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
48
class InvalidURL(errors.PathError):
50
_fmt = 'Invalid url supplied to transport: "%(path)s"%(extra)s'
53
class InvalidURLJoin(errors.PathError):
55
_fmt = "Invalid URL join request: %(reason)s: %(base)r + %(join_args)r"
57
def __init__(self, reason, base, join_args):
60
self.join_args = join_args
61
errors.PathError.__init__(self, base, reason)
64
class InvalidRebaseURLs(errors.PathError):
66
_fmt = "URLs differ by more than path: %(from_)r and %(to)r"
68
def __init__(self, from_, to):
71
errors.PathError.__init__(
72
self, from_, 'URLs differ by more than path.')
75
def basename(url, exclude_trailing_slash=True):
76
"""Return the last component of a URL.
78
:param url: The URL in question
79
:param exclude_trailing_slash: If the url looks like "path/to/foo/"
80
ignore the final slash and return 'foo' rather than ''
81
:return: Just the final component of the URL. This can return ''
82
if you don't exclude_trailing_slash, or if you are at the
85
return split(url, exclude_trailing_slash=exclude_trailing_slash)[1]
88
def dirname(url, exclude_trailing_slash=True):
89
"""Return the parent directory of the given path.
91
:param url: Relative or absolute URL
92
:param exclude_trailing_slash: Remove a final slash
93
(treat http://host/foo/ as http://host/foo, but
94
http://host/ stays http://host/)
95
:return: Everything in the URL except the last path chunk
97
# TODO: jam 20060502 This was named dirname to be consistent
98
# with the os functions, but maybe "parent" would be better
99
return split(url, exclude_trailing_slash=exclude_trailing_slash)[0]
103
quote_from_bytes = urlparse.quote_from_bytes
104
quote = urlparse.quote
105
unquote_to_bytes = urlparse.unquote_to_bytes
107
# Private copies of quote and unquote, copied from Python's urllib module
108
# because urllib unconditionally imports socket, which imports ssl.
110
always_safe = ('ABCDEFGHIJKLMNOPQRSTUVWXYZ'
111
'abcdefghijklmnopqrstuvwxyz'
114
for i, c in zip(range(256), ''.join(map(chr, range(256)))):
115
_safe_map[c] = c if (
116
i < 128 and c in always_safe) else '%{0:02X}'.format(i)
119
def quote_from_bytes(s, safe='/'):
120
"""quote('abc def') -> 'abc%20def'
122
Each part of a URL, e.g. the path info, the query, etc., has a
123
different set of reserved characters that must be quoted.
125
RFC 2396 Uniform Resource Identifiers (URI): Generic Syntax lists
126
the following reserved characters.
128
reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" |
131
Each of these characters is reserved in some component of a URL,
132
but not necessarily in all of them.
134
By default, the quote function is intended for quoting the path
135
section of a URL. Thus, it will not encode '/'. This character
136
is reserved, but in typical usage the quote function is being
137
called on a path where the existing slash characters are used as
143
raise TypeError('None object cannot be quoted')
145
cachekey = (safe, always_safe)
147
(quoter, safe) = _safe_quoters[cachekey]
149
safe_map = _safe_map.copy()
150
safe_map.update([(c, c) for c in safe])
151
quoter = safe_map.__getitem__
152
safe = always_safe + safe
153
_safe_quoters[cachekey] = (quoter, safe)
154
if not s.rstrip(safe):
156
return ''.join(map(quoter, s))
158
quote = quote_from_bytes
159
unquote_to_bytes = urlparse.unquote
162
unquote = urlparse.unquote
165
def escape(relpath, safe='/~'):
166
"""Escape relpath to be a valid url."""
167
if not isinstance(relpath, str) and sys.version_info[0] == 2:
168
# GZ 2019-06-16: Should use _fs_enc instead here really?
169
relpath = relpath.encode('utf-8')
170
return quote(relpath, safe=safe)
173
def file_relpath(base, path):
174
"""Compute just the relative sub-portion of a url
176
This assumes that both paths are already fully specified file:// URLs.
178
if len(base) < MIN_ABS_FILEURL_LENGTH:
179
raise ValueError('Length of base (%r) must equal or'
180
' exceed the platform minimum url length (which is %d)' %
181
(base, MIN_ABS_FILEURL_LENGTH))
182
base = osutils.normpath(local_path_from_url(base))
183
path = osutils.normpath(local_path_from_url(path))
184
return escape(osutils.relpath(base, path))
187
def _find_scheme_and_separator(url):
188
"""Find the scheme separator (://) and the first path separator
190
This is just a helper functions for other path utilities.
191
It could probably be replaced by urlparse
193
m = _url_scheme_re.match(url)
197
scheme = m.group('scheme')
198
path = m.group('path')
200
# Find the path separating slash
201
# (first slash after the ://)
202
first_path_slash = path.find('/')
203
if first_path_slash == -1:
204
return len(scheme), None
205
return len(scheme), first_path_slash + m.start('path')
209
"""Tests whether a URL is in actual fact a URL."""
210
return _url_scheme_re.match(url) is not None
213
def join(base, *args):
214
"""Create a URL by joining sections.
216
This will normalize '..', assuming that paths are absolute
217
(it assumes no symlinks in either path)
219
If any of *args is an absolute URL, it will be treated correctly.
221
join('http://foo', 'http://bar') => 'http://bar'
222
join('http://foo', 'bar') => 'http://foo/bar'
223
join('http://foo', 'bar', '../baz') => 'http://foo/baz'
227
scheme_end, path_start = _find_scheme_and_separator(base)
228
if scheme_end is None and path_start is None:
230
elif path_start is None:
231
path_start = len(base)
232
path = base[path_start:]
234
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
235
if arg_scheme_end is None and arg_path_start is None:
237
elif arg_path_start is None:
238
arg_path_start = len(arg)
239
if arg_scheme_end is not None:
241
path = arg[arg_path_start:]
242
scheme_end = arg_scheme_end
243
path_start = arg_path_start
245
path = joinpath(path, arg)
246
return base[:path_start] + path
249
def joinpath(base, *args):
250
"""Join URL path segments to a URL path segment.
252
This is somewhat like osutils.joinpath, but intended for URLs.
254
XXX: this duplicates some normalisation logic, and also duplicates a lot of
255
path handling logic that already exists in some Transport implementations.
256
We really should try to have exactly one place in the code base responsible
257
for combining paths of URLs.
259
path = base.split('/')
260
if len(path) > 1 and path[-1] == '':
261
# If the path ends in a trailing /, remove it.
264
if arg.startswith('/'):
266
for chunk in arg.split('/'):
271
raise InvalidURLJoin('Cannot go above root',
279
return '/'.join(path)
282
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
283
def _posix_local_path_from_url(url):
284
"""Convert a url like file:///path/to/foo into /path/to/foo"""
285
url = strip_segment_parameters(url)
286
file_localhost_prefix = 'file://localhost/'
287
if url.startswith(file_localhost_prefix):
288
path = url[len(file_localhost_prefix) - 1:]
289
elif not url.startswith('file:///'):
291
url, 'local urls must start with file:/// or file://localhost/')
293
path = url[len('file://'):]
294
# We only strip off 2 slashes
295
return unescape(path)
298
def _posix_local_path_to_url(path):
299
"""Convert a local path like ./foo into a URL like file:///path/to/foo
301
This also handles transforming escaping unicode characters, etc.
303
# importing directly from posixpath allows us to test this
304
# on non-posix platforms
305
return 'file://' + escape(osutils._posix_abspath(path))
308
def _win32_local_path_from_url(url):
309
"""Convert a url like file:///C:/path/to/foo into C:/path/to/foo"""
310
if not url.startswith('file://'):
311
raise InvalidURL(url, 'local urls must start with file:///, '
312
'UNC path urls must start with file://')
313
url = strip_segment_parameters(url)
314
# We strip off all 3 slashes
315
win32_url = url[len('file:'):]
316
# check for UNC path: //HOST/path
317
if not win32_url.startswith('///'):
318
if (win32_url[2] == '/'
319
or win32_url[3] in '|:'):
320
raise InvalidURL(url, 'Win32 UNC path urls'
321
' have form file://HOST/path')
322
return unescape(win32_url)
324
# allow empty paths so we can serve all roots
325
if win32_url == '///':
328
# usual local path with drive letter
329
if (len(win32_url) < 6
330
or win32_url[3] not in ('abcdefghijklmnopqrstuvwxyz'
331
'ABCDEFGHIJKLMNOPQRSTUVWXYZ') or
332
win32_url[4] not in '|:'
333
or win32_url[5] != '/'):
334
raise InvalidURL(url, 'Win32 file urls start with'
335
' file:///x:/, where x is a valid drive letter')
336
return win32_url[3].upper() + u':' + unescape(win32_url[5:])
339
def _win32_local_path_to_url(path):
340
"""Convert a local path like ./foo into a URL like file:///C:/path/to/foo
342
This also handles transforming escaping unicode characters, etc.
344
# importing directly from ntpath allows us to test this
345
# on non-win32 platform
346
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
347
# which actually strips trailing space characters.
348
# The worst part is that on linux ntpath.abspath has different
349
# semantics, since 'nt' is not an available module.
353
win32_path = osutils._win32_abspath(path)
354
# check for UNC path \\HOST\path
355
if win32_path.startswith('//'):
356
return 'file:' + escape(win32_path)
357
return ('file:///' + str(win32_path[0].upper()) + ':' +
358
escape(win32_path[2:]))
361
local_path_to_url = _posix_local_path_to_url
362
local_path_from_url = _posix_local_path_from_url
363
MIN_ABS_FILEURL_LENGTH = len('file:///')
364
WIN32_MIN_ABS_FILEURL_LENGTH = len('file:///C:/')
366
if sys.platform == 'win32':
367
local_path_to_url = _win32_local_path_to_url
368
local_path_from_url = _win32_local_path_from_url
370
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH
373
_url_scheme_re = re.compile('^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$')
374
_url_hex_escapes_re = re.compile('(%[0-9a-fA-F]{2})')
377
def _unescape_safe_chars(matchobj):
378
"""re.sub callback to convert hex-escapes to plain characters (if safe).
380
e.g. '%7E' will be converted to '~'.
382
hex_digits = matchobj.group(0)[1:]
383
char = chr(int(hex_digits, 16))
384
if char in _url_dont_escape_characters:
387
return matchobj.group(0).upper()
390
def normalize_url(url):
391
"""Make sure that a path string is in fully normalized URL form.
393
This handles URLs which have unicode characters, spaces,
394
special characters, etc.
396
It has two basic modes of operation, depending on whether the
397
supplied string starts with a url specifier (scheme://) or not.
398
If it does not have a specifier it is considered a local path,
399
and will be converted into a file:/// url. Non-ascii characters
400
will be encoded using utf-8.
401
If it does have a url specifier, it will be treated as a "hybrid"
402
URL. Basically, a URL that should have URL special characters already
403
escaped (like +?&# etc), but may have unicode characters, etc
404
which would not be valid in a real URL.
406
:param url: Either a hybrid URL or a local path
407
:return: A normalized URL which only includes 7-bit ASCII characters.
409
scheme_end, path_start = _find_scheme_and_separator(url)
410
if scheme_end is None:
411
return local_path_to_url(url)
412
prefix = url[:path_start]
413
path = url[path_start:]
414
if not isinstance(url, text_type):
416
if c not in _url_safe_characters:
417
raise InvalidURL(url, 'URLs can only contain specific'
418
' safe characters (not %r)' % c)
419
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
420
return str(prefix + ''.join(path))
422
# We have a unicode (hybrid) url
423
path_chars = list(path)
425
for i in range(len(path_chars)):
426
if path_chars[i] not in _url_safe_characters:
427
path_chars[i] = ''.join(
428
['%%%02X' % c for c in bytearray(path_chars[i].encode('utf-8'))])
429
path = ''.join(path_chars)
430
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
431
return str(prefix + path)
434
def relative_url(base, other):
435
"""Return a path to other from base.
437
If other is unrelated to base, return other. Else return a relative path.
438
This assumes no symlinks as part of the url.
440
dummy, base_first_slash = _find_scheme_and_separator(base)
441
if base_first_slash is None:
444
dummy, other_first_slash = _find_scheme_and_separator(other)
445
if other_first_slash is None:
448
# this takes care of differing schemes or hosts
449
base_scheme = base[:base_first_slash]
450
other_scheme = other[:other_first_slash]
451
if base_scheme != other_scheme:
453
elif sys.platform == 'win32' and base_scheme == 'file://':
454
base_drive = base[base_first_slash + 1:base_first_slash + 3]
455
other_drive = other[other_first_slash + 1:other_first_slash + 3]
456
if base_drive != other_drive:
459
base_path = base[base_first_slash + 1:]
460
other_path = other[other_first_slash + 1:]
462
if base_path.endswith('/'):
463
base_path = base_path[:-1]
465
base_sections = base_path.split('/')
466
other_sections = other_path.split('/')
468
if base_sections == ['']:
470
if other_sections == ['']:
474
for b, o in zip(base_sections, other_sections):
477
output_sections.append(b)
479
match_len = len(output_sections)
480
output_sections = ['..' for x in base_sections[match_len:]]
481
output_sections.extend(other_sections[match_len:])
483
return "/".join(output_sections) or "."
486
def _win32_extract_drive_letter(url_base, path):
487
"""On win32 the drive letter needs to be added to the url base."""
488
# Strip off the drive letter
489
# path is currently /C:/foo
490
if len(path) < 4 or path[2] not in ':|' or path[3] != '/':
491
raise InvalidURL(url_base + path,
492
'win32 file:/// paths need a drive letter')
493
url_base += path[0:3] # file:// + /C:
494
path = path[3:] # /foo
495
return url_base, path
498
def split(url, exclude_trailing_slash=True):
499
"""Split a URL into its parent directory and a child directory.
501
:param url: A relative or absolute URL
502
:param exclude_trailing_slash: Strip off a final '/' if it is part
503
of the path (but not if it is part of the protocol specification)
505
:return: (parent_url, child_dir). child_dir may be the empty string if
508
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
510
if first_path_slash is None:
511
# We have either a relative path, or no separating slash
512
if scheme_loc is None:
514
if exclude_trailing_slash and url.endswith('/'):
516
return _posix_split(url)
518
# Scheme with no path
521
# We have a fully defined path
522
url_base = url[:first_path_slash] # http://host, file://
523
path = url[first_path_slash:] # /file/foo
525
if sys.platform == 'win32' and url.startswith('file:///'):
526
# Strip off the drive letter
527
# url_base is currently file://
528
# path is currently /C:/foo
529
url_base, path = _win32_extract_drive_letter(url_base, path)
530
# now it should be file:///C: and /foo
532
if exclude_trailing_slash and len(path) > 1 and path.endswith('/'):
534
head, tail = _posix_split(path)
535
return url_base + head, tail
538
def split_segment_parameters_raw(url):
539
"""Split the subsegment of the last segment of a URL.
541
:param url: A relative or absolute URL
542
:return: (url, subsegments)
544
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
545
# operates on urls not url+segments, and Transport classes
546
# should not be blindly adding slashes in the first place.
547
lurl = strip_trailing_slash(url)
548
# Segments begin at first comma after last forward slash, if one exists
549
segment_start = lurl.find(",", lurl.rfind("/") + 1)
550
if segment_start == -1:
552
return (lurl[:segment_start],
553
[str(s) for s in lurl[segment_start + 1:].split(",")])
556
def split_segment_parameters(url):
557
"""Split the segment parameters of the last segment of a URL.
559
:param url: A relative or absolute URL
560
:return: (url, segment_parameters)
562
(base_url, subsegments) = split_segment_parameters_raw(url)
564
for subsegment in subsegments:
566
(key, value) = subsegment.split("=", 1)
568
raise InvalidURL(url, "missing = in subsegment")
569
if not isinstance(key, str):
571
if not isinstance(value, str):
572
raise TypeError(value)
573
parameters[key] = value
574
return (base_url, parameters)
577
def strip_segment_parameters(url):
578
"""Strip the segment parameters from a URL.
580
:param url: A relative or absolute URL
583
base_url, subsegments = split_segment_parameters_raw(url)
587
def join_segment_parameters_raw(base, *subsegments):
588
"""Create a new URL by adding subsegments to an existing one.
590
This adds the specified subsegments to the last path in the specified
591
base URL. The subsegments should be bytestrings.
593
:note: You probably want to use join_segment_parameters instead.
597
for subsegment in subsegments:
598
if not isinstance(subsegment, str):
599
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
600
if "," in subsegment:
601
raise InvalidURLJoin(", exists in subsegments",
603
return ",".join((base,) + subsegments)
606
def join_segment_parameters(url, parameters):
607
"""Create a new URL by adding segment parameters to an existing one.
609
The parameters of the last segment in the URL will be updated; if a
610
parameter with the same key already exists it will be overwritten.
612
:param url: A URL, as string
613
:param parameters: Dictionary of parameters, keys and values as bytestrings
615
(base, existing_parameters) = split_segment_parameters(url)
617
new_parameters.update(existing_parameters)
618
for key, value in parameters.items():
619
if not isinstance(key, str):
620
raise TypeError("parameter key %r is not a str" % key)
621
if not isinstance(value, str):
622
raise TypeError("parameter value %r for %r is not a str" %
625
raise InvalidURLJoin("= exists in parameter key", url,
627
new_parameters[key] = value
628
return join_segment_parameters_raw(
629
base, *["%s=%s" % item for item in sorted(new_parameters.items())])
632
def _win32_strip_local_trailing_slash(url):
633
"""Strip slashes after the drive letter"""
634
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
640
def strip_trailing_slash(url):
641
"""Strip trailing slash, except for root paths.
643
The definition of 'root path' is platform-dependent.
644
This assumes that all URLs are valid netloc urls, such that they
647
It searches for ://, and then refuses to remove the next '/'.
648
It can also handle relative paths
650
path/to/foo => path/to/foo
651
path/to/foo/ => path/to/foo
652
http://host/path/ => http://host/path
653
http://host/path => http://host/path
654
http://host/ => http://host/
656
file:///foo/ => file:///foo
657
# This is unique on win32 platforms, and is the only URL
658
# format which does it differently.
659
file:///c|/ => file:///c:/
661
if not url.endswith('/'):
664
if sys.platform == 'win32' and url.startswith('file://'):
665
return _win32_strip_local_trailing_slash(url)
667
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
668
if scheme_loc is None:
669
# This is a relative path, as it has no scheme
670
# so just chop off the last character
673
if first_path_slash is None or first_path_slash == len(url) - 1:
674
# Don't chop off anything if the only slash is the path
682
"""Unescape relpath from url format.
684
This returns a Unicode path from a URL
686
# jam 20060427 URLs are supposed to be ASCII only strings
687
# If they are passed in as unicode, unquote
688
# will return a UNICODE string, which actually contains
689
# utf-8 bytes. So we have to ensure that they are
690
# plain ASCII strings, or the final .decode will
691
# try to encode the UNICODE => ASCII, and then decode
695
if isinstance(url, text_type):
698
except UnicodeError as e:
700
url, 'URL was not a plain ASCII url: %s' % (e,))
701
return urlparse.unquote(url)
703
if isinstance(url, text_type):
705
url = url.encode("ascii")
706
except UnicodeError as e:
708
url, 'URL was not a plain ASCII url: %s' % (e,))
709
unquoted = unquote(url)
711
unicode_path = unquoted.decode('utf-8')
712
except UnicodeError as e:
714
url, 'Unable to encode the URL as utf-8: %s' % (e,))
718
# These are characters that if escaped, should stay that way
719
_no_decode_chars = ';/?:@&=+$,#'
720
_no_decode_ords = [ord(c) for c in _no_decode_chars]
721
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
722
+ ['%02X' % o for o in _no_decode_ords])
723
_hex_display_map = dict(([('%02x' % o, int2byte(o)) for o in range(256)]
724
+ [('%02X' % o, int2byte(o)) for o in range(256)]))
725
# These entries get mapped to themselves
726
_hex_display_map.update((hex, b'%' + hex.encode('ascii'))
727
for hex in _no_decode_hex)
729
# These characters shouldn't be percent-encoded, and it's always safe to
730
# unencode them if they are.
731
_url_dont_escape_characters = set(
732
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
733
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
734
"0123456789" # Numbers
735
"-._~" # Unreserved characters
738
# These characters should not be escaped
739
_url_safe_characters = set(
740
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
741
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
742
"0123456789" # Numbers
743
"_.-!~*'()" # Unreserved characters
744
"/;?:@&=+$," # Reserved characters
745
"%#" # Extra reserved characters
749
def _unescape_segment_for_display(segment, encoding):
750
"""Unescape a segment for display.
752
Helper for unescape_for_display
754
:param url: A 7-bit ASCII URL
755
:param encoding: The final output encoding
757
:return: A unicode string which can be safely encoded into the
760
escaped_chunks = segment.split('%')
761
escaped_chunks[0] = escaped_chunks[0].encode('utf-8')
762
for j in range(1, len(escaped_chunks)):
763
item = escaped_chunks[j]
765
escaped_chunks[j] = _hex_display_map[item[:2]]
767
# Put back the percent symbol
768
escaped_chunks[j] = b'%' + \
769
(item[:2].encode('utf-8') if PY3 else item[:2])
770
except UnicodeDecodeError:
771
escaped_chunks[j] = unichr(int(item[:2], 16)).encode('utf-8')
772
escaped_chunks[j] += (item[2:].encode('utf-8') if PY3 else item[2:])
773
unescaped = b''.join(escaped_chunks)
775
decoded = unescaped.decode('utf-8')
776
except UnicodeDecodeError:
777
# If this path segment cannot be properly utf-8 decoded
778
# after doing unescaping we will just leave it alone
782
decoded.encode(encoding)
783
except UnicodeEncodeError:
784
# If this chunk cannot be encoded in the local
785
# encoding, then we should leave it alone
788
# Otherwise take the url decoded one
792
def unescape_for_display(url, encoding):
793
"""Decode what you can for a URL, so that we get a nice looking path.
795
This will turn file:// urls into local paths, and try to decode
796
any portions of a http:// style url that it can.
798
Any sections of the URL which can't be represented in the encoding or
799
need to stay as escapes are left alone.
801
:param url: A 7-bit ASCII URL
802
:param encoding: The final output encoding
804
:return: A unicode string which can be safely encoded into the
808
raise ValueError('you cannot specify None for the display encoding')
809
if url.startswith('file://'):
811
path = local_path_from_url(url)
812
path.encode(encoding)
817
# Split into sections to try to decode utf-8
819
for i in range(1, len(res)):
820
res[i] = _unescape_segment_for_display(res[i], encoding)
821
return u'/'.join(res)
824
def derive_to_location(from_location):
825
"""Derive a TO_LOCATION given a FROM_LOCATION.
827
The normal case is a FROM_LOCATION of http://foo/bar => bar.
828
The Right Thing for some logical destinations may differ though
829
because no / may be present at all. In that case, the result is
830
the full name without the scheme indicator, e.g. lp:foo-bar => foo-bar.
831
This latter case also applies when a Windows drive
832
is used without a path, e.g. c:foo-bar => foo-bar.
833
If no /, path separator or : is found, the from_location is returned.
835
from_location = strip_segment_parameters(from_location)
836
if from_location.find("/") >= 0 or from_location.find(os.sep) >= 0:
837
return os.path.basename(from_location.rstrip("/\\"))
839
sep = from_location.find(":")
841
return from_location[sep + 1:]
846
def _is_absolute(url):
847
return (osutils.pathjoin('/foo', url) == url)
850
def rebase_url(url, old_base, new_base):
851
"""Convert a relative path from an old base URL to a new base URL.
853
The result will be a relative path.
854
Absolute paths and full URLs are returned unaltered.
856
scheme, separator = _find_scheme_and_separator(url)
857
if scheme is not None:
859
if _is_absolute(url):
861
old_parsed = urlparse.urlparse(old_base)
862
new_parsed = urlparse.urlparse(new_base)
863
if (old_parsed[:2]) != (new_parsed[:2]):
864
raise InvalidRebaseURLs(old_base, new_base)
865
return determine_relative_path(new_parsed[2],
866
join(old_parsed[2], url))
869
def determine_relative_path(from_path, to_path):
870
"""Determine a relative path from from_path to to_path."""
871
from_segments = osutils.splitpath(from_path)
872
to_segments = osutils.splitpath(to_path)
874
for count, (from_element, to_element) in enumerate(zip(from_segments,
876
if from_element != to_element:
880
unique_from = from_segments[count:]
881
unique_to = to_segments[count:]
882
segments = (['..'] * len(unique_from) + unique_to)
883
if len(segments) == 0:
885
return osutils.pathjoin(*segments)
891
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
894
self.quoted_host = quoted_host
895
self.host = unquote(self.quoted_host)
896
self.quoted_user = quoted_user
897
if self.quoted_user is not None:
898
self.user = unquote(self.quoted_user)
901
self.quoted_password = quoted_password
902
if self.quoted_password is not None:
903
self.password = unquote(self.quoted_password)
907
self.quoted_path = _url_hex_escapes_re.sub(
908
_unescape_safe_chars, quoted_path)
909
self.path = unquote(self.quoted_path)
911
def __eq__(self, other):
912
return (isinstance(other, self.__class__) and
913
self.scheme == other.scheme and
914
self.host == other.host and
915
self.user == other.user and
916
self.password == other.password and
917
self.path == other.path)
920
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
921
self.__class__.__name__,
922
self.scheme, self.quoted_user, self.quoted_password,
923
self.quoted_host, self.port, self.quoted_path)
926
def from_string(cls, url):
927
"""Create a URL object from a string.
929
:param url: URL as bytestring
931
# GZ 2017-06-09: Actually validate ascii-ness
932
# pad.lv/1696545: For the moment, accept both native strings and
934
if isinstance(url, str):
936
elif isinstance(url, text_type):
939
except UnicodeEncodeError:
940
raise InvalidURL(url)
942
raise InvalidURL(url)
943
(scheme, netloc, path, params,
944
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
945
user = password = host = port = None
947
user, host = netloc.rsplit('@', 1)
949
user, password = user.split(':', 1)
953
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
955
host, port = host.rsplit(':', 1)
960
raise InvalidURL('invalid port number %s in url:\n%s' %
964
if host != "" and host[0] == '[' and host[-1] == ']': # IPv6
967
return cls(scheme, user, password, host, port, path)
970
netloc = self.quoted_host
972
netloc = "[%s]" % netloc
973
if self.quoted_user is not None:
974
# Note that we don't put the password back even if we
975
# have one so that it doesn't get accidentally
977
netloc = '%s@%s' % (self.quoted_user, netloc)
978
if self.port is not None:
979
netloc = '%s:%d' % (netloc, self.port)
980
return urlparse.urlunparse(
981
(self.scheme, netloc, self.quoted_path, None, None, None))
984
def _combine_paths(base_path, relpath):
985
"""Transform a Transport-relative path to a remote absolute path.
987
This does not handle substitution of ~ but does handle '..' and '.'
992
t._combine_paths('/home/sarah', 'project/foo')
993
=> '/home/sarah/project/foo'
994
t._combine_paths('/home/sarah', '../../etc')
996
t._combine_paths('/home/sarah', '/etc')
999
:param base_path: base path
1000
:param relpath: relative url string for relative part of remote path.
1001
:return: urlencoded string for final path.
1003
# pad.lv/1696545: For the moment, accept both native strings and
1005
if isinstance(relpath, str):
1007
elif isinstance(relpath, text_type):
1009
relpath = relpath.encode()
1010
except UnicodeEncodeError:
1011
raise InvalidURL(relpath)
1013
raise InvalidURL(relpath)
1014
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
1015
if relpath.startswith('/'):
1018
base_parts = base_path.split('/')
1019
if len(base_parts) > 0 and base_parts[-1] == '':
1020
base_parts = base_parts[:-1]
1021
for p in relpath.split('/'):
1023
if len(base_parts) == 0:
1024
# In most filesystems, a request for the parent
1025
# of root, just returns root.
1031
base_parts.append(p)
1032
path = '/'.join(base_parts)
1033
if not path.startswith('/'):
1037
def clone(self, offset=None):
1038
"""Return a new URL for a path relative to this URL.
1040
:param offset: A relative path, already urlencoded
1041
:return: `URL` instance
1043
if offset is not None:
1044
relative = unescape(offset)
1045
if sys.version_info[0] == 2:
1046
relative = relative.encode('utf-8')
1047
path = self._combine_paths(self.path, relative)
1048
path = quote(path, safe="/~")
1050
path = self.quoted_path
1051
return self.__class__(self.scheme, self.quoted_user,
1052
self.quoted_password, self.quoted_host, self.port,
1057
"""Extract the server address, the credentials and the path from the url.
1059
user, password, host and path should be quoted if they contain reserved
1062
:param url: an quoted url
1063
:return: (scheme, user, password, host, port, path) tuple, all fields
1066
parsed_url = URL.from_string(url)
1067
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
1068
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
breezy/urlutils.py
