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
166
"""Escape relpath to be a valid url."""
167
if not isinstance(relpath, str) and sys.version_info[0] == 2:
168
relpath = relpath.encode('utf-8')
169
return quote(relpath, safe='/~')
172
def file_relpath(base, path):
173
"""Compute just the relative sub-portion of a url
175
This assumes that both paths are already fully specified file:// URLs.
177
if len(base) < MIN_ABS_FILEURL_LENGTH:
178
raise ValueError('Length of base (%r) must equal or'
179
' exceed the platform minimum url length (which is %d)' %
180
(base, MIN_ABS_FILEURL_LENGTH))
181
base = osutils.normpath(local_path_from_url(base))
182
path = osutils.normpath(local_path_from_url(path))
183
return escape(osutils.relpath(base, path))
186
def _find_scheme_and_separator(url):
187
"""Find the scheme separator (://) and the first path separator
189
This is just a helper functions for other path utilities.
190
It could probably be replaced by urlparse
192
m = _url_scheme_re.match(url)
196
scheme = m.group('scheme')
197
path = m.group('path')
199
# Find the path separating slash
200
# (first slash after the ://)
201
first_path_slash = path.find('/')
202
if first_path_slash == -1:
203
return len(scheme), None
204
return len(scheme), first_path_slash + m.start('path')
208
"""Tests whether a URL is in actual fact a URL."""
209
return _url_scheme_re.match(url) is not None
212
def join(base, *args):
213
"""Create a URL by joining sections.
215
This will normalize '..', assuming that paths are absolute
216
(it assumes no symlinks in either path)
218
If any of *args is an absolute URL, it will be treated correctly.
220
join('http://foo', 'http://bar') => 'http://bar'
221
join('http://foo', 'bar') => 'http://foo/bar'
222
join('http://foo', 'bar', '../baz') => 'http://foo/baz'
226
scheme_end, path_start = _find_scheme_and_separator(base)
227
if scheme_end is None and path_start is None:
229
elif path_start is None:
230
path_start = len(base)
231
path = base[path_start:]
233
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg)
234
if arg_scheme_end is None and arg_path_start is None:
236
elif arg_path_start is None:
237
arg_path_start = len(arg)
238
if arg_scheme_end is not None:
240
path = arg[arg_path_start:]
241
scheme_end = arg_scheme_end
242
path_start = arg_path_start
244
path = joinpath(path, arg)
245
return base[:path_start] + path
248
def joinpath(base, *args):
249
"""Join URL path segments to a URL path segment.
251
This is somewhat like osutils.joinpath, but intended for URLs.
253
XXX: this duplicates some normalisation logic, and also duplicates a lot of
254
path handling logic that already exists in some Transport implementations.
255
We really should try to have exactly one place in the code base responsible
256
for combining paths of URLs.
258
path = base.split('/')
259
if len(path) > 1 and path[-1] == '':
260
# If the path ends in a trailing /, remove it.
263
if arg.startswith('/'):
265
for chunk in arg.split('/'):
270
raise InvalidURLJoin('Cannot go above root',
278
return '/'.join(path)
281
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
282
def _posix_local_path_from_url(url):
283
"""Convert a url like file:///path/to/foo into /path/to/foo"""
284
url = split_segment_parameters_raw(url)[0]
285
file_localhost_prefix = 'file://localhost/'
286
if url.startswith(file_localhost_prefix):
287
path = url[len(file_localhost_prefix) - 1:]
288
elif not url.startswith('file:///'):
290
url, 'local urls must start with file:/// or file://localhost/')
292
path = url[len('file://'):]
293
# We only strip off 2 slashes
294
return unescape(path)
297
def _posix_local_path_to_url(path):
298
"""Convert a local path like ./foo into a URL like file:///path/to/foo
300
This also handles transforming escaping unicode characters, etc.
302
# importing directly from posixpath allows us to test this
303
# on non-posix platforms
304
return 'file://' + escape(osutils._posix_abspath(path))
307
def _win32_local_path_from_url(url):
308
"""Convert a url like file:///C:/path/to/foo into C:/path/to/foo"""
309
if not url.startswith('file://'):
310
raise InvalidURL(url, 'local urls must start with file:///, '
311
'UNC path urls must start with file://')
312
url = split_segment_parameters_raw(url)[0]
313
# We strip off all 3 slashes
314
win32_url = url[len('file:'):]
315
# check for UNC path: //HOST/path
316
if not win32_url.startswith('///'):
317
if (win32_url[2] == '/'
318
or win32_url[3] in '|:'):
319
raise InvalidURL(url, 'Win32 UNC path urls'
320
' have form file://HOST/path')
321
return unescape(win32_url)
323
# allow empty paths so we can serve all roots
324
if win32_url == '///':
327
# usual local path with drive letter
328
if (len(win32_url) < 6
329
or win32_url[3] not in ('abcdefghijklmnopqrstuvwxyz'
330
'ABCDEFGHIJKLMNOPQRSTUVWXYZ') or
331
win32_url[4] not in '|:'
332
or win32_url[5] != '/'):
333
raise InvalidURL(url, 'Win32 file urls start with'
334
' file:///x:/, where x is a valid drive letter')
335
return win32_url[3].upper() + u':' + unescape(win32_url[5:])
338
def _win32_local_path_to_url(path):
339
"""Convert a local path like ./foo into a URL like file:///C:/path/to/foo
341
This also handles transforming escaping unicode characters, etc.
343
# importing directly from ntpath allows us to test this
344
# on non-win32 platform
345
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
346
# which actually strips trailing space characters.
347
# The worst part is that on linux ntpath.abspath has different
348
# semantics, since 'nt' is not an available module.
352
win32_path = osutils._win32_abspath(path)
353
# check for UNC path \\HOST\path
354
if win32_path.startswith('//'):
355
return 'file:' + escape(win32_path)
356
return ('file:///' + str(win32_path[0].upper()) + ':' +
357
escape(win32_path[2:]))
360
local_path_to_url = _posix_local_path_to_url
361
local_path_from_url = _posix_local_path_from_url
362
MIN_ABS_FILEURL_LENGTH = len('file:///')
363
WIN32_MIN_ABS_FILEURL_LENGTH = len('file:///C:/')
365
if sys.platform == 'win32':
366
local_path_to_url = _win32_local_path_to_url
367
local_path_from_url = _win32_local_path_from_url
369
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH
372
_url_scheme_re = re.compile('^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$')
373
_url_hex_escapes_re = re.compile('(%[0-9a-fA-F]{2})')
376
def _unescape_safe_chars(matchobj):
377
"""re.sub callback to convert hex-escapes to plain characters (if safe).
379
e.g. '%7E' will be converted to '~'.
381
hex_digits = matchobj.group(0)[1:]
382
char = chr(int(hex_digits, 16))
383
if char in _url_dont_escape_characters:
386
return matchobj.group(0).upper()
389
def normalize_url(url):
390
"""Make sure that a path string is in fully normalized URL form.
392
This handles URLs which have unicode characters, spaces,
393
special characters, etc.
395
It has two basic modes of operation, depending on whether the
396
supplied string starts with a url specifier (scheme://) or not.
397
If it does not have a specifier it is considered a local path,
398
and will be converted into a file:/// url. Non-ascii characters
399
will be encoded using utf-8.
400
If it does have a url specifier, it will be treated as a "hybrid"
401
URL. Basically, a URL that should have URL special characters already
402
escaped (like +?&# etc), but may have unicode characters, etc
403
which would not be valid in a real URL.
405
:param url: Either a hybrid URL or a local path
406
:return: A normalized URL which only includes 7-bit ASCII characters.
408
scheme_end, path_start = _find_scheme_and_separator(url)
409
if scheme_end is None:
410
return local_path_to_url(url)
411
prefix = url[:path_start]
412
path = url[path_start:]
413
if not isinstance(url, text_type):
415
if c not in _url_safe_characters:
416
raise InvalidURL(url, 'URLs can only contain specific'
417
' safe characters (not %r)' % c)
418
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
419
return str(prefix + ''.join(path))
421
# We have a unicode (hybrid) url
422
path_chars = list(path)
424
for i in range(len(path_chars)):
425
if path_chars[i] not in _url_safe_characters:
426
path_chars[i] = ''.join(
427
['%%%02X' % c for c in bytearray(path_chars[i].encode('utf-8'))])
428
path = ''.join(path_chars)
429
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path)
430
return str(prefix + path)
433
def relative_url(base, other):
434
"""Return a path to other from base.
436
If other is unrelated to base, return other. Else return a relative path.
437
This assumes no symlinks as part of the url.
439
dummy, base_first_slash = _find_scheme_and_separator(base)
440
if base_first_slash is None:
443
dummy, other_first_slash = _find_scheme_and_separator(other)
444
if other_first_slash is None:
447
# this takes care of differing schemes or hosts
448
base_scheme = base[:base_first_slash]
449
other_scheme = other[:other_first_slash]
450
if base_scheme != other_scheme:
452
elif sys.platform == 'win32' and base_scheme == 'file://':
453
base_drive = base[base_first_slash + 1:base_first_slash + 3]
454
other_drive = other[other_first_slash + 1:other_first_slash + 3]
455
if base_drive != other_drive:
458
base_path = base[base_first_slash + 1:]
459
other_path = other[other_first_slash + 1:]
461
if base_path.endswith('/'):
462
base_path = base_path[:-1]
464
base_sections = base_path.split('/')
465
other_sections = other_path.split('/')
467
if base_sections == ['']:
469
if other_sections == ['']:
473
for b, o in zip(base_sections, other_sections):
476
output_sections.append(b)
478
match_len = len(output_sections)
479
output_sections = ['..' for x in base_sections[match_len:]]
480
output_sections.extend(other_sections[match_len:])
482
return "/".join(output_sections) or "."
485
def _win32_extract_drive_letter(url_base, path):
486
"""On win32 the drive letter needs to be added to the url base."""
487
# Strip off the drive letter
488
# path is currently /C:/foo
489
if len(path) < 4 or path[2] not in ':|' or path[3] != '/':
490
raise InvalidURL(url_base + path,
491
'win32 file:/// paths need a drive letter')
492
url_base += path[0:3] # file:// + /C:
493
path = path[3:] # /foo
494
return url_base, path
497
def split(url, exclude_trailing_slash=True):
498
"""Split a URL into its parent directory and a child directory.
500
:param url: A relative or absolute URL
501
:param exclude_trailing_slash: Strip off a final '/' if it is part
502
of the path (but not if it is part of the protocol specification)
504
:return: (parent_url, child_dir). child_dir may be the empty string if
507
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
509
if first_path_slash is None:
510
# We have either a relative path, or no separating slash
511
if scheme_loc is None:
513
if exclude_trailing_slash and url.endswith('/'):
515
return _posix_split(url)
517
# Scheme with no path
520
# We have a fully defined path
521
url_base = url[:first_path_slash] # http://host, file://
522
path = url[first_path_slash:] # /file/foo
524
if sys.platform == 'win32' and url.startswith('file:///'):
525
# Strip off the drive letter
526
# url_base is currently file://
527
# path is currently /C:/foo
528
url_base, path = _win32_extract_drive_letter(url_base, path)
529
# now it should be file:///C: and /foo
531
if exclude_trailing_slash and len(path) > 1 and path.endswith('/'):
533
head, tail = _posix_split(path)
534
return url_base + head, tail
537
def split_segment_parameters_raw(url):
538
"""Split the subsegment of the last segment of a URL.
540
:param url: A relative or absolute URL
541
:return: (url, subsegments)
543
# GZ 2011-11-18: Dodgy removing the terminal slash like this, function
544
# operates on urls not url+segments, and Transport classes
545
# should not be blindly adding slashes in the first place.
546
lurl = strip_trailing_slash(url)
547
# Segments begin at first comma after last forward slash, if one exists
548
segment_start = lurl.find(",", lurl.rfind("/") + 1)
549
if segment_start == -1:
551
return (lurl[:segment_start],
552
[str(s) for s in lurl[segment_start + 1:].split(",")])
555
def split_segment_parameters(url):
556
"""Split the segment parameters of the last segment of a URL.
558
:param url: A relative or absolute URL
559
:return: (url, segment_parameters)
561
(base_url, subsegments) = split_segment_parameters_raw(url)
563
for subsegment in subsegments:
565
(key, value) = subsegment.split("=", 1)
567
raise InvalidURL(url, "missing = in subsegment")
568
if not isinstance(key, str):
570
if not isinstance(value, str):
571
raise TypeError(value)
572
parameters[key] = value
573
return (base_url, parameters)
576
def join_segment_parameters_raw(base, *subsegments):
577
"""Create a new URL by adding subsegments to an existing one.
579
This adds the specified subsegments to the last path in the specified
580
base URL. The subsegments should be bytestrings.
582
:note: You probably want to use join_segment_parameters instead.
586
for subsegment in subsegments:
587
if not isinstance(subsegment, str):
588
raise TypeError("Subsegment %r is not a bytestring" % subsegment)
589
if "," in subsegment:
590
raise InvalidURLJoin(", exists in subsegments",
592
return ",".join((base,) + subsegments)
595
def join_segment_parameters(url, parameters):
596
"""Create a new URL by adding segment parameters to an existing one.
598
The parameters of the last segment in the URL will be updated; if a
599
parameter with the same key already exists it will be overwritten.
601
:param url: A URL, as string
602
:param parameters: Dictionary of parameters, keys and values as bytestrings
604
(base, existing_parameters) = split_segment_parameters(url)
606
new_parameters.update(existing_parameters)
607
for key, value in parameters.items():
608
if not isinstance(key, str):
609
raise TypeError("parameter key %r is not a str" % key)
610
if not isinstance(value, str):
611
raise TypeError("parameter value %r for %r is not a str" %
614
raise InvalidURLJoin("= exists in parameter key", url,
616
new_parameters[key] = value
617
return join_segment_parameters_raw(
618
base, *["%s=%s" % item for item in sorted(new_parameters.items())])
621
def _win32_strip_local_trailing_slash(url):
622
"""Strip slashes after the drive letter"""
623
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH:
629
def strip_trailing_slash(url):
630
"""Strip trailing slash, except for root paths.
632
The definition of 'root path' is platform-dependent.
633
This assumes that all URLs are valid netloc urls, such that they
636
It searches for ://, and then refuses to remove the next '/'.
637
It can also handle relative paths
639
path/to/foo => path/to/foo
640
path/to/foo/ => path/to/foo
641
http://host/path/ => http://host/path
642
http://host/path => http://host/path
643
http://host/ => http://host/
645
file:///foo/ => file:///foo
646
# This is unique on win32 platforms, and is the only URL
647
# format which does it differently.
648
file:///c|/ => file:///c:/
650
if not url.endswith('/'):
653
if sys.platform == 'win32' and url.startswith('file://'):
654
return _win32_strip_local_trailing_slash(url)
656
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
657
if scheme_loc is None:
658
# This is a relative path, as it has no scheme
659
# so just chop off the last character
662
if first_path_slash is None or first_path_slash == len(url) - 1:
663
# Don't chop off anything if the only slash is the path
671
"""Unescape relpath from url format.
673
This returns a Unicode path from a URL
675
# jam 20060427 URLs are supposed to be ASCII only strings
676
# If they are passed in as unicode, unquote
677
# will return a UNICODE string, which actually contains
678
# utf-8 bytes. So we have to ensure that they are
679
# plain ASCII strings, or the final .decode will
680
# try to encode the UNICODE => ASCII, and then decode
684
if isinstance(url, text_type):
687
except UnicodeError as e:
689
url, 'URL was not a plain ASCII url: %s' % (e,))
690
return urlparse.unquote(url)
692
if isinstance(url, text_type):
694
url = url.encode("ascii")
695
except UnicodeError as e:
697
url, 'URL was not a plain ASCII url: %s' % (e,))
698
unquoted = unquote(url)
700
unicode_path = unquoted.decode('utf-8')
701
except UnicodeError as e:
703
url, 'Unable to encode the URL as utf-8: %s' % (e,))
707
# These are characters that if escaped, should stay that way
708
_no_decode_chars = ';/?:@&=+$,#'
709
_no_decode_ords = [ord(c) for c in _no_decode_chars]
710
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
711
+ ['%02X' % o for o in _no_decode_ords])
712
_hex_display_map = dict(([('%02x' % o, int2byte(o)) for o in range(256)]
713
+ [('%02X' % o, int2byte(o)) for o in range(256)]))
714
# These entries get mapped to themselves
715
_hex_display_map.update((hex, b'%' + hex.encode('ascii'))
716
for hex in _no_decode_hex)
718
# These characters shouldn't be percent-encoded, and it's always safe to
719
# unencode them if they are.
720
_url_dont_escape_characters = set(
721
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
722
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
723
"0123456789" # Numbers
724
"-._~" # Unreserved characters
727
# These characters should not be escaped
728
_url_safe_characters = set(
729
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha
730
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha
731
"0123456789" # Numbers
732
"_.-!~*'()" # Unreserved characters
733
"/;?:@&=+$," # Reserved characters
734
"%#" # Extra reserved characters
738
def _unescape_segment_for_display(segment, encoding):
739
"""Unescape a segment for display.
741
Helper for unescape_for_display
743
:param url: A 7-bit ASCII URL
744
:param encoding: The final output encoding
746
:return: A unicode string which can be safely encoded into the
749
escaped_chunks = segment.split('%')
750
escaped_chunks[0] = escaped_chunks[0].encode('utf-8')
751
for j in range(1, len(escaped_chunks)):
752
item = escaped_chunks[j]
754
escaped_chunks[j] = _hex_display_map[item[:2]]
756
# Put back the percent symbol
757
escaped_chunks[j] = b'%' + \
758
(item[:2].encode('utf-8') if PY3 else item[:2])
759
except UnicodeDecodeError:
760
escaped_chunks[j] = unichr(int(item[:2], 16)).encode('utf-8')
761
escaped_chunks[j] += (item[2:].encode('utf-8') if PY3 else item[2:])
762
unescaped = b''.join(escaped_chunks)
764
decoded = unescaped.decode('utf-8')
765
except UnicodeDecodeError:
766
# If this path segment cannot be properly utf-8 decoded
767
# after doing unescaping we will just leave it alone
771
decoded.encode(encoding)
772
except UnicodeEncodeError:
773
# If this chunk cannot be encoded in the local
774
# encoding, then we should leave it alone
777
# Otherwise take the url decoded one
781
def unescape_for_display(url, encoding):
782
"""Decode what you can for a URL, so that we get a nice looking path.
784
This will turn file:// urls into local paths, and try to decode
785
any portions of a http:// style url that it can.
787
Any sections of the URL which can't be represented in the encoding or
788
need to stay as escapes are left alone.
790
:param url: A 7-bit ASCII URL
791
:param encoding: The final output encoding
793
:return: A unicode string which can be safely encoded into the
797
raise ValueError('you cannot specify None for the display encoding')
798
if url.startswith('file://'):
800
path = local_path_from_url(url)
801
path.encode(encoding)
806
# Split into sections to try to decode utf-8
808
for i in range(1, len(res)):
809
res[i] = _unescape_segment_for_display(res[i], encoding)
810
return u'/'.join(res)
813
def derive_to_location(from_location):
814
"""Derive a TO_LOCATION given a FROM_LOCATION.
816
The normal case is a FROM_LOCATION of http://foo/bar => bar.
817
The Right Thing for some logical destinations may differ though
818
because no / may be present at all. In that case, the result is
819
the full name without the scheme indicator, e.g. lp:foo-bar => foo-bar.
820
This latter case also applies when a Windows drive
821
is used without a path, e.g. c:foo-bar => foo-bar.
822
If no /, path separator or : is found, the from_location is returned.
824
from_location, unused_params = split_segment_parameters(from_location)
825
if from_location.find("/") >= 0 or from_location.find(os.sep) >= 0:
826
return os.path.basename(from_location.rstrip("/\\"))
828
sep = from_location.find(":")
830
return from_location[sep + 1:]
835
def _is_absolute(url):
836
return (osutils.pathjoin('/foo', url) == url)
839
def rebase_url(url, old_base, new_base):
840
"""Convert a relative path from an old base URL to a new base URL.
842
The result will be a relative path.
843
Absolute paths and full URLs are returned unaltered.
845
scheme, separator = _find_scheme_and_separator(url)
846
if scheme is not None:
848
if _is_absolute(url):
850
old_parsed = urlparse.urlparse(old_base)
851
new_parsed = urlparse.urlparse(new_base)
852
if (old_parsed[:2]) != (new_parsed[:2]):
853
raise InvalidRebaseURLs(old_base, new_base)
854
return determine_relative_path(new_parsed[2],
855
join(old_parsed[2], url))
858
def determine_relative_path(from_path, to_path):
859
"""Determine a relative path from from_path to to_path."""
860
from_segments = osutils.splitpath(from_path)
861
to_segments = osutils.splitpath(to_path)
863
for count, (from_element, to_element) in enumerate(zip(from_segments,
865
if from_element != to_element:
869
unique_from = from_segments[count:]
870
unique_to = to_segments[count:]
871
segments = (['..'] * len(unique_from) + unique_to)
872
if len(segments) == 0:
874
return osutils.pathjoin(*segments)
880
def __init__(self, scheme, quoted_user, quoted_password, quoted_host,
883
self.quoted_host = quoted_host
884
self.host = unquote(self.quoted_host)
885
self.quoted_user = quoted_user
886
if self.quoted_user is not None:
887
self.user = unquote(self.quoted_user)
890
self.quoted_password = quoted_password
891
if self.quoted_password is not None:
892
self.password = unquote(self.quoted_password)
896
self.quoted_path = _url_hex_escapes_re.sub(
897
_unescape_safe_chars, quoted_path)
898
self.path = unquote(self.quoted_path)
900
def __eq__(self, other):
901
return (isinstance(other, self.__class__) and
902
self.scheme == other.scheme and
903
self.host == other.host and
904
self.user == other.user and
905
self.password == other.password and
906
self.path == other.path)
909
return "<%s(%r, %r, %r, %r, %r, %r)>" % (
910
self.__class__.__name__,
911
self.scheme, self.quoted_user, self.quoted_password,
912
self.quoted_host, self.port, self.quoted_path)
915
def from_string(cls, url):
916
"""Create a URL object from a string.
918
:param url: URL as bytestring
920
# GZ 2017-06-09: Actually validate ascii-ness
921
# pad.lv/1696545: For the moment, accept both native strings and
923
if isinstance(url, str):
925
elif isinstance(url, text_type):
928
except UnicodeEncodeError:
929
raise InvalidURL(url)
931
raise InvalidURL(url)
932
(scheme, netloc, path, params,
933
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
934
user = password = host = port = None
936
user, host = netloc.rsplit('@', 1)
938
user, password = user.split(':', 1)
942
if ':' in host and not (host[0] == '[' and host[-1] == ']'):
944
host, port = host.rsplit(':', 1)
949
raise InvalidURL('invalid port number %s in url:\n%s' %
953
if host != "" and host[0] == '[' and host[-1] == ']': # IPv6
956
return cls(scheme, user, password, host, port, path)
959
netloc = self.quoted_host
961
netloc = "[%s]" % netloc
962
if self.quoted_user is not None:
963
# Note that we don't put the password back even if we
964
# have one so that it doesn't get accidentally
966
netloc = '%s@%s' % (self.quoted_user, netloc)
967
if self.port is not None:
968
netloc = '%s:%d' % (netloc, self.port)
969
return urlparse.urlunparse(
970
(self.scheme, netloc, self.quoted_path, None, None, None))
973
def _combine_paths(base_path, relpath):
974
"""Transform a Transport-relative path to a remote absolute path.
976
This does not handle substitution of ~ but does handle '..' and '.'
981
t._combine_paths('/home/sarah', 'project/foo')
982
=> '/home/sarah/project/foo'
983
t._combine_paths('/home/sarah', '../../etc')
985
t._combine_paths('/home/sarah', '/etc')
988
:param base_path: base path
989
:param relpath: relative url string for relative part of remote path.
990
:return: urlencoded string for final path.
992
# pad.lv/1696545: For the moment, accept both native strings and
994
if isinstance(relpath, str):
996
elif isinstance(relpath, text_type):
998
relpath = relpath.encode()
999
except UnicodeEncodeError:
1000
raise InvalidURL(relpath)
1002
raise InvalidURL(relpath)
1003
relpath = _url_hex_escapes_re.sub(_unescape_safe_chars, relpath)
1004
if relpath.startswith('/'):
1007
base_parts = base_path.split('/')
1008
if len(base_parts) > 0 and base_parts[-1] == '':
1009
base_parts = base_parts[:-1]
1010
for p in relpath.split('/'):
1012
if len(base_parts) == 0:
1013
# In most filesystems, a request for the parent
1014
# of root, just returns root.
1020
base_parts.append(p)
1021
path = '/'.join(base_parts)
1022
if not path.startswith('/'):
1026
def clone(self, offset=None):
1027
"""Return a new URL for a path relative to this URL.
1029
:param offset: A relative path, already urlencoded
1030
:return: `URL` instance
1032
if offset is not None:
1033
relative = unescape(offset)
1034
if sys.version_info[0] == 2:
1035
relative = relative.encode('utf-8')
1036
path = self._combine_paths(self.path, relative)
1037
path = quote(path, safe="/~")
1039
path = self.quoted_path
1040
return self.__class__(self.scheme, self.quoted_user,
1041
self.quoted_password, self.quoted_host, self.port,
1046
"""Extract the server address, the credentials and the path from the url.
1048
user, password, host and path should be quoted if they contain reserved
1051
:param url: an quoted url
1052
:return: (scheme, user, password, host, port, path) tuple, all fields
1055
parsed_url = URL.from_string(url)
1056
return (parsed_url.scheme, parsed_url.user, parsed_url.password,
1057
parsed_url.host, parsed_url.port, parsed_url.path)
added
removed
breezy/urlutils.py
