In regard to: Re: EPIPE, Nick Lamb said (at 2:57am on May 11, 2000):
>On Wed, May 10, 2000 at 07:15:57PM -0500, Tim Mooney wrote:
>> Finding a processor/OS combo where sizeof(pointer) != sizeof(int) is pretty
>> easy, however. How does this change your thinking?
>Wouldn't atomicity guarantees be a processor feature, and hence tied
>to word size (probably pointer width if you are taking full advantage
>of your CPU) rather than whatever CC might think sizeof(int) is ?
>If I've completely forgotten my architecture course, don't hesitate
>to write me a long rant, off-list of course...
You may very well be completely correct, but it's essentially circumstantial
to the point I was trying to make. My whole goal with the text you quoted
was trying to dissuade anyone from making the "All the world's a ILP32 OS"
assumption, which seemed where things might be headed. These days, recent
versions of Unix from Sun, HP, SGI, IBM, and Compaq (as well as other's I'm
not familiar with) on recent hardware are all LP64 or LP64-capable. Making
pointer size assumptions was never a good idea, but it's especially problematic
Linux and some of the BSD derivatives also have plans to be LP64 some day too,
so even Linux-only developers want to keep a mindful eye on code portability to
Tim Mooney [EMAIL PROTECTED]
Information Technology Services (701) 231-1076 (Voice)
Room 242-J1, IACC Building (701) 231-8541 (Fax)
North Dakota State University, Fargo, ND 58105-5164