Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Fri, 27 Dec 2013 11:22:10 -0500 (EST)
Subject: Re: CVE issues with recent python flaws

Hash: SHA1

> One CVE (CVE-2013-1752) as assigned to all of these, which would have
> been perfectly reasonable if they had _all_ been fixed simultaneously
> (or at least in the same version).

> Unfortunately, upstream assigned a single CVE to all of these

MITRE is responsible for the CVE-2013-1752 assignment. (In other
words, the upstream Python vendor was not responsible for deciding how
many CVEs were originally assigned.)

The background is that, months ago, MITRE received a request from an
authoritative upstream vendor contact indicating that bugs
16037/16038/16039/16040/16041/16042 needed vulnerability remediation
and that a CVE mapping was desired. At the time of the request, no fix
had been released for any of the six bugs. As far as we could tell,
all of the six pieces of affected code were shipped in the primary
downloadable Python distributions. (In other words, they weren't third
party components distributed elsewhere for use with Python.) The six
pieces of code seemed to have the same type of implementation issue,
so one CVE was assigned.

We do realize that many CVE consumers rely on separate CVE IDs for
vulnerabilities with different affected versions, and thus MITRE can
generate the appropriate number of new CVE IDs (apparently either 4 or
5), and include explanatory references to the new CVEs within the
upcoming "REJECT" entry at:

The decision between 4 and 5 only depends on:

  imaplib [9] (not yet fixed in 2.7.x, fixed in 2.6.9 [10], 3.3.3 [11])
  poplib [16] (not yet fixed in 2.7.x, fixed in 2.6.9 [17], 3.3.3 [18])

If anyone knows that the upstream vendor is specifically planning to
fix imaplib and poplib in different 2.7.x versions, then these two
will have separate CVE IDs, and the total will be 5.

Otherwise (e.g., if both fixes are probably going to be in 2.7.7, or
if just no one knows when the fixes will be shipped), then we assign
the same CVE ID for imaplib and poplib, and the total will be 4.

Finally, we had one comment about this:

> Date: Thu, 26 Dec 2013 21:32:27 -0700
> From: Kurt Seifried
> my personal take: these are very different code modules (different
> protocols) so CVE split

Differences in the implemented protocol typically aren't something
that MITRE looks at in deciding how many CVE IDs are needed, as long
as each protocol provides an opportunity for untrusted input to

- -- 
CVE assignment team, MITRE CVE Numbering Authority
M/S M300
202 Burlington Road, Bedford, MA 01730 USA
[ PGP key available through ]
Version: GnuPG v1.4.14 (SunOS)


Powered by blists - more mailing lists

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.