Microsoft JPEG Hoax!

John Bean wrote:

My knowledge about programming tells me that writing data to non-allocated
buffer is a bad thing and needs to be fixed!

It is a buffer overrun in the simple sense of the word: Data is written
*beyond* the end of an allocated memory buffer!

Whatever. You are simply wrong - writing to a null pointer is *not* the same
as a buffer overflow. You're making it up as you go along.

I have never said that there is a writing to a null pointer, that was
(mistakenly) suggested by another guy.
If an allocated buffer overruns, then you are writing to *non-allocated*
memory past the end of the buffer. That is what I said.

Regards
Guido

Roland Karlsson wrote:
Bruce Murphy wrote in
:

If you want to do something useful, try to stop people
using C.

Thats correct. But what alternative?


/Roland

A Pascal-like language, e.g. Borland's Delphi, has much stronger
type-checking and a somewhat wider range of data types. These features
can significantly improve the quality of the resulting code. Ada may go
even further.

Cheers,
David

Roland Karlsson wrote:
Bruce Murphy wrote in
:

If you want to do something useful, try to stop people
using C.

Thats correct. But what alternative?


/Roland

A Pascal-like language, e.g. Borland's Delphi, has much stronger
type-checking and a somewhat wider range of data types. These features
can significantly improve the quality of the resulting code. Ada may go
even further.

Cheers,
David

Guido Vollbeding writes:

Bruce Murphy wrote:

How do you know? Have you seen the code?

That isn't the point.
The point is that common programs still crash with this problem,
and it doesn't get fixed.

Actually the *point* is your completely unsubstantiated claimn that it was
an exploitable buffer overrun.

B

Guido Vollbeding writes:

Bruce Murphy wrote:

How do you know? Have you seen the code?

That isn't the point.
The point is that common programs still crash with this problem,
and it doesn't get fixed.

Actually the *point* is your completely unsubstantiated claimn that it was
an exploitable buffer overrun.

B

Guido Vollbeding writes:

John Bean wrote:

My knowledge about programming tells me that writing data to non-allocated
buffer is a bad thing and needs to be fixed!

It is a buffer overrun in the simple sense of the word: Data is written
*beyond* the end of an allocated memory buffer!

Whatever. You are simply wrong - writing to a null pointer is *not* the same
as a buffer overflow. You're making it up as you go along.

I have never said that there is a writing to a null pointer, that was
(mistakenly) suggested by another guy.
If an allocated buffer overruns, then you are writing to *non-allocated*
memory past the end of the buffer. That is what I said.

Not necessarily. And in fact, writing to non-allocated memory is
completely harmless. I see that you completely fail to comprehend why
buffer overruns are bad.

B

Guido Vollbeding writes:

John Bean wrote:

My knowledge about programming tells me that writing data to non-allocated
buffer is a bad thing and needs to be fixed!

It is a buffer overrun in the simple sense of the word: Data is written
*beyond* the end of an allocated memory buffer!

Whatever. You are simply wrong - writing to a null pointer is *not* the same
as a buffer overflow. You're making it up as you go along.

I have never said that there is a writing to a null pointer, that was
(mistakenly) suggested by another guy.
If an allocated buffer overruns, then you are writing to *non-allocated*
memory past the end of the buffer. That is what I said.

Not necessarily. And in fact, writing to non-allocated memory is
completely harmless. I see that you completely fail to comprehend why
buffer overruns are bad.

B

"David J Taylor" wrote in :

A Pascal-like language, e.g. Borland's Delphi, has much stronger
type-checking and a somewhat wider range of data types. These features
can significantly improve the quality of the resulting code. Ada may go
even further.

Agree ... but Borland Delphi is a propriarity version of Object Pascal.
I would rule that out as a general programming language. You could use
ordinary Pascal of course. Don't know the status for compilers and
portability. Ada ... strange beast .. might be good ... but no one uses
it over here in Sweden.

Unforunately C++ is the clear winner. It is the only useful language
that programmers knows today.


/Roland

"David J Taylor" wrote in :

A Pascal-like language, e.g. Borland's Delphi, has much stronger
type-checking and a somewhat wider range of data types. These features
can significantly improve the quality of the resulting code. Ada may go
even further.

Agree ... but Borland Delphi is a propriarity version of Object Pascal.
I would rule that out as a general programming language. You could use
ordinary Pascal of course. Don't know the status for compilers and
portability. Ada ... strange beast .. might be good ... but no one uses
it over here in Sweden.

Unforunately C++ is the clear winner. It is the only useful language
that programmers knows today.


/Roland

On 24 Sep 2004 16:53:00 GMT, Roland Karlsson wrote:

"David J Taylor" wrote in :

A Pascal-like language, e.g. Borland's Delphi, has much stronger
type-checking and a somewhat wider range of data types. These features
can significantly improve the quality of the resulting code. Ada may go
even further.

Agree ... but Borland Delphi is a propriarity version of Object Pascal.
I would rule that out as a general programming language. You could use
ordinary Pascal of course. Don't know the status for compilers and
portability. Ada ... strange beast .. might be good ... but no one uses
it over here in Sweden.

I'd be surprised if Saab don't use it in their aerospace systems. Just
getting an Ada program to compile without errors is a major achievement.
Terrible language for general purposes, goes too far the other way.


--
John Bean

All programmers are playwrights and all computers are lousy actors (Anon)