--- deHTMLxs.xs.orig	2005-06-23 08:24:04.000000000 -0600
+++ deHTMLxs.xs	2005-06-23 08:29:27.000000000 -0600
@@ -112,6 +112,7 @@
             text = SvRV(scalarref);
             raw = SvPV(text,size);
 
+            *(raw + size - 1) = '\0';
             if ( (cleaned = malloc(size+1)) && 
                  (res = html_strip(raw, cleaned))  // html_strip will memset cleaned to 0
                  ) {

On JUNE 22nd, Jordan Ritter (a razor developer) responds:

Hey folks,

    We really appreciate the effort from everyone.  We now have
    email.txt and are adding it to the unit tests, plus are in the
    process of fixing the bug itself (permanently this time!).  Please
    be patient, and we'll roll a new release shortly.

    If there are any other problems or issues, razor agents is getting
    some attention so now would be a good time to speak up, on list or
    off (direct to Vipul and myself).

Best,
--jordan

In a subsequent conversation with Jordan:

# Correct me if I'm wrong, but I believe this problem applies to all
# versions between (and including) 2.61 to 2.72.

Well, AFAICT the offending line of code was introduced in the most
recent release cycle.  This is not to say that the original crash was
not present from a different bug(s) in the code; just that this
particular single byte overwrite past the end of the buffer is brand
new to the 2.7x series.

# Jordan, has any thought been given to publishing patches for major
# older versions? I know that Debian 3.1 (Sarge stable) uses 2.67 and
# something like that might be of value to package maintainers of
# distros like Debian that maintain older versions.

Only enough thought to dismiss the idea altogether.

The old versions have serious problems, many of which were finally
addressed in this latest release cycle (see 2.7x release notes for
full list).  Plus, frankly, the older code is overcomplicated and
painfully difficult to maintain or extend (current is only an
iterative improvement, if that).  Finally, the latest versions have
significant accuracy improvements.

So, aside from this recent (admittedly serious) mea culpa which we
will address shortly, there is no real reason for people not to
upgrade.  It's Better.

Best,
--jordan