🚀 go-pugleaf

In article jeh@alpha.phys.msu.su, krotoff@boy.nmd.msu.ru (Alexsander Krotoff) writes:

>In the september working paper section 18.4 (Dynamic memory management)
>
>i	Header <new> synopsis
>
>	#include <cstdlib>	// for size_t
>	^^^^^^^^^^^^^^^^^^^^
>It looks wrong. size_t must be defined in the cstdef header, not
>cstdlib.

According to the ISO C standard, size_t is defined in several headers,
one of which is <stdlib.h>, so it is not wrong to specify <cstdlib>.
But if the only definition that is required is size_t, it would
be more appropriate to include <cstddef> instead, since it has
fewer extraneous declarations.

Apart from that, it was proposed in Tokyo that no standard header
could include another (as is already the case in C). There were a
few problems discovered with that proposal, and it is still being
worked on, but some form of it may pass. In that case, this point
I expect would become moot.

---
Steve Clamage, stephen.clamage@eng.sun.com



[ comp.std.c++ is moderated.  Submission address: std-c++@ncar.ucar.edu.
  Contact address: std-c++-request@ncar.ucar.edu.  The moderation policy
  is summarized in http://dogbert.lbl.gov/~matt/std-c++/policy.html. ]