On Mon, 08 Sep 2008 18:00:19 -0500
René Berber <[EMAIL PROTECTED]> wrote:

> The idea of Cygwin is to be able to build with the same sources used
> under Linux.  Of course there are problem areas, those are few, and the
> Cygwin maintainers/package porters should take care of them.
> 
> My recommendation for ClamAV is to get rid of all the #ifdef CL_CYGWIN
> blocks, and any others related to Cygwin.

This is what we did in 0.94. We simply dropped the special support for
Cygwin (that is all Cygwin related code entries) as this code was not
working properly. The major problem here was similar to the one described
in my previous message. If something was not working with a particular 
version of the Cygwin DLL and we implemented a work-around for the
issue then a little while later it was turning out that this workaround is
not compatible with a new version of Cygwin. This definitely wasn't the
way to go.

This morning I was doing various tests of ClamAV 0.93.1 provided by
Cygwin. I must say the results were pretty much surprising to me as
I didn't find any problems (10000+ infected files in various formats got
detected properly; also there were no false positives). Now I'm curious
how 0.94 (which comes without any special Cygwin code) would pass
the same test.

> A native build will not work with Exim (which has sections written
> explicitly for Cygwin and builds out of the box, just like ClamAV) or
> any other application that uses sockets; the socket file is not
> compatible with native applications.

Doesn't the AV code in Exim support TCP sockets?

-- 
   oo    .....         Tomasz Kojm <[EMAIL PROTECTED]>
  (\/)\.........         http://www.ClamAV.net/gpg/tkojm.gpg
     \..........._         0DCA5A08407D5288279DB43454822DC8985A444B
       //\   /\              Tue Sep  9 12:15:08 CEST 2008
_______________________________________________
Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net
http://www.clamav.net/support/ml

Reply via email to