Mercurial > cpython-lru_cache-weakref
annotate Tools/scripts/parseentities.py @ 84007:dbfbdf2b5c19 3.3
Fix #17967: For ftp urls CWD to target instead of hopping to each directory
towards target. This fixes a bug where target is accessible, but parent
directories are restricted.
| author | Senthil Kumaran <senthil@uthcode.com> |
|---|---|
| date | Sat, 01 Jun 2013 08:27:06 -0700 |
| parents | 62dde5dd475e |
| children |
| rev | line source |
|---|---|
|
59522
fcfe324663a6
convert shebang lines: python -> python3
Benjamin Peterson <benjamin@python.org>
parents:
47100
diff
changeset
|
1 #!/usr/bin/env python3 |
|
11210
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
2 """ Utility for parsing HTML entity definitions available from: |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
3 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
4 http://www.w3.org/ as e.g. |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
5 http://www.w3.org/TR/REC-html40/HTMLlat1.ent |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
6 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
7 Input is read from stdin, output is written to stdout in form of a |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
8 Python snippet defining a dictionary "entitydefs" mapping literal |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
9 entity name to character or numeric entity. |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
10 |
|
15930
14a27acc755a
Whitespace normalization.
Tim Peters <tim.peters@gmail.com>
parents:
14657
diff
changeset
|
11 Marc-Andre Lemburg, mal@lemburg.com, 1999. |
|
11210
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
12 Use as you like. NO WARRANTIES. |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
13 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
14 """ |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
15 import re,sys |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
16 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
17 entityRE = re.compile('<!ENTITY +(\w+) +CDATA +"([^"]+)" +-- +((?:.|\n)+?) *-->') |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
18 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
19 def parse(text,pos=0,endpos=None): |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
20 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
21 pos = 0 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
22 if endpos is None: |
| 14657 | 23 endpos = len(text) |
|
11210
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
24 d = {} |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
25 while 1: |
| 14657 | 26 m = entityRE.search(text,pos,endpos) |
| 27 if not m: | |
| 28 break | |
| 29 name,charcode,comment = m.groups() | |
| 30 d[name] = charcode,comment | |
| 31 pos = m.end() | |
|
11210
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
32 return d |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
33 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
34 def writefile(f,defs): |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
35 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
36 f.write("entitydefs = {\n") |
|
47100
893a4dc17a91
Ran 2to3 over scripts directory.
Georg Brandl <georg@python.org>
parents:
15930
diff
changeset
|
37 items = sorted(defs.items()) |
|
893a4dc17a91
Ran 2to3 over scripts directory.
Georg Brandl <georg@python.org>
parents:
15930
diff
changeset
|
38 for name, (charcode,comment) in items: |
| 14657 | 39 if charcode[:2] == '&#': |
| 40 code = int(charcode[2:-1]) | |
| 41 if code < 256: | |
| 42 charcode = "'\%o'" % code | |
| 43 else: | |
| 44 charcode = repr(charcode) | |
| 45 else: | |
| 46 charcode = repr(charcode) | |
|
76106
62dde5dd475e
#14490, #14491: add 'sundry'-style import tests for Tools/scripts.
R David Murray <rdmurray@bitdance.com>
parents:
59522
diff
changeset
|
47 comment = ' '.join(comment.split()) |
| 14657 | 48 f.write(" '%s':\t%s, \t# %s\n" % (name,charcode,comment)) |
|
11210
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
49 f.write('\n}\n') |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
50 |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
51 if __name__ == '__main__': |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
52 if len(sys.argv) > 1: |
| 14657 | 53 infile = open(sys.argv[1]) |
|
11210
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
54 else: |
| 14657 | 55 infile = sys.stdin |
|
11210
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
56 if len(sys.argv) > 2: |
| 14657 | 57 outfile = open(sys.argv[2],'w') |
|
11210
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
58 else: |
| 14657 | 59 outfile = sys.stdout |
|
11210
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
60 text = infile.read() |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
61 defs = parse(text) |
|
cf7cd7b04fa6
Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff
changeset
|
62 writefile(outfile,defs) |
