1. RE: What are Win2000 Diferance
Tony Steward wrote:
> What do I need to be aware of for my progs to run on win2000 and I can
> only
> test on win98se.
>
> I have found that result variables need to be atoms as integers will
> cause
> crashes.
Are you sure about this? I haven't really noticed anything like this
with Windows 2000, but I might not be doing what you're doing.
Could you supply a small example with alternatives for win2k and
win98, please?
Thanks, Tony.
-Ron T.
2. RE: What are Win2000 Diferance
Actually, this should be the case with all Windows programs. It's just
that you'll hit this issue more frequently on an NT based system. I
can't think of anything other than that. You can always ask for testers
if you want to be sure...
-- Brian
Tony Steward wrote:
> Hi All,
> What do I need to be aware of for my progs to run on win2000 and I can
> only
> test on win98se.
>
> I have found that result variables need to be atoms as integers will
> cause
> crashes.
>
> What else?
>
> Thanks
> Tony
>
>
3. RE: What are Win2000 Diferance
Which results are you guys talking about? I've never run into this
(at least, if I did, I didn't recognize it).
-Ron T.
Brian Broker wrote:
> Actually, this should be the case with all Windows programs. It's just
> that you'll hit this issue more frequently on an NT based system. I
> can't think of anything other than that. You can always ask for testers
>
> if you want to be sure...
>
> -- Brian
>
> Tony Steward wrote:
> > Hi All,
> > What do I need to be aware of for my progs to run on win2000 and I can
> > only
> > test on win98se.
> >
> > I have found that result variables need to be atoms as integers will
> > cause
> > crashes.
> >
> > What else?
> >
> > Thanks
> > Tony
> >
> >
4. RE: What are Win2000 Diferance
Oh, right. I remember hearing about the 16-bit int used by Windows.
At the time I was using an Amiga where the int was 32-bit; I've
always thought of int's as being 32-bit and that's why
it didn't register.
Thanks, David.
-Ron T.
David Cuny wrote:
> Under Win95, most pointers returned in Win32Lib tended to be low memory
> addresses. So if you wrote a code like:
>
> int myVar
> myVar = getSomeWin32Pointer()
>
> under Win95, it would tend to work, because the addresses returned were
> small
> enough that they would fit into an int. It happened to me quite often:
> I'd
> test a program under Win95 and it would run fine, but under WinNT, it
> would
> crash.
>
> The error under WinNT would be attempting to store a large address into
> an
> int. That's because WinNT (and WinMe, Win2000 and even Wine under Linux)
>
> would tend to hand back pointer addresses that were too large to store
> in an
> int. The proper way to code the example would have been to use an atom
> instead of an int:
>
> atom myVar
>
> So basically, it was *always* an error to use an int, but because Win95
> tended to parcel out lower memory addresses, it was less likely to crash
>
> under Win95.
>
> -- David Cuny
>
> -- David Cuny
>
>
