This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author gnuthor
Recipients gnuthor
Date 2010-07-29.11:09:15
SpamBayes Score 0.0010446255
Marked as misclassified No
Message-id <1280401758.9.0.252357713623.issue9414@psf.upfronthosting.co.za>
In-reply-to
Content
The follow code hangs on vanilla compile python 2.7 on Ubuntu 10.04 x86_64 and all other versions of Python that I could find (i386/x86_64, 2.6.5/2.5.2/2.2):

>>> import re
>>> regex = re.compile(r'^((?:\.\d+)+|(?:\.?\w+(?:\-*\w+)+)+)\.(\d*?)$')
>>> match = regex.match('lldpLocChassisIdSubtype')
<infinite loop or possibly exponential time>

the regex is taken from file client.py of libsnmp-python ( version 5.4.2.1 ~dfsg0ubuntu1-0ubuntu2.1 in Ubuntu 10.04 )

libsnmp-python ( 5.4.1~dfsg-12  in Debian Lenny) contains
a previous version of this regex which does not produce an infinite loop

>>> regex = re.compile(r'^((?:\.\d+)+|(?:\.?\w+(?:\-*\w+)+)+)\.?(.*)$')
>>> match = regex.match('lldpLocChassisIdSubtype')
>>> print match
<_sre.SRE_Match object at 0x7f5d2abebc68>


Perl 5.10.1 can run both of the regular expressions without problems

$ perl -e '$x="lldpLocChassisIdSubtype"; print "$1" if $x =~ m/^((?:\.\d+)+|(?:\.?\w+(?:\-*\w+)+)+)\.(\d*?)$/;'
$ perl -e '$x="lldpLocChassisIdSubtype"; print "$1" if $x =~ m/^((?:\.\d+)+|(?:\.?\w+(?:\-*\w+)+)+)\.?(.*)$/;'
lldpLocChassisIdSubtype

I realise that these two regular expression might not particularly sensible and are certainly not equivalent semantically, but they are syntactically correct, as far as I can tell, and do not contain back references etc. and thus the matching process should be able to always
terminate.
History
Date User Action Args
2010-07-29 11:09:19gnuthorsetrecipients: + gnuthor
2010-07-29 11:09:18gnuthorsetmessageid: <1280401758.9.0.252357713623.issue9414@psf.upfronthosting.co.za>
2010-07-29 11:09:16gnuthorlinkissue9414 messages
2010-07-29 11:09:15gnuthorcreate