Quoting Michael Van Canneyt <[EMAIL PROTECTED]>:

> 
> 
> On Mon, 22 Aug 2005, Mattias Gaertner wrote:
> 
> > On Mon, 22 Aug 2005 10:45:04 +0300
> > Ido Kanner <[EMAIL PROTECTED]> wrote:
> >
> >> Hello all,
> >>
> >> There is a security advisory regarding SynEdit.
> >>
> >> Don't warry it's not that bad :)
> >
> > Yes it is.

Well in security issues, it is not bad I would rate it low, and if I had a lower
rating it was there instead :)

The reason is because its just hide information... it does not do anything else.
It is not a DoS, or a Buffer Overflow or hurt anything ... only hide 
information...

> >
> >
> >> It seems that by placing NULL Zerrow chars inside a text file, you can
> >> hide from that point, the rest of the file content. That way I can give
> >> you a code that may seems like implemention something X but hide more
> code
> >> that will be compiled at the end by a programming language etc...
> >>
> >> The advisory btw was reported at: http://rgod.altervista.org/syn.html
> >>
> >> BTW I hope that there will be much more securiy advisory for Pascal based
> >> programs/components. That way we will know that more and more people uses
> >> this type of programs (Now I open Pandora's box) :)
> >
> > I fixed TSynPasSyn and TSynPHPSyn. Probably the other highlighters also
> have
> > the problem.
> >
> > But what more troubling is, that the FCL TStrings, TStringList stop at #0
> > and some parts of synedit too. Because of this you can loose code and
> that's
> > pretty bad.

There should be a filter for a null terminited string that will convert it to #0
 string or will just remove it... in PHP it is more sevear.

> 
> I don't see how you can loose code. If there is a #0 somewhere in your
> source,
> the compiler won't compile it, this is for sure.

You do not loose code... only you do not see it. Lets assume you wrote a PHP
code for example:
<?php echo "hello world"; php?> #0 <?php `rm -rf /`; php?>

You will not see the "`" chars and the execution itself...

> 
> What is more, delphi has the same behaviour.

It's the string way behavior. In delphi it meant that you can not concat string
with "#0" without remove that char. And it seems that FPC is the same with that
behavior.

> 
> But the FCL should be fixed, this is for sure.
> 
> Michael.

Ido 





_________________________________________________________________
     To unsubscribe: mail [EMAIL PROTECTED] with
                "unsubscribe" as the Subject
   archives at http://www.lazarus.freepascal.org/mailarchives

Reply via email to