Roy,

I've tried the same test project with FullDebugMode turned on for FastMM4,
and it isn't detecting any memory overwrites either.

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim


Annoying isn't it

I'm at a loss as to what to try to isolate it. There's obviously some sort of horrible complex interaction (or very simple cock up) going on.

Roy Lambert

Tim


It looks as though 2.01b4 has removed whatever caused the effect. Either that or some subtle code change of mine has removed whatever was influencing your code.

My bet is its going to be one of those great unsolved mysteries.

Roy Lambert

Roy,

<< It looks as though 2.01b4 has removed whatever caused the effect. Either
that or some subtle code change of mine has removed whatever was influencing
your code.

My bet is its going to be one of those great unsolved mysteries. >>

My guess is that it was this issue:

http://www.elevatesoft.com/incident?action=viewrep&category=edb&release=2.01&type=f&incident=2722

The problem would only manifest itself if the memory manager happened to
re-use the exact same block of memory for the cursor being returned from
each script or procedure execution.  So, depending upon how it was executed,
the script or procedure could experience the error on your machine/setup but
not on mine, or vice-versa.  Ulrich just so happened to put together a
scenario that would allow it to happen all of the time (God bless him, too,
because this would have been a doozy to find manually .

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim


>The problem would only manifest itself if the memory manager happened to
>re-use the exact same block of memory for the cursor being returned from
>each script or procedure execution. So, depending upon how it was executed,
>the script or procedure could experience the error on your machine/setup but
>not on mine, or vice-versa. Ulrich just so happened to put together a
>scenario that would allow it to happen all of the time (God bless him, too,
>because this would have been a doozy to find manually .

Which could also explain why I could get it 100% of the time in the app but when I tried to create a demo by copying the form nada.

Roy Lambert

ps - I never want him to assign me a password

Roy,

<< Which could also explain why I could get it 100% of the time in the app
but when I tried to create a demo by copying the form nada. >>

Yep, exactly.

<< ps - I never want him to assign me a password >>

Well, they say long passwords are more secure.

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim

>Well, they say long passwords are more secure.

Actually I would say they are intrinsically less secure, at least if they are like Ulrich's. Impossible to memorise which means either written down on a bit of paper (probably stuck next to the monitor) or held in a file (probably called passwords.txt) in plain text so you can cut'n'paste.

Roy Lambert

Roy,

<< Actually I would say they are intrinsically less secure, at least if they
are like Ulrich's. Impossible to memorise which means either written down on
a bit of paper (probably stuck next to the monitor) or held in a file
(probably called passwords.txt) in plain text so you can cut'n'paste. >>

That's certainly one way of looking at it - the social engineering side of
things.  I was speaking more to the brute-force attack side of things.

--
Tim Young
Elevate Software
www.elevatesoft.com

Tim


><< Actually I would say they are intrinsically less secure, at least if they
>are like Ulrich's. Impossible to memorise which means either written down on
>a bit of paper (probably stuck next to the monitor) or held in a file
>(probably called passwords.txt) in plain text so you can cut'n'paste. >>
>
>That's certainly one way of looking at it - the social engineering side of
>things. I was speaking more to the brute-force attack side of things.

I have noticed that we tend to have differing viewpoints possibly described as human interface level vs code operation level

Roy Lambert