Re: [evolution-patches] small patch for security

[Not Zed <notzed ximian com>]

Possible, but it would be better to be optional since it isn't free.  Actually now i think about it it should have been a camel-object property.  That way it could be propagated to related objects.

On Fri, 2004-07-09 at 12:05 -0400, Jeffrey Stedfast wrote:
> Looks good. I wonder if we might just always want to clear memory used
> by camel for certain objects all the time tho? CamelMimeFilter could
> just always clear the backup buffer and such. We could walways clear
> memstream objects, etc.
>
> I dunno, just a thought.
>
> Jeff
>
> On Fri, 2004-07-09 at 13:19 +0800, Not Zed wrote:
> > 
> > As the comment says, this doesn't really make a huge security
> > difference, but provides a first-level attempt, and the start of an
> > api for it.
> > 
> > It Just munges memory of decrypted parts after we're finished with
> > them, rather than leaving the data around in memory for bugs to
> > potentially expose the data.  Of course this doesn't affect any
> > mimefilters or gtkhtml data either ...
> > 
> > (i left the other outstanding camel bug in the diff 'cause i'm lazy)
> > 
> > -- 
> > 
> > Michael Zucchi <notzed ximian com>
> > 
> > Novell's Evolution and Free
> > Software Developer

--

Michael Zucchi <notzed ximian com> Novell's Evolution and Free Software Developer