On Saturday, 29 April 2017 at 09:24:35 UTC, Ola Fosheim Grøstad
wrote:
On Saturday, 29 April 2017 at 08:45:26 UTC, Moritz Maxeiner
wrote:
On Saturday, 29 April 2017 at 07:26:45 UTC, Timon Gehr wrote:
I don't doubt that, but the implicit generalization is
"multiple pointer types are necessarily always a royal PITA".
The "implicit generalization" is your interpretation, though.
No, it was brought up in a thread as an argument against having
multiple pointer types.
It was brought up in that thread as an example of multiple
pointer types having (I assume unintended) negative consequences.
That's not the same as implying that *all* occurrences of
multiple pointer types *will* have such negative consequences; at
most, it implies that you have to be very careful when designing
them, so as to avoid such consequences (and this latter part is
*my* interpretation).
The thread was not about near/far pointers or segmented memory
models.
I am aware; it was originally about the viability of automatic
reference counting in D, and its potential benefits/drawbacks
compared to garbage collection.
