Mercurial > cpython-lru_cache-weakref
annotate Lib/urlparse.py @ 16413:8819867f14ef legacy-trunk
Patch derived from Trent's 101162: a Python/C API testing framework.
STILL NEEDS UNIX BUILD CHANGES.
| author | Tim Peters <tim.peters@gmail.com> |
|---|---|
| date | Fri, 02 Feb 2001 05:57:15 +0000 |
| parents | 4c01ee0a8073 |
| children | 0feb63490093 |
| rev | line source |
|---|---|
|
11707
6f4146846337
The third and final doc-string sweep by Ka-Ping Yee.
Guido van Rossum <guido@python.org>
parents:
10830
diff
changeset
|
1 """Parse (absolute and relative) URLs. |
|
6f4146846337
The third and final doc-string sweep by Ka-Ping Yee.
Guido van Rossum <guido@python.org>
parents:
10830
diff
changeset
|
2 |
|
6f4146846337
The third and final doc-string sweep by Ka-Ping Yee.
Guido van Rossum <guido@python.org>
parents:
10830
diff
changeset
|
3 See RFC 1808: "Relative Uniform Resource Locators", by R. Fielding, |
|
6f4146846337
The third and final doc-string sweep by Ka-Ping Yee.
Guido van Rossum <guido@python.org>
parents:
10830
diff
changeset
|
4 UC Irvine, June 1995. |
|
6f4146846337
The third and final doc-string sweep by Ka-Ping Yee.
Guido van Rossum <guido@python.org>
parents:
10830
diff
changeset
|
5 """ |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
6 |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
7 # A classification of schemes ('' means apply by default) |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
8 uses_relative = ['ftp', 'http', 'gopher', 'nntp', 'wais', 'file', |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
9 'https', 'shttp', |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
10 'prospero', 'rtsp', 'rtspu', ''] |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
11 uses_netloc = ['ftp', 'http', 'gopher', 'nntp', 'telnet', 'wais', |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
12 'file', |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
13 'https', 'shttp', 'snews', |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
14 'prospero', 'rtsp', 'rtspu', ''] |
|
4853
cf90745381bf
Added characteristics of shttp, https, and snews.
Guido van Rossum <guido@python.org>
parents:
4831
diff
changeset
|
15 non_hierarchical = ['gopher', 'hdl', 'mailto', 'news', 'telnet', 'wais', |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
16 'snews', 'sip', |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
17 ] |
|
4853
cf90745381bf
Added characteristics of shttp, https, and snews.
Guido van Rossum <guido@python.org>
parents:
4831
diff
changeset
|
18 uses_params = ['ftp', 'hdl', 'prospero', 'http', |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
19 'https', 'shttp', 'rtsp', 'rtspu', 'sip', |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
20 ''] |
|
4853
cf90745381bf
Added characteristics of shttp, https, and snews.
Guido van Rossum <guido@python.org>
parents:
4831
diff
changeset
|
21 uses_query = ['http', 'wais', |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
22 'https', 'shttp', |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
23 'gopher', 'rtsp', 'rtspu', 'sip', |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
24 ''] |
|
3641
e4faadb4abe2
added hdl protocol properties
Guido van Rossum <guido@python.org>
parents:
3482
diff
changeset
|
25 uses_fragment = ['ftp', 'hdl', 'http', 'gopher', 'news', 'nntp', 'wais', |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
26 'https', 'shttp', 'snews', |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
27 'file', 'prospero', ''] |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
28 |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
29 # Characters valid in scheme names |
|
15623
db69c092c906
Be explicit about scheme_chars -- string.letters is locale dependent
Guido van Rossum <guido@python.org>
parents:
12967
diff
changeset
|
30 scheme_chars = ('abcdefghijklmnopqrstuvwxyz' |
|
db69c092c906
Be explicit about scheme_chars -- string.letters is locale dependent
Guido van Rossum <guido@python.org>
parents:
12967
diff
changeset
|
31 'ABCDEFGHIJKLMNOPQRSTUVWXYZ' |
|
db69c092c906
Be explicit about scheme_chars -- string.letters is locale dependent
Guido van Rossum <guido@python.org>
parents:
12967
diff
changeset
|
32 '0123456789' |
|
db69c092c906
Be explicit about scheme_chars -- string.letters is locale dependent
Guido van Rossum <guido@python.org>
parents:
12967
diff
changeset
|
33 '+-.') |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
34 |
|
5785
b247297be9a0
After some discussion with Jeremy and Fred, decided to limit the
Guido van Rossum <guido@python.org>
parents:
5776
diff
changeset
|
35 MAX_CACHE_SIZE = 20 |
|
3652
728a39a3e0d5
optimizations due to Fred Drake; added urldefrag() function
Guido van Rossum <guido@python.org>
parents:
3641
diff
changeset
|
36 _parse_cache = {} |
|
728a39a3e0d5
optimizations due to Fred Drake; added urldefrag() function
Guido van Rossum <guido@python.org>
parents:
3641
diff
changeset
|
37 |
|
728a39a3e0d5
optimizations due to Fred Drake; added urldefrag() function
Guido van Rossum <guido@python.org>
parents:
3641
diff
changeset
|
38 def clear_cache(): |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
39 """Clear the parse cache.""" |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
40 global _parse_cache |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
41 _parse_cache = {} |
|
3652
728a39a3e0d5
optimizations due to Fred Drake; added urldefrag() function
Guido van Rossum <guido@python.org>
parents:
3641
diff
changeset
|
42 |
|
728a39a3e0d5
optimizations due to Fred Drake; added urldefrag() function
Guido van Rossum <guido@python.org>
parents:
3641
diff
changeset
|
43 |
|
9051
839c6622d15a
fix typo in keyword argument 'allow_frament' should be 'allow_fragment'
Jeremy Hylton <jeremy@alum.mit.edu>
parents:
7125
diff
changeset
|
44 def urlparse(url, scheme = '', allow_fragments = 1): |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
45 """Parse a URL into 6 components: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
46 <scheme>://<netloc>/<path>;<params>?<query>#<fragment> |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
47 Return a 6-tuple: (scheme, netloc, path, params, query, fragment). |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
48 Note that we don't break the components up in smaller bits |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
49 (e.g. netloc is a single string) and we don't expand % escapes.""" |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
50 key = url, scheme, allow_fragments |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
51 cached = _parse_cache.get(key, None) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
52 if cached: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
53 return cached |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
54 if len(_parse_cache) >= MAX_CACHE_SIZE: # avoid runaway growth |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
55 clear_cache() |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
56 netloc = path = params = query = fragment = '' |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
57 i = url.find(':') |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
58 if i > 0: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
59 if url[:i] == 'http': # optimize the common case |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
60 scheme = url[:i].lower() |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
61 url = url[i+1:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
62 if url[:2] == '//': |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
63 i = url.find('/', 2) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
64 if i < 0: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
65 i = len(url) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
66 netloc = url[2:i] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
67 url = url[i:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
68 if allow_fragments: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
69 i = url.rfind('#') |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
70 if i >= 0: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
71 fragment = url[i+1:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
72 url = url[:i] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
73 i = url.find('?') |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
74 if i >= 0: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
75 query = url[i+1:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
76 url = url[:i] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
77 i = url.find(';') |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
78 if i >= 0: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
79 params = url[i+1:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
80 url = url[:i] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
81 tuple = scheme, netloc, url, params, query, fragment |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
82 _parse_cache[key] = tuple |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
83 return tuple |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
84 for c in url[:i]: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
85 if c not in scheme_chars: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
86 break |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
87 else: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
88 scheme, url = url[:i].lower(), url[i+1:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
89 if scheme in uses_netloc: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
90 if url[:2] == '//': |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
91 i = url.find('/', 2) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
92 if i < 0: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
93 i = len(url) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
94 netloc, url = url[2:i], url[i:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
95 if allow_fragments and scheme in uses_fragment: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
96 i = url.rfind('#') |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
97 if i >= 0: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
98 url, fragment = url[:i], url[i+1:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
99 if scheme in uses_query: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
100 i = url.find('?') |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
101 if i >= 0: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
102 url, query = url[:i], url[i+1:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
103 if scheme in uses_params: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
104 i = url.find(';') |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
105 if i >= 0: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
106 url, params = url[:i], url[i+1:] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
107 tuple = scheme, netloc, url, params, query, fragment |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
108 _parse_cache[key] = tuple |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
109 return tuple |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
110 |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
111 def urlunparse((scheme, netloc, url, params, query, fragment)): |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
112 """Put a parsed URL back together again. This may result in a |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
113 slightly different, but equivalent URL, if the URL that was parsed |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
114 originally had redundant delimiters, e.g. a ? with an empty query |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
115 (the draft states that these are equivalent).""" |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
116 if netloc or (scheme in uses_netloc and url[:2] == '//'): |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
117 if url and url[:1] != '/': url = '/' + url |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
118 url = '//' + (netloc or '') + url |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
119 if scheme: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
120 url = scheme + ':' + url |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
121 if params: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
122 url = url + ';' + params |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
123 if query: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
124 url = url + '?' + query |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
125 if fragment: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
126 url = url + '#' + fragment |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
127 return url |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
128 |
|
9051
839c6622d15a
fix typo in keyword argument 'allow_frament' should be 'allow_fragment'
Jeremy Hylton <jeremy@alum.mit.edu>
parents:
7125
diff
changeset
|
129 def urljoin(base, url, allow_fragments = 1): |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
130 """Join a base URL and a possibly relative URL to form an absolute |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
131 interpretation of the latter.""" |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
132 if not base: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
133 return url |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
134 if not url: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
135 return base |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
136 bscheme, bnetloc, bpath, bparams, bquery, bfragment = \ |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
137 urlparse(base, '', allow_fragments) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
138 scheme, netloc, path, params, query, fragment = \ |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
139 urlparse(url, bscheme, allow_fragments) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
140 if scheme != bscheme or scheme not in uses_relative: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
141 return url |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
142 if scheme in uses_netloc: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
143 if netloc: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
144 return urlunparse((scheme, netloc, path, |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
145 params, query, fragment)) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
146 netloc = bnetloc |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
147 if path[:1] == '/': |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
148 return urlunparse((scheme, netloc, path, |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
149 params, query, fragment)) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
150 if not path: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
151 if not params: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
152 params = bparams |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
153 if not query: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
154 query = bquery |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
155 return urlunparse((scheme, netloc, bpath, |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
156 params, query, fragment)) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
157 segments = bpath.split('/')[:-1] + path.split('/') |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
158 # XXX The stuff below is bogus in various ways... |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
159 if segments[-1] == '.': |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
160 segments[-1] = '' |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
161 while '.' in segments: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
162 segments.remove('.') |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
163 while 1: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
164 i = 1 |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
165 n = len(segments) - 1 |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
166 while i < n: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
167 if (segments[i] == '..' |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
168 and segments[i-1] not in ('', '..')): |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
169 del segments[i-1:i+1] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
170 break |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
171 i = i+1 |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
172 else: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
173 break |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
174 if segments == ['', '..']: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
175 segments[-1] = '' |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
176 elif len(segments) >= 2 and segments[-1] == '..': |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
177 segments[-2:] = [''] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
178 return urlunparse((scheme, netloc, '/'.join(segments), |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
179 params, query, fragment)) |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
180 |
|
3652
728a39a3e0d5
optimizations due to Fred Drake; added urldefrag() function
Guido van Rossum <guido@python.org>
parents:
3641
diff
changeset
|
181 def urldefrag(url): |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
182 """Removes any existing fragment from URL. |
|
3652
728a39a3e0d5
optimizations due to Fred Drake; added urldefrag() function
Guido van Rossum <guido@python.org>
parents:
3641
diff
changeset
|
183 |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
184 Returns a tuple of the defragmented URL and the fragment. If |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
185 the URL contained no fragments, the second element is the |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
186 empty string. |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
187 """ |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
188 s, n, p, a, q, frag = urlparse(url) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
189 defrag = urlunparse((s, n, p, a, q, '')) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
190 return defrag, frag |
|
3652
728a39a3e0d5
optimizations due to Fred Drake; added urldefrag() function
Guido van Rossum <guido@python.org>
parents:
3641
diff
changeset
|
191 |
|
728a39a3e0d5
optimizations due to Fred Drake; added urldefrag() function
Guido van Rossum <guido@python.org>
parents:
3641
diff
changeset
|
192 |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
193 test_input = """ |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
194 http://a/b/c/d |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
195 |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
196 g:h = <URL:g:h> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
197 http:g = <URL:http://a/b/c/g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
198 http: = <URL:http://a/b/c/d> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
199 g = <URL:http://a/b/c/g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
200 ./g = <URL:http://a/b/c/g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
201 g/ = <URL:http://a/b/c/g/> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
202 /g = <URL:http://a/g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
203 //g = <URL:http://g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
204 ?y = <URL:http://a/b/c/d?y> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
205 g?y = <URL:http://a/b/c/g?y> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
206 g?y/./x = <URL:http://a/b/c/g?y/./x> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
207 . = <URL:http://a/b/c/> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
208 ./ = <URL:http://a/b/c/> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
209 .. = <URL:http://a/b/> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
210 ../ = <URL:http://a/b/> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
211 ../g = <URL:http://a/b/g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
212 ../.. = <URL:http://a/> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
213 ../../g = <URL:http://a/g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
214 ../../../g = <URL:http://a/../g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
215 ./../g = <URL:http://a/b/g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
216 ./g/. = <URL:http://a/b/c/g/> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
217 /./g = <URL:http://a/./g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
218 g/./h = <URL:http://a/b/c/g/h> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
219 g/../h = <URL:http://a/b/c/h> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
220 http:g = <URL:http://a/b/c/g> |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
221 http: = <URL:http://a/b/c/d> |
|
9793
737d05a9bb5e
Fixed bug in the common-case code for HTTP URLs; it would lose the query,
Andrew M. Kuchling <amk@amk.ca>
parents:
9665
diff
changeset
|
222 http:?y = <URL:http://a/b/c/d?y> |
|
737d05a9bb5e
Fixed bug in the common-case code for HTTP URLs; it would lose the query,
Andrew M. Kuchling <amk@amk.ca>
parents:
9665
diff
changeset
|
223 http:g?y = <URL:http://a/b/c/g?y> |
|
737d05a9bb5e
Fixed bug in the common-case code for HTTP URLs; it would lose the query,
Andrew M. Kuchling <amk@amk.ca>
parents:
9665
diff
changeset
|
224 http:g?y/./x = <URL:http://a/b/c/g?y/./x> |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
225 """ |
|
9665
e11d75005a5a
Add XXX comment about a test that doesn't seem right -- no time to
Guido van Rossum <guido@python.org>
parents:
9068
diff
changeset
|
226 # XXX The result for //g is actually http://g/; is this a problem? |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
227 |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
228 def test(): |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
229 import sys |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
230 base = '' |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
231 if sys.argv[1:]: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
232 fn = sys.argv[1] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
233 if fn == '-': |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
234 fp = sys.stdin |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
235 else: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
236 fp = open(fn) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
237 else: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
238 import StringIO |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
239 fp = StringIO.StringIO(test_input) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
240 while 1: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
241 line = fp.readline() |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
242 if not line: break |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
243 words = line.split() |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
244 if not words: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
245 continue |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
246 url = words[0] |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
247 parts = urlparse(url) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
248 print '%-10s : %s' % (url, parts) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
249 abs = urljoin(base, url) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
250 if not base: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
251 base = abs |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
252 wrapped = '<URL:%s>' % abs |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
253 print '%-10s = %s' % (url, wrapped) |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
254 if len(words) == 3 and words[1] == '=': |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
255 if wrapped != words[2]: |
|
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
256 print 'EXPECTED', words[2], '!!!!!!!!!!' |
|
1875
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
257 |
|
8c9cc054c5ed
New tty/pty modules by Steen; new urlparser.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
258 if __name__ == '__main__': |
|
15874
4c01ee0a8073
Whitespace normalization. Top level of Lib now fixed-point for reindent.py!
Tim Peters <tim.peters@gmail.com>
parents:
15741
diff
changeset
|
259 test() |
