Mercurial > traipse_dev

comparison orpg/orpg_xml.py @ 23:551cd440acce traipse_dev

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Final Dev build of the Day. Should reduce RAM usage more.
author sirebral
date Sat, 25 Jul 2009 21:05:18 -0500
parents 4385a7d0efd1
children c54768cffbd4
comparison
equal deleted inserted replaced
22:88cea66228d6 23:551cd440acce
51 return t_node 51 return t_node
52 52
53 def strip_unicode(txt): 53 def strip_unicode(txt):
54 for i in xrange(len(txt)): 54 for i in xrange(len(txt)):
55 if txt[i] not in string.printable: 55 if txt[i] not in string.printable:
56 try: 56 try: txt = txt.replace(txt[i], '&#' + str(ord(txt[i])) + ';')
57 txt = txt.replace(txt[i], '&#' + str(ord(txt[i])) + ';') 57 except: txt = txt.replace(txt[i], '{?}')
58 except:
59 txt = txt.replace(txt[i], '{?}')
60 return txt 58 return txt
61 59
62 def strip_text(txt): 60 def strip_text(txt):
63 # The following block strips out 8-bit characters 61 # The following block strips out 8-bit characters
64 u_txt = "" 62 u_txt = ""
65 bad_txt_found = 0 63 bad_txt_found = 0
66 txt = strip_unicode(txt) 64 txt = strip_unicode(txt)
67 for c in txt: 65 for c in txt:
68 if ord(c) < 128: 66 if ord(c) < 128: u_txt += c
69 u_txt += c 67 else: bad_txt_found = 1
70 else: 68 if bad_txt_found: print "Some non 7-bit ASCII characters found and stripped"
71 bad_txt_found = 1
72 if bad_txt_found:
73 print "Some non 7-bit ASCII characters found and stripped"
74 return u_txt 69 return u_txt