[softwarelist] Re: OvnPro 2.77 type 5 errors

In message <Mov5T9Ii56aJFwM$@pilling.demon.co.uk> you wrote:

> In message <79cba21c50.Jo@xxxxxxxxxxxxxxxxxxxxxxxx>, John Tytgat 
> <John.Tytgat@xxxxxxxx> writes
> >In message <4jgEhyBeezaJFw6c@xxxxxxxxxxxxxxxxxxx> you wrote:
> >
> >> There's a case to be made for filling all allocated memory with zero,
> >> this hides bugs like the above or at least gives consistent results -
> >> Windows does it for security reasons [...]
> >
> >I don't think this is true.  You might think that's being the case when
> >running programs in debug mode but that's not happening in release mode.
> 
> What is not true?
> 
> Windows won't give your process new memory full of random data from 
> another process because it is a security hazard.

That might be true, but when you allocate memory you don't know whether
it is memory freshly allocated from another process to your process, or
wheter you're reusing memory pages you've deallocated before or a
combination.

> I may well be wrong in the sense that malloc() the C language allocator 
> can hand me memory I've used before and contains random data. But that's 
> an issue internal to my program.

Sure, but are you know implying that you don't use standard C memory
allocation routines and directly use Window memory management in Ovation Pro
on Windows ? Otherwise, this feels like a moot point.

And even then, the prefilled new process memory pages can contain any
value, even pseudo random data, that shouldn't be any security problem (as
long it doesn't reveal information from other program running).

John.
-- 
John Tytgat, in his comfy chair at home                                 BASS
John.Tytgat@xxxxxxxx                             ARM powered, RISC OS driven
To unsubscribe or subscribe goto: http://www.freelists.org/list/davidpilling

Other related posts: