1
# Bazaar -- distributed version control
3
# Copyright (C) 2006 Canonical Ltd
1
# Copyright (C) 2006, 2008 Canonical Ltd
5
3
# This program is free software; you can redistribute it and/or modify
6
4
# it under the terms of the GNU General Public License as published by
15
13
# You should have received a copy of the GNU General Public License
16
14
# along with this program; if not, write to the Free Software
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
15
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19
17
"""A collection of function for handling URL operations."""
68
66
relpath = relpath.encode('utf-8')
69
67
# After quoting and encoding, the path should be perfectly
70
68
# safe as a plain ASCII string, str() just enforces this
71
return str(urllib.quote(relpath))
69
return str(urllib.quote(relpath, safe='/~'))
74
72
def file_relpath(base, path):
75
73
"""Compute just the relative sub-portion of a url
77
75
This assumes that both paths are already fully specified file:// URLs.
79
77
if len(base) < MIN_ABS_FILEURL_LENGTH:
80
raise ValueError('Length of base must be equal or'
78
raise ValueError('Length of base (%r) must equal or'
81
79
' exceed the platform minimum url length (which is %d)' %
82
MIN_ABS_FILEURL_LENGTH)
80
(base, MIN_ABS_FILEURL_LENGTH))
83
81
base = local_path_from_url(base)
84
82
path = local_path_from_url(path)
85
83
return escape(osutils.relpath(base, path))
186
184
def joinpath(base, *args):
187
185
"""Join URL path segments to a URL path segment.
189
187
This is somewhat like osutils.joinpath, but intended for URLs.
191
189
XXX: this duplicates some normalisation logic, and also duplicates a lot of
231
229
This also handles transforming escaping unicode characters, etc.
233
# importing directly from posixpath allows us to test this
231
# importing directly from posixpath allows us to test this
234
232
# on non-posix platforms
235
233
return 'file://' + escape(_posix_normpath(
236
234
osutils._posix_abspath(path)))
271
269
This also handles transforming escaping unicode characters, etc.
273
# importing directly from ntpath allows us to test this
271
# importing directly from ntpath allows us to test this
274
272
# on non-win32 platform
275
273
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
276
274
# which actually strips trailing space characters.
370
368
dummy, base_first_slash = _find_scheme_and_separator(base)
371
369
if base_first_slash is None:
374
372
dummy, other_first_slash = _find_scheme_and_separator(other)
375
373
if other_first_slash is None:
418
416
# Strip off the drive letter
419
417
# path is currently /C:/foo
420
418
if len(path) < 3 or path[2] not in ':|' or path[3] != '/':
421
raise errors.InvalidURL(url_base + path,
419
raise errors.InvalidURL(url_base + path,
422
420
'win32 file:/// paths need a drive letter')
423
421
url_base += path[0:3] # file:// + /C:
424
422
path = path[3:] # /foo
432
430
:param exclude_trailing_slash: Strip off a final '/' if it is part
433
431
of the path (but not if it is part of the protocol specification)
435
:return: (parent_url, child_dir). child_dir may be the empty string if we're at
433
:return: (parent_url, child_dir). child_dir may be the empty string if we're at
438
436
scheme_loc, first_path_slash = _find_scheme_and_separator(url)
542
540
# These are characters that if escaped, should stay that way
543
541
_no_decode_chars = ';/?:@&=+$,#'
544
542
_no_decode_ords = [ord(c) for c in _no_decode_chars]
545
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
543
_no_decode_hex = (['%02x' % o for o in _no_decode_ords]
546
544
+ ['%02X' % o for o in _no_decode_ords])
547
545
_hex_display_map = dict(([('%02x' % o, chr(o)) for o in range(256)]
548
546
+ [('%02X' % o, chr(o)) for o in range(256)]))
574
572
This will turn file:// urls into local paths, and try to decode
575
573
any portions of a http:// style url that it can.
577
Any sections of the URL which can't be represented in the encoding or
575
Any sections of the URL which can't be represented in the encoding or
578
576
need to stay as escapes are left alone.
580
578
:param url: A 7-bit ASCII URL
581
579
:param encoding: The final output encoding
583
:return: A unicode string which can be safely encoded into the
581
:return: A unicode string which can be safely encoded into the
584
582
specified encoding.
586
584
if encoding is None:
687
685
if len(segments) == 0:
689
687
return osutils.pathjoin(*segments)
692
"""Extract the server address, the credentials and the path from the url.
694
user, password, host and path should be quoted if they contain reserved
697
:param url: an quoted url
699
:return: (scheme, user, password, host, port, path) tuple, all fields
702
if isinstance(url, unicode):
703
raise errors.InvalidURL('should be ascii:\n%r' % url)
704
url = url.encode('utf-8')
705
(scheme, netloc, path, params,
706
query, fragment) = urlparse.urlparse(url, allow_fragments=False)
707
user = password = host = port = None
709
user, host = netloc.rsplit('@', 1)
711
user, password = user.split(':', 1)
712
password = urllib.unquote(password)
713
user = urllib.unquote(user)
717
if ':' in host and not (host[0] == '[' and host[-1] == ']'): #there *is* port
718
host, port = host.rsplit(':',1)
722
raise errors.InvalidURL('invalid port number %s in url:\n%s' %
724
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6
727
host = urllib.unquote(host)
728
path = urllib.unquote(path)
730
return (scheme, user, password, host, port, path)