Re: COBOL Compiler for Windows

On May 2, 4:58=A0pm, Binyamin Dissen <postin...@[EMAIL PROTECTED]
>
wrote:
> On Thu, 01 May 2008 21:24:38 -0500 Robert <n...@[EMAIL PROTECTED]
> wrote:
>
> :>On Thu, 1 May 2008 10:39:46 +1200, "Pete Dashwood"
<dashw...@[EMAIL PROTECTED]
>
> :>wrote:
>
> :>>Point 2 is largely an illusion. Certainly, there are no viruses on
> :>>mainframes, but there will be once they are opened up to the network;
i=
t is
> :>>only a matter of time. (No, I don't have mainframe virus writing on
my
> :>>"TODO" list, but it would be an "interesting" exercise... :-)).
>
> :>How does a virus get into a database server? You can't send it
executabl=
e code.
>
> Buffer overflow.

In order for a 'Buffer Overflow' to work then first there needs to be
a buffer that can overflow, that is a data block that is larger than
the buffer must be able to get to it unchecked. If the gateway or I/O
system checks the data block size then buffer overflow won't happen.

Next it needs to overflow from a data area into an area that can be
executed. If the memory is segmented so that data segments cannot be
executed then the buffer overflow may corrupt data but it won't
execute code.

Then if it can get executed it needs to be able to establish itelf in
the system, by for example adding some code to an executable on the
disk so that it can be run each time that program is started. If the
file system priviledge prevents this then it will only be a data
corruption in a single program run.

So why are the vast majority of viruses aimed at Windows ? It is
because MS makes it easy and other systems make it hard.