@@ -18,8 +18,61 @@ def capfirst(x):
18 18
return x and str(x)[0].upper() + str(x)[1:]
19 19 20 20 21 -
# Set up regular expressions
22 -
re_words = _lazy_re_compile(r'<[^>]+?>|([^<>\s]+)', re.S)
21 +
# ----- Begin security-related performance workaround -----
22 + 23 +
# We used to have, below
24 +
#
25 +
# re_words = _lazy_re_compile(r"<[^>]+?>|([^<>\s]+)", re.S)
26 +
#
27 +
# But it was shown that this regex, in the way we use it here, has some
28 +
# catastrophic edge-case performance features. Namely, when it is applied to
29 +
# text with only open brackets "<<<...". The class below provides the services
30 +
# and correct answers for the use cases, but in these edge cases does it much
31 +
# faster.
32 +
re_notag = _lazy_re_compile(r"([^<>\s]+)", re.S)
33 +
re_prt = _lazy_re_compile(r"<|([^<>\s]+)", re.S)
34 + 35 + 36 +
class WordsRegex:
37 +
@staticmethod
38 +
def search(text, pos):
39 +
# Look for "<" or a non-tag word.
40 +
partial = re_prt.search(text, pos)
41 +
if partial is None or partial[1] is not None:
42 +
return partial
43 + 44 +
# "<" was found, look for a closing ">".
45 +
end = text.find(">", partial.end(0))
46 +
if end < 0:
47 +
# ">" cannot be found, look for a word.
48 +
return re_notag.search(text, pos + 1)
49 +
else:
50 +
# "<" followed by a ">" was found -- fake a match.
51 +
end += 1
52 +
return FakeMatch(text[partial.start(0): end], end)
53 + 54 + 55 +
class FakeMatch:
56 +
__slots__ = ["_text", "_end"]
57 + 58 +
def end(self, group=0):
59 +
assert group == 0, "This specific object takes only group=0"
60 +
return self._end
61 + 62 +
def __getitem__(self, group):
63 +
if group == 1:
64 +
return None
65 +
assert group == 0, "This specific object takes only group in {0,1}"
66 +
return self._text
67 + 68 +
def __init__(self, text, end):
69 +
self._text, self._end = text, end
70 + 71 + 72 +
# ----- End security-related performance workaround -----
73 + 74 +
# Set up regular expressions.
75 +
re_words = WordsRegex
23 76
re_chars = _lazy_re_compile(r'<[^>]+?>|(.)', re.S)
24 77
re_tag = _lazy_re_compile(r'<(/)?(\S+?)(?:(\s*/)|\s.*?)?>', re.S)
25 78
re_newlines = _lazy_re_compile(r'\r\n|\r') # Used in normalize_newlines
RetroSearch is an open source project built by @garambo | Open a GitHub Issue
Search and Browse the WWW like it's 1997 | Search results from DuckDuckGo
HTML:
3.2
| Encoding:
UTF-8
| Version:
0.7.4