.join(escaped_chunks)
754
decoded = unescaped.decode('utf-8')
755
except UnicodeDecodeError:
756
# If this path segment cannot be properly utf-8 decoded
757
# after doing unescaping we will just leave it alone
761
decoded.encode(encoding)
762
except UnicodeEncodeError:
763
# If this chunk cannot be encoded in the local
764
# encoding, then we should leave it alone
767
# Otherwise take the url decoded one
771
def unescape_for_display(url, encoding):
772
"""Decode what you can for a URL, so that we get a nice looking path.
774
This will turn file:// urls into local paths, and try to decode
775
any portions of a http:// style url that it can.
777
Any sections of the URL which can't be represented in the encoding or
778
need to stay as escapes are left alone.
780
:param url: A 7-bit ASCII URL
781
:param encoding: The final output encoding
783
:return: A unicode string which can be safely encoded into the
787
raise ValueError('you cannot specify None for the display encoding')
788
if url.startswith('file://'):
790
path = local_path_from_url(url)
791
path.encode(encoding)
796
# Split into sections to try to decode utf-8
798
for i in range(1, len(res)):
799
res[i] = _unescape_segment_for_display(res[i], encoding)
800
return u'/'.join(res)
803
def derive_to_location(from_location):
804
"""Derive a TO_LOCATION given a FROM_LOCATION.
806
The normal case is a FROM_LOCATION of http://foo/bar => bar.
807
The Right Thing for some logical destinations may differ though
808
because no / may be present at all. In that case, the result is
809
the full name without the scheme indicator, e.g. lp:foo-bar => foo-bar.
810
This latter case also applies when a Windows drive
811
is used without a path, e.g. c:foo-bar => foo-bar.
812
If no /, path separator or : is found, the from_location is returned.
814
from_location, unused_params = split_segment_parameters(from_location)
815
if from_location.find("/") >= 0 or from_location.find(os.sep) >= 0:
816
return os.path.basename(from_location.rstrip("/\\"))
818
sep = from_location.find(":")
820
return from_location[sep+1:]
825
def _is_absolute(url):
826
return (osutils.pathjoin('/foo', url) == url)
829
def rebase_url(url, old_base, new_base):
830
"""Convert a relative path from an old base URL to a new base URL.
832
The result will be a relative path.
833
Absolute paths and full URLs are returned unaltered.
835
scheme, separator = _find_scheme_and_separator(url)
836
if scheme is not None:
838
if _is_absolute(url):
840
old_parsed = urlparse.urlparse(old_base)
841
new_parsed = urlparse.urlparse(new_base)
842
if (old_parsed[:2]) != (new_parsed[:2]):
843
raise InvalidRebaseURLs(old_base, new_base)
844
return determine_relative_path(new_parsed[2],
845
join(old_parsed[2], url))
848
def determine_relative_path(from_path, to_path):
849
"""Determine a relative path from from_path to to_path."""
850
from_segments = osutils.splitpath(from_path)
851
to_segments = osutils.splitpath(to_path)
853
for count, (from_element, to_element) in enumerate(zip(from_segments,
855
if from_element != to_element:
859
unique_from = from_segments[count:]
860
unique_to = to_segments[count:]
861
segments = (['..'] * len(unique_from) + unique_to)
862
if len(segments) == 0:
864
return osutils.pathjoin(*segments)
870
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
873
self.quoted_host = quoted_host
874
self.host = unquote(self.quoted_host)
875
self.quoted_user = quoted_user
876
if self.quoted_user is not None:
877
self.user = unquote(self.quoted_user)
880
self.quoted_password = quoted_password
881
if self.quoted_password is not None:
882
self.password = unquote(self.quoted_password)
886
self.quoted_path = _url_hex_escapes_re.sub(_unescape_safe_chars, quoted_path)
887
self.path = unquote(self.quoted_path)
889
def __eq__(self, other):
890
return (isinstance(other, self.__class__) and
891
self.scheme == other.scheme and
892
self.host == other.host and
893
self.user == other.user and
894
self.password == other.password and
895
self.path == other.path)
898
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
899
self.__class__.__name__,
900
self.scheme, self.quoted_user, self.quoted_password,
901
self.quoted_host, self.port, self.quoted_path)
904
def from_string(cls, url):
905
"""Create a URL object from a string.
907
:param url: URL as bytestring
909
# GZ 2017-06-09: Actually validate ascii-ness
910
# pad.lv/1696545: For the moment, accept both native strings and unicode.
911
if isinstance(url, str):
913
elif isinstance(url, text_type):
916
except UnicodeEncodeError:
917
raise InvalidURL(url)
919
raise InvalidURL(url)
920
(scheme, netloc, path, params,
921
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
922
user = password = host = port = None
924
user, host = netloc.rsplit('@', 1)
926
user, password = user.split(':', 1)
930
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
932
host, port = host.rsplit(':', 1)
937
raise InvalidURL('invalid port number %s in url:\n%s' %
941
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
944
return cls(scheme, user, password, host, port, path)
947
netloc = self.quoted_host
949
netloc = "[%s]" % netloc
950
if self.quoted_user is not None:
951
# Note that we don't put the password back even if we
952
# have one so that it doesn't get accidentally
954
netloc = '%s@%s' % (self.quoted_user, netloc)
955
if self.port is not None:
956
netloc = '%s:%d' % (netloc, self.port)
957
return urlparse.urlunparse(
958
(self.scheme, netloc, self.quoted_path, None, None, None))
961
def _combine_paths(base_path, relpath):
962
"""Transform a Transport-relative path to a remote absolute path.
964
This does not handle substitution of ~ but does handle '..' and '.'
969
t._combine_paths('/home/sarah', 'project/foo')
970
=> '/home/sarah/project/foo'
971
t._combine_paths('/home/sarah', '../../etc')
973
t._combine_paths('/home/sarah', '/etc')
976
:param base_path: base path
977
:param relpath: relative url string for relative part of remote path.
978
:return: urlencoded string for final path.
980
# pad.lv/1696545: For the moment, accept both native strings and unicode.
981
if isinstance(relpath, str):
983
elif isinstance(relpath, text_type):
985
relpath = relpath.encode()
986
except UnicodeEncodeError:
987
raise InvalidURL(relpath)
989
raise InvalidURL(relpath)
990
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
991
if relpath.startswith('/'):
994
base_parts = base_path.split('/')
995
if len(base_parts) > 0 and base_parts[-1] == '':
996
base_parts = base_parts[:-1]
997
for p in relpath.split('/'):
999
if len(base_parts) == 0:
1000
# In most filesystems, a request for the parent
1001
# of root, just returns root.
1007
base_parts.append(p)
1008
path = '/'.join(base_parts)
1009
if not path.startswith('/'):
1013
def clone(self, offset=None):
1014
"""Return a new URL for a path relative to this URL.
1016
:param offset: A relative path, already urlencoded
1017
:return: `URL` instance
1019
if offset is not None:
1020
relative = unescape(offset)
1021
if sys.version_info[0] == 2:
1022
relative = relative.encode('utf-8')
1023
path = self._combine_paths(self.path, relative)
1024
path = quote(path, safe="/~")
1026
path = self.quoted_path
1027
return self.__class__(self.scheme, self.quoted_user,
1028
self.quoted_password, self.quoted_host, self.port,
1033
"""Extract the server address, the credentials and the path from the url.
1035
user, password, host and path should be quoted if they contain reserved
1038
:param url: an quoted url
1039
:return: (scheme, user, password, host, port, path) tuple, all fields
1042
parsed_url = URL.from_string(url)
1043
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
1044
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
breezy/urlutils.py
