On 07-Apr-05, 15:22 (CDT), Pierre THIERRY <[EMAIL PROTECTED]> wrote:
> Scribit Steve Greenland dies 06/04/2005 hora 17:37:
> > There's a long history of people relying on explicitly unspecified
> > behaviour, and then bitching when that behaviour changes.
>
> For the sake of my curiosity again, could you point me some precise
> examples?
Off the top of my head:
i = i++;
Any number of things to do with the C pre-processor
Assuming that sizeof(int) = some specific value
Assumptions about which order the linux kernel will load modules
Assumptions about variables being initialized.
Assumptions about underlying sort algorithms.
...
And yes, I've seen (well, read) people in each case demanding that the
vendor "fix" the compiler/kernel/whatever because "it used to work".
Think about how much effort MS puts into ensuring that older software
runs on new versions of the OS. We all point and laugh at e.g XP SP2,
and at Windows reliability in general, but they've got huge testing labs
to deal with the fact that people may read the published API, but often
rely on some subtle side-effect behaviour.
Steve
--
Steve Greenland
The irony is that Bill Gates claims to be making a stable operating
system and Linus Torvalds claims to be trying to take over the
world. -- seen on the net
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
