On Fri, 29 Apr 2011 15:22:16 +0200 lukn <lukn...@gmail.com> wrote:
> Dear Clamav-List
> 
> I'm trying to build an md5-sigs from HTML-files. My procedure is as
> follows:
> 
> mkdir /tmp/clamsig
> cd /tmp/clamsig
> sigtool --html-normalise=/data/foo.html
> mv nocomment.html foo.html
> sigtool --md5 foo.html >> /tmp/htmlfiles.hdb
> 
> Then I verified the file using
> clamscan --database=/tmp/htmlfiles.hdb --leave-temps --debug /data/foo.html
> But the signature did not match.
> 
> I investigated the leftover tempfiles from clamscan and it seems that
> sigtool and clamscan normalize differently. sigtool apparently converts
> & (ampersand) even in URLs to &amp; where clamscan leaves ampersands
> intact. This can produce different files and therefore different md5
> hashes.
> 
> I am not absolutely sure, but I think this is new since clamav 0.97. My
> procedure worked in previous versions of clamav.
> Can anybody confirm this? Is this indended behaviour? If so, what's the
> recommended way of creating md5 signatures from HTML-files?
> 
> My current version of clamav is 0.97 from Debian repositories:
> clamscan --version
> ClamAV 0.97/13022/Fri Apr 29 08:03:10 2011

Please open a bug report at bugs.clamav.net and attach the HTML file if
possible.

-- 
   oo    .....         Tomasz Kojm <tk...@clamav.net>
  (\/)\.........         http://www.ClamAV.net/gpg/tkojm.gpg
     \..........._         0DCA5A08407D5288279DB43454822DC8985A444B
       //\   /\              Fri Apr 29 16:44:24 CEST 2011
_______________________________________________
Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net
http://www.clamav.net/support/ml

Reply via email to