@@ -23,8 +23,61 @@ def capfirst(x):
23 23
return x[0].upper() + x[1:]
24 24 25 25 26 -
# Set up regular expressions
27 -
re_words = _lazy_re_compile(r"<[^>]+?>|([^<>\s]+)", re.S)
26 +
# ----- Begin security-related performance workaround -----
27 + 28 +
# We used to have, below
29 +
#
30 +
# re_words = _lazy_re_compile(r"<[^>]+?>|([^<>\s]+)", re.S)
31 +
#
32 +
# But it was shown that this regex, in the way we use it here, has some
33 +
# catastrophic edge-case performance features. Namely, when it is applied to
34 +
# text with only open brackets "<<<...". The class below provides the services
35 +
# and correct answers for the use cases, but in these edge cases does it much
36 +
# faster.
37 +
re_notag = _lazy_re_compile(r"([^<>\s]+)", re.S)
38 +
re_prt = _lazy_re_compile(r"<|([^<>\s]+)", re.S)
39 + 40 + 41 +
class WordsRegex:
42 +
@staticmethod
43 +
def search(text, pos):
44 +
# Look for "<" or a non-tag word.
45 +
partial = re_prt.search(text, pos)
46 +
if partial is None or partial[1] is not None:
47 +
return partial
48 + 49 +
# "<" was found, look for a closing ">".
50 +
end = text.find(">", partial.end(0))
51 +
if end < 0:
52 +
# ">" cannot be found, look for a word.
53 +
return re_notag.search(text, pos + 1)
54 +
else:
55 +
# "<" followed by a ">" was found -- fake a match.
56 +
end += 1
57 +
return FakeMatch(text[partial.start(0) : end], end)
58 + 59 + 60 +
class FakeMatch:
61 +
__slots__ = ["_text", "_end"]
62 + 63 +
def end(self, group=0):
64 +
assert group == 0, "This specific object takes only group=0"
65 +
return self._end
66 + 67 +
def __getitem__(self, group):
68 +
if group == 1:
69 +
return None
70 +
assert group == 0, "This specific object takes only group in {0,1}"
71 +
return self._text
72 + 73 +
def __init__(self, text, end):
74 +
self._text, self._end = text, end
75 + 76 + 77 +
# ----- End security-related performance workaround -----
78 + 79 +
# Set up regular expressions.
80 +
re_words = WordsRegex
28 81
re_chars = _lazy_re_compile(r"<[^>]+?>|(.)", re.S)
29 82
re_tag = _lazy_re_compile(r"<(/)?(\S+?)(?:(\s*/)|\s.*?)?>", re.S)
30 83
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