D pointers

[Michael Pyne]

On Sunday 02 October 2005 05:11, Alexander Neundorf wrote:
> There's always much talk about how open source software might help
> especially in poor countries. But if the software requires up-to-date
> hardware this argument is mainly void :-/

That's a valid point, but if running one more malloc() is seriously that much 
of a burden then why are we splitting hairs about D pointers when there is a 
whole slew of code in kdelibs that could probably be optimized?  People have 
already pointed out that D pointers may not be a good idea in classes that 
*have to be* speed-intensive.  But for the vast majority of classes one extra 
malloc() will not be noticable.

As far as the indirection cost goes, what I remember from my CS courses seems 
to indicate that it should be an amortized cost of about zero on modern 
systems (and by modern I mean > i486).  Otherwise wouldn't we be having 
massive problems with the dynamic PLT tables and other fun things that .so 
files use nowadays?

Regards,
 - Michael Pyne
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/kde-core-devel/attachments/20051002/e62db1be/attachment.sig>