4763.2.4
by John Arbash Meinel
merge bzr.2.1 in preparation for NEWS entry. |
1 |
# Copyright (C) 2006-2010 Canonical Ltd
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
2 |
#
|
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.
|
|
7 |
#
|
|
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.
|
|
12 |
#
|
|
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
|
|
4183.7.1
by Sabin Iacob
update FSF mailing address |
15 |
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
16 |
|
17 |
"""A collection of function for handling URL operations."""
|
|
18 |
||
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
19 |
import os |
1685.1.50
by John Arbash Meinel
Added an re for handling scheme paths. |
20 |
import re |
21 |
import sys |
|
1996.3.12
by John Arbash Meinel
Change how 'revision' is imported to avoid problems later |
22 |
|
23 |
from bzrlib.lazy_import import lazy_import |
|
24 |
lazy_import(globals(), """ |
|
25 |
from posixpath import split as _posix_split, normpath as _posix_normpath
|
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
26 |
import urllib
|
3242.3.26
by Aaron Bentley
Implement rebase_url |
27 |
import urlparse
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
28 |
|
1996.3.12
by John Arbash Meinel
Change how 'revision' is imported to avoid problems later |
29 |
from bzrlib import (
|
30 |
errors,
|
|
31 |
osutils,
|
|
32 |
)
|
|
33 |
""") |
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
34 |
|
35 |
||
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
36 |
def basename(url, exclude_trailing_slash=True): |
37 |
"""Return the last component of a URL.
|
|
38 |
||
39 |
:param url: The URL in question
|
|
40 |
:param exclude_trailing_slash: If the url looks like "path/to/foo/"
|
|
41 |
ignore the final slash and return 'foo' rather than ''
|
|
42 |
:return: Just the final component of the URL. This can return ''
|
|
43 |
if you don't exclude_trailing_slash, or if you are at the
|
|
44 |
root of the URL.
|
|
45 |
"""
|
|
46 |
return split(url, exclude_trailing_slash=exclude_trailing_slash)[1] |
|
47 |
||
48 |
||
49 |
def dirname(url, exclude_trailing_slash=True): |
|
50 |
"""Return the parent directory of the given path.
|
|
51 |
||
52 |
:param url: Relative or absolute URL
|
|
53 |
:param exclude_trailing_slash: Remove a final slash
|
|
54 |
(treat http://host/foo/ as http://host/foo, but
|
|
55 |
http://host/ stays http://host/)
|
|
56 |
:return: Everything in the URL except the last path chunk
|
|
57 |
"""
|
|
58 |
# TODO: jam 20060502 This was named dirname to be consistent
|
|
59 |
# with the os functions, but maybe "parent" would be better
|
|
60 |
return split(url, exclude_trailing_slash=exclude_trailing_slash)[0] |
|
61 |
||
62 |
||
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
63 |
def escape(relpath): |
64 |
"""Escape relpath to be a valid url."""
|
|
65 |
if isinstance(relpath, unicode): |
|
66 |
relpath = relpath.encode('utf-8') |
|
67 |
# After quoting and encoding, the path should be perfectly
|
|
68 |
# safe as a plain ASCII string, str() just enforces this
|
|
4098.3.1
by Jonathan Lange
Don't escape tildes |
69 |
return str(urllib.quote(relpath, safe='/~')) |
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
70 |
|
71 |
||
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
72 |
def file_relpath(base, path): |
73 |
"""Compute just the relative sub-portion of a url
|
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
74 |
|
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
75 |
This assumes that both paths are already fully specified file:// URLs.
|
76 |
"""
|
|
3376.2.4
by Martin Pool
Remove every assert statement from bzrlib! |
77 |
if len(base) < MIN_ABS_FILEURL_LENGTH: |
4539.1.1
by Andrew Bennetts
Improve error message in osutils.file_relpath. |
78 |
raise ValueError('Length of base (%r) must equal or' |
3376.2.4
by Martin Pool
Remove every assert statement from bzrlib! |
79 |
' exceed the platform minimum url length (which is %d)' % |
4539.1.1
by Andrew Bennetts
Improve error message in osutils.file_relpath. |
80 |
(base, MIN_ABS_FILEURL_LENGTH)) |
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
81 |
base = local_path_from_url(base) |
82 |
path = local_path_from_url(path) |
|
1996.3.12
by John Arbash Meinel
Change how 'revision' is imported to avoid problems later |
83 |
return escape(osutils.relpath(base, path)) |
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
84 |
|
85 |
||
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
86 |
def _find_scheme_and_separator(url): |
87 |
"""Find the scheme separator (://) and the first path separator
|
|
88 |
||
89 |
This is just a helper functions for other path utilities.
|
|
90 |
It could probably be replaced by urlparse
|
|
91 |
"""
|
|
1685.1.51
by John Arbash Meinel
Working on getting normalize_url working. |
92 |
m = _url_scheme_re.match(url) |
93 |
if not m: |
|
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
94 |
return None, None |
95 |
||
1685.1.51
by John Arbash Meinel
Working on getting normalize_url working. |
96 |
scheme = m.group('scheme') |
97 |
path = m.group('path') |
|
98 |
||
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
99 |
# Find the path separating slash
|
100 |
# (first slash after the ://)
|
|
1685.1.51
by John Arbash Meinel
Working on getting normalize_url working. |
101 |
first_path_slash = path.find('/') |
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
102 |
if first_path_slash == -1: |
1685.1.56
by John Arbash Meinel
Fixing _find_scheme_and_separator |
103 |
return len(scheme), None |
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
104 |
return len(scheme), first_path_slash+m.start('path') |
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
105 |
|
106 |
||
5254.2.1
by Gordon Tyler
Fixed how get_transport's convert_path_to_url tests whether a path is actually a URL. |
107 |
def is_url(url): |
108 |
"""Tests whether a URL is in actual fact a URL."""
|
|
109 |
return _url_scheme_re.match(url) is not None |
|
110 |
||
111 |
||
1685.1.55
by John Arbash Meinel
Adding bzrlib.urlutils.join() to handle joining URLs |
112 |
def join(base, *args): |
113 |
"""Create a URL by joining sections.
|
|
114 |
||
115 |
This will normalize '..', assuming that paths are absolute
|
|
116 |
(it assumes no symlinks in either path)
|
|
117 |
||
118 |
If any of *args is an absolute URL, it will be treated correctly.
|
|
119 |
Example:
|
|
120 |
join('http://foo', 'http://bar') => 'http://bar'
|
|
121 |
join('http://foo', 'bar') => 'http://foo/bar'
|
|
122 |
join('http://foo', 'bar', '../baz') => 'http://foo/baz'
|
|
123 |
"""
|
|
2018.5.100
by Andrew Bennetts
Fix IndexError in urlutils.join with 'http://host/a' and '../../b'. |
124 |
if not args: |
125 |
return base |
|
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
126 |
scheme_end, path_start = _find_scheme_and_separator(base) |
127 |
if scheme_end is None and path_start is None: |
|
128 |
path_start = 0 |
|
129 |
elif path_start is None: |
|
130 |
path_start = len(base) |
|
131 |
path = base[path_start:] |
|
1685.1.55
by John Arbash Meinel
Adding bzrlib.urlutils.join() to handle joining URLs |
132 |
for arg in args: |
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
133 |
arg_scheme_end, arg_path_start = _find_scheme_and_separator(arg) |
134 |
if arg_scheme_end is None and arg_path_start is None: |
|
135 |
arg_path_start = 0 |
|
136 |
elif arg_path_start is None: |
|
137 |
arg_path_start = len(arg) |
|
5254.1.5
by Gordon Tyler
Fixes according to spiv's review. |
138 |
if arg_scheme_end is not None: |
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
139 |
base = arg |
140 |
path = arg[arg_path_start:] |
|
141 |
scheme_end = arg_scheme_end |
|
142 |
path_start = arg_path_start |
|
1685.1.55
by John Arbash Meinel
Adding bzrlib.urlutils.join() to handle joining URLs |
143 |
else: |
2018.5.54
by Andrew Bennetts
Fix ChrootTransportDecorator's abspath method to be consistent with its clone |
144 |
path = joinpath(path, arg) |
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
145 |
return base[:path_start] + path |
1685.1.55
by John Arbash Meinel
Adding bzrlib.urlutils.join() to handle joining URLs |
146 |
|
147 |
||
2018.5.46
by Andrew Bennetts
Fix ChrootTransportDecorator's clone to pass less surprising offsets to the decorated transport's clone. |
148 |
def joinpath(base, *args): |
149 |
"""Join URL path segments to a URL path segment.
|
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
150 |
|
2018.5.46
by Andrew Bennetts
Fix ChrootTransportDecorator's clone to pass less surprising offsets to the decorated transport's clone. |
151 |
This is somewhat like osutils.joinpath, but intended for URLs.
|
152 |
||
153 |
XXX: this duplicates some normalisation logic, and also duplicates a lot of
|
|
154 |
path handling logic that already exists in some Transport implementations.
|
|
155 |
We really should try to have exactly one place in the code base responsible
|
|
156 |
for combining paths of URLs.
|
|
157 |
"""
|
|
2018.5.100
by Andrew Bennetts
Fix IndexError in urlutils.join with 'http://host/a' and '../../b'. |
158 |
path = base.split('/') |
159 |
if len(path) > 1 and path[-1] == '': |
|
160 |
#If the path ends in a trailing /, remove it.
|
|
161 |
path.pop() |
|
2018.5.46
by Andrew Bennetts
Fix ChrootTransportDecorator's clone to pass less surprising offsets to the decorated transport's clone. |
162 |
for arg in args: |
163 |
if arg.startswith('/'): |
|
164 |
path = [] |
|
165 |
for chunk in arg.split('/'): |
|
166 |
if chunk == '.': |
|
167 |
continue
|
|
168 |
elif chunk == '..': |
|
169 |
if path == ['']: |
|
170 |
raise errors.InvalidURLJoin('Cannot go above root', |
|
171 |
base, args) |
|
172 |
path.pop() |
|
173 |
else: |
|
174 |
path.append(chunk) |
|
175 |
if path == ['']: |
|
176 |
return '/' |
|
177 |
else: |
|
178 |
return '/'.join(path) |
|
179 |
||
180 |
||
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
181 |
# jam 20060502 Sorted to 'l' because the final target is 'local_path_from_url'
|
182 |
def _posix_local_path_from_url(url): |
|
183 |
"""Convert a url like file:///path/to/foo into /path/to/foo"""
|
|
4828.1.1
by Michael Hudson
test and fix |
184 |
file_localhost_prefix = 'file://localhost/' |
185 |
if url.startswith(file_localhost_prefix): |
|
186 |
path = url[len(file_localhost_prefix) - 1:] |
|
187 |
elif not url.startswith('file:///'): |
|
188 |
raise errors.InvalidURL( |
|
189 |
url, 'local urls must start with file:/// or file://localhost/') |
|
190 |
else: |
|
191 |
path = url[len('file://'):] |
|
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
192 |
# We only strip off 2 slashes
|
4828.1.1
by Michael Hudson
test and fix |
193 |
return unescape(path) |
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
194 |
|
195 |
||
196 |
def _posix_local_path_to_url(path): |
|
197 |
"""Convert a local path like ./foo into a URL like file:///path/to/foo
|
|
198 |
||
199 |
This also handles transforming escaping unicode characters, etc.
|
|
200 |
"""
|
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
201 |
# importing directly from posixpath allows us to test this
|
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
202 |
# on non-posix platforms
|
1711.4.5
by John Arbash Meinel
the _posix_* routines should use posixpath not os.path, so tests pass on win32 |
203 |
return 'file://' + escape(_posix_normpath( |
1996.3.12
by John Arbash Meinel
Change how 'revision' is imported to avoid problems later |
204 |
osutils._posix_abspath(path))) |
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
205 |
|
206 |
||
207 |
def _win32_local_path_from_url(url): |
|
1711.4.4
by John Arbash Meinel
Fix some broken tests because of stupid ntpath.abspath behavior |
208 |
"""Convert a url like file:///C:/path/to/foo into C:/path/to/foo"""
|
2162.2.7
by Alexander Belchenko
Win32 UNC path \\HOST\path mapped to URL file://HOST/path |
209 |
if not url.startswith('file://'): |
210 |
raise errors.InvalidURL(url, 'local urls must start with file:///, ' |
|
211 |
'UNC path urls must start with file://') |
|
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
212 |
# We strip off all 3 slashes
|
2162.2.7
by Alexander Belchenko
Win32 UNC path \\HOST\path mapped to URL file://HOST/path |
213 |
win32_url = url[len('file:'):] |
2162.2.2
by Alexander Belchenko
Support for win32 UNC path (like: \\HOST\path) |
214 |
# check for UNC path: //HOST/path
|
2162.2.7
by Alexander Belchenko
Win32 UNC path \\HOST\path mapped to URL file://HOST/path |
215 |
if not win32_url.startswith('///'): |
2162.2.2
by Alexander Belchenko
Support for win32 UNC path (like: \\HOST\path) |
216 |
if (win32_url[2] == '/' |
217 |
or win32_url[3] in '|:'): |
|
218 |
raise errors.InvalidURL(url, 'Win32 UNC path urls' |
|
2162.2.7
by Alexander Belchenko
Win32 UNC path \\HOST\path mapped to URL file://HOST/path |
219 |
' have form file://HOST/path') |
2162.2.2
by Alexander Belchenko
Support for win32 UNC path (like: \\HOST\path) |
220 |
return unescape(win32_url) |
3503.1.2
by adwi2
Permits Windows to serve all paths on all drives. |
221 |
|
222 |
# allow empty paths so we can serve all roots
|
|
223 |
if win32_url == '///': |
|
224 |
return '/' |
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
225 |
|
2162.2.2
by Alexander Belchenko
Support for win32 UNC path (like: \\HOST\path) |
226 |
# usual local path with drive letter
|
5510.2.3
by Jared Bunting
Changed _win32_local_path_from_url to not allow "file:///C:" form. |
227 |
if (len(win32_url) < 6 |
5510.2.1
by Jared Bunting
Modified _win32_local_path_from_url to: |
228 |
or win32_url[3] not in ('abcdefghijklmnopqrstuvwxyz' |
229 |
'ABCDEFGHIJKLMNOPQRSTUVWXYZ') |
|
2162.2.7
by Alexander Belchenko
Win32 UNC path \\HOST\path mapped to URL file://HOST/path |
230 |
or win32_url[4] not in '|:' |
5510.2.3
by Jared Bunting
Changed _win32_local_path_from_url to not allow "file:///C:" form. |
231 |
or win32_url[5] != '/'): |
1711.4.4
by John Arbash Meinel
Fix some broken tests because of stupid ntpath.abspath behavior |
232 |
raise errors.InvalidURL(url, 'Win32 file urls start with' |
1711.4.8
by John Arbash Meinel
switch to prefering lowercase drive letters, since that matches os.getcwd() drive letters |
233 |
' file:///x:/, where x is a valid drive letter') |
2162.2.7
by Alexander Belchenko
Win32 UNC path \\HOST\path mapped to URL file://HOST/path |
234 |
return win32_url[3].upper() + u':' + unescape(win32_url[5:]) |
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
235 |
|
236 |
||
237 |
def _win32_local_path_to_url(path): |
|
1711.4.4
by John Arbash Meinel
Fix some broken tests because of stupid ntpath.abspath behavior |
238 |
"""Convert a local path like ./foo into a URL like file:///C:/path/to/foo
|
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
239 |
|
240 |
This also handles transforming escaping unicode characters, etc.
|
|
241 |
"""
|
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
242 |
# importing directly from ntpath allows us to test this
|
1711.4.4
by John Arbash Meinel
Fix some broken tests because of stupid ntpath.abspath behavior |
243 |
# on non-win32 platform
|
244 |
# FIXME: It turns out that on nt, ntpath.abspath uses nt._getfullpathname
|
|
245 |
# which actually strips trailing space characters.
|
|
5278.1.5
by Martin Pool
Correct more sloppy use of the term 'Linux' |
246 |
# The worst part is that on linux ntpath.abspath has different
|
1711.4.4
by John Arbash Meinel
Fix some broken tests because of stupid ntpath.abspath behavior |
247 |
# semantics, since 'nt' is not an available module.
|
3503.1.1
by Adrian Wilkins
Add a couple of special cases to urlutils._win32_path_(from|to)_url |
248 |
if path == '/': |
3503.1.2
by adwi2
Permits Windows to serve all paths on all drives. |
249 |
return 'file:///' |
3503.1.1
by Adrian Wilkins
Add a couple of special cases to urlutils._win32_path_(from|to)_url |
250 |
|
2279.4.2
by Alexander Belchenko
Don't do normpath after abspath, because this function is called inside abspath |
251 |
win32_path = osutils._win32_abspath(path) |
2162.2.2
by Alexander Belchenko
Support for win32 UNC path (like: \\HOST\path) |
252 |
# check for UNC path \\HOST\path
|
253 |
if win32_path.startswith('//'): |
|
2162.2.7
by Alexander Belchenko
Win32 UNC path \\HOST\path mapped to URL file://HOST/path |
254 |
return 'file:' + escape(win32_path) |
3234.3.1
by Alexander Belchenko
ensure that local_path_to_url() always returns plain string, not unicode. |
255 |
return ('file:///' + str(win32_path[0].upper()) + ':' + |
256 |
escape(win32_path[2:])) |
|
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
257 |
|
258 |
||
259 |
local_path_to_url = _posix_local_path_to_url |
|
260 |
local_path_from_url = _posix_local_path_from_url |
|
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
261 |
MIN_ABS_FILEURL_LENGTH = len('file:///') |
1711.4.17
by John Arbash Meinel
[merge] bzr.dev 1790 |
262 |
WIN32_MIN_ABS_FILEURL_LENGTH = len('file:///C:/') |
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
263 |
|
264 |
if sys.platform == 'win32': |
|
265 |
local_path_to_url = _win32_local_path_to_url |
|
266 |
local_path_from_url = _win32_local_path_from_url |
|
267 |
||
1711.2.44
by John Arbash Meinel
Factor out another win32 special case and add platform independent tests for it. |
268 |
MIN_ABS_FILEURL_LENGTH = WIN32_MIN_ABS_FILEURL_LENGTH |
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
269 |
|
270 |
||
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
271 |
_url_scheme_re = re.compile(r'^(?P<scheme>[^:/]{2,}):(//)?(?P<path>.*)$') |
2208.4.1
by Andrew Bennetts
normalize_url should normalise escaping of unreserved characters, like '~'. |
272 |
_url_hex_escapes_re = re.compile(r'(%[0-9a-fA-F]{2})') |
273 |
||
274 |
||
275 |
def _unescape_safe_chars(matchobj): |
|
276 |
"""re.sub callback to convert hex-escapes to plain characters (if safe).
|
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
277 |
|
2208.4.1
by Andrew Bennetts
normalize_url should normalise escaping of unreserved characters, like '~'. |
278 |
e.g. '%7E' will be converted to '~'.
|
279 |
"""
|
|
280 |
hex_digits = matchobj.group(0)[1:] |
|
281 |
char = chr(int(hex_digits, 16)) |
|
282 |
if char in _url_dont_escape_characters: |
|
283 |
return char |
|
284 |
else: |
|
285 |
return matchobj.group(0).upper() |
|
1685.1.50
by John Arbash Meinel
Added an re for handling scheme paths. |
286 |
|
287 |
||
288 |
def normalize_url(url): |
|
289 |
"""Make sure that a path string is in fully normalized URL form.
|
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
290 |
|
2208.4.1
by Andrew Bennetts
normalize_url should normalise escaping of unreserved characters, like '~'. |
291 |
This handles URLs which have unicode characters, spaces,
|
1685.1.50
by John Arbash Meinel
Added an re for handling scheme paths. |
292 |
special characters, etc.
|
293 |
||
294 |
It has two basic modes of operation, depending on whether the
|
|
295 |
supplied string starts with a url specifier (scheme://) or not.
|
|
296 |
If it does not have a specifier it is considered a local path,
|
|
297 |
and will be converted into a file:/// url. Non-ascii characters
|
|
298 |
will be encoded using utf-8.
|
|
299 |
If it does have a url specifier, it will be treated as a "hybrid"
|
|
300 |
URL. Basically, a URL that should have URL special characters already
|
|
301 |
escaped (like +?&# etc), but may have unicode characters, etc
|
|
302 |
which would not be valid in a real URL.
|
|
303 |
||
304 |
:param url: Either a hybrid URL or a local path
|
|
305 |
:return: A normalized URL which only includes 7-bit ASCII characters.
|
|
306 |
"""
|
|
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
307 |
scheme_end, path_start = _find_scheme_and_separator(url) |
308 |
if scheme_end is None: |
|
1685.1.51
by John Arbash Meinel
Working on getting normalize_url working. |
309 |
return local_path_to_url(url) |
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
310 |
prefix = url[:path_start] |
311 |
path = url[path_start:] |
|
1685.1.51
by John Arbash Meinel
Working on getting normalize_url working. |
312 |
if not isinstance(url, unicode): |
313 |
for c in url: |
|
314 |
if c not in _url_safe_characters: |
|
1685.1.53
by John Arbash Meinel
Updated normalize_url |
315 |
raise errors.InvalidURL(url, 'URLs can only contain specific' |
316 |
' safe characters (not %r)' % c) |
|
2208.4.1
by Andrew Bennetts
normalize_url should normalise escaping of unreserved characters, like '~'. |
317 |
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path) |
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
318 |
return str(prefix + ''.join(path)) |
2208.4.1
by Andrew Bennetts
normalize_url should normalise escaping of unreserved characters, like '~'. |
319 |
|
1685.1.51
by John Arbash Meinel
Working on getting normalize_url working. |
320 |
# We have a unicode (hybrid) url
|
2208.4.1
by Andrew Bennetts
normalize_url should normalise escaping of unreserved characters, like '~'. |
321 |
path_chars = list(path) |
1685.1.50
by John Arbash Meinel
Added an re for handling scheme paths. |
322 |
|
2208.4.1
by Andrew Bennetts
normalize_url should normalise escaping of unreserved characters, like '~'. |
323 |
for i in xrange(len(path_chars)): |
324 |
if path_chars[i] not in _url_safe_characters: |
|
325 |
chars = path_chars[i].encode('utf-8') |
|
326 |
path_chars[i] = ''.join( |
|
327 |
['%%%02X' % ord(c) for c in path_chars[i].encode('utf-8')]) |
|
328 |
path = ''.join(path_chars) |
|
329 |
path = _url_hex_escapes_re.sub(_unescape_safe_chars, path) |
|
5254.1.1
by Gordon Tyler
Added support to urlutils for URLs such as Launchpad's lp:foobar. |
330 |
return str(prefix + path) |
1685.1.50
by John Arbash Meinel
Added an re for handling scheme paths. |
331 |
|
332 |
||
1685.1.70
by Wouter van Heyst
working on get_parent, set_parent and relative urls, broken |
333 |
def relative_url(base, other): |
334 |
"""Return a path to other from base.
|
|
335 |
||
336 |
If other is unrelated to base, return other. Else return a relative path.
|
|
337 |
This assumes no symlinks as part of the url.
|
|
338 |
"""
|
|
1685.1.71
by Wouter van Heyst
change branch.{get,set}_parent to store a relative path but return full urls |
339 |
dummy, base_first_slash = _find_scheme_and_separator(base) |
340 |
if base_first_slash is None: |
|
1685.1.70
by Wouter van Heyst
working on get_parent, set_parent and relative urls, broken |
341 |
return other |
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
342 |
|
1685.1.71
by Wouter van Heyst
change branch.{get,set}_parent to store a relative path but return full urls |
343 |
dummy, other_first_slash = _find_scheme_and_separator(other) |
344 |
if other_first_slash is None: |
|
345 |
return other |
|
346 |
||
347 |
# this takes care of differing schemes or hosts
|
|
348 |
base_scheme = base[:base_first_slash] |
|
349 |
other_scheme = other[:other_first_slash] |
|
350 |
if base_scheme != other_scheme: |
|
351 |
return other |
|
3139.2.1
by Alexander Belchenko
bugfix #90847: fix problem with parent location on another logical drive |
352 |
elif sys.platform == 'win32' and base_scheme == 'file://': |
353 |
base_drive = base[base_first_slash+1:base_first_slash+3] |
|
354 |
other_drive = other[other_first_slash+1:other_first_slash+3] |
|
355 |
if base_drive != other_drive: |
|
356 |
return other |
|
1685.1.71
by Wouter van Heyst
change branch.{get,set}_parent to store a relative path but return full urls |
357 |
|
358 |
base_path = base[base_first_slash+1:] |
|
359 |
other_path = other[other_first_slash+1:] |
|
360 |
||
361 |
if base_path.endswith('/'): |
|
362 |
base_path = base_path[:-1] |
|
1685.1.70
by Wouter van Heyst
working on get_parent, set_parent and relative urls, broken |
363 |
|
364 |
base_sections = base_path.split('/') |
|
365 |
other_sections = other_path.split('/') |
|
1685.1.71
by Wouter van Heyst
change branch.{get,set}_parent to store a relative path but return full urls |
366 |
|
367 |
if base_sections == ['']: |
|
368 |
base_sections = [] |
|
369 |
if other_sections == ['']: |
|
370 |
other_sections = [] |
|
1685.1.70
by Wouter van Heyst
working on get_parent, set_parent and relative urls, broken |
371 |
|
372 |
output_sections = [] |
|
373 |
for b, o in zip(base_sections, other_sections): |
|
374 |
if b != o: |
|
375 |
break
|
|
376 |
output_sections.append(b) |
|
1685.1.71
by Wouter van Heyst
change branch.{get,set}_parent to store a relative path but return full urls |
377 |
|
1685.1.70
by Wouter van Heyst
working on get_parent, set_parent and relative urls, broken |
378 |
match_len = len(output_sections) |
1685.1.71
by Wouter van Heyst
change branch.{get,set}_parent to store a relative path but return full urls |
379 |
output_sections = ['..' for x in base_sections[match_len:]] |
1685.1.70
by Wouter van Heyst
working on get_parent, set_parent and relative urls, broken |
380 |
output_sections.extend(other_sections[match_len:]) |
381 |
||
382 |
return "/".join(output_sections) or "." |
|
383 |
||
384 |
||
1711.2.43
by John Arbash Meinel
Split out win32 specific code so that it can be tested on all platforms. |
385 |
def _win32_extract_drive_letter(url_base, path): |
386 |
"""On win32 the drive letter needs to be added to the url base."""
|
|
387 |
# Strip off the drive letter
|
|
388 |
# path is currently /C:/foo
|
|
389 |
if len(path) < 3 or path[2] not in ':|' or path[3] != '/': |
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
390 |
raise errors.InvalidURL(url_base + path, |
1711.2.43
by John Arbash Meinel
Split out win32 specific code so that it can be tested on all platforms. |
391 |
'win32 file:/// paths need a drive letter') |
392 |
url_base += path[0:3] # file:// + /C: |
|
393 |
path = path[3:] # /foo |
|
394 |
return url_base, path |
|
395 |
||
396 |
||
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
397 |
def split(url, exclude_trailing_slash=True): |
398 |
"""Split a URL into its parent directory and a child directory.
|
|
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
399 |
|
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
400 |
:param url: A relative or absolute URL
|
401 |
:param exclude_trailing_slash: Strip off a final '/' if it is part
|
|
402 |
of the path (but not if it is part of the protocol specification)
|
|
1685.1.61
by Martin Pool
[broken] Change BzrDir._make_tail to use urlutils.split |
403 |
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
404 |
:return: (parent_url, child_dir). child_dir may be the empty string if we're at
|
1685.1.61
by Martin Pool
[broken] Change BzrDir._make_tail to use urlutils.split |
405 |
the root.
|
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
406 |
"""
|
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
407 |
scheme_loc, first_path_slash = _find_scheme_and_separator(url) |
408 |
||
409 |
if first_path_slash is None: |
|
410 |
# We have either a relative path, or no separating slash
|
|
411 |
if scheme_loc is None: |
|
412 |
# Relative path
|
|
413 |
if exclude_trailing_slash and url.endswith('/'): |
|
414 |
url = url[:-1] |
|
415 |
return _posix_split(url) |
|
416 |
else: |
|
417 |
# Scheme with no path
|
|
418 |
return url, '' |
|
419 |
||
420 |
# We have a fully defined path
|
|
421 |
url_base = url[:first_path_slash] # http://host, file:// |
|
422 |
path = url[first_path_slash:] # /file/foo |
|
423 |
||
424 |
if sys.platform == 'win32' and url.startswith('file:///'): |
|
425 |
# Strip off the drive letter
|
|
1711.2.43
by John Arbash Meinel
Split out win32 specific code so that it can be tested on all platforms. |
426 |
# url_base is currently file://
|
1711.2.39
by John Arbash Meinel
Fix bzrlib.urlutils.split() to work properly on win32 local paths. |
427 |
# path is currently /C:/foo
|
1711.2.43
by John Arbash Meinel
Split out win32 specific code so that it can be tested on all platforms. |
428 |
url_base, path = _win32_extract_drive_letter(url_base, path) |
429 |
# now it should be file:///C: and /foo
|
|
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
430 |
|
431 |
if exclude_trailing_slash and len(path) > 1 and path.endswith('/'): |
|
432 |
path = path[:-1] |
|
433 |
head, tail = _posix_split(path) |
|
434 |
return url_base + head, tail |
|
435 |
||
1685.1.46
by John Arbash Meinel
Sorting functions by name. |
436 |
|
5163.2.5
by Jelmer Vernooij
rename {split,join}_subsegments -> {split,join}_segment_parameters_raw and add more tests. |
437 |
def split_segment_parameters_raw(url): |
5163.2.1
by Jelmer Vernooij
Add urlutils.split_subsegments. |
438 |
"""Split the subsegment of the last segment of a URL.
|
439 |
||
440 |
:param url: A relative or absolute URL
|
|
441 |
:return: (url, subsegments)
|
|
442 |
"""
|
|
443 |
(parent_url, child_dir) = split(url) |
|
444 |
subsegments = child_dir.split(",") |
|
445 |
if len(subsegments) == 1: |
|
446 |
return (url, []) |
|
447 |
return (join(parent_url, subsegments[0]), subsegments[1:]) |
|
448 |
||
449 |
||
5163.2.3
by Jelmer Vernooij
Add join_segment_parameters / split_segment_parameters. |
450 |
def split_segment_parameters(url): |
451 |
"""Split the segment parameters of the last segment of a URL.
|
|
452 |
||
453 |
:param url: A relative or absolute URL
|
|
454 |
:return: (url, segment_parameters)
|
|
455 |
"""
|
|
5163.2.5
by Jelmer Vernooij
rename {split,join}_subsegments -> {split,join}_segment_parameters_raw and add more tests. |
456 |
(base_url, subsegments) = split_segment_parameters_raw(url) |
5163.2.3
by Jelmer Vernooij
Add join_segment_parameters / split_segment_parameters. |
457 |
parameters = {} |
458 |
for subsegment in subsegments: |
|
459 |
(key, value) = subsegment.split("=", 1) |
|
460 |
parameters[key] = value |
|
461 |
return (base_url, parameters) |
|
462 |
||
463 |
||
5163.2.5
by Jelmer Vernooij
rename {split,join}_subsegments -> {split,join}_segment_parameters_raw and add more tests. |
464 |
def join_segment_parameters_raw(base, *subsegments): |
5163.2.7
by Jelmer Vernooij
Add type checking. |
465 |
"""Create a new URL by adding subsegments to an existing one.
|
466 |
||
467 |
This adds the specified subsegments to the last path in the specified
|
|
468 |
base URL. The subsegments should be bytestrings.
|
|
5163.2.2
by Jelmer Vernooij
Add bzrlib.urlutils.join_subsegments. |
469 |
|
5163.2.5
by Jelmer Vernooij
rename {split,join}_subsegments -> {split,join}_segment_parameters_raw and add more tests. |
470 |
:note: You probably want to use join_segment_parameters instead.
|
5163.2.2
by Jelmer Vernooij
Add bzrlib.urlutils.join_subsegments. |
471 |
"""
|
472 |
if not subsegments: |
|
473 |
return base |
|
474 |
for subsegment in subsegments: |
|
5163.2.7
by Jelmer Vernooij
Add type checking. |
475 |
if type(subsegment) is not str: |
476 |
raise TypeError("Subsegment %r is not a bytestring" % subsegment) |
|
5163.2.2
by Jelmer Vernooij
Add bzrlib.urlutils.join_subsegments. |
477 |
if "," in subsegment: |
478 |
raise errors.InvalidURLJoin(", exists in subsegments", |
|
479 |
base, subsegments) |
|
480 |
return ",".join((base,) + subsegments) |
|
481 |
||
482 |
||
5163.2.3
by Jelmer Vernooij
Add join_segment_parameters / split_segment_parameters. |
483 |
def join_segment_parameters(url, parameters): |
484 |
"""Create a new URL by adding segment parameters to an existing one.
|
|
485 |
||
5163.2.7
by Jelmer Vernooij
Add type checking. |
486 |
The parameters of the last segment in the URL will be updated; if a
|
487 |
parameter with the same key already exists it will be overwritten.
|
|
488 |
||
5163.2.3
by Jelmer Vernooij
Add join_segment_parameters / split_segment_parameters. |
489 |
:param url: A URL, as string
|
5163.2.7
by Jelmer Vernooij
Add type checking. |
490 |
:param parameters: Dictionary of parameters, keys and values as bytestrings
|
5163.2.3
by Jelmer Vernooij
Add join_segment_parameters / split_segment_parameters. |
491 |
"""
|
492 |
(base, existing_parameters) = split_segment_parameters(url) |
|
493 |
new_parameters = {} |
|
494 |
new_parameters.update(existing_parameters) |
|
495 |
for key, value in parameters.iteritems(): |
|
5163.2.7
by Jelmer Vernooij
Add type checking. |
496 |
if type(key) is not str: |
497 |
raise TypeError("parameter key %r is not a bytestring" % key) |
|
498 |
if type(value) is not str: |
|
499 |
raise TypeError("parameter value %r for %s is not a bytestring" % |
|
500 |
(key, value)) |
|
5163.2.3
by Jelmer Vernooij
Add join_segment_parameters / split_segment_parameters. |
501 |
if "=" in key: |
502 |
raise errors.InvalidURLJoin("= exists in parameter key", url, |
|
503 |
parameters) |
|
504 |
new_parameters[key] = value |
|
5163.2.5
by Jelmer Vernooij
rename {split,join}_subsegments -> {split,join}_segment_parameters_raw and add more tests. |
505 |
return join_segment_parameters_raw(base, |
5163.2.6
by Jelmer Vernooij
Fix example names in tests. |
506 |
*["%s=%s" % item for item in sorted(new_parameters.items())]) |
5163.2.3
by Jelmer Vernooij
Add join_segment_parameters / split_segment_parameters. |
507 |
|
508 |
||
1711.2.44
by John Arbash Meinel
Factor out another win32 special case and add platform independent tests for it. |
509 |
def _win32_strip_local_trailing_slash(url): |
510 |
"""Strip slashes after the drive letter"""
|
|
511 |
if len(url) > WIN32_MIN_ABS_FILEURL_LENGTH: |
|
512 |
return url[:-1] |
|
513 |
else: |
|
514 |
return url |
|
515 |
||
516 |
||
1685.1.47
by John Arbash Meinel
s comes before u |
517 |
def strip_trailing_slash(url): |
518 |
"""Strip trailing slash, except for root paths.
|
|
519 |
||
520 |
The definition of 'root path' is platform-dependent.
|
|
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
521 |
This assumes that all URLs are valid netloc urls, such that they
|
522 |
form:
|
|
523 |
scheme://host/path
|
|
524 |
It searches for ://, and then refuses to remove the next '/'.
|
|
525 |
It can also handle relative paths
|
|
526 |
Examples:
|
|
527 |
path/to/foo => path/to/foo
|
|
528 |
path/to/foo/ => path/to/foo
|
|
529 |
http://host/path/ => http://host/path
|
|
530 |
http://host/path => http://host/path
|
|
531 |
http://host/ => http://host/
|
|
532 |
file:/// => file:///
|
|
533 |
file:///foo/ => file:///foo
|
|
534 |
# This is unique on win32 platforms, and is the only URL
|
|
535 |
# format which does it differently.
|
|
1711.4.8
by John Arbash Meinel
switch to prefering lowercase drive letters, since that matches os.getcwd() drive letters |
536 |
file:///c|/ => file:///c:/
|
1685.1.47
by John Arbash Meinel
s comes before u |
537 |
"""
|
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
538 |
if not url.endswith('/'): |
539 |
# Nothing to do
|
|
540 |
return url |
|
2245.6.1
by Alexander Belchenko
win32 UNC path: recursive cloning UNC path to root stops on //HOST, not on // |
541 |
if sys.platform == 'win32' and url.startswith('file://'): |
1711.2.44
by John Arbash Meinel
Factor out another win32 special case and add platform independent tests for it. |
542 |
return _win32_strip_local_trailing_slash(url) |
1685.1.80
by Wouter van Heyst
more code cleanup |
543 |
|
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
544 |
scheme_loc, first_path_slash = _find_scheme_and_separator(url) |
545 |
if scheme_loc is None: |
|
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
546 |
# This is a relative path, as it has no scheme
|
547 |
# so just chop off the last character
|
|
1685.1.47
by John Arbash Meinel
s comes before u |
548 |
return url[:-1] |
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
549 |
|
1685.1.49
by John Arbash Meinel
Added bzrlib.urlutils.split and basename + dirname |
550 |
if first_path_slash is None or first_path_slash == len(url)-1: |
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
551 |
# Don't chop off anything if the only slash is the path
|
552 |
# separating slash
|
|
1685.1.47
by John Arbash Meinel
s comes before u |
553 |
return url |
554 |
||
1685.1.48
by John Arbash Meinel
Updated strip_trailing_slash to support lots more url stuff, added tests |
555 |
return url[:-1] |
556 |
||
1685.1.47
by John Arbash Meinel
s comes before u |
557 |
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
558 |
def unescape(url): |
559 |
"""Unescape relpath from url format.
|
|
560 |
||
561 |
This returns a Unicode path from a URL
|
|
562 |
"""
|
|
563 |
# jam 20060427 URLs are supposed to be ASCII only strings
|
|
564 |
# If they are passed in as unicode, urllib.unquote
|
|
565 |
# will return a UNICODE string, which actually contains
|
|
566 |
# utf-8 bytes. So we have to ensure that they are
|
|
567 |
# plain ASCII strings, or the final .decode will
|
|
568 |
# try to encode the UNICODE => ASCII, and then decode
|
|
569 |
# it into utf-8.
|
|
570 |
try: |
|
571 |
url = str(url) |
|
572 |
except UnicodeError, e: |
|
573 |
raise errors.InvalidURL(url, 'URL was not a plain ASCII url: %s' % (e,)) |
|
1685.1.80
by Wouter van Heyst
more code cleanup |
574 |
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
575 |
unquoted = urllib.unquote(url) |
576 |
try: |
|
577 |
unicode_path = unquoted.decode('utf-8') |
|
578 |
except UnicodeError, e: |
|
579 |
raise errors.InvalidURL(url, 'Unable to encode the URL as utf-8: %s' % (e,)) |
|
580 |
return unicode_path |
|
581 |
||
582 |
||
583 |
# These are characters that if escaped, should stay that way
|
|
584 |
_no_decode_chars = ';/?:@&=+$,#' |
|
585 |
_no_decode_ords = [ord(c) for c in _no_decode_chars] |
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
586 |
_no_decode_hex = (['%02x' % o for o in _no_decode_ords] |
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
587 |
+ ['%02X' % o for o in _no_decode_ords]) |
1685.1.50
by John Arbash Meinel
Added an re for handling scheme paths. |
588 |
_hex_display_map = dict(([('%02x' % o, chr(o)) for o in range(256)] |
589 |
+ [('%02X' % o, chr(o)) for o in range(256)])) |
|
1685.1.51
by John Arbash Meinel
Working on getting normalize_url working. |
590 |
#These entries get mapped to themselves
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
591 |
_hex_display_map.update((hex,'%'+hex) for hex in _no_decode_hex) |
1685.1.51
by John Arbash Meinel
Working on getting normalize_url working. |
592 |
|
2208.4.1
by Andrew Bennetts
normalize_url should normalise escaping of unreserved characters, like '~'. |
593 |
# These characters shouldn't be percent-encoded, and it's always safe to
|
594 |
# unencode them if they are.
|
|
595 |
_url_dont_escape_characters = set( |
|
596 |
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha |
|
597 |
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha |
|
598 |
"0123456789" # Numbers |
|
599 |
"-._~" # Unreserved characters |
|
600 |
)
|
|
601 |
||
1685.1.51
by John Arbash Meinel
Working on getting normalize_url working. |
602 |
# These characters should not be escaped
|
2167.2.2
by Aaron Bentley
Update safe character list |
603 |
_url_safe_characters = set( |
604 |
"abcdefghijklmnopqrstuvwxyz" # Lowercase alpha |
|
605 |
"ABCDEFGHIJKLMNOPQRSTUVWXYZ" # Uppercase alpha |
|
606 |
"0123456789" # Numbers |
|
607 |
"_.-!~*'()" # Unreserved characters |
|
608 |
"/;?:@&=+$," # Reserved characters |
|
609 |
"%#" # Extra reserved characters |
|
610 |
)
|
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
611 |
|
1685.1.54
by John Arbash Meinel
url_for_display now makes sure output can be properly encoded. |
612 |
def unescape_for_display(url, encoding): |
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
613 |
"""Decode what you can for a URL, so that we get a nice looking path.
|
614 |
||
615 |
This will turn file:// urls into local paths, and try to decode
|
|
616 |
any portions of a http:// style url that it can.
|
|
1685.1.54
by John Arbash Meinel
url_for_display now makes sure output can be properly encoded. |
617 |
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
618 |
Any sections of the URL which can't be represented in the encoding or
|
1685.1.58
by Martin Pool
urlutils.unescape_for_display should return Unicode |
619 |
need to stay as escapes are left alone.
|
620 |
||
1685.1.54
by John Arbash Meinel
url_for_display now makes sure output can be properly encoded. |
621 |
:param url: A 7-bit ASCII URL
|
622 |
:param encoding: The final output encoding
|
|
1685.1.58
by Martin Pool
urlutils.unescape_for_display should return Unicode |
623 |
|
3943.8.1
by Marius Kruger
remove all trailing whitespace from bzr source |
624 |
:return: A unicode string which can be safely encoded into the
|
1685.1.58
by Martin Pool
urlutils.unescape_for_display should return Unicode |
625 |
specified encoding.
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
626 |
"""
|
3376.2.4
by Martin Pool
Remove every assert statement from bzrlib! |
627 |
if encoding is None: |
628 |
raise ValueError('you cannot specify None for the display encoding') |
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
629 |
if url.startswith('file://'): |
1685.1.54
by John Arbash Meinel
url_for_display now makes sure output can be properly encoded. |
630 |
try: |
631 |
path = local_path_from_url(url) |
|
1685.1.58
by Martin Pool
urlutils.unescape_for_display should return Unicode |
632 |
path.encode(encoding) |
633 |
return path |
|
1685.1.54
by John Arbash Meinel
url_for_display now makes sure output can be properly encoded. |
634 |
except UnicodeError: |
635 |
return url |
|
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
636 |
|
637 |
# Split into sections to try to decode utf-8
|
|
638 |
res = url.split('/') |
|
639 |
for i in xrange(1, len(res)): |
|
640 |
escaped_chunks = res[i].split('%') |
|
641 |
for j in xrange(1, len(escaped_chunks)): |
|
642 |
item = escaped_chunks[j] |
|
643 |
try: |
|
644 |
escaped_chunks[j] = _hex_display_map[item[:2]] + item[2:] |
|
645 |
except KeyError: |
|
646 |
# Put back the percent symbol
|
|
647 |
escaped_chunks[j] = '%' + item |
|
648 |
except UnicodeDecodeError: |
|
649 |
escaped_chunks[j] = unichr(int(item[:2], 16)) + item[2:] |
|
650 |
unescaped = ''.join(escaped_chunks) |
|
651 |
try: |
|
1685.1.54
by John Arbash Meinel
url_for_display now makes sure output can be properly encoded. |
652 |
decoded = unescaped.decode('utf-8') |
1685.1.45
by John Arbash Meinel
Moved url functions into bzrlib.urlutils |
653 |
except UnicodeDecodeError: |
654 |
# If this path segment cannot be properly utf-8 decoded
|
|
655 |
# after doing unescaping we will just leave it alone
|
|
656 |
pass
|
|
1685.1.54
by John Arbash Meinel
url_for_display now makes sure output can be properly encoded. |
657 |
else: |
658 |
try: |
|
1685.1.58
by Martin Pool
urlutils.unescape_for_display should return Unicode |
659 |
decoded.encode(encoding) |
1685.1.54
by John Arbash Meinel
url_for_display now makes sure output can be properly encoded. |
660 |
except UnicodeEncodeError: |
661 |
# If this chunk cannot be encoded in the local
|
|
662 |
# encoding, then we should leave it alone
|
|
663 |
pass
|
|
1685.1.58
by Martin Pool
urlutils.unescape_for_display should return Unicode |
664 |
else: |
665 |
# Otherwise take the url decoded one
|
|
666 |
res[i] = decoded |
|
667 |
return u'/'.join(res) |
|
2512.4.1
by Ian Clatworthy
Fixes #115491 - 'branch lp:projname' now creates ./projname as exected |
668 |
|
669 |
||
670 |
def derive_to_location(from_location): |
|
671 |
"""Derive a TO_LOCATION given a FROM_LOCATION.
|
|
672 |
||
673 |
The normal case is a FROM_LOCATION of http://foo/bar => bar.
|
|
674 |
The Right Thing for some logical destinations may differ though
|
|
675 |
because no / may be present at all. In that case, the result is
|
|
676 |
the full name without the scheme indicator, e.g. lp:foo-bar => foo-bar.
|
|
677 |
This latter case also applies when a Windows drive
|
|
678 |
is used without a path, e.g. c:foo-bar => foo-bar.
|
|
679 |
If no /, path separator or : is found, the from_location is returned.
|
|
680 |
"""
|
|
681 |
if from_location.find("/") >= 0 or from_location.find(os.sep) >= 0: |
|
682 |
return os.path.basename(from_location.rstrip("/\\")) |
|
683 |
else: |
|
684 |
sep = from_location.find(":") |
|
685 |
if sep > 0: |
|
686 |
return from_location[sep+1:] |
|
687 |
else: |
|
688 |
return from_location |
|
3242.3.26
by Aaron Bentley
Implement rebase_url |
689 |
|
3242.3.35
by Aaron Bentley
Cleanups and documentation |
690 |
|
3242.3.26
by Aaron Bentley
Implement rebase_url |
691 |
def _is_absolute(url): |
692 |
return (osutils.pathjoin('/foo', url) == url) |
|
693 |
||
3242.3.35
by Aaron Bentley
Cleanups and documentation |
694 |
|
3242.3.26
by Aaron Bentley
Implement rebase_url |
695 |
def rebase_url(url, old_base, new_base): |
696 |
"""Convert a relative path from an old base URL to a new base URL.
|
|
697 |
||
698 |
The result will be a relative path.
|
|
699 |
Absolute paths and full URLs are returned unaltered.
|
|
700 |
"""
|
|
701 |
scheme, separator = _find_scheme_and_separator(url) |
|
702 |
if scheme is not None: |
|
703 |
return url |
|
704 |
if _is_absolute(url): |
|
705 |
return url |
|
706 |
old_parsed = urlparse.urlparse(old_base) |
|
707 |
new_parsed = urlparse.urlparse(new_base) |
|
708 |
if (old_parsed[:2]) != (new_parsed[:2]): |
|
3242.3.33
by Aaron Bentley
Handle relative URL stacking cleanly |
709 |
raise errors.InvalidRebaseURLs(old_base, new_base) |
3242.3.36
by Aaron Bentley
Updates from review comments |
710 |
return determine_relative_path(new_parsed[2], |
3567.2.1
by Michael Hudson
urlutils.rebase_url handles '..' path segments in 'url' |
711 |
join(old_parsed[2], url)) |
3242.3.26
by Aaron Bentley
Implement rebase_url |
712 |
|
713 |
||
714 |
def determine_relative_path(from_path, to_path): |
|
715 |
"""Determine a relative path from from_path to to_path."""
|
|
716 |
from_segments = osutils.splitpath(from_path) |
|
717 |
to_segments = osutils.splitpath(to_path) |
|
718 |
count = -1 |
|
719 |
for count, (from_element, to_element) in enumerate(zip(from_segments, |
|
720 |
to_segments)): |
|
721 |
if from_element != to_element: |
|
722 |
break
|
|
723 |
else: |
|
724 |
count += 1 |
|
725 |
unique_from = from_segments[count:] |
|
726 |
unique_to = to_segments[count:] |
|
727 |
segments = (['..'] * len(unique_from) + unique_to) |
|
728 |
if len(segments) == 0: |
|
729 |
return '.' |
|
730 |
return osutils.pathjoin(*segments) |
|
3873.3.1
by Martin Pool
Move Transport._split_url to urlutils, and ad a simple test |
731 |
|
732 |
||
6055.2.7
by Jelmer Vernooij
Change parse_url to URL.from_string. |
733 |
class URL(object): |
6055.2.1
by Jelmer Vernooij
Add UnparsedUrl. |
734 |
"""Parsed URL."""
|
735 |
||
6055.2.7
by Jelmer Vernooij
Change parse_url to URL.from_string. |
736 |
def __init__(self, scheme, quoted_user, quoted_password, quoted_host, |
737 |
port, quoted_path): |
|
6055.2.1
by Jelmer Vernooij
Add UnparsedUrl. |
738 |
self.scheme = scheme |
739 |
self.quoted_host = quoted_host |
|
740 |
self.host = urllib.unquote(self.quoted_host) |
|
741 |
self.quoted_user = quoted_user |
|
742 |
if self.quoted_user is not None: |
|
743 |
self.user = urllib.unquote(self.quoted_user) |
|
744 |
else: |
|
745 |
self.user = None |
|
746 |
self.quoted_password = quoted_password |
|
747 |
if self.quoted_password is not None: |
|
748 |
self.password = urllib.unquote(self.quoted_password) |
|
749 |
else: |
|
750 |
self.password = None |
|
751 |
self.port = port |
|
752 |
self.quoted_path = quoted_path |
|
753 |
self.path = urllib.unquote(self.quoted_path) |
|
754 |
||
6055.2.8
by Jelmer Vernooij
Add repr() |
755 |
def __eq__(self, other): |
756 |
return (isinstance(other, self.__class__) and |
|
757 |
self.scheme == other.scheme and |
|
758 |
self.host == other.host and |
|
759 |
self.user == other.user and |
|
760 |
self.password == other.password and |
|
761 |
self.path == other.path) |
|
762 |
||
763 |
def __repr__(self): |
|
6055.2.11
by Jelmer Vernooij
Fix tests. |
764 |
return "<%s(%r, %r, %r, %r, %r, %r)>" % ( |
6055.2.8
by Jelmer Vernooij
Add repr() |
765 |
self.__class__.__name__, |
6055.2.11
by Jelmer Vernooij
Fix tests. |
766 |
self.scheme, self.quoted_user, self.quoted_password, |
767 |
self.quoted_host, self.port, self.quoted_path) |
|
6055.2.8
by Jelmer Vernooij
Add repr() |
768 |
|
6055.2.6
by Jelmer Vernooij
Split out parse_url. |
769 |
@classmethod
|
6055.2.7
by Jelmer Vernooij
Change parse_url to URL.from_string. |
770 |
def from_string(cls, url): |
771 |
"""Create a URL object from a string.
|
|
6055.2.6
by Jelmer Vernooij
Split out parse_url. |
772 |
|
773 |
:param url: URL as bytestring
|
|
774 |
"""
|
|
775 |
if isinstance(url, unicode): |
|
776 |
raise errors.InvalidURL('should be ascii:\n%r' % url) |
|
777 |
url = url.encode('utf-8') |
|
778 |
(scheme, netloc, path, params, |
|
779 |
query, fragment) = urlparse.urlparse(url, allow_fragments=False) |
|
780 |
user = password = host = port = None |
|
781 |
if '@' in netloc: |
|
782 |
user, host = netloc.rsplit('@', 1) |
|
783 |
if ':' in user: |
|
784 |
user, password = user.split(':', 1) |
|
785 |
else: |
|
786 |
host = netloc |
|
787 |
||
788 |
if ':' in host and not (host[0] == '[' and host[-1] == ']'): #there *is* port |
|
789 |
host, port = host.rsplit(':',1) |
|
790 |
try: |
|
791 |
port = int(port) |
|
792 |
except ValueError: |
|
793 |
raise errors.InvalidURL('invalid port number %s in url:\n%s' % |
|
794 |
(port, url)) |
|
795 |
if host != "" and host[0] == '[' and host[-1] == ']': #IPv6 |
|
796 |
host = host[1:-1] |
|
797 |
||
798 |
return cls(scheme, user, password, host, port, path) |
|
799 |
||
3873.3.1
by Martin Pool
Move Transport._split_url to urlutils, and ad a simple test |
800 |
|
801 |
def parse_url(url): |
|
802 |
"""Extract the server address, the credentials and the path from the url.
|
|
803 |
||
804 |
user, password, host and path should be quoted if they contain reserved
|
|
805 |
chars.
|
|
806 |
||
807 |
:param url: an quoted url
|
|
808 |
:return: (scheme, user, password, host, port, path) tuple, all fields
|
|
809 |
are unquoted.
|
|
810 |
"""
|
|
6055.2.7
by Jelmer Vernooij
Change parse_url to URL.from_string. |
811 |
parsed_url = URL.from_string(url) |
6055.2.6
by Jelmer Vernooij
Split out parse_url. |
812 |
return (parsed_url.scheme, parsed_url.user, parsed_url.password, |
813 |
parsed_url.host, parsed_url.port, parsed_url.path) |