Sequence points

[Dave Theese]

Hello all,

I have read the definition of a sequence point in the standard and can
follow it mechanically, but for some reason, I'm having a hard time grasping
the conceptual meaning and significance. Can someone out there offer good
intuition on sequence points???

Thank you!
Dave

[Alf P. Steinbach]

Re: Sequence points

On Tue, 2 Sep 2003 19:19:55 -0700, "Dave Theese" <cheeser_1998@y ahoo.com> wrote:
[color=blue]
>I have read the definition of a sequence point in the standard and can
>follow it mechanically, but for some reason, I'm having a hard time grasping
>the conceptual meaning and significance. Can someone out there offer good
>intuition on sequence points???[/color]

Goodest intuition: you don't need to think about them if you avoid
expressions with side-effects.

That also generally gives much more clear code.

Leave it to the compiler to do the trivial optimizations.

[Karl Heinz Buchegger]

Re: Sequence points



Dave Theese wrote:[color=blue]
>
> Hello all,
>
> I have read the definition of a sequence point in the standard and can
> follow it mechanically, but for some reason, I'm having a hard time grasping
> the conceptual meaning and significance. Can someone out there offer good
> intuition on sequence points???[/color]

Sequence points are a somewhat artificial concept. Basically a sequencepoint
is a spot in processing where you can be sure that every computation has finished
and the results are where they should be. Between 2 sequencepoints the compiler
is free to rearrange things any way it pleases him. So you cant relay on anything.
Only after a sequencepoint is passed, it is safe to expect results where they
should be.

--
Karl Heinz Buchegger
kbuchegg@gascad .at