Changes by Serhiy Storchaka storch...@gmail.com:
--
resolution: - wont fix
stage: - resolved
status: open - closed
___
Python tracker rep...@bugs.python.org
http://bugs.python.org/issue23541
___
New submission from Ceridwen:
Attached is a three-line script whose third line, a call to match() for a
compiled regex, fails to halt on Python 2.7 and on 3.4 (after changing ur in
front of the regex string to r to accommodate the change in Unicode handling).
I found it by stepping through
Serhiy Storchaka added the comment:
Minimal example for your case is
re.match(r'(\w+\s*)+\s+\d+', 'Compiling Regular Patterns to Sequential
Machines')
It doesn't halt, it just needs too long time to finish. Your should rewrite the
regular expression to avoid catastrophic backtracking
Matthew Barnett added the comment:
The problem is with the (\w+\s*)+. \s* can match an empty string, so when
matching a word it's like (\w+)+.
If you have n letters, there are 2**n ways it could match, and if what follows
never matches, it'll try all of them.
It _will_ finish, eventually...
