Re: C file recoginzed as image file

["Juanma Barranquero" <lekktu@gmail.com>]

On 1/9/07, Chris Moore <dooglus@gmail.com> wrote:

> then that's tautological.  I can't believe that's what you meant to
> say.

Yes, I was stating a tautology. That was the idea. The interesting
part was the bit about the ".c" extension, of course. Somehow people
seems to think (or at least, seem to be expressing in this thread the
thought) that files with .c extension (or whatever) can be *anything*
and we can do *nothing* to reliably detect them, just because some
contrived, absolutely non-realworld examples can be found. I was
saying: of course you can reliably detect many things.

The point is that contents detection is not infallible, but it is much
more reliable (if done correctly) than extension matching. I don't see
why we should treat the latter as more significant (or even equally
significant) that the former.

But all this is irrelevant to whether we should auto-detect images or
not, I think. The issue there is not that we cannot auto-detect them
reliably (we can, for any significant value of "reliably"), but that
we cannot detect whether they contain viruses.

                    /L/e/k/t/u