Re: [Exim] exiscan-acl-4.31-16

Bruce Richardson wrote:
>
> On Wed, Mar 31, 2004 at 11:36:52AM -0600, Edgar Lovecraft wrote:
> > > I prefer not to filter against the message body like that. There is
> > > an unavoidable risk of false positives.
> > >
> >
> > Anything is possible, but two notes,
> > 1) regex aside, how do you think MIME parsers find such information?
>
> The MIME parser won't even be looking for it unless there was a matching
> content-type declaration in the header. If the exiscan code were no
> more sophisticated than a crude regex match against the body then I
> wouldn't use it.
>

You misunderstood my point 'regex aside' (forget the regex), I never said
that exiscan was using a 'crude regex' to find content-type boundaries, my
point was, any MIME parser does, in essence, some kind of 'matching' or
parsing, on the message text to find where each part should begin and or
end, find what the file names are, etc. So I got currious and went looking
over Toms code and I am correct, a bunch of string compares are being done.
That was my only point, yes, my regex after looking at it again could use
a little tweaking (to match even better as there are a few things I miss),
but it is the same principal.

Don't take offense Tom I am not knocking your code in any way :)

--

--EAL--