Mercurial > public > sg101

comparison gpp/legacy/phpbb.py @ 294:254db4cb6a86

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Changes / scripts to import forums. Other tweaks and moving other import scripts to the legacy application.
author Brian Neal <bgneal@gmail.com>
date Wed, 05 Jan 2011 04:09:35 +0000
parents 64c188a9d31f
children
comparison
equal deleted inserted replaced
293:c92fb89dbc7d 294:254db4cb6a86
46 pass 46 pass
47 return text # leave as is 47 return text # leave as is
48 return re.sub("&#?\w+;", fixup, text) 48 return re.sub("&#?\w+;", fixup, text)
49 49
50 50
51 def unphpbb(s): 51 def unphpbb(s, encoding='latin-1'):
52 """Converts BBCode from phpBB database data into 'pure' BBCode. 52 """Converts BBCode from phpBB database data into 'pure' BBCode.
53 53
54 phpBB doesn't store plain BBCode in its database. The BBCode tags have 54 phpBB doesn't store plain BBCode in its database. The BBCode tags have
55 "uids" added to them and the data has already been HTML entity'ized. 55 "uids" added to them and the data has already been HTML entity'ized.
56 This function removes the uid stuff and undoes the entity'ification and 56 This function removes the uid stuff and undoes the entity'ification and
57 returns the result as a unicode string. 57 returns the result as a unicode string.
58 58
59 If the input 's' is not already unicode, it will be decoded using the
60 supplied encoding.
61
59 """ 62 """
60 if not isinstance(s, unicode): 63 if not isinstance(s, unicode):
61 s = s.decode('utf-8', 'replace') 64 s = s.decode(encoding, 'replace')
62 for start, end in BBCODE_RES: 65 for start, end in BBCODE_RES:
63 s = re.sub(start, r'\1', s, re.MULTILINE) 66 s = re.sub(start, r'\1', s, re.MULTILINE)
64 s = re.sub(end, r'\1]', s, re.MULTILINE) 67 s = re.sub(end, r'\1]', s, re.MULTILINE)
65 return unescape(s) 68 return unescape(s)